diff --git "a/checkpoint-10000/trainer_state.json" "b/checkpoint-10000/trainer_state.json" new file mode 100644--- /dev/null +++ "b/checkpoint-10000/trainer_state.json" @@ -0,0 +1,72224 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 0.19859395480001588, + "global_step": 10000, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.0, + "learning_rate": 9.92063492063492e-07, + "loss": 10.5363, + "theoretical_loss": 20.841284642178216, + "tokens_seen": 65536 + }, + { + "epoch": 0.0, + "learning_rate": 1.984126984126984e-06, + "loss": 10.5427, + "theoretical_loss": 17.59466794495971, + "tokens_seen": 131072 + }, + { + "epoch": 0.0, + "learning_rate": 2.9761904761904763e-06, + "loss": 10.4533, + "theoretical_loss": 15.967943933467634, + "tokens_seen": 196608 + }, + { + "epoch": 0.0, + "learning_rate": 3.968253968253968e-06, + "loss": 10.3058, + "theoretical_loss": 14.920783596619636, + "tokens_seen": 262144 + }, + { + "epoch": 0.0, + "learning_rate": 4.96031746031746e-06, + "loss": 10.1096, + "theoretical_loss": 14.164683778394041, + "tokens_seen": 327680 + }, + { + "epoch": 0.0, + "learning_rate": 5.9523809523809525e-06, + "loss": 9.8845, + "theoretical_loss": 13.581028313181289, + "tokens_seen": 393216 + }, + { + "epoch": 0.0, + "learning_rate": 6.944444444444444e-06, + "loss": 9.6505, + "theoretical_loss": 13.110275842032255, + "tokens_seen": 458752 + }, + { + "epoch": 0.0, + "learning_rate": 7.936507936507936e-06, + "loss": 9.4442, + "theoretical_loss": 12.71859646611439, + "tokens_seen": 524288 + }, + { + "epoch": 0.0, + "learning_rate": 8.928571428571428e-06, + "loss": 9.2734, + "theoretical_loss": 12.38505930450469, + "tokens_seen": 589824 + }, + { + "epoch": 0.0, + "learning_rate": 9.92063492063492e-06, + "loss": 9.1025, + "theoretical_loss": 12.095879447666144, + "tokens_seen": 655360 + }, + { + "epoch": 0.0, + "learning_rate": 1.0912698412698412e-05, + "loss": 9.0262, + "theoretical_loss": 11.841533071732742, + "tokens_seen": 720896 + }, + { + "epoch": 0.0, + "learning_rate": 1.1904761904761905e-05, + "loss": 8.9331, + "theoretical_loss": 11.615186049337796, + "tokens_seen": 786432 + }, + { + "epoch": 0.0, + "learning_rate": 1.2896825396825396e-05, + "loss": 8.7148, + "theoretical_loss": 11.411780980817854, + "tokens_seen": 851968 + }, + { + "epoch": 0.0, + "learning_rate": 1.3888888888888888e-05, + "loss": 8.6914, + "theoretical_loss": 11.227478542742938, + "tokens_seen": 917504 + }, + { + "epoch": 0.0, + "learning_rate": 1.4880952380952381e-05, + "loss": 8.6474, + "theoretical_loss": 11.059300758608979, + "tokens_seen": 983040 + }, + { + "epoch": 0.0, + "learning_rate": 1.5873015873015872e-05, + "loss": 8.5482, + "theoretical_loss": 10.904894927088016, + "tokens_seen": 1048576 + }, + { + "epoch": 0.0, + "learning_rate": 1.6865079365079364e-05, + "loss": 8.5071, + "theoretical_loss": 10.76237258069379, + "tokens_seen": 1114112 + }, + { + "epoch": 0.0, + "learning_rate": 1.7857142857142855e-05, + "loss": 8.5023, + "theoretical_loss": 10.630196716861345, + "tokens_seen": 1179648 + }, + { + "epoch": 0.0, + "learning_rate": 1.884920634920635e-05, + "loss": 8.3973, + "theoretical_loss": 10.507101012987707, + "tokens_seen": 1245184 + }, + { + "epoch": 0.0, + "learning_rate": 1.984126984126984e-05, + "loss": 8.4645, + "theoretical_loss": 10.392030784394397, + "tokens_seen": 1310720 + }, + { + "epoch": 0.0, + "learning_rate": 2.0833333333333333e-05, + "loss": 8.3747, + "theoretical_loss": 10.284099061527812, + "tokens_seen": 1376256 + }, + { + "epoch": 0.0, + "learning_rate": 2.1825396825396824e-05, + "loss": 8.2515, + "theoretical_loss": 10.182553393901085, + "tokens_seen": 1441792 + }, + { + "epoch": 0.0, + "learning_rate": 2.2817460317460315e-05, + "loss": 8.2989, + "theoretical_loss": 10.086750402363664, + "tokens_seen": 1507328 + }, + { + "epoch": 0.0, + "learning_rate": 2.380952380952381e-05, + "loss": 8.1898, + "theoretical_loss": 9.996136019471344, + "tokens_seen": 1572864 + }, + { + "epoch": 0.0, + "learning_rate": 2.48015873015873e-05, + "loss": 8.1327, + "theoretical_loss": 9.910229967024176, + "tokens_seen": 1638400 + }, + { + "epoch": 0.0, + "learning_rate": 2.5793650793650793e-05, + "loss": 8.1546, + "theoretical_loss": 9.828613432171625, + "tokens_seen": 1703936 + }, + { + "epoch": 0.0, + "learning_rate": 2.6785714285714284e-05, + "loss": 8.0295, + "theoretical_loss": 9.75091918753851, + "tokens_seen": 1769472 + }, + { + "epoch": 0.0, + "learning_rate": 2.7777777777777776e-05, + "loss": 7.9978, + "theoretical_loss": 9.676823599712613, + "tokens_seen": 1835008 + }, + { + "epoch": 0.0, + "learning_rate": 2.876984126984127e-05, + "loss": 8.0568, + "theoretical_loss": 9.606040111776705, + "tokens_seen": 1900544 + }, + { + "epoch": 0.0, + "learning_rate": 2.9761904761904762e-05, + "loss": 7.9578, + "theoretical_loss": 9.538313887395919, + "tokens_seen": 1966080 + }, + { + "epoch": 0.0, + "learning_rate": 3.075396825396825e-05, + "loss": 7.8925, + "theoretical_loss": 9.473417378263001, + "tokens_seen": 2031616 + }, + { + "epoch": 0.0, + "learning_rate": 3.1746031746031745e-05, + "loss": 7.7494, + "theoretical_loss": 9.411146631541524, + "tokens_seen": 2097152 + }, + { + "epoch": 0.0, + "learning_rate": 3.273809523809524e-05, + "loss": 7.7455, + "theoretical_loss": 9.351318194865048, + "tokens_seen": 2162688 + }, + { + "epoch": 0.0, + "learning_rate": 3.373015873015873e-05, + "loss": 7.7254, + "theoretical_loss": 9.293766507291341, + "tokens_seen": 2228224 + }, + { + "epoch": 0.0, + "learning_rate": 3.472222222222222e-05, + "loss": 7.6986, + "theoretical_loss": 9.238341688075822, + "tokens_seen": 2293760 + }, + { + "epoch": 0.0, + "learning_rate": 3.571428571428571e-05, + "loss": 7.6191, + "theoretical_loss": 9.184907653139359, + "tokens_seen": 2359296 + }, + { + "epoch": 0.0, + "learning_rate": 3.670634920634921e-05, + "loss": 7.6517, + "theoretical_loss": 9.133340503044415, + "tokens_seen": 2424832 + }, + { + "epoch": 0.0, + "learning_rate": 3.76984126984127e-05, + "loss": 7.5377, + "theoretical_loss": 9.0835271371648, + "tokens_seen": 2490368 + }, + { + "epoch": 0.0, + "learning_rate": 3.8690476190476195e-05, + "loss": 7.4805, + "theoretical_loss": 9.035364057274894, + "tokens_seen": 2555904 + }, + { + "epoch": 0.0, + "learning_rate": 3.968253968253968e-05, + "loss": 7.4399, + "theoretical_loss": 8.988756330540422, + "tokens_seen": 2621440 + }, + { + "epoch": 0.0, + "learning_rate": 4.067460317460318e-05, + "loss": 7.3525, + "theoretical_loss": 8.943616687271678, + "tokens_seen": 2686976 + }, + { + "epoch": 0.0, + "learning_rate": 4.1666666666666665e-05, + "loss": 7.4896, + "theoretical_loss": 8.89986473310929, + "tokens_seen": 2752512 + }, + { + "epoch": 0.0, + "learning_rate": 4.265873015873016e-05, + "loss": 7.4314, + "theoretical_loss": 8.857426258784564, + "tokens_seen": 2818048 + }, + { + "epoch": 0.0, + "learning_rate": 4.365079365079365e-05, + "loss": 7.329, + "theoretical_loss": 8.816232633409479, + "tokens_seen": 2883584 + }, + { + "epoch": 0.0, + "learning_rate": 4.464285714285714e-05, + "loss": 7.3404, + "theoretical_loss": 8.776220269542522, + "tokens_seen": 2949120 + }, + { + "epoch": 0.0, + "learning_rate": 4.563492063492063e-05, + "loss": 7.1576, + "theoretical_loss": 8.737330150151898, + "tokens_seen": 3014656 + }, + { + "epoch": 0.0, + "learning_rate": 4.6626984126984126e-05, + "loss": 7.2459, + "theoretical_loss": 8.699507409139912, + "tokens_seen": 3080192 + }, + { + "epoch": 0.0, + "learning_rate": 4.761904761904762e-05, + "loss": 7.1942, + "theoretical_loss": 8.662700958366539, + "tokens_seen": 3145728 + }, + { + "epoch": 0.0, + "learning_rate": 4.8611111111111115e-05, + "loss": 7.0679, + "theoretical_loss": 8.62686315516739, + "tokens_seen": 3211264 + }, + { + "epoch": 0.0, + "objective/train/docs_used": 8966, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 6.998983383178711, + "objective/train/theoretical_loss": 8.591949505242134, + "objective/train/tokens_used": 3276800, + "theoretical_loss": 8.591949505242134, + "tokens_seen": 3276800 + }, + { + "epoch": 0.0, + "learning_rate": 4.96031746031746e-05, + "loss": 7.0724, + "theoretical_loss": 8.591949505242134, + "tokens_seen": 3276800 + }, + { + "epoch": 0.0, + "learning_rate": 5.05952380952381e-05, + "loss": 7.0441, + "theoretical_loss": 8.557918396526182, + "tokens_seen": 3342336 + }, + { + "epoch": 0.0, + "learning_rate": 5.1587301587301586e-05, + "loss": 6.8523, + "theoretical_loss": 8.524730860277067, + "tokens_seen": 3407872 + }, + { + "epoch": 0.0, + "learning_rate": 5.257936507936508e-05, + "loss": 6.9204, + "theoretical_loss": 8.492350356128252, + "tokens_seen": 3473408 + }, + { + "epoch": 0.0, + "learning_rate": 5.357142857142857e-05, + "loss": 6.8266, + "theoretical_loss": 8.460742578303845, + "tokens_seen": 3538944 + }, + { + "epoch": 0.0, + "learning_rate": 5.4563492063492063e-05, + "loss": 6.9001, + "theoretical_loss": 8.429875280561783, + "tokens_seen": 3604480 + }, + { + "epoch": 0.0, + "learning_rate": 5.555555555555555e-05, + "loss": 6.6689, + "theoretical_loss": 8.399718117751275, + "tokens_seen": 3670016 + }, + { + "epoch": 0.0, + "learning_rate": 5.6547619047619046e-05, + "loss": 6.8371, + "theoretical_loss": 8.370242502142142, + "tokens_seen": 3735552 + }, + { + "epoch": 0.0, + "learning_rate": 5.753968253968254e-05, + "loss": 6.7857, + "theoretical_loss": 8.341421472916394, + "tokens_seen": 3801088 + }, + { + "epoch": 0.0, + "learning_rate": 5.8531746031746036e-05, + "loss": 6.7543, + "theoretical_loss": 8.313229577412248, + "tokens_seen": 3866624 + }, + { + "epoch": 0.0, + "learning_rate": 5.9523809523809524e-05, + "loss": 6.6209, + "theoretical_loss": 8.28564276288293, + "tokens_seen": 3932160 + }, + { + "epoch": 0.0, + "learning_rate": 6.051587301587302e-05, + "loss": 6.6031, + "theoretical_loss": 8.258638277681179, + "tokens_seen": 3997696 + }, + { + "epoch": 0.0, + "learning_rate": 6.15079365079365e-05, + "loss": 6.7098, + "theoretical_loss": 8.232194580909036, + "tokens_seen": 4063232 + }, + { + "epoch": 0.0, + "learning_rate": 6.25e-05, + "loss": 6.6293, + "theoretical_loss": 8.20629125968411, + "tokens_seen": 4128768 + }, + { + "epoch": 0.0, + "learning_rate": 6.349206349206349e-05, + "loss": 6.3992, + "theoretical_loss": 8.180908953270682, + "tokens_seen": 4194304 + }, + { + "epoch": 0.0, + "learning_rate": 6.448412698412699e-05, + "loss": 6.4486, + "theoretical_loss": 8.156029283408564, + "tokens_seen": 4259840 + }, + { + "epoch": 0.0, + "learning_rate": 6.547619047619048e-05, + "loss": 6.325, + "theoretical_loss": 8.131634790246775, + "tokens_seen": 4325376 + }, + { + "epoch": 0.0, + "learning_rate": 6.646825396825397e-05, + "loss": 6.4583, + "theoretical_loss": 8.107708873353706, + "tokens_seen": 4390912 + }, + { + "epoch": 0.0, + "learning_rate": 6.746031746031745e-05, + "loss": 6.3865, + "theoretical_loss": 8.084235737332481, + "tokens_seen": 4456448 + }, + { + "epoch": 0.0, + "learning_rate": 6.845238095238096e-05, + "loss": 6.2288, + "theoretical_loss": 8.061200341620074, + "tokens_seen": 4521984 + }, + { + "epoch": 0.0, + "learning_rate": 6.944444444444444e-05, + "loss": 6.2213, + "theoretical_loss": 8.038588354092902, + "tokens_seen": 4587520 + }, + { + "epoch": 0.0, + "learning_rate": 7.043650793650793e-05, + "loss": 6.1839, + "theoretical_loss": 8.016386108140445, + "tokens_seen": 4653056 + }, + { + "epoch": 0.0, + "learning_rate": 7.142857142857142e-05, + "loss": 6.2489, + "theoretical_loss": 7.994580562902867, + "tokens_seen": 4718592 + }, + { + "epoch": 0.0, + "learning_rate": 7.242063492063492e-05, + "loss": 6.1256, + "theoretical_loss": 7.973159266399054, + "tokens_seen": 4784128 + }, + { + "epoch": 0.0, + "learning_rate": 7.341269841269842e-05, + "loss": 6.0849, + "theoretical_loss": 7.952110321298584, + "tokens_seen": 4849664 + }, + { + "epoch": 0.0, + "learning_rate": 7.440476190476191e-05, + "loss": 6.0181, + "theoretical_loss": 7.931422353115133, + "tokens_seen": 4915200 + }, + { + "epoch": 0.0, + "learning_rate": 7.53968253968254e-05, + "loss": 6.0866, + "theoretical_loss": 7.911084480620269, + "tokens_seen": 4980736 + }, + { + "epoch": 0.0, + "learning_rate": 7.63888888888889e-05, + "loss": 6.0697, + "theoretical_loss": 7.891086288295694, + "tokens_seen": 5046272 + }, + { + "epoch": 0.0, + "learning_rate": 7.738095238095239e-05, + "loss": 6.0702, + "theoretical_loss": 7.871417800659003, + "tokens_seen": 5111808 + }, + { + "epoch": 0.0, + "learning_rate": 7.837301587301588e-05, + "loss": 5.8881, + "theoretical_loss": 7.8520694583133785, + "tokens_seen": 5177344 + }, + { + "epoch": 0.0, + "learning_rate": 7.936507936507937e-05, + "loss": 6.0891, + "theoretical_loss": 7.833032095585231, + "tokens_seen": 5242880 + }, + { + "epoch": 0.0, + "learning_rate": 8.035714285714287e-05, + "loss": 6.0062, + "theoretical_loss": 7.8142969196261385, + "tokens_seen": 5308416 + }, + { + "epoch": 0.0, + "learning_rate": 8.134920634920635e-05, + "loss": 5.8641, + "theoretical_loss": 7.79585549086643, + "tokens_seen": 5373952 + }, + { + "epoch": 0.0, + "learning_rate": 8.234126984126984e-05, + "loss": 5.8038, + "theoretical_loss": 7.777699704717685, + "tokens_seen": 5439488 + }, + { + "epoch": 0.0, + "learning_rate": 8.333333333333333e-05, + "loss": 5.8129, + "theoretical_loss": 7.759821774430384, + "tokens_seen": 5505024 + }, + { + "epoch": 0.0, + "learning_rate": 8.432539682539683e-05, + "loss": 5.9048, + "theoretical_loss": 7.742214215020986, + "tokens_seen": 5570560 + }, + { + "epoch": 0.0, + "learning_rate": 8.531746031746032e-05, + "loss": 5.7387, + "theoretical_loss": 7.72486982819, + "tokens_seen": 5636096 + }, + { + "epoch": 0.0, + "learning_rate": 8.630952380952381e-05, + "loss": 5.8788, + "theoretical_loss": 7.707781688159247, + "tokens_seen": 5701632 + }, + { + "epoch": 0.0, + "learning_rate": 8.73015873015873e-05, + "loss": 5.8354, + "theoretical_loss": 7.690943128362394, + "tokens_seen": 5767168 + }, + { + "epoch": 0.0, + "learning_rate": 8.82936507936508e-05, + "loss": 5.7952, + "theoretical_loss": 7.674347728928345, + "tokens_seen": 5832704 + }, + { + "epoch": 0.0, + "learning_rate": 8.928571428571429e-05, + "loss": 5.815, + "theoretical_loss": 7.6579893049019265, + "tokens_seen": 5898240 + }, + { + "epoch": 0.0, + "learning_rate": 9.027777777777777e-05, + "loss": 5.7962, + "theoretical_loss": 7.641861895150762, + "tokens_seen": 5963776 + }, + { + "epoch": 0.0, + "learning_rate": 9.126984126984126e-05, + "loss": 5.6841, + "theoretical_loss": 7.625959751911346, + "tokens_seen": 6029312 + }, + { + "epoch": 0.0, + "learning_rate": 9.226190476190476e-05, + "loss": 5.687, + "theoretical_loss": 7.610277330930954, + "tokens_seen": 6094848 + }, + { + "epoch": 0.0, + "learning_rate": 9.325396825396825e-05, + "loss": 5.6526, + "theoretical_loss": 7.5948092821654365, + "tokens_seen": 6160384 + }, + { + "epoch": 0.0, + "learning_rate": 9.424603174603175e-05, + "loss": 5.5096, + "theoretical_loss": 7.579550440996015, + "tokens_seen": 6225920 + }, + { + "epoch": 0.0, + "learning_rate": 9.523809523809524e-05, + "loss": 5.6188, + "theoretical_loss": 7.564495819930985, + "tokens_seen": 6291456 + }, + { + "epoch": 0.0, + "learning_rate": 9.623015873015874e-05, + "loss": 5.5871, + "theoretical_loss": 7.549640600760828, + "tokens_seen": 6356992 + }, + { + "epoch": 0.0, + "learning_rate": 9.722222222222223e-05, + "loss": 5.5818, + "theoretical_loss": 7.534980127137563, + "tokens_seen": 6422528 + }, + { + "epoch": 0.0, + "learning_rate": 9.821428571428572e-05, + "loss": 5.6187, + "theoretical_loss": 7.520509897551315, + "tokens_seen": 6488064 + }, + { + "epoch": 0.0, + "objective/train/docs_used": 10124, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 5.656030654907227, + "objective/train/theoretical_loss": 7.5062255586790965, + "objective/train/tokens_used": 6553600, + "theoretical_loss": 7.5062255586790965, + "tokens_seen": 6553600 + }, + { + "epoch": 0.0, + "learning_rate": 9.92063492063492e-05, + "loss": 5.5868, + "theoretical_loss": 7.5062255586790965, + "tokens_seen": 6553600 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010019841269841271, + "loss": 5.6162, + "theoretical_loss": 7.492122899082528, + "tokens_seen": 6619136 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001011904761904762, + "loss": 5.6289, + "theoretical_loss": 7.478197843232982, + "tokens_seen": 6684672 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010218253968253968, + "loss": 5.558, + "theoretical_loss": 7.464446445844085, + "tokens_seen": 6750208 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010317460317460317, + "loss": 5.5461, + "theoretical_loss": 7.450864886492971, + "tokens_seen": 6815744 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010416666666666667, + "loss": 5.4898, + "theoretical_loss": 7.437449464512941, + "tokens_seen": 6881280 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010515873015873016, + "loss": 5.4514, + "theoretical_loss": 7.424196594141411, + "tokens_seen": 6946816 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010615079365079365, + "loss": 5.4857, + "theoretical_loss": 7.4111027999080985, + "tokens_seen": 7012352 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010714285714285714, + "loss": 5.5841, + "theoretical_loss": 7.398164712249439, + "tokens_seen": 7077888 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010813492063492064, + "loss": 5.6012, + "theoretical_loss": 7.3853790633361545, + "tokens_seen": 7143424 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010912698412698413, + "loss": 5.375, + "theoretical_loss": 7.3727426831017535, + "tokens_seen": 7208960 + }, + { + "epoch": 0.0, + "learning_rate": 0.00011011904761904761, + "loss": 5.5218, + "theoretical_loss": 7.360252495460585, + "tokens_seen": 7274496 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001111111111111111, + "loss": 5.3903, + "theoretical_loss": 7.347905514704743, + "tokens_seen": 7340032 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001121031746031746, + "loss": 5.4841, + "theoretical_loss": 7.33569884206989, + "tokens_seen": 7405568 + }, + { + "epoch": 0.0, + "learning_rate": 0.00011309523809523809, + "loss": 5.3972, + "theoretical_loss": 7.32362966246063, + "tokens_seen": 7471104 + }, + { + "epoch": 0.0, + "learning_rate": 0.00011408730158730158, + "loss": 5.3545, + "theoretical_loss": 7.311695241326718, + "tokens_seen": 7536640 + }, + { + "epoch": 0.0, + "learning_rate": 0.00011507936507936508, + "loss": 5.4828, + "theoretical_loss": 7.299892921681898, + "tokens_seen": 7602176 + }, + { + "epoch": 0.0, + "learning_rate": 0.00011607142857142858, + "loss": 5.4892, + "theoretical_loss": 7.288220121257689, + "tokens_seen": 7667712 + }, + { + "epoch": 0.0, + "learning_rate": 0.00011706349206349207, + "loss": 5.4534, + "theoretical_loss": 7.276674329784936, + "tokens_seen": 7733248 + }, + { + "epoch": 0.0, + "learning_rate": 0.00011805555555555556, + "loss": 5.3461, + "theoretical_loss": 7.265253106396328, + "tokens_seen": 7798784 + }, + { + "epoch": 0.0, + "learning_rate": 0.00011904761904761905, + "loss": 5.2743, + "theoretical_loss": 7.253954077143563, + "tokens_seen": 7864320 + }, + { + "epoch": 0.0, + "learning_rate": 0.00012003968253968255, + "loss": 5.4591, + "theoretical_loss": 7.242774932623186, + "tokens_seen": 7929856 + }, + { + "epoch": 0.0, + "learning_rate": 0.00012103174603174604, + "loss": 5.1881, + "theoretical_loss": 7.2317134257054665, + "tokens_seen": 7995392 + }, + { + "epoch": 0.0, + "learning_rate": 0.00012202380952380953, + "loss": 5.2861, + "theoretical_loss": 7.220767369361081, + "tokens_seen": 8060928 + }, + { + "epoch": 0.0, + "learning_rate": 0.000123015873015873, + "loss": 5.1446, + "theoretical_loss": 7.2099346345805815, + "tokens_seen": 8126464 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001240079365079365, + "loss": 5.4553, + "theoretical_loss": 7.1992131483820305, + "tokens_seen": 8192000 + }, + { + "epoch": 0.0, + "learning_rate": 0.000125, + "loss": 5.3148, + "theoretical_loss": 7.188600891902338, + "tokens_seen": 8257536 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001259920634920635, + "loss": 5.3311, + "theoretical_loss": 7.178095898568204, + "tokens_seen": 8323072 + }, + { + "epoch": 0.0, + "learning_rate": 0.00012698412698412698, + "loss": 5.4596, + "theoretical_loss": 7.167696252342704, + "tokens_seen": 8388608 + }, + { + "epoch": 0.0, + "learning_rate": 0.00012797619047619048, + "loss": 5.2787, + "theoretical_loss": 7.15740008604386, + "tokens_seen": 8454144 + }, + { + "epoch": 0.0, + "learning_rate": 0.00012896825396825398, + "loss": 5.36, + "theoretical_loss": 7.147205579731682, + "tokens_seen": 8519680 + }, + { + "epoch": 0.0, + "learning_rate": 0.00012996031746031748, + "loss": 5.2469, + "theoretical_loss": 7.1371109591604105, + "tokens_seen": 8585216 + }, + { + "epoch": 0.0, + "learning_rate": 0.00013095238095238096, + "loss": 5.3925, + "theoretical_loss": 7.127114494292836, + "tokens_seen": 8650752 + }, + { + "epoch": 0.0, + "learning_rate": 0.00013194444444444446, + "loss": 5.1935, + "theoretical_loss": 7.117214497873759, + "tokens_seen": 8716288 + }, + { + "epoch": 0.0, + "learning_rate": 0.00013293650793650793, + "loss": 5.1444, + "theoretical_loss": 7.107409324059815, + "tokens_seen": 8781824 + }, + { + "epoch": 0.0, + "learning_rate": 0.00013392857142857144, + "loss": 5.2861, + "theoretical_loss": 7.097697367103032, + "tokens_seen": 8847360 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001349206349206349, + "loss": 5.3107, + "theoretical_loss": 7.0880770600856335, + "tokens_seen": 8912896 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001359126984126984, + "loss": 5.2902, + "theoretical_loss": 7.0785468737037345, + "tokens_seen": 8978432 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001369047619047619, + "loss": 5.0761, + "theoretical_loss": 7.069105315097691, + "tokens_seen": 9043968 + }, + { + "epoch": 0.0, + "learning_rate": 0.00013789682539682541, + "loss": 5.135, + "theoretical_loss": 7.059750926726999, + "tokens_seen": 9109504 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001388888888888889, + "loss": 5.068, + "theoretical_loss": 7.050482285287744, + "tokens_seen": 9175040 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001398809523809524, + "loss": 5.0888, + "theoretical_loss": 7.041298000670675, + "tokens_seen": 9240576 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014087301587301586, + "loss": 5.1133, + "theoretical_loss": 7.032196714958137, + "tokens_seen": 9306112 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014186507936507937, + "loss": 5.1895, + "theoretical_loss": 7.023177101458123, + "tokens_seen": 9371648 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014285714285714284, + "loss": 5.2873, + "theoretical_loss": 7.014237863773846, + "tokens_seen": 9437184 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014384920634920634, + "loss": 5.1419, + "theoretical_loss": 7.005377734907255, + "tokens_seen": 9502720 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014484126984126984, + "loss": 5.1221, + "theoretical_loss": 6.996595476395081, + "tokens_seen": 9568256 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014583333333333335, + "loss": 5.1015, + "theoretical_loss": 6.98788987747596, + "tokens_seen": 9633792 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014682539682539685, + "loss": 5.1013, + "theoretical_loss": 6.979259754287375, + "tokens_seen": 9699328 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014781746031746032, + "loss": 5.1131, + "theoretical_loss": 6.970703949091095, + "tokens_seen": 9764864 + }, + { + "epoch": 0.0, + "objective/train/docs_used": 11267, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 5.158783912658691, + "objective/train/theoretical_loss": 6.962221329525966, + "objective/train/tokens_used": 9830400, + "theoretical_loss": 6.962221329525966, + "tokens_seen": 9830400 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014880952380952382, + "loss": 5.1608, + "theoretical_loss": 6.962221329525966, + "tokens_seen": 9830400 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001498015873015873, + "loss": 5.1, + "theoretical_loss": 6.953810787886883, + "tokens_seen": 9895936 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001507936507936508, + "loss": 5.018, + "theoretical_loss": 6.945471240428864, + "tokens_seen": 9961472 + }, + { + "epoch": 0.0, + "learning_rate": 0.00015178571428571427, + "loss": 4.9871, + "theoretical_loss": 6.937201626695213, + "tokens_seen": 10027008 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001527777777777778, + "loss": 5.1259, + "theoretical_loss": 6.929000908868754, + "tokens_seen": 10092544 + }, + { + "epoch": 0.0, + "learning_rate": 0.00015376984126984128, + "loss": 5.1534, + "theoretical_loss": 6.920868071145226, + "tokens_seen": 10158080 + }, + { + "epoch": 0.0, + "learning_rate": 0.00015476190476190478, + "loss": 5.1829, + "theoretical_loss": 6.912802119127937, + "tokens_seen": 10223616 + }, + { + "epoch": 0.0, + "learning_rate": 0.00015575396825396825, + "loss": 4.9816, + "theoretical_loss": 6.904802079242816, + "tokens_seen": 10289152 + }, + { + "epoch": 0.0, + "learning_rate": 0.00015674603174603175, + "loss": 5.0054, + "theoretical_loss": 6.896866998173063, + "tokens_seen": 10354688 + }, + { + "epoch": 0.0, + "learning_rate": 0.00015773809523809523, + "loss": 5.0656, + "theoretical_loss": 6.888995942312617, + "tokens_seen": 10420224 + }, + { + "epoch": 0.0, + "learning_rate": 0.00015873015873015873, + "loss": 4.9665, + "theoretical_loss": 6.881187997237696, + "tokens_seen": 10485760 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001597222222222222, + "loss": 5.0084, + "theoretical_loss": 6.873442267195708, + "tokens_seen": 10551296 + }, + { + "epoch": 0.0, + "learning_rate": 0.00016071428571428573, + "loss": 5.1402, + "theoretical_loss": 6.8657578746108605, + "tokens_seen": 10616832 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001617063492063492, + "loss": 5.0535, + "theoretical_loss": 6.858133959605815, + "tokens_seen": 10682368 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001626984126984127, + "loss": 5.2164, + "theoretical_loss": 6.8505696795387845, + "tokens_seen": 10747904 + }, + { + "epoch": 0.0, + "learning_rate": 0.00016369047619047618, + "loss": 4.9745, + "theoretical_loss": 6.843064208555461, + "tokens_seen": 10813440 + }, + { + "epoch": 0.0, + "learning_rate": 0.00016468253968253969, + "loss": 5.0974, + "theoretical_loss": 6.835616737155247, + "tokens_seen": 10878976 + }, + { + "epoch": 0.0, + "learning_rate": 0.00016567460317460316, + "loss": 4.9074, + "theoretical_loss": 6.828226471771211, + "tokens_seen": 10944512 + }, + { + "epoch": 0.0, + "learning_rate": 0.00016666666666666666, + "loss": 5.0744, + "theoretical_loss": 6.8208926343632905, + "tokens_seen": 11010048 + }, + { + "epoch": 0.0, + "learning_rate": 0.00016765873015873016, + "loss": 5.066, + "theoretical_loss": 6.813614462024214, + "tokens_seen": 11075584 + }, + { + "epoch": 0.0, + "learning_rate": 0.00016865079365079366, + "loss": 4.9987, + "theoretical_loss": 6.806391206597704, + "tokens_seen": 11141120 + }, + { + "epoch": 0.0, + "learning_rate": 0.00016964285714285717, + "loss": 5.0226, + "theoretical_loss": 6.799222134308489, + "tokens_seen": 11206656 + }, + { + "epoch": 0.0, + "learning_rate": 0.00017063492063492064, + "loss": 4.9594, + "theoretical_loss": 6.792106525403691, + "tokens_seen": 11272192 + }, + { + "epoch": 0.0, + "learning_rate": 0.00017162698412698414, + "loss": 5.0722, + "theoretical_loss": 6.785043673805189, + "tokens_seen": 11337728 + }, + { + "epoch": 0.0, + "learning_rate": 0.00017261904761904762, + "loss": 4.9588, + "theoretical_loss": 6.778032886772552, + "tokens_seen": 11403264 + }, + { + "epoch": 0.0, + "learning_rate": 0.00017361111111111112, + "loss": 4.8418, + "theoretical_loss": 6.771073484576156, + "tokens_seen": 11468800 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001746031746031746, + "loss": 5.0703, + "theoretical_loss": 6.764164800180141, + "tokens_seen": 11534336 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001755952380952381, + "loss": 4.8795, + "theoretical_loss": 6.757306178934827, + "tokens_seen": 11599872 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001765873015873016, + "loss": 5.0713, + "theoretical_loss": 6.750496978278292, + "tokens_seen": 11665408 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001775793650793651, + "loss": 4.8703, + "theoretical_loss": 6.743736567446755, + "tokens_seen": 11730944 + }, + { + "epoch": 0.0, + "learning_rate": 0.00017857142857142857, + "loss": 4.9792, + "theoretical_loss": 6.737024327193478, + "tokens_seen": 11796480 + }, + { + "epoch": 0.0, + "learning_rate": 0.00017956349206349207, + "loss": 4.9943, + "theoretical_loss": 6.73035964951589, + "tokens_seen": 11862016 + }, + { + "epoch": 0.0, + "learning_rate": 0.00018055555555555555, + "loss": 4.9053, + "theoretical_loss": 6.723741937390625, + "tokens_seen": 11927552 + }, + { + "epoch": 0.0, + "learning_rate": 0.00018154761904761905, + "loss": 4.8668, + "theoretical_loss": 6.717170604516246, + "tokens_seen": 11993088 + }, + { + "epoch": 0.0, + "learning_rate": 0.00018253968253968252, + "loss": 4.8883, + "theoretical_loss": 6.71064507506334, + "tokens_seen": 12058624 + }, + { + "epoch": 0.0, + "learning_rate": 0.00018353174603174602, + "loss": 4.9096, + "theoretical_loss": 6.704164783431781, + "tokens_seen": 12124160 + }, + { + "epoch": 0.0, + "learning_rate": 0.00018452380952380953, + "loss": 4.8376, + "theoretical_loss": 6.6977291740148805, + "tokens_seen": 12189696 + }, + { + "epoch": 0.0, + "learning_rate": 0.00018551587301587303, + "loss": 4.8575, + "theoretical_loss": 6.691337700970224, + "tokens_seen": 12255232 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001865079365079365, + "loss": 4.8381, + "theoretical_loss": 6.684989827996945, + "tokens_seen": 12320768 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001875, + "loss": 5.0694, + "theoretical_loss": 6.6786850281192365, + "tokens_seen": 12386304 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001884920634920635, + "loss": 4.8019, + "theoretical_loss": 6.672422783475897, + "tokens_seen": 12451840 + }, + { + "epoch": 0.0, + "learning_rate": 0.00018948412698412698, + "loss": 4.9003, + "theoretical_loss": 6.666202585115692, + "tokens_seen": 12517376 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019047619047619048, + "loss": 4.904, + "theoretical_loss": 6.66002393279837, + "tokens_seen": 12582912 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019146825396825398, + "loss": 4.7957, + "theoretical_loss": 6.653886334801115, + "tokens_seen": 12648448 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019246031746031748, + "loss": 4.9042, + "theoretical_loss": 6.647789307730289, + "tokens_seen": 12713984 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019345238095238096, + "loss": 4.877, + "theoretical_loss": 6.641732376338277, + "tokens_seen": 12779520 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019444444444444446, + "loss": 4.9934, + "theoretical_loss": 6.6357150733452785, + "tokens_seen": 12845056 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019543650793650793, + "loss": 4.8432, + "theoretical_loss": 6.629736939265876, + "tokens_seen": 12910592 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019642857142857144, + "loss": 5.0274, + "theoretical_loss": 6.623797522240246, + "tokens_seen": 12976128 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001974206349206349, + "loss": 5.0072, + "theoretical_loss": 6.617896377869851, + "tokens_seen": 13041664 + }, + { + "epoch": 0.0, + "objective/train/docs_used": 12399, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 4.916722297668457, + "objective/train/theoretical_loss": 6.61203306905748, + "objective/train/tokens_used": 13107200, + "theoretical_loss": 6.61203306905748, + "tokens_seen": 13107200 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001984126984126984, + "loss": 4.8694, + "theoretical_loss": 6.61203306905748, + "tokens_seen": 13107200 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019940476190476191, + "loss": 4.9185, + "theoretical_loss": 6.6062071658514965, + "tokens_seen": 13172736 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020039682539682542, + "loss": 4.889, + "theoretical_loss": 6.600418245294164, + "tokens_seen": 13238272 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002013888888888889, + "loss": 4.7748, + "theoretical_loss": 6.594665891273929, + "tokens_seen": 13303808 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002023809523809524, + "loss": 4.9759, + "theoretical_loss": 6.588949694381529, + "tokens_seen": 13369344 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020337301587301587, + "loss": 4.8182, + "theoretical_loss": 6.583269251769811, + "tokens_seen": 13434880 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020436507936507937, + "loss": 4.7935, + "theoretical_loss": 6.577624167017157, + "tokens_seen": 13500416 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020535714285714284, + "loss": 4.79, + "theoretical_loss": 6.5720140499943955, + "tokens_seen": 13565952 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020634920634920634, + "loss": 4.7915, + "theoretical_loss": 6.566438516735092, + "tokens_seen": 13631488 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020734126984126985, + "loss": 4.7613, + "theoretical_loss": 6.560897189309134, + "tokens_seen": 13697024 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020833333333333335, + "loss": 4.7846, + "theoretical_loss": 6.5553896956994855, + "tokens_seen": 13762560 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020932539682539685, + "loss": 4.9138, + "theoretical_loss": 6.5499156696820435, + "tokens_seen": 13828096 + }, + { + "epoch": 0.0, + "learning_rate": 0.00021031746031746032, + "loss": 4.7997, + "theoretical_loss": 6.544474750708482, + "tokens_seen": 13893632 + }, + { + "epoch": 0.0, + "learning_rate": 0.00021130952380952382, + "loss": 4.7965, + "theoretical_loss": 6.539066583792016, + "tokens_seen": 13959168 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002123015873015873, + "loss": 4.7926, + "theoretical_loss": 6.533690819395976, + "tokens_seen": 14024704 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002132936507936508, + "loss": 4.7842, + "theoretical_loss": 6.528347113325134, + "tokens_seen": 14090240 + }, + { + "epoch": 0.0, + "learning_rate": 0.00021428571428571427, + "loss": 4.9612, + "theoretical_loss": 6.523035126619684, + "tokens_seen": 14155776 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002152777777777778, + "loss": 4.7489, + "theoretical_loss": 6.517754525451811, + "tokens_seen": 14221312 + }, + { + "epoch": 0.0, + "learning_rate": 0.00021626984126984128, + "loss": 4.9433, + "theoretical_loss": 6.512504981024765, + "tokens_seen": 14286848 + }, + { + "epoch": 0.0, + "learning_rate": 0.00021726190476190478, + "loss": 4.8904, + "theoretical_loss": 6.507286169474376, + "tokens_seen": 14352384 + }, + { + "epoch": 0.0, + "learning_rate": 0.00021825396825396825, + "loss": 4.7815, + "theoretical_loss": 6.502097771772935, + "tokens_seen": 14417920 + }, + { + "epoch": 0.0, + "learning_rate": 0.00021924603174603176, + "loss": 4.7427, + "theoretical_loss": 6.496939473635374, + "tokens_seen": 14483456 + }, + { + "epoch": 0.0, + "learning_rate": 0.00022023809523809523, + "loss": 4.7958, + "theoretical_loss": 6.491810965427682, + "tokens_seen": 14548992 + }, + { + "epoch": 0.0, + "learning_rate": 0.00022123015873015873, + "loss": 4.6655, + "theoretical_loss": 6.486711942077496, + "tokens_seen": 14614528 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002222222222222222, + "loss": 4.6956, + "theoretical_loss": 6.481642102986796, + "tokens_seen": 14680064 + }, + { + "epoch": 0.0, + "learning_rate": 0.00022321428571428573, + "loss": 4.73, + "theoretical_loss": 6.476601151946659, + "tokens_seen": 14745600 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002242063492063492, + "loss": 4.8086, + "theoretical_loss": 6.471588797054005, + "tokens_seen": 14811136 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002251984126984127, + "loss": 4.8188, + "theoretical_loss": 6.46660475063028, + "tokens_seen": 14876672 + }, + { + "epoch": 0.0, + "learning_rate": 0.00022619047619047618, + "loss": 4.6631, + "theoretical_loss": 6.46164872914203, + "tokens_seen": 14942208 + }, + { + "epoch": 0.0, + "learning_rate": 0.00022718253968253969, + "loss": 4.8528, + "theoretical_loss": 6.456720453123305, + "tokens_seen": 15007744 + }, + { + "epoch": 0.0, + "learning_rate": 0.00022817460317460316, + "loss": 4.715, + "theoretical_loss": 6.4518196470998515, + "tokens_seen": 15073280 + }, + { + "epoch": 0.0, + "learning_rate": 0.00022916666666666666, + "loss": 4.8023, + "theoretical_loss": 6.446946039515035, + "tokens_seen": 15138816 + }, + { + "epoch": 0.0, + "learning_rate": 0.00023015873015873016, + "loss": 4.6724, + "theoretical_loss": 6.442099362657457, + "tokens_seen": 15204352 + }, + { + "epoch": 0.0, + "learning_rate": 0.00023115079365079367, + "loss": 4.6309, + "theoretical_loss": 6.437279352590206, + "tokens_seen": 15269888 + }, + { + "epoch": 0.0, + "learning_rate": 0.00023214285714285717, + "loss": 4.7337, + "theoretical_loss": 6.432485749081721, + "tokens_seen": 15335424 + }, + { + "epoch": 0.0, + "learning_rate": 0.00023313492063492064, + "loss": 4.776, + "theoretical_loss": 6.427718295538199, + "tokens_seen": 15400960 + }, + { + "epoch": 0.0, + "learning_rate": 0.00023412698412698414, + "loss": 4.7802, + "theoretical_loss": 6.422976738937518, + "tokens_seen": 15466496 + }, + { + "epoch": 0.0, + "learning_rate": 0.00023511904761904762, + "loss": 4.7229, + "theoretical_loss": 6.418260829764641, + "tokens_seen": 15532032 + }, + { + "epoch": 0.0, + "learning_rate": 0.00023611111111111112, + "loss": 4.6853, + "theoretical_loss": 6.413570321948453, + "tokens_seen": 15597568 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002371031746031746, + "loss": 4.6681, + "theoretical_loss": 6.40890497279999, + "tokens_seen": 15663104 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002380952380952381, + "loss": 4.6305, + "theoretical_loss": 6.404264542952033, + "tokens_seen": 15728640 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002390873015873016, + "loss": 4.6353, + "theoretical_loss": 6.399648796300032, + "tokens_seen": 15794176 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002400793650793651, + "loss": 4.7157, + "theoretical_loss": 6.395057499944313, + "tokens_seen": 15859712 + }, + { + "epoch": 0.0, + "learning_rate": 0.00024107142857142857, + "loss": 4.6684, + "theoretical_loss": 6.390490424133546, + "tokens_seen": 15925248 + }, + { + "epoch": 0.0, + "learning_rate": 0.00024206349206349207, + "loss": 4.8077, + "theoretical_loss": 6.3859473422094375, + "tokens_seen": 15990784 + }, + { + "epoch": 0.0, + "learning_rate": 0.00024305555555555555, + "loss": 4.7314, + "theoretical_loss": 6.381428030552619, + "tokens_seen": 16056320 + }, + { + "epoch": 0.0, + "learning_rate": 0.00024404761904761905, + "loss": 4.5715, + "theoretical_loss": 6.376932268529695, + "tokens_seen": 16121856 + }, + { + "epoch": 0.0, + "learning_rate": 0.00024503968253968255, + "loss": 4.8244, + "theoretical_loss": 6.372459838441432, + "tokens_seen": 16187392 + }, + { + "epoch": 0.0, + "learning_rate": 0.000246031746031746, + "loss": 4.7778, + "theoretical_loss": 6.368010525472036, + "tokens_seen": 16252928 + }, + { + "epoch": 0.0, + "learning_rate": 0.00024702380952380955, + "loss": 4.592, + "theoretical_loss": 6.363584117639526, + "tokens_seen": 16318464 + }, + { + "epoch": 0.0, + "objective/train/docs_used": 13625, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 4.724667072296143, + "objective/train/theoretical_loss": 6.35918040574715, + "objective/train/tokens_used": 16384000, + "theoretical_loss": 6.35918040574715, + "tokens_seen": 16384000 + }, + { + "epoch": 0.0, + "learning_rate": 0.000248015873015873, + "loss": 4.635, + "theoretical_loss": 6.35918040574715, + "tokens_seen": 16384000 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002490079365079365, + "loss": 4.8013, + "theoretical_loss": 6.354799183335816, + "tokens_seen": 16449536 + }, + { + "epoch": 0.01, + "learning_rate": 0.00025, + "loss": 4.6564, + "theoretical_loss": 6.350440246637536, + "tokens_seen": 16515072 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002509920634920635, + "loss": 4.5694, + "theoretical_loss": 6.3461033945298295, + "tokens_seen": 16580608 + }, + { + "epoch": 0.01, + "learning_rate": 0.000251984126984127, + "loss": 4.5373, + "theoretical_loss": 6.341788428491087, + "tokens_seen": 16646144 + }, + { + "epoch": 0.01, + "learning_rate": 0.00025297619047619046, + "loss": 4.6861, + "theoretical_loss": 6.337495152556851, + "tokens_seen": 16711680 + }, + { + "epoch": 0.01, + "learning_rate": 0.00025396825396825396, + "loss": 4.6677, + "theoretical_loss": 6.333223373277006, + "tokens_seen": 16777216 + }, + { + "epoch": 0.01, + "learning_rate": 0.00025496031746031746, + "loss": 4.6807, + "theoretical_loss": 6.32897289967384, + "tokens_seen": 16842752 + }, + { + "epoch": 0.01, + "learning_rate": 0.00025595238095238096, + "loss": 4.6581, + "theoretical_loss": 6.324743543200984, + "tokens_seen": 16908288 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002569444444444444, + "loss": 4.5691, + "theoretical_loss": 6.320535117703164, + "tokens_seen": 16973824 + }, + { + "epoch": 0.01, + "learning_rate": 0.00025793650793650796, + "loss": 4.5734, + "theoretical_loss": 6.316347439376797, + "tokens_seen": 17039360 + }, + { + "epoch": 0.01, + "learning_rate": 0.00025892857142857146, + "loss": 4.643, + "theoretical_loss": 6.312180326731362, + "tokens_seen": 17104896 + }, + { + "epoch": 0.01, + "learning_rate": 0.00025992063492063497, + "loss": 4.5753, + "theoretical_loss": 6.308033600551573, + "tokens_seen": 17170432 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002609126984126984, + "loss": 4.7378, + "theoretical_loss": 6.303907083860294, + "tokens_seen": 17235968 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002619047619047619, + "loss": 4.7538, + "theoretical_loss": 6.2998006018822075, + "tokens_seen": 17301504 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002628968253968254, + "loss": 4.6408, + "theoretical_loss": 6.295713982008203, + "tokens_seen": 17367040 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002638888888888889, + "loss": 4.5742, + "theoretical_loss": 6.2916470537604745, + "tokens_seen": 17432576 + }, + { + "epoch": 0.01, + "learning_rate": 0.00026488095238095237, + "loss": 4.6161, + "theoretical_loss": 6.287599648758312, + "tokens_seen": 17498112 + }, + { + "epoch": 0.01, + "learning_rate": 0.00026587301587301587, + "loss": 4.5134, + "theoretical_loss": 6.283571600684564, + "tokens_seen": 17563648 + }, + { + "epoch": 0.01, + "learning_rate": 0.00026686507936507937, + "loss": 4.6065, + "theoretical_loss": 6.279562745252756, + "tokens_seen": 17629184 + }, + { + "epoch": 0.01, + "learning_rate": 0.00026785714285714287, + "loss": 4.6903, + "theoretical_loss": 6.2755729201748665, + "tokens_seen": 17694720 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002688492063492063, + "loss": 4.4869, + "theoretical_loss": 6.271601965129717, + "tokens_seen": 17760256 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002698412698412698, + "loss": 4.5323, + "theoretical_loss": 6.267649721731981, + "tokens_seen": 17825792 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002708333333333333, + "loss": 4.6906, + "theoretical_loss": 6.263716033501797, + "tokens_seen": 17891328 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002718253968253968, + "loss": 4.5328, + "theoretical_loss": 6.259800745834964, + "tokens_seen": 17956864 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002728174603174603, + "loss": 4.5172, + "theoretical_loss": 6.255903705973709, + "tokens_seen": 18022400 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002738095238095238, + "loss": 4.5649, + "theoretical_loss": 6.252024762978021, + "tokens_seen": 18087936 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002748015873015873, + "loss": 4.4082, + "theoretical_loss": 6.248163767697523, + "tokens_seen": 18153472 + }, + { + "epoch": 0.01, + "learning_rate": 0.00027579365079365083, + "loss": 4.6068, + "theoretical_loss": 6.244320572743888, + "tokens_seen": 18219008 + }, + { + "epoch": 0.01, + "learning_rate": 0.00027678571428571433, + "loss": 4.4057, + "theoretical_loss": 6.240495032463768, + "tokens_seen": 18284544 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002777777777777778, + "loss": 4.465, + "theoretical_loss": 6.236687002912243, + "tokens_seen": 18350080 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002787698412698413, + "loss": 4.647, + "theoretical_loss": 6.232896341826766, + "tokens_seen": 18415616 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002797619047619048, + "loss": 4.5281, + "theoretical_loss": 6.229122908601596, + "tokens_seen": 18481152 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002807539682539683, + "loss": 4.4868, + "theoretical_loss": 6.225366564262705, + "tokens_seen": 18546688 + }, + { + "epoch": 0.01, + "learning_rate": 0.00028174603174603173, + "loss": 4.5948, + "theoretical_loss": 6.2216271714431635, + "tokens_seen": 18612224 + }, + { + "epoch": 0.01, + "learning_rate": 0.00028273809523809523, + "loss": 4.4644, + "theoretical_loss": 6.217904594358968, + "tokens_seen": 18677760 + }, + { + "epoch": 0.01, + "learning_rate": 0.00028373015873015873, + "loss": 4.5604, + "theoretical_loss": 6.214198698785328, + "tokens_seen": 18743296 + }, + { + "epoch": 0.01, + "learning_rate": 0.00028472222222222223, + "loss": 4.4102, + "theoretical_loss": 6.210509352033379, + "tokens_seen": 18808832 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002857142857142857, + "loss": 4.4662, + "theoretical_loss": 6.206836422927337, + "tokens_seen": 18874368 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002867063492063492, + "loss": 4.5813, + "theoretical_loss": 6.203179781782049, + "tokens_seen": 18939904 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002876984126984127, + "loss": 4.4194, + "theoretical_loss": 6.199539300380979, + "tokens_seen": 19005440 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002886904761904762, + "loss": 4.4881, + "theoretical_loss": 6.195914851954575, + "tokens_seen": 19070976 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002896825396825397, + "loss": 4.4943, + "theoretical_loss": 6.19230631115903, + "tokens_seen": 19136512 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002906746031746032, + "loss": 4.4646, + "theoretical_loss": 6.188713554055438, + "tokens_seen": 19202048 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002916666666666667, + "loss": 4.4328, + "theoretical_loss": 6.185136458089309, + "tokens_seen": 19267584 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002926587301587302, + "loss": 4.4352, + "theoretical_loss": 6.181574902070463, + "tokens_seen": 19333120 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002936507936507937, + "loss": 4.4054, + "theoretical_loss": 6.178028766153277, + "tokens_seen": 19398656 + }, + { + "epoch": 0.01, + "learning_rate": 0.00029464285714285714, + "loss": 4.4233, + "theoretical_loss": 6.174497931817287, + "tokens_seen": 19464192 + }, + { + "epoch": 0.01, + "learning_rate": 0.00029563492063492064, + "loss": 4.4473, + "theoretical_loss": 6.17098228184813, + "tokens_seen": 19529728 + }, + { + "epoch": 0.01, + "learning_rate": 0.00029662698412698414, + "loss": 4.4886, + "theoretical_loss": 6.167481700318827, + "tokens_seen": 19595264 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 14832, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 4.4568328857421875, + "objective/train/theoretical_loss": 6.163996072571392, + "objective/train/tokens_used": 19660800, + "theoretical_loss": 6.163996072571392, + "tokens_seen": 19660800 + }, + { + "epoch": 0.01, + "learning_rate": 0.00029761904761904765, + "loss": 4.4754, + "theoretical_loss": 6.163996072571392, + "tokens_seen": 19660800 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002986111111111111, + "loss": 4.4103, + "theoretical_loss": 6.160525285198771, + "tokens_seen": 19726336 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002996031746031746, + "loss": 4.5743, + "theoretical_loss": 6.157069226027087, + "tokens_seen": 19791872 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003005952380952381, + "loss": 4.6146, + "theoretical_loss": 6.153627784098204, + "tokens_seen": 19857408 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003015873015873016, + "loss": 4.3293, + "theoretical_loss": 6.150200849652586, + "tokens_seen": 19922944 + }, + { + "epoch": 0.01, + "learning_rate": 0.00030257936507936505, + "loss": 4.3663, + "theoretical_loss": 6.146788314112464, + "tokens_seen": 19988480 + }, + { + "epoch": 0.01, + "learning_rate": 0.00030357142857142855, + "loss": 4.3699, + "theoretical_loss": 6.143390070065279, + "tokens_seen": 20054016 + }, + { + "epoch": 0.01, + "learning_rate": 0.00030456349206349205, + "loss": 4.4736, + "theoretical_loss": 6.140006011247426, + "tokens_seen": 20119552 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003055555555555556, + "loss": 4.4497, + "theoretical_loss": 6.136636032528262, + "tokens_seen": 20185088 + }, + { + "epoch": 0.01, + "learning_rate": 0.00030654761904761905, + "loss": 4.3873, + "theoretical_loss": 6.133280029894401, + "tokens_seen": 20250624 + }, + { + "epoch": 0.01, + "learning_rate": 0.00030753968253968255, + "loss": 4.4295, + "theoretical_loss": 6.129937900434269, + "tokens_seen": 20316160 + }, + { + "epoch": 0.01, + "learning_rate": 0.00030853174603174605, + "loss": 4.3107, + "theoretical_loss": 6.126609542322921, + "tokens_seen": 20381696 + }, + { + "epoch": 0.01, + "learning_rate": 0.00030952380952380956, + "loss": 4.3457, + "theoretical_loss": 6.123294854807119, + "tokens_seen": 20447232 + }, + { + "epoch": 0.01, + "learning_rate": 0.000310515873015873, + "loss": 4.3982, + "theoretical_loss": 6.11999373819066, + "tokens_seen": 20512768 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003115079365079365, + "loss": 4.4875, + "theoretical_loss": 6.116706093819951, + "tokens_seen": 20578304 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003125, + "loss": 4.1838, + "theoretical_loss": 6.113431824069829, + "tokens_seen": 20643840 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003134920634920635, + "loss": 4.3586, + "theoretical_loss": 6.110170832329612, + "tokens_seen": 20709376 + }, + { + "epoch": 0.01, + "learning_rate": 0.000314484126984127, + "loss": 4.4453, + "theoretical_loss": 6.10692302298939, + "tokens_seen": 20774912 + }, + { + "epoch": 0.01, + "learning_rate": 0.00031547619047619046, + "loss": 4.5773, + "theoretical_loss": 6.103688301426538, + "tokens_seen": 20840448 + }, + { + "epoch": 0.01, + "learning_rate": 0.00031646825396825396, + "loss": 4.3212, + "theoretical_loss": 6.100466573992453, + "tokens_seen": 20905984 + }, + { + "epoch": 0.01, + "learning_rate": 0.00031746031746031746, + "loss": 4.4159, + "theoretical_loss": 6.097257747999514, + "tokens_seen": 20971520 + }, + { + "epoch": 0.01, + "learning_rate": 0.00031845238095238096, + "loss": 4.4369, + "theoretical_loss": 6.094061731708257, + "tokens_seen": 21037056 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003194444444444444, + "loss": 4.4332, + "theoretical_loss": 6.090878434314753, + "tokens_seen": 21102592 + }, + { + "epoch": 0.01, + "learning_rate": 0.00032043650793650796, + "loss": 4.21, + "theoretical_loss": 6.0877077659382, + "tokens_seen": 21168128 + }, + { + "epoch": 0.01, + "learning_rate": 0.00032142857142857147, + "loss": 4.4939, + "theoretical_loss": 6.084549637608715, + "tokens_seen": 21233664 + }, + { + "epoch": 0.01, + "learning_rate": 0.00032242063492063497, + "loss": 4.2141, + "theoretical_loss": 6.081403961255324, + "tokens_seen": 21299200 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003234126984126984, + "loss": 4.2429, + "theoretical_loss": 6.078270649694149, + "tokens_seen": 21364736 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003244047619047619, + "loss": 4.3184, + "theoretical_loss": 6.07514961661678, + "tokens_seen": 21430272 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003253968253968254, + "loss": 4.2534, + "theoretical_loss": 6.072040776578845, + "tokens_seen": 21495808 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003263888888888889, + "loss": 4.3538, + "theoretical_loss": 6.068944044988752, + "tokens_seen": 21561344 + }, + { + "epoch": 0.01, + "learning_rate": 0.00032738095238095237, + "loss": 4.2514, + "theoretical_loss": 6.065859338096619, + "tokens_seen": 21626880 + }, + { + "epoch": 0.01, + "learning_rate": 0.00032837301587301587, + "loss": 4.1773, + "theoretical_loss": 6.0627865729833745, + "tokens_seen": 21692416 + }, + { + "epoch": 0.01, + "learning_rate": 0.00032936507936507937, + "loss": 4.3498, + "theoretical_loss": 6.059725667550044, + "tokens_seen": 21757952 + }, + { + "epoch": 0.01, + "learning_rate": 0.00033035714285714287, + "loss": 4.3238, + "theoretical_loss": 6.0566765405071905, + "tokens_seen": 21823488 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003313492063492063, + "loss": 4.2048, + "theoretical_loss": 6.0536391113645305, + "tokens_seen": 21889024 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003323412698412698, + "loss": 4.4113, + "theoretical_loss": 6.050613300420713, + "tokens_seen": 21954560 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003333333333333333, + "loss": 4.2596, + "theoretical_loss": 6.047599028753265, + "tokens_seen": 22020096 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003343253968253968, + "loss": 4.2868, + "theoretical_loss": 6.04459621820868, + "tokens_seen": 22085632 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003353174603174603, + "loss": 4.3576, + "theoretical_loss": 6.041604791392675, + "tokens_seen": 22151168 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003363095238095238, + "loss": 4.2732, + "theoretical_loss": 6.038624671660598, + "tokens_seen": 22216704 + }, + { + "epoch": 0.01, + "learning_rate": 0.00033730158730158733, + "loss": 4.3441, + "theoretical_loss": 6.035655783107973, + "tokens_seen": 22282240 + }, + { + "epoch": 0.01, + "learning_rate": 0.00033829365079365083, + "loss": 4.3334, + "theoretical_loss": 6.0326980505612084, + "tokens_seen": 22347776 + }, + { + "epoch": 0.01, + "learning_rate": 0.00033928571428571433, + "loss": 4.3301, + "theoretical_loss": 6.029751399568434, + "tokens_seen": 22413312 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003402777777777778, + "loss": 4.2055, + "theoretical_loss": 6.026815756390491, + "tokens_seen": 22478848 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003412698412698413, + "loss": 4.1341, + "theoretical_loss": 6.023891047992053, + "tokens_seen": 22544384 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003422619047619048, + "loss": 4.3806, + "theoretical_loss": 6.020977202032884, + "tokens_seen": 22609920 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003432539682539683, + "loss": 4.2677, + "theoretical_loss": 6.018074146859234, + "tokens_seen": 22675456 + }, + { + "epoch": 0.01, + "learning_rate": 0.00034424603174603173, + "loss": 4.2443, + "theoretical_loss": 6.015181811495358, + "tokens_seen": 22740992 + }, + { + "epoch": 0.01, + "learning_rate": 0.00034523809523809523, + "loss": 4.3245, + "theoretical_loss": 6.012300125635177, + "tokens_seen": 22806528 + }, + { + "epoch": 0.01, + "learning_rate": 0.00034623015873015873, + "loss": 4.247, + "theoretical_loss": 6.009429019634049, + "tokens_seen": 22872064 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 16074, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 4.314324855804443, + "objective/train/theoretical_loss": 6.006568424500674, + "objective/train/tokens_used": 22937600, + "theoretical_loss": 6.006568424500674, + "tokens_seen": 22937600 + }, + { + "epoch": 0.01, + "learning_rate": 0.00034722222222222224, + "loss": 4.2902, + "theoretical_loss": 6.006568424500674, + "tokens_seen": 22937600 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003482142857142857, + "loss": 4.2128, + "theoretical_loss": 6.003718271889123, + "tokens_seen": 23003136 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003492063492063492, + "loss": 4.2511, + "theoretical_loss": 6.000878494090979, + "tokens_seen": 23068672 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003501984126984127, + "loss": 4.3115, + "theoretical_loss": 5.998049024027599, + "tokens_seen": 23134208 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003511904761904762, + "loss": 4.1709, + "theoretical_loss": 5.995229795242498, + "tokens_seen": 23199744 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003521825396825397, + "loss": 4.1992, + "theoretical_loss": 5.99242074189383, + "tokens_seen": 23265280 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003531746031746032, + "loss": 4.181, + "theoretical_loss": 5.989621798747002, + "tokens_seen": 23330816 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003541666666666667, + "loss": 4.2081, + "theoretical_loss": 5.98683290116738, + "tokens_seen": 23396352 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003551587301587302, + "loss": 4.2125, + "theoretical_loss": 5.9840539851131105, + "tokens_seen": 23461888 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003561507936507937, + "loss": 4.1751, + "theoretical_loss": 5.981284987128044, + "tokens_seen": 23527424 + }, + { + "epoch": 0.01, + "learning_rate": 0.00035714285714285714, + "loss": 4.1139, + "theoretical_loss": 5.97852584433477, + "tokens_seen": 23592960 + }, + { + "epoch": 0.01, + "learning_rate": 0.00035813492063492064, + "loss": 4.3265, + "theoretical_loss": 5.975776494427746, + "tokens_seen": 23658496 + }, + { + "epoch": 0.01, + "learning_rate": 0.00035912698412698415, + "loss": 4.1817, + "theoretical_loss": 5.973036875666525, + "tokens_seen": 23724032 + }, + { + "epoch": 0.01, + "learning_rate": 0.00036011904761904765, + "loss": 4.1412, + "theoretical_loss": 5.970306926869093, + "tokens_seen": 23789568 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003611111111111111, + "loss": 4.2317, + "theoretical_loss": 5.967586587405295, + "tokens_seen": 23855104 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003621031746031746, + "loss": 4.2076, + "theoretical_loss": 5.96487579719035, + "tokens_seen": 23920640 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003630952380952381, + "loss": 4.072, + "theoretical_loss": 5.962174496678481, + "tokens_seen": 23986176 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003640873015873016, + "loss": 4.029, + "theoretical_loss": 5.959482626856606, + "tokens_seen": 24051712 + }, + { + "epoch": 0.01, + "learning_rate": 0.00036507936507936505, + "loss": 4.1314, + "theoretical_loss": 5.956800129238152, + "tokens_seen": 24117248 + }, + { + "epoch": 0.01, + "learning_rate": 0.00036607142857142855, + "loss": 4.1585, + "theoretical_loss": 5.954126945856933, + "tokens_seen": 24182784 + }, + { + "epoch": 0.01, + "learning_rate": 0.00036706349206349205, + "loss": 4.1577, + "theoretical_loss": 5.951463019261125, + "tokens_seen": 24248320 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003680555555555556, + "loss": 4.3404, + "theoretical_loss": 5.948808292507333, + "tokens_seen": 24313856 + }, + { + "epoch": 0.01, + "learning_rate": 0.00036904761904761905, + "loss": 4.3076, + "theoretical_loss": 5.946162709154724, + "tokens_seen": 24379392 + }, + { + "epoch": 0.01, + "learning_rate": 0.00037003968253968255, + "loss": 4.1079, + "theoretical_loss": 5.943526213259268, + "tokens_seen": 24444928 + }, + { + "epoch": 0.01, + "learning_rate": 0.00037103174603174606, + "loss": 4.1351, + "theoretical_loss": 5.940898749368037, + "tokens_seen": 24510464 + }, + { + "epoch": 0.01, + "learning_rate": 0.00037202380952380956, + "loss": 4.1368, + "theoretical_loss": 5.9382802625136, + "tokens_seen": 24576000 + }, + { + "epoch": 0.01, + "learning_rate": 0.000373015873015873, + "loss": 4.1858, + "theoretical_loss": 5.935670698208488, + "tokens_seen": 24641536 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003740079365079365, + "loss": 4.1378, + "theoretical_loss": 5.933070002439743, + "tokens_seen": 24707072 + }, + { + "epoch": 0.01, + "learning_rate": 0.000375, + "loss": 4.1491, + "theoretical_loss": 5.930478121663538, + "tokens_seen": 24772608 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003759920634920635, + "loss": 4.2101, + "theoretical_loss": 5.927895002799874, + "tokens_seen": 24838144 + }, + { + "epoch": 0.01, + "learning_rate": 0.000376984126984127, + "loss": 4.231, + "theoretical_loss": 5.925320593227351, + "tokens_seen": 24903680 + }, + { + "epoch": 0.01, + "learning_rate": 0.00037797619047619046, + "loss": 4.0721, + "theoretical_loss": 5.922754840778015, + "tokens_seen": 24969216 + }, + { + "epoch": 0.01, + "learning_rate": 0.00037896825396825396, + "loss": 4.1095, + "theoretical_loss": 5.9201976937322645, + "tokens_seen": 25034752 + }, + { + "epoch": 0.01, + "learning_rate": 0.00037996031746031746, + "loss": 4.0962, + "theoretical_loss": 5.917649100813846, + "tokens_seen": 25100288 + }, + { + "epoch": 0.01, + "learning_rate": 0.00038095238095238096, + "loss": 4.2504, + "theoretical_loss": 5.915109011184899, + "tokens_seen": 25165824 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003819444444444444, + "loss": 4.1581, + "theoretical_loss": 5.9125773744410814, + "tokens_seen": 25231360 + }, + { + "epoch": 0.01, + "learning_rate": 0.00038293650793650797, + "loss": 4.167, + "theoretical_loss": 5.91005414060676, + "tokens_seen": 25296896 + }, + { + "epoch": 0.01, + "learning_rate": 0.00038392857142857147, + "loss": 4.0207, + "theoretical_loss": 5.90753926013026, + "tokens_seen": 25362432 + }, + { + "epoch": 0.01, + "learning_rate": 0.00038492063492063497, + "loss": 4.2021, + "theoretical_loss": 5.90503268387919, + "tokens_seen": 25427968 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003859126984126984, + "loss": 4.051, + "theoretical_loss": 5.9025343631358185, + "tokens_seen": 25493504 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003869047619047619, + "loss": 4.1718, + "theoretical_loss": 5.900044249592523, + "tokens_seen": 25559040 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003878968253968254, + "loss": 4.0689, + "theoretical_loss": 5.897562295347297, + "tokens_seen": 25624576 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003888888888888889, + "loss": 4.0832, + "theoretical_loss": 5.895088452899312, + "tokens_seen": 25690112 + }, + { + "epoch": 0.01, + "learning_rate": 0.00038988095238095237, + "loss": 4.0056, + "theoretical_loss": 5.892622675144553, + "tokens_seen": 25755648 + }, + { + "epoch": 0.01, + "learning_rate": 0.00039087301587301587, + "loss": 3.9442, + "theoretical_loss": 5.890164915371495, + "tokens_seen": 25821184 + }, + { + "epoch": 0.01, + "learning_rate": 0.00039186507936507937, + "loss": 4.1518, + "theoretical_loss": 5.887715127256854, + "tokens_seen": 25886720 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003928571428571429, + "loss": 4.0461, + "theoretical_loss": 5.885273264861381, + "tokens_seen": 25952256 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003938492063492063, + "loss": 4.1153, + "theoretical_loss": 5.882839282625721, + "tokens_seen": 26017792 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003948412698412698, + "loss": 4.1185, + "theoretical_loss": 5.880413135366323, + "tokens_seen": 26083328 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003958333333333333, + "loss": 4.1858, + "theoretical_loss": 5.877994778271411, + "tokens_seen": 26148864 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 17681, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 4.1578049659729, + "objective/train/theoretical_loss": 5.8755841668969895, + "objective/train/tokens_used": 26214400, + "theoretical_loss": 5.8755841668969895, + "tokens_seen": 26214400 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003968253968253968, + "loss": 4.0358, + "theoretical_loss": 5.8755841668969895, + "tokens_seen": 26214400 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003978174603174603, + "loss": 4.1146, + "theoretical_loss": 5.873181257162926, + "tokens_seen": 26279936 + }, + { + "epoch": 0.01, + "learning_rate": 0.00039880952380952383, + "loss": 4.0096, + "theoretical_loss": 5.870786005349071, + "tokens_seen": 26345472 + }, + { + "epoch": 0.01, + "learning_rate": 0.00039980158730158733, + "loss": 4.012, + "theoretical_loss": 5.868398368091427, + "tokens_seen": 26411008 + }, + { + "epoch": 0.01, + "learning_rate": 0.00040079365079365083, + "loss": 4.0778, + "theoretical_loss": 5.866018302378377, + "tokens_seen": 26476544 + }, + { + "epoch": 0.01, + "learning_rate": 0.00040178571428571433, + "loss": 3.9905, + "theoretical_loss": 5.863645765546956, + "tokens_seen": 26542080 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004027777777777778, + "loss": 3.9379, + "theoretical_loss": 5.861280715279168, + "tokens_seen": 26607616 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004037698412698413, + "loss": 4.0855, + "theoretical_loss": 5.858923109598366, + "tokens_seen": 26673152 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004047619047619048, + "loss": 3.948, + "theoretical_loss": 5.8565729068656545, + "tokens_seen": 26738688 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004057539682539683, + "loss": 3.9088, + "theoretical_loss": 5.85423006577636, + "tokens_seen": 26804224 + }, + { + "epoch": 0.01, + "learning_rate": 0.00040674603174603173, + "loss": 3.9533, + "theoretical_loss": 5.851894545356539, + "tokens_seen": 26869760 + }, + { + "epoch": 0.01, + "learning_rate": 0.00040773809523809523, + "loss": 4.0346, + "theoretical_loss": 5.849566304959527, + "tokens_seen": 26935296 + }, + { + "epoch": 0.01, + "learning_rate": 0.00040873015873015874, + "loss": 3.915, + "theoretical_loss": 5.84724530426254, + "tokens_seen": 27000832 + }, + { + "epoch": 0.01, + "learning_rate": 0.00040972222222222224, + "loss": 3.9503, + "theoretical_loss": 5.844931503263309, + "tokens_seen": 27066368 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004107142857142857, + "loss": 4.0508, + "theoretical_loss": 5.842624862276773, + "tokens_seen": 27131904 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004117063492063492, + "loss": 3.9759, + "theoretical_loss": 5.840325341931795, + "tokens_seen": 27197440 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004126984126984127, + "loss": 3.9598, + "theoretical_loss": 5.838032903167934, + "tokens_seen": 27262976 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004136904761904762, + "loss": 4.013, + "theoretical_loss": 5.835747507232257, + "tokens_seen": 27328512 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004146825396825397, + "loss": 3.9878, + "theoretical_loss": 5.833469115676177, + "tokens_seen": 27394048 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004156746031746032, + "loss": 4.0391, + "theoretical_loss": 5.8311976903523535, + "tokens_seen": 27459584 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004166666666666667, + "loss": 4.0397, + "theoretical_loss": 5.828933193411611, + "tokens_seen": 27525120 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004176587301587302, + "loss": 4.0979, + "theoretical_loss": 5.826675587299913, + "tokens_seen": 27590656 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004186507936507937, + "loss": 3.9448, + "theoretical_loss": 5.824424834755357, + "tokens_seen": 27656192 + }, + { + "epoch": 0.01, + "learning_rate": 0.00041964285714285714, + "loss": 4.0773, + "theoretical_loss": 5.8221808988052235, + "tokens_seen": 27721728 + }, + { + "epoch": 0.01, + "learning_rate": 0.00042063492063492065, + "loss": 3.9358, + "theoretical_loss": 5.819943742763051, + "tokens_seen": 27787264 + }, + { + "epoch": 0.01, + "learning_rate": 0.00042162698412698415, + "loss": 3.8817, + "theoretical_loss": 5.817713330225754, + "tokens_seen": 27852800 + }, + { + "epoch": 0.01, + "learning_rate": 0.00042261904761904765, + "loss": 3.9642, + "theoretical_loss": 5.815489625070766, + "tokens_seen": 27918336 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004236111111111111, + "loss": 3.9413, + "theoretical_loss": 5.813272591453236, + "tokens_seen": 27983872 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004246031746031746, + "loss": 3.8514, + "theoretical_loss": 5.811062193803242, + "tokens_seen": 28049408 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004255952380952381, + "loss": 3.8665, + "theoretical_loss": 5.808858396823043, + "tokens_seen": 28114944 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004265873015873016, + "loss": 3.9022, + "theoretical_loss": 5.806661165484377, + "tokens_seen": 28180480 + }, + { + "epoch": 0.01, + "learning_rate": 0.00042757936507936505, + "loss": 3.9555, + "theoretical_loss": 5.804470465025777, + "tokens_seen": 28246016 + }, + { + "epoch": 0.01, + "learning_rate": 0.00042857142857142855, + "loss": 3.8925, + "theoretical_loss": 5.802286260949924, + "tokens_seen": 28311552 + }, + { + "epoch": 0.01, + "learning_rate": 0.00042956349206349205, + "loss": 3.9545, + "theoretical_loss": 5.8001085190210375, + "tokens_seen": 28377088 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004305555555555556, + "loss": 3.8417, + "theoretical_loss": 5.797937205262292, + "tokens_seen": 28442624 + }, + { + "epoch": 0.01, + "learning_rate": 0.00043154761904761905, + "loss": 4.0024, + "theoretical_loss": 5.795772285953266, + "tokens_seen": 28508160 + }, + { + "epoch": 0.01, + "learning_rate": 0.00043253968253968256, + "loss": 3.925, + "theoretical_loss": 5.793613727627429, + "tokens_seen": 28573696 + }, + { + "epoch": 0.01, + "learning_rate": 0.00043353174603174606, + "loss": 4.0179, + "theoretical_loss": 5.791461497069644, + "tokens_seen": 28639232 + }, + { + "epoch": 0.01, + "learning_rate": 0.00043452380952380956, + "loss": 3.8686, + "theoretical_loss": 5.789315561313717, + "tokens_seen": 28704768 + }, + { + "epoch": 0.01, + "learning_rate": 0.000435515873015873, + "loss": 3.9179, + "theoretical_loss": 5.787175887639959, + "tokens_seen": 28770304 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004365079365079365, + "loss": 3.9392, + "theoretical_loss": 5.785042443572798, + "tokens_seen": 28835840 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004375, + "loss": 3.9636, + "theoretical_loss": 5.782915196878396, + "tokens_seen": 28901376 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004384920634920635, + "loss": 3.9243, + "theoretical_loss": 5.780794115562314, + "tokens_seen": 28966912 + }, + { + "epoch": 0.01, + "learning_rate": 0.000439484126984127, + "loss": 4.035, + "theoretical_loss": 5.778679167867196, + "tokens_seen": 29032448 + }, + { + "epoch": 0.01, + "learning_rate": 0.00044047619047619046, + "loss": 3.8345, + "theoretical_loss": 5.7765703222704765, + "tokens_seen": 29097984 + }, + { + "epoch": 0.01, + "learning_rate": 0.00044146825396825396, + "loss": 3.8599, + "theoretical_loss": 5.7744675474821285, + "tokens_seen": 29163520 + }, + { + "epoch": 0.01, + "learning_rate": 0.00044246031746031746, + "loss": 3.8147, + "theoretical_loss": 5.772370812442426, + "tokens_seen": 29229056 + }, + { + "epoch": 0.01, + "learning_rate": 0.00044345238095238096, + "loss": 3.9394, + "theoretical_loss": 5.770280086319733, + "tokens_seen": 29294592 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004444444444444444, + "loss": 3.9621, + "theoretical_loss": 5.768195338508333, + "tokens_seen": 29360128 + }, + { + "epoch": 0.01, + "learning_rate": 0.00044543650793650797, + "loss": 4.0098, + "theoretical_loss": 5.766116538626263, + "tokens_seen": 29425664 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 18880, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 3.9492247104644775, + "objective/train/theoretical_loss": 5.7640436565131905, + "objective/train/tokens_used": 29491200, + "theoretical_loss": 5.7640436565131905, + "tokens_seen": 29491200 + }, + { + "epoch": 0.01, + "learning_rate": 0.00044642857142857147, + "loss": 3.9917, + "theoretical_loss": 5.7640436565131905, + "tokens_seen": 29491200 + }, + { + "epoch": 0.01, + "learning_rate": 0.00044742063492063497, + "loss": 3.8488, + "theoretical_loss": 5.761976662228307, + "tokens_seen": 29556736 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004484126984126984, + "loss": 4.0269, + "theoretical_loss": 5.759915526048243, + "tokens_seen": 29622272 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004494047619047619, + "loss": 3.8713, + "theoretical_loss": 5.757860218465023, + "tokens_seen": 29687808 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004503968253968254, + "loss": 3.9387, + "theoretical_loss": 5.755810710184019, + "tokens_seen": 29753344 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004513888888888889, + "loss": 3.8531, + "theoretical_loss": 5.753766972121954, + "tokens_seen": 29818880 + }, + { + "epoch": 0.01, + "learning_rate": 0.00045238095238095237, + "loss": 3.9661, + "theoretical_loss": 5.751728975404911, + "tokens_seen": 29884416 + }, + { + "epoch": 0.01, + "learning_rate": 0.00045337301587301587, + "loss": 3.7998, + "theoretical_loss": 5.7496966913663705, + "tokens_seen": 29949952 + }, + { + "epoch": 0.01, + "learning_rate": 0.00045436507936507937, + "loss": 3.9405, + "theoretical_loss": 5.747670091545274, + "tokens_seen": 30015488 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004553571428571429, + "loss": 3.8252, + "theoretical_loss": 5.745649147684103, + "tokens_seen": 30081024 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004563492063492063, + "loss": 3.799, + "theoretical_loss": 5.743633831726987, + "tokens_seen": 30146560 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004573412698412698, + "loss": 3.7585, + "theoretical_loss": 5.741624115817827, + "tokens_seen": 30212096 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004583333333333333, + "loss": 3.8208, + "theoretical_loss": 5.739619972298446, + "tokens_seen": 30277632 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004593253968253968, + "loss": 3.746, + "theoretical_loss": 5.737621373706753, + "tokens_seen": 30343168 + }, + { + "epoch": 0.01, + "learning_rate": 0.00046031746031746033, + "loss": 3.9784, + "theoretical_loss": 5.735628292774945, + "tokens_seen": 30408704 + }, + { + "epoch": 0.01, + "learning_rate": 0.00046130952380952383, + "loss": 3.7694, + "theoretical_loss": 5.7336407024276985, + "tokens_seen": 30474240 + }, + { + "epoch": 0.01, + "learning_rate": 0.00046230158730158733, + "loss": 3.8574, + "theoretical_loss": 5.731658575780418, + "tokens_seen": 30539776 + }, + { + "epoch": 0.01, + "learning_rate": 0.00046329365079365083, + "loss": 3.8273, + "theoretical_loss": 5.729681886137476, + "tokens_seen": 30605312 + }, + { + "epoch": 0.01, + "learning_rate": 0.00046428571428571433, + "loss": 3.8341, + "theoretical_loss": 5.727710606990488, + "tokens_seen": 30670848 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004652777777777778, + "loss": 3.7075, + "theoretical_loss": 5.725744712016601, + "tokens_seen": 30736384 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004662698412698413, + "loss": 3.8238, + "theoretical_loss": 5.723784175076802, + "tokens_seen": 30801920 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004672619047619048, + "loss": 3.7753, + "theoretical_loss": 5.72182897021425, + "tokens_seen": 30867456 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004682539682539683, + "loss": 3.8618, + "theoretical_loss": 5.719879071652615, + "tokens_seen": 30932992 + }, + { + "epoch": 0.01, + "learning_rate": 0.00046924603174603173, + "loss": 3.7777, + "theoretical_loss": 5.717934453794454, + "tokens_seen": 30998528 + }, + { + "epoch": 0.01, + "learning_rate": 0.00047023809523809523, + "loss": 3.8204, + "theoretical_loss": 5.715995091219584, + "tokens_seen": 31064064 + }, + { + "epoch": 0.01, + "learning_rate": 0.00047123015873015874, + "loss": 3.8684, + "theoretical_loss": 5.7140609586834925, + "tokens_seen": 31129600 + }, + { + "epoch": 0.01, + "learning_rate": 0.00047222222222222224, + "loss": 3.7264, + "theoretical_loss": 5.712132031115749, + "tokens_seen": 31195136 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004732142857142857, + "loss": 3.7874, + "theoretical_loss": 5.710208283618447, + "tokens_seen": 31260672 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004742063492063492, + "loss": 3.814, + "theoretical_loss": 5.708289691464657, + "tokens_seen": 31326208 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004751984126984127, + "loss": 3.8277, + "theoretical_loss": 5.706376230096897, + "tokens_seen": 31391744 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004761904761904762, + "loss": 3.8048, + "theoretical_loss": 5.704467875125619, + "tokens_seen": 31457280 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004771825396825397, + "loss": 3.7792, + "theoretical_loss": 5.702564602327719, + "tokens_seen": 31522816 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004781746031746032, + "loss": 3.717, + "theoretical_loss": 5.700666387645049, + "tokens_seen": 31588352 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004791666666666667, + "loss": 3.824, + "theoretical_loss": 5.698773207182963, + "tokens_seen": 31653888 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004801587301587302, + "loss": 3.7964, + "theoretical_loss": 5.696885037208865, + "tokens_seen": 31719424 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004811507936507937, + "loss": 3.8602, + "theoretical_loss": 5.695001854150782, + "tokens_seen": 31784960 + }, + { + "epoch": 0.01, + "learning_rate": 0.00048214285714285715, + "loss": 3.7472, + "theoretical_loss": 5.693123634595938, + "tokens_seen": 31850496 + }, + { + "epoch": 0.01, + "learning_rate": 0.00048313492063492065, + "loss": 3.7976, + "theoretical_loss": 5.69125035528937, + "tokens_seen": 31916032 + }, + { + "epoch": 0.01, + "learning_rate": 0.00048412698412698415, + "loss": 3.7707, + "theoretical_loss": 5.689381993132532, + "tokens_seen": 31981568 + }, + { + "epoch": 0.01, + "learning_rate": 0.00048511904761904765, + "loss": 3.6974, + "theoretical_loss": 5.687518525181924, + "tokens_seen": 32047104 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004861111111111111, + "loss": 3.7914, + "theoretical_loss": 5.685659928647743, + "tokens_seen": 32112640 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004871031746031746, + "loss": 3.7266, + "theoretical_loss": 5.68380618089254, + "tokens_seen": 32178176 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004880952380952381, + "loss": 3.8905, + "theoretical_loss": 5.6819572594298915, + "tokens_seen": 32243712 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004890873015873016, + "loss": 3.8071, + "theoretical_loss": 5.680113141923088, + "tokens_seen": 32309248 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004900793650793651, + "loss": 3.7795, + "theoretical_loss": 5.67827380618384, + "tokens_seen": 32374784 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004910714285714286, + "loss": 3.7937, + "theoretical_loss": 5.6764392301709865, + "tokens_seen": 32440320 + }, + { + "epoch": 0.01, + "learning_rate": 0.000492063492063492, + "loss": 3.7568, + "theoretical_loss": 5.674609391989232, + "tokens_seen": 32505856 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004930555555555556, + "loss": 3.7979, + "theoretical_loss": 5.672784269887885, + "tokens_seen": 32571392 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004940476190476191, + "loss": 3.754, + "theoretical_loss": 5.670963842259615, + "tokens_seen": 32636928 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004950396825396826, + "loss": 3.7097, + "theoretical_loss": 5.669148087639222, + "tokens_seen": 32702464 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 20023, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 3.64802885055542, + "objective/train/theoretical_loss": 5.66733698470242, + "objective/train/tokens_used": 32768000, + "theoretical_loss": 5.66733698470242, + "tokens_seen": 32768000 + }, + { + "epoch": 0.01, + "learning_rate": 0.000496031746031746, + "loss": 3.7888, + "theoretical_loss": 5.66733698470242, + "tokens_seen": 32768000 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004970238095238095, + "loss": 3.6564, + "theoretical_loss": 5.665530512264631, + "tokens_seen": 32833536 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498015873015873, + "loss": 3.8402, + "theoretical_loss": 5.663728649279793, + "tokens_seen": 32899072 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004990079365079365, + "loss": 3.5983, + "theoretical_loss": 5.6619313748391855, + "tokens_seen": 32964608 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005, + "loss": 3.6989, + "theoretical_loss": 5.6601386681702515, + "tokens_seen": 33030144 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004999899699097292, + "loss": 3.7376, + "theoretical_loss": 5.658350508635454, + "tokens_seen": 33095680 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004999799398194584, + "loss": 3.7389, + "theoretical_loss": 5.656566875731127, + "tokens_seen": 33161216 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004999699097291876, + "loss": 3.6364, + "theoretical_loss": 5.654787749086345, + "tokens_seen": 33226752 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004999598796389167, + "loss": 3.7201, + "theoretical_loss": 5.6530131084618045, + "tokens_seen": 33292288 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499949849548646, + "loss": 3.7642, + "theoretical_loss": 5.651242933748718, + "tokens_seen": 33357824 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004999398194583751, + "loss": 3.8021, + "theoretical_loss": 5.649477204967717, + "tokens_seen": 33423360 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004999297893681044, + "loss": 3.7472, + "theoretical_loss": 5.647715902267764, + "tokens_seen": 33488896 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004999197592778335, + "loss": 3.6083, + "theoretical_loss": 5.645959005925086, + "tokens_seen": 33554432 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004999097291875627, + "loss": 3.6859, + "theoretical_loss": 5.644206496342103, + "tokens_seen": 33619968 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004998996990972919, + "loss": 3.5522, + "theoretical_loss": 5.642458354046386, + "tokens_seen": 33685504 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004998896690070211, + "loss": 3.6177, + "theoretical_loss": 5.6407145596896076, + "tokens_seen": 33751040 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004998796389167503, + "loss": 3.6185, + "theoretical_loss": 5.638975094046518, + "tokens_seen": 33816576 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004998696088264795, + "loss": 3.7041, + "theoretical_loss": 5.637239938013924, + "tokens_seen": 33882112 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004998595787362087, + "loss": 3.6024, + "theoretical_loss": 5.6355090726096755, + "tokens_seen": 33947648 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004998495486459378, + "loss": 3.8661, + "theoretical_loss": 5.633782478971672, + "tokens_seen": 34013184 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499839518555667, + "loss": 3.6809, + "theoretical_loss": 5.632060138356873, + "tokens_seen": 34078720 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004998294884653962, + "loss": 3.6527, + "theoretical_loss": 5.630342032140315, + "tokens_seen": 34144256 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004998194583751254, + "loss": 3.7485, + "theoretical_loss": 5.628628141814151, + "tokens_seen": 34209792 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004998094282848546, + "loss": 3.7409, + "theoretical_loss": 5.62691844898668, + "tokens_seen": 34275328 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004997993981945837, + "loss": 3.503, + "theoretical_loss": 5.625212935381407, + "tokens_seen": 34340864 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004997893681043129, + "loss": 3.558, + "theoretical_loss": 5.6235115828361035, + "tokens_seen": 34406400 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004997793380140421, + "loss": 3.7493, + "theoretical_loss": 5.621814373301866, + "tokens_seen": 34471936 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004997693079237714, + "loss": 3.6965, + "theoretical_loss": 5.620121288842209, + "tokens_seen": 34537472 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004997592778335005, + "loss": 3.6443, + "theoretical_loss": 5.618432311632143, + "tokens_seen": 34603008 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004997492477432298, + "loss": 3.5529, + "theoretical_loss": 5.616747423957278, + "tokens_seen": 34668544 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004997392176529588, + "loss": 3.7107, + "theoretical_loss": 5.615066608212926, + "tokens_seen": 34734080 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004997291875626881, + "loss": 3.6327, + "theoretical_loss": 5.613389846903216, + "tokens_seen": 34799616 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004997191574724173, + "loss": 3.708, + "theoretical_loss": 5.611717122640225, + "tokens_seen": 34865152 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004997091273821465, + "loss": 3.8699, + "theoretical_loss": 5.610048418143098, + "tokens_seen": 34930688 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004996990972918757, + "loss": 3.651, + "theoretical_loss": 5.6083837162372, + "tokens_seen": 34996224 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004996890672016048, + "loss": 3.7231, + "theoretical_loss": 5.606722999853261, + "tokens_seen": 35061760 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499679037111334, + "loss": 3.5881, + "theoretical_loss": 5.605066252026537, + "tokens_seen": 35127296 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004996690070210632, + "loss": 3.6909, + "theoretical_loss": 5.603413455895973, + "tokens_seen": 35192832 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004996589769307924, + "loss": 3.65, + "theoretical_loss": 5.601764594703377, + "tokens_seen": 35258368 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004996489468405216, + "loss": 3.664, + "theoretical_loss": 5.600119651792607, + "tokens_seen": 35323904 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004996389167502507, + "loss": 3.7232, + "theoretical_loss": 5.598478610608758, + "tokens_seen": 35389440 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004996288866599799, + "loss": 3.518, + "theoretical_loss": 5.596841454697365, + "tokens_seen": 35454976 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004996188565697091, + "loss": 3.7024, + "theoretical_loss": 5.595208167703601, + "tokens_seen": 35520512 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004996088264794383, + "loss": 3.6958, + "theoretical_loss": 5.5935787333714995, + "tokens_seen": 35586048 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004995987963891675, + "loss": 3.6133, + "theoretical_loss": 5.59195313554317, + "tokens_seen": 35651584 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004995887662988968, + "loss": 3.7519, + "theoretical_loss": 5.590331358158036, + "tokens_seen": 35717120 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004995787362086258, + "loss": 3.6175, + "theoretical_loss": 5.58871338525206, + "tokens_seen": 35782656 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004995687061183551, + "loss": 3.6638, + "theoretical_loss": 5.587099200956995, + "tokens_seen": 35848192 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004995586760280842, + "loss": 3.5958, + "theoretical_loss": 5.585488789499637, + "tokens_seen": 35913728 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004995486459378135, + "loss": 3.6626, + "theoretical_loss": 5.583882135201078, + "tokens_seen": 35979264 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 21333, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 3.7413370609283447, + "objective/train/theoretical_loss": 5.582279222475973, + "objective/train/tokens_used": 36044800, + "theoretical_loss": 5.582279222475973, + "tokens_seen": 36044800 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004995386158475427, + "loss": 3.6368, + "theoretical_loss": 5.582279222475973, + "tokens_seen": 36044800 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004995285857572718, + "loss": 3.3716, + "theoretical_loss": 5.580680035831821, + "tokens_seen": 36110336 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499518555667001, + "loss": 3.5671, + "theoretical_loss": 5.579084559868232, + "tokens_seen": 36175872 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004995085255767302, + "loss": 3.6344, + "theoretical_loss": 5.577492779276222, + "tokens_seen": 36241408 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004994984954864594, + "loss": 3.5464, + "theoretical_loss": 5.575904678837501, + "tokens_seen": 36306944 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004994884653961886, + "loss": 3.7069, + "theoretical_loss": 5.574320243423783, + "tokens_seen": 36372480 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004994784353059178, + "loss": 3.5112, + "theoretical_loss": 5.572739457996079, + "tokens_seen": 36438016 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004994684052156469, + "loss": 3.4806, + "theoretical_loss": 5.5711623076040215, + "tokens_seen": 36503552 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004994583751253761, + "loss": 3.4531, + "theoretical_loss": 5.569588777385177, + "tokens_seen": 36569088 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004994483450351053, + "loss": 3.5561, + "theoretical_loss": 5.568018852564379, + "tokens_seen": 36634624 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004994383149448345, + "loss": 3.738, + "theoretical_loss": 5.566452518453051, + "tokens_seen": 36700160 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004994282848545637, + "loss": 3.4876, + "theoretical_loss": 5.564889760448561, + "tokens_seen": 36765696 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004994182547642928, + "loss": 3.5183, + "theoretical_loss": 5.5633305640335475, + "tokens_seen": 36831232 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004994082246740221, + "loss": 3.5745, + "theoretical_loss": 5.561774914775286, + "tokens_seen": 36896768 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004993981945837512, + "loss": 3.5997, + "theoretical_loss": 5.5602227983250385, + "tokens_seen": 36962304 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004993881644934805, + "loss": 3.7304, + "theoretical_loss": 5.558674200417419, + "tokens_seen": 37027840 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004993781344032096, + "loss": 3.6739, + "theoretical_loss": 5.557129106869764, + "tokens_seen": 37093376 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004993681043129389, + "loss": 3.485, + "theoretical_loss": 5.555587503581506, + "tokens_seen": 37158912 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499358074222668, + "loss": 3.4551, + "theoretical_loss": 5.554049376533554, + "tokens_seen": 37224448 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004993480441323972, + "loss": 3.5917, + "theoretical_loss": 5.552514711787683, + "tokens_seen": 37289984 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004993380140421264, + "loss": 3.5172, + "theoretical_loss": 5.550983495485925, + "tokens_seen": 37355520 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004993279839518556, + "loss": 3.5729, + "theoretical_loss": 5.549455713849963, + "tokens_seen": 37421056 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004993179538615848, + "loss": 3.4699, + "theoretical_loss": 5.547931353180543, + "tokens_seen": 37486592 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004993079237713139, + "loss": 3.4621, + "theoretical_loss": 5.546410399856875, + "tokens_seen": 37552128 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004992978936810431, + "loss": 3.5364, + "theoretical_loss": 5.544892840336053, + "tokens_seen": 37617664 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004992878635907723, + "loss": 3.5406, + "theoretical_loss": 5.543378661152472, + "tokens_seen": 37683200 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004992778335005015, + "loss": 3.5673, + "theoretical_loss": 5.5418678489172555, + "tokens_seen": 37748736 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004992678034102307, + "loss": 3.5267, + "theoretical_loss": 5.540360390317682, + "tokens_seen": 37814272 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004992577733199598, + "loss": 3.5996, + "theoretical_loss": 5.538856272116626, + "tokens_seen": 37879808 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499247743229689, + "loss": 3.4475, + "theoretical_loss": 5.5373554811519945, + "tokens_seen": 37945344 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004992377131394183, + "loss": 3.6489, + "theoretical_loss": 5.535858004336175, + "tokens_seen": 38010880 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004992276830491475, + "loss": 3.3855, + "theoretical_loss": 5.534363828655488, + "tokens_seen": 38076416 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004992176529588767, + "loss": 3.3802, + "theoretical_loss": 5.532872941169639, + "tokens_seen": 38141952 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004992076228686059, + "loss": 3.5816, + "theoretical_loss": 5.531385329011181, + "tokens_seen": 38207488 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499197592778335, + "loss": 3.4592, + "theoretical_loss": 5.529900979384985, + "tokens_seen": 38273024 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004991875626880642, + "loss": 3.5131, + "theoretical_loss": 5.5284198795677, + "tokens_seen": 38338560 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004991775325977934, + "loss": 3.4417, + "theoretical_loss": 5.526942016907242, + "tokens_seen": 38404096 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004991675025075226, + "loss": 3.6538, + "theoretical_loss": 5.5254673788222615, + "tokens_seen": 38469632 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004991574724172518, + "loss": 3.4595, + "theoretical_loss": 5.523995952801634, + "tokens_seen": 38535168 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499147442326981, + "loss": 3.4381, + "theoretical_loss": 5.522527726403953, + "tokens_seen": 38600704 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004991374122367101, + "loss": 3.5432, + "theoretical_loss": 5.521062687257016, + "tokens_seen": 38666240 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004991273821464393, + "loss": 3.5434, + "theoretical_loss": 5.519600823057332, + "tokens_seen": 38731776 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004991173520561685, + "loss": 3.557, + "theoretical_loss": 5.518142121569613, + "tokens_seen": 38797312 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004991073219658977, + "loss": 3.4499, + "theoretical_loss": 5.516686570626293, + "tokens_seen": 38862848 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499097291875627, + "loss": 3.5188, + "theoretical_loss": 5.515234158127032, + "tokens_seen": 38928384 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499087261785356, + "loss": 3.5154, + "theoretical_loss": 5.5137848720382365, + "tokens_seen": 38993920 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004990772316950853, + "loss": 3.65, + "theoretical_loss": 5.5123387003925775, + "tokens_seen": 39059456 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004990672016048144, + "loss": 3.5513, + "theoretical_loss": 5.510895631288515, + "tokens_seen": 39124992 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004990571715145437, + "loss": 3.5139, + "theoretical_loss": 5.50945565288983, + "tokens_seen": 39190528 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004990471414242729, + "loss": 3.4521, + "theoretical_loss": 5.508018753425152, + "tokens_seen": 39256064 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 22534, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 3.509709358215332, + "objective/train/theoretical_loss": 5.506584921187503, + "objective/train/tokens_used": 39321600, + "theoretical_loss": 5.506584921187503, + "tokens_seen": 39321600 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499037111334002, + "loss": 3.5696, + "theoretical_loss": 5.506584921187503, + "tokens_seen": 39321600 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004990270812437312, + "loss": 3.4769, + "theoretical_loss": 5.505154144533835, + "tokens_seen": 39387136 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004990170511534604, + "loss": 3.402, + "theoretical_loss": 5.503726411884568, + "tokens_seen": 39452672 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004990070210631896, + "loss": 3.4346, + "theoretical_loss": 5.502301711723156, + "tokens_seen": 39518208 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004989969909729188, + "loss": 3.3808, + "theoretical_loss": 5.5008800325956235, + "tokens_seen": 39583744 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498986960882648, + "loss": 3.472, + "theoretical_loss": 5.499461363110136, + "tokens_seen": 39649280 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004989769307923771, + "loss": 3.5127, + "theoretical_loss": 5.498045691936548, + "tokens_seen": 39714816 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004989669007021063, + "loss": 3.3674, + "theoretical_loss": 5.496633007805977, + "tokens_seen": 39780352 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004989568706118355, + "loss": 3.315, + "theoretical_loss": 5.495223299510372, + "tokens_seen": 39845888 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004989468405215647, + "loss": 3.5025, + "theoretical_loss": 5.493816555902079, + "tokens_seen": 39911424 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004989368104312939, + "loss": 3.5393, + "theoretical_loss": 5.492412765893421, + "tokens_seen": 39976960 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498926780341023, + "loss": 3.4322, + "theoretical_loss": 5.49101191845628, + "tokens_seen": 40042496 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004989167502507523, + "loss": 3.424, + "theoretical_loss": 5.489614002621677, + "tokens_seen": 40108032 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004989067201604814, + "loss": 3.4955, + "theoretical_loss": 5.4882190074793575, + "tokens_seen": 40173568 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004988966900702107, + "loss": 3.4164, + "theoretical_loss": 5.486826922177387, + "tokens_seen": 40239104 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004988866599799398, + "loss": 3.5892, + "theoretical_loss": 5.4854377359217406, + "tokens_seen": 40304640 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004988766298896691, + "loss": 3.4174, + "theoretical_loss": 5.484051437975901, + "tokens_seen": 40370176 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004988665997993982, + "loss": 3.4968, + "theoretical_loss": 5.482668017660462, + "tokens_seen": 40435712 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004988565697091274, + "loss": 3.3865, + "theoretical_loss": 5.481287464352727, + "tokens_seen": 40501248 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004988465396188566, + "loss": 3.4209, + "theoretical_loss": 5.479909767486326, + "tokens_seen": 40566784 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004988365095285858, + "loss": 3.403, + "theoretical_loss": 5.478534916550817, + "tokens_seen": 40632320 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498826479438315, + "loss": 3.4289, + "theoretical_loss": 5.477162901091306, + "tokens_seen": 40697856 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004988164493480441, + "loss": 3.5679, + "theoretical_loss": 5.475793710708061, + "tokens_seen": 40763392 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004988064192577733, + "loss": 3.3113, + "theoretical_loss": 5.474427335056138, + "tokens_seen": 40828928 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004987963891675025, + "loss": 3.3464, + "theoretical_loss": 5.473063763844998, + "tokens_seen": 40894464 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004987863590772317, + "loss": 3.3822, + "theoretical_loss": 5.471702986838139, + "tokens_seen": 40960000 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004987763289869609, + "loss": 3.3161, + "theoretical_loss": 5.470344993852729, + "tokens_seen": 41025536 + }, + { + "epoch": 0.01, + "learning_rate": 0.00049876629889669, + "loss": 3.2148, + "theoretical_loss": 5.46898977475923, + "tokens_seen": 41091072 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004987562688064192, + "loss": 3.5694, + "theoretical_loss": 5.467637319481042, + "tokens_seen": 41156608 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004987462387161484, + "loss": 3.1787, + "theoretical_loss": 5.466287617994144, + "tokens_seen": 41222144 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004987362086258777, + "loss": 3.3672, + "theoretical_loss": 5.464940660326731, + "tokens_seen": 41287680 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004987261785356068, + "loss": 3.3874, + "theoretical_loss": 5.463596436558861, + "tokens_seen": 41353216 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004987161484453361, + "loss": 3.2971, + "theoretical_loss": 5.462254936822105, + "tokens_seen": 41418752 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004987061183550651, + "loss": 3.4361, + "theoretical_loss": 5.4609161512992, + "tokens_seen": 41484288 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004986960882647944, + "loss": 3.3206, + "theoretical_loss": 5.4595800702237005, + "tokens_seen": 41549824 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004986860581745236, + "loss": 3.4073, + "theoretical_loss": 5.458246683879636, + "tokens_seen": 41615360 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004986760280842528, + "loss": 3.3945, + "theoretical_loss": 5.456915982601173, + "tokens_seen": 41680896 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498665997993982, + "loss": 3.3378, + "theoretical_loss": 5.455587956772277, + "tokens_seen": 41746432 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004986559679037111, + "loss": 3.3531, + "theoretical_loss": 5.454262596826377, + "tokens_seen": 41811968 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004986459378134403, + "loss": 3.2848, + "theoretical_loss": 5.452939893246034, + "tokens_seen": 41877504 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004986359077231695, + "loss": 3.4117, + "theoretical_loss": 5.451619836562617, + "tokens_seen": 41943040 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004986258776328987, + "loss": 3.3215, + "theoretical_loss": 5.450302417355973, + "tokens_seen": 42008576 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004986158475426279, + "loss": 3.2556, + "theoretical_loss": 5.448987626254097, + "tokens_seen": 42074112 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498605817452357, + "loss": 3.2977, + "theoretical_loss": 5.447675453932829, + "tokens_seen": 42139648 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004985957873620862, + "loss": 3.3764, + "theoretical_loss": 5.446365891115516, + "tokens_seen": 42205184 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004985857572718154, + "loss": 3.4593, + "theoretical_loss": 5.445058928572712, + "tokens_seen": 42270720 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004985757271815446, + "loss": 3.4414, + "theoretical_loss": 5.443754557121853, + "tokens_seen": 42336256 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004985656970912738, + "loss": 3.3623, + "theoretical_loss": 5.442452767626957, + "tokens_seen": 42401792 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004985556670010031, + "loss": 3.2816, + "theoretical_loss": 5.441153550998311, + "tokens_seen": 42467328 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004985456369107321, + "loss": 3.3794, + "theoretical_loss": 5.439856898192163, + "tokens_seen": 42532864 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 23643, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 3.6165151596069336, + "objective/train/theoretical_loss": 5.438562800210427, + "objective/train/tokens_used": 42598400, + "theoretical_loss": 5.438562800210427, + "tokens_seen": 42598400 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004985356068204614, + "loss": 3.43, + "theoretical_loss": 5.438562800210427, + "tokens_seen": 42598400 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004985255767301905, + "loss": 3.3635, + "theoretical_loss": 5.437271248100375, + "tokens_seen": 42663936 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004985155466399198, + "loss": 3.3825, + "theoretical_loss": 5.435982232954342, + "tokens_seen": 42729472 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498505516549649, + "loss": 3.3531, + "theoretical_loss": 5.434695745909433, + "tokens_seen": 42795008 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984954864593782, + "loss": 3.2909, + "theoretical_loss": 5.433411778147226, + "tokens_seen": 42860544 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984854563691073, + "loss": 3.3706, + "theoretical_loss": 5.432130320893483, + "tokens_seen": 42926080 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984754262788365, + "loss": 3.2709, + "theoretical_loss": 5.430851365417862, + "tokens_seen": 42991616 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984653961885657, + "loss": 3.4062, + "theoretical_loss": 5.429574903033627, + "tokens_seen": 43057152 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984553660982949, + "loss": 3.1786, + "theoretical_loss": 5.428300925097373, + "tokens_seen": 43122688 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984453360080241, + "loss": 3.1838, + "theoretical_loss": 5.427029423008732, + "tokens_seen": 43188224 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984353059177532, + "loss": 3.3475, + "theoretical_loss": 5.425760388210108, + "tokens_seen": 43253760 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984252758274825, + "loss": 3.2032, + "theoretical_loss": 5.424493812186386, + "tokens_seen": 43319296 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984152457372116, + "loss": 3.2263, + "theoretical_loss": 5.423229686464667, + "tokens_seen": 43384832 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984052156469409, + "loss": 3.3253, + "theoretical_loss": 5.421968002613992, + "tokens_seen": 43450368 + }, + { + "epoch": 0.01, + "learning_rate": 0.00049839518555667, + "loss": 3.1964, + "theoretical_loss": 5.420708752245071, + "tokens_seen": 43515904 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004983851554663993, + "loss": 3.2344, + "theoretical_loss": 5.419451927010014, + "tokens_seen": 43581440 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004983751253761284, + "loss": 3.1453, + "theoretical_loss": 5.418197518602068, + "tokens_seen": 43646976 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004983650952858576, + "loss": 3.4177, + "theoretical_loss": 5.41694551875535, + "tokens_seen": 43712512 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004983550651955868, + "loss": 3.2278, + "theoretical_loss": 5.415695919244586, + "tokens_seen": 43778048 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498345035105316, + "loss": 3.4052, + "theoretical_loss": 5.41444871188485, + "tokens_seen": 43843584 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004983350050150452, + "loss": 3.2505, + "theoretical_loss": 5.413203888531308, + "tokens_seen": 43909120 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004983249749247743, + "loss": 3.3522, + "theoretical_loss": 5.411961441078963, + "tokens_seen": 43974656 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004983149448345035, + "loss": 3.2557, + "theoretical_loss": 5.410721361462395, + "tokens_seen": 44040192 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004983049147442327, + "loss": 3.0895, + "theoretical_loss": 5.409483641655516, + "tokens_seen": 44105728 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004982948846539619, + "loss": 3.1288, + "theoretical_loss": 5.408248273671317, + "tokens_seen": 44171264 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004982848545636911, + "loss": 3.2691, + "theoretical_loss": 5.407015249561624, + "tokens_seen": 44236800 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004982748244734202, + "loss": 3.1949, + "theoretical_loss": 5.405784561416844, + "tokens_seen": 44302336 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004982647943831494, + "loss": 3.3006, + "theoretical_loss": 5.404556201365727, + "tokens_seen": 44367872 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004982547642928786, + "loss": 3.3115, + "theoretical_loss": 5.403330161575123, + "tokens_seen": 44433408 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004982447342026079, + "loss": 3.372, + "theoretical_loss": 5.402106434249742, + "tokens_seen": 44498944 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498234704112337, + "loss": 3.14, + "theoretical_loss": 5.400885011631912, + "tokens_seen": 44564480 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004982246740220663, + "loss": 3.2636, + "theoretical_loss": 5.399665886001349, + "tokens_seen": 44630016 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004982146439317953, + "loss": 3.2508, + "theoretical_loss": 5.398449049674916, + "tokens_seen": 44695552 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004982046138415246, + "loss": 3.048, + "theoretical_loss": 5.397234495006392, + "tokens_seen": 44761088 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004981945837512538, + "loss": 3.146, + "theoretical_loss": 5.396022214386246, + "tokens_seen": 44826624 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498184553660983, + "loss": 3.2554, + "theoretical_loss": 5.394812200241393, + "tokens_seen": 44892160 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004981745235707122, + "loss": 3.2569, + "theoretical_loss": 5.393604445034988, + "tokens_seen": 44957696 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004981644934804413, + "loss": 3.1528, + "theoretical_loss": 5.392398941266183, + "tokens_seen": 45023232 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004981544633901705, + "loss": 3.1717, + "theoretical_loss": 5.391195681469908, + "tokens_seen": 45088768 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004981444332998997, + "loss": 3.1977, + "theoretical_loss": 5.389994658216651, + "tokens_seen": 45154304 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004981344032096289, + "loss": 3.167, + "theoretical_loss": 5.388795864112235, + "tokens_seen": 45219840 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004981243731193581, + "loss": 3.1497, + "theoretical_loss": 5.387599291797599, + "tokens_seen": 45285376 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004981143430290873, + "loss": 3.0773, + "theoretical_loss": 5.386404933948584, + "tokens_seen": 45350912 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004981043129388164, + "loss": 3.1171, + "theoretical_loss": 5.38521278327571, + "tokens_seen": 45416448 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004980942828485456, + "loss": 3.1576, + "theoretical_loss": 5.384022832523971, + "tokens_seen": 45481984 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004980842527582748, + "loss": 3.2231, + "theoretical_loss": 5.382835074472616, + "tokens_seen": 45547520 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498074222668004, + "loss": 3.1688, + "theoretical_loss": 5.381649501934939, + "tokens_seen": 45613056 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004980641925777333, + "loss": 3.3166, + "theoretical_loss": 5.3804661077580755, + "tokens_seen": 45678592 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004980541624874623, + "loss": 3.1347, + "theoretical_loss": 5.379284884822787, + "tokens_seen": 45744128 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004980441323971916, + "loss": 3.286, + "theoretical_loss": 5.37810582604326, + "tokens_seen": 45809664 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 24846, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 3.291611671447754, + "objective/train/theoretical_loss": 5.376928924366901, + "objective/train/tokens_used": 45875200, + "theoretical_loss": 5.376928924366901, + "tokens_seen": 45875200 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004980341023069207, + "loss": 3.2372, + "theoretical_loss": 5.376928924366901, + "tokens_seen": 45875200 + }, + { + "epoch": 0.01, + "learning_rate": 0.00049802407221665, + "loss": 3.1136, + "theoretical_loss": 5.37575417277413, + "tokens_seen": 45940736 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004980140421263792, + "loss": 3.1323, + "theoretical_loss": 5.374581564278184, + "tokens_seen": 46006272 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004980040120361084, + "loss": 3.1167, + "theoretical_loss": 5.373411091924918, + "tokens_seen": 46071808 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004979939819458375, + "loss": 3.1561, + "theoretical_loss": 5.372242748792598, + "tokens_seen": 46137344 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004979839518555667, + "loss": 2.9741, + "theoretical_loss": 5.371076527991711, + "tokens_seen": 46202880 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004979739217652959, + "loss": 3.0963, + "theoretical_loss": 5.36991242266477, + "tokens_seen": 46268416 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004979638916750251, + "loss": 3.1019, + "theoretical_loss": 5.36875042598612, + "tokens_seen": 46333952 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004979538615847543, + "loss": 3.2754, + "theoretical_loss": 5.3675905311617385, + "tokens_seen": 46399488 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004979438314944834, + "loss": 3.1641, + "theoretical_loss": 5.366432731429054, + "tokens_seen": 46465024 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004979338014042126, + "loss": 3.1938, + "theoretical_loss": 5.365277020056752, + "tokens_seen": 46530560 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004979237713139418, + "loss": 3.1788, + "theoretical_loss": 5.364123390344582, + "tokens_seen": 46596096 + }, + { + "epoch": 0.01, + "learning_rate": 0.000497913741223671, + "loss": 3.0928, + "theoretical_loss": 5.362971835623181, + "tokens_seen": 46661632 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004979037111334002, + "loss": 3.0689, + "theoretical_loss": 5.361822349253881, + "tokens_seen": 46727168 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004978936810431293, + "loss": 3.1448, + "theoretical_loss": 5.360674924628524, + "tokens_seen": 46792704 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004978836509528586, + "loss": 3.1729, + "theoretical_loss": 5.359529555169287, + "tokens_seen": 46858240 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004978736208625877, + "loss": 3.1299, + "theoretical_loss": 5.358386234328487, + "tokens_seen": 46923776 + }, + { + "epoch": 0.01, + "learning_rate": 0.000497863590772317, + "loss": 3.0612, + "theoretical_loss": 5.357244955588417, + "tokens_seen": 46989312 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004978535606820461, + "loss": 3.0475, + "theoretical_loss": 5.356105712461154, + "tokens_seen": 47054848 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004978435305917754, + "loss": 3.0856, + "theoretical_loss": 5.354968498488391, + "tokens_seen": 47120384 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004978335005015045, + "loss": 3.1141, + "theoretical_loss": 5.353833307241256, + "tokens_seen": 47185920 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004978234704112337, + "loss": 3.2587, + "theoretical_loss": 5.352700132320135, + "tokens_seen": 47251456 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004978134403209629, + "loss": 3.0138, + "theoretical_loss": 5.351568967354504, + "tokens_seen": 47316992 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004978034102306921, + "loss": 3.0683, + "theoretical_loss": 5.350439806002757, + "tokens_seen": 47382528 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004977933801404213, + "loss": 3.1979, + "theoretical_loss": 5.349312641952025, + "tokens_seen": 47448064 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004977833500501504, + "loss": 3.1036, + "theoretical_loss": 5.348187468918022, + "tokens_seen": 47513600 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004977733199598796, + "loss": 2.9684, + "theoretical_loss": 5.347064280644861, + "tokens_seen": 47579136 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004977632898696088, + "loss": 3.0675, + "theoretical_loss": 5.345943070904897, + "tokens_seen": 47644672 + }, + { + "epoch": 0.01, + "learning_rate": 0.000497753259779338, + "loss": 3.1384, + "theoretical_loss": 5.344823833498557, + "tokens_seen": 47710208 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004977432296890672, + "loss": 3.1572, + "theoretical_loss": 5.343706562254177, + "tokens_seen": 47775744 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004977331995987965, + "loss": 3.1656, + "theoretical_loss": 5.342591251027833, + "tokens_seen": 47841280 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004977231695085255, + "loss": 3.0261, + "theoretical_loss": 5.3414778937031855, + "tokens_seen": 47906816 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004977131394182548, + "loss": 3.0898, + "theoretical_loss": 5.340366484191316, + "tokens_seen": 47972352 + }, + { + "epoch": 0.01, + "learning_rate": 0.000497703109327984, + "loss": 3.0282, + "theoretical_loss": 5.339257016430563, + "tokens_seen": 48037888 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004976930792377132, + "loss": 3.0171, + "theoretical_loss": 5.338149484386367, + "tokens_seen": 48103424 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004976830491474424, + "loss": 3.155, + "theoretical_loss": 5.337043882051111, + "tokens_seen": 48168960 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004976730190571715, + "loss": 3.1338, + "theoretical_loss": 5.3359402034439665, + "tokens_seen": 48234496 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004976629889669007, + "loss": 3.0354, + "theoretical_loss": 5.334838442610732, + "tokens_seen": 48300032 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004976529588766299, + "loss": 2.9301, + "theoretical_loss": 5.333738593623686, + "tokens_seen": 48365568 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004976429287863591, + "loss": 3.178, + "theoretical_loss": 5.332640650581424, + "tokens_seen": 48431104 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004976328986960883, + "loss": 3.0741, + "theoretical_loss": 5.331544607608718, + "tokens_seen": 48496640 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004976228686058175, + "loss": 3.0649, + "theoretical_loss": 5.330450458856356, + "tokens_seen": 48562176 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004976128385155466, + "loss": 2.9932, + "theoretical_loss": 5.329358198500994, + "tokens_seen": 48627712 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004976028084252758, + "loss": 2.9332, + "theoretical_loss": 5.328267820745011, + "tokens_seen": 48693248 + }, + { + "epoch": 0.01, + "learning_rate": 0.000497592778335005, + "loss": 3.0347, + "theoretical_loss": 5.327179319816354, + "tokens_seen": 48758784 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004975827482447342, + "loss": 3.0627, + "theoretical_loss": 5.326092689968397, + "tokens_seen": 48824320 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004975727181544635, + "loss": 3.057, + "theoretical_loss": 5.3250079254797935, + "tokens_seen": 48889856 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004975626880641925, + "loss": 3.0463, + "theoretical_loss": 5.323925020654329, + "tokens_seen": 48955392 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004975526579739218, + "loss": 2.985, + "theoretical_loss": 5.322843969820781, + "tokens_seen": 49020928 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004975426278836509, + "loss": 2.9775, + "theoretical_loss": 5.321764767332772, + "tokens_seen": 49086464 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 26071, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 3.0261588096618652, + "objective/train/theoretical_loss": 5.320687407568633, + "objective/train/tokens_used": 49152000, + "theoretical_loss": 5.320687407568633, + "tokens_seen": 49152000 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004975325977933802, + "loss": 2.9649, + "theoretical_loss": 5.320687407568633, + "tokens_seen": 49152000 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004975225677031094, + "loss": 3.006, + "theoretical_loss": 5.3196118849312555, + "tokens_seen": 49217536 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004975125376128386, + "loss": 3.1159, + "theoretical_loss": 5.31853819384796, + "tokens_seen": 49283072 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004975025075225677, + "loss": 2.9316, + "theoretical_loss": 5.31746632877035, + "tokens_seen": 49348608 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004974924774322969, + "loss": 3.0383, + "theoretical_loss": 5.316396284174177, + "tokens_seen": 49414144 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004974824473420261, + "loss": 2.9737, + "theoretical_loss": 5.3153280545592025, + "tokens_seen": 49479680 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004974724172517553, + "loss": 3.1439, + "theoretical_loss": 5.314261634449066, + "tokens_seen": 49545216 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004974623871614845, + "loss": 3.1478, + "theoretical_loss": 5.313197018391143, + "tokens_seen": 49610752 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004974523570712136, + "loss": 3.0056, + "theoretical_loss": 5.312134200956416, + "tokens_seen": 49676288 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004974423269809428, + "loss": 2.9945, + "theoretical_loss": 5.3110731767393435, + "tokens_seen": 49741824 + }, + { + "epoch": 0.02, + "learning_rate": 0.000497432296890672, + "loss": 2.9396, + "theoretical_loss": 5.31001394035772, + "tokens_seen": 49807360 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004974222668004012, + "loss": 3.0714, + "theoretical_loss": 5.308956486452549, + "tokens_seen": 49872896 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004974122367101304, + "loss": 2.8905, + "theoretical_loss": 5.307900809687913, + "tokens_seen": 49938432 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004974022066198595, + "loss": 2.9574, + "theoretical_loss": 5.306846904750847, + "tokens_seen": 50003968 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004973921765295888, + "loss": 2.9504, + "theoretical_loss": 5.305794766351202, + "tokens_seen": 50069504 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004973821464393179, + "loss": 2.8367, + "theoretical_loss": 5.304744389221523, + "tokens_seen": 50135040 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004973721163490472, + "loss": 2.967, + "theoretical_loss": 5.303695768116921, + "tokens_seen": 50200576 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004973620862587763, + "loss": 2.9679, + "theoretical_loss": 5.302648897814946, + "tokens_seen": 50266112 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004973520561685056, + "loss": 2.9198, + "theoretical_loss": 5.301603773115465, + "tokens_seen": 50331648 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004973420260782347, + "loss": 2.9651, + "theoretical_loss": 5.300560388840532, + "tokens_seen": 50397184 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004973319959879639, + "loss": 2.9582, + "theoretical_loss": 5.299518739834273, + "tokens_seen": 50462720 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004973219658976931, + "loss": 2.9574, + "theoretical_loss": 5.298478820962751, + "tokens_seen": 50528256 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004973119358074223, + "loss": 2.9645, + "theoretical_loss": 5.297440627113859, + "tokens_seen": 50593792 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004973019057171515, + "loss": 2.9299, + "theoretical_loss": 5.296404153197189, + "tokens_seen": 50659328 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004972918756268806, + "loss": 2.925, + "theoretical_loss": 5.295369394143913, + "tokens_seen": 50724864 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004972818455366098, + "loss": 2.9526, + "theoretical_loss": 5.294336344906668, + "tokens_seen": 50790400 + }, + { + "epoch": 0.02, + "learning_rate": 0.000497271815446339, + "loss": 2.9956, + "theoretical_loss": 5.293305000459435, + "tokens_seen": 50855936 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004972617853560682, + "loss": 2.8395, + "theoretical_loss": 5.292275355797421, + "tokens_seen": 50921472 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004972517552657974, + "loss": 2.997, + "theoretical_loss": 5.291247405936941, + "tokens_seen": 50987008 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004972417251755266, + "loss": 2.92, + "theoretical_loss": 5.290221145915307, + "tokens_seen": 51052544 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004972316950852557, + "loss": 2.9459, + "theoretical_loss": 5.289196570790708, + "tokens_seen": 51118080 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004972216649949849, + "loss": 2.8445, + "theoretical_loss": 5.288173675642094, + "tokens_seen": 51183616 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004972116349047142, + "loss": 2.8193, + "theoretical_loss": 5.287152455569069, + "tokens_seen": 51249152 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004972016048144433, + "loss": 2.9131, + "theoretical_loss": 5.286132905691774, + "tokens_seen": 51314688 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004971915747241726, + "loss": 2.6976, + "theoretical_loss": 5.285115021150775, + "tokens_seen": 51380224 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004971815446339017, + "loss": 2.8598, + "theoretical_loss": 5.284098797106948, + "tokens_seen": 51445760 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004971715145436309, + "loss": 3.0003, + "theoretical_loss": 5.283084228741377, + "tokens_seen": 51511296 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004971614844533601, + "loss": 2.828, + "theoretical_loss": 5.282071311255235, + "tokens_seen": 51576832 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004971514543630893, + "loss": 2.9938, + "theoretical_loss": 5.2810600398696845, + "tokens_seen": 51642368 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004971414242728185, + "loss": 2.8898, + "theoretical_loss": 5.2800504098257575, + "tokens_seen": 51707904 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004971313941825477, + "loss": 2.8169, + "theoretical_loss": 5.279042416384257, + "tokens_seen": 51773440 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004971213640922768, + "loss": 2.8384, + "theoretical_loss": 5.278036054825645, + "tokens_seen": 51838976 + }, + { + "epoch": 0.02, + "learning_rate": 0.000497111334002006, + "loss": 2.8136, + "theoretical_loss": 5.277031320449941, + "tokens_seen": 51904512 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004971013039117352, + "loss": 2.9588, + "theoretical_loss": 5.27602820857661, + "tokens_seen": 51970048 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004970912738214644, + "loss": 2.8456, + "theoretical_loss": 5.275026714544463, + "tokens_seen": 52035584 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004970812437311936, + "loss": 2.7619, + "theoretical_loss": 5.27402683371155, + "tokens_seen": 52101120 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004970712136409227, + "loss": 2.8112, + "theoretical_loss": 5.273028561455056, + "tokens_seen": 52166656 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004970611835506519, + "loss": 2.8869, + "theoretical_loss": 5.272031893171203, + "tokens_seen": 52232192 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004970511534603811, + "loss": 2.9362, + "theoretical_loss": 5.27103682427514, + "tokens_seen": 52297728 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004970411233701103, + "loss": 2.9259, + "theoretical_loss": 5.270043350200851, + "tokens_seen": 52363264 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 27101, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.81410551071167, + "objective/train/theoretical_loss": 5.269051466401043, + "objective/train/tokens_used": 52428800, + "theoretical_loss": 5.269051466401043, + "tokens_seen": 52428800 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004970310932798396, + "loss": 2.7535, + "theoretical_loss": 5.269051466401043, + "tokens_seen": 52428800 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004970210631895686, + "loss": 2.9203, + "theoretical_loss": 5.268061168347058, + "tokens_seen": 52494336 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004970110330992979, + "loss": 2.7378, + "theoretical_loss": 5.267072451528764, + "tokens_seen": 52559872 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004970010030090271, + "loss": 2.6628, + "theoretical_loss": 5.266085311454461, + "tokens_seen": 52625408 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004969909729187563, + "loss": 2.8847, + "theoretical_loss": 5.265099743650779, + "tokens_seen": 52690944 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004969809428284855, + "loss": 2.8459, + "theoretical_loss": 5.264115743662588, + "tokens_seen": 52756480 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004969709127382147, + "loss": 2.9361, + "theoretical_loss": 5.26313330705289, + "tokens_seen": 52822016 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004969608826479438, + "loss": 2.779, + "theoretical_loss": 5.262152429402734, + "tokens_seen": 52887552 + }, + { + "epoch": 0.02, + "learning_rate": 0.000496950852557673, + "loss": 2.8922, + "theoretical_loss": 5.261173106311116, + "tokens_seen": 52953088 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004969408224674022, + "loss": 2.7433, + "theoretical_loss": 5.2601953333948765, + "tokens_seen": 53018624 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004969307923771314, + "loss": 2.8532, + "theoretical_loss": 5.25921910628862, + "tokens_seen": 53084160 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004969207622868606, + "loss": 2.8244, + "theoretical_loss": 5.258244420644611, + "tokens_seen": 53149696 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004969107321965897, + "loss": 2.9254, + "theoretical_loss": 5.257271272132685, + "tokens_seen": 53215232 + }, + { + "epoch": 0.02, + "learning_rate": 0.000496900702106319, + "loss": 2.8583, + "theoretical_loss": 5.256299656440154, + "tokens_seen": 53280768 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004968906720160481, + "loss": 2.8722, + "theoretical_loss": 5.255329569271717, + "tokens_seen": 53346304 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004968806419257774, + "loss": 2.7011, + "theoretical_loss": 5.254361006349367, + "tokens_seen": 53411840 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004968706118355065, + "loss": 2.7875, + "theoretical_loss": 5.2533939634122975, + "tokens_seen": 53477376 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004968605817452358, + "loss": 2.9007, + "theoretical_loss": 5.252428436216821, + "tokens_seen": 53542912 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004968505516549649, + "loss": 2.8122, + "theoretical_loss": 5.251464420536269, + "tokens_seen": 53608448 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004968405215646941, + "loss": 2.82, + "theoretical_loss": 5.250501912160911, + "tokens_seen": 53673984 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004968304914744233, + "loss": 2.8594, + "theoretical_loss": 5.24954090689786, + "tokens_seen": 53739520 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004968204613841525, + "loss": 2.6814, + "theoretical_loss": 5.248581400570989, + "tokens_seen": 53805056 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004968104312938817, + "loss": 2.7851, + "theoretical_loss": 5.247623389020841, + "tokens_seen": 53870592 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004968004012036108, + "loss": 2.7072, + "theoretical_loss": 5.246666868104547, + "tokens_seen": 53936128 + }, + { + "epoch": 0.02, + "learning_rate": 0.00049679037111334, + "loss": 2.7256, + "theoretical_loss": 5.245711833695731, + "tokens_seen": 54001664 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004967803410230692, + "loss": 2.7786, + "theoretical_loss": 5.244758281684433, + "tokens_seen": 54067200 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004967703109327984, + "loss": 2.7225, + "theoretical_loss": 5.24380620797702, + "tokens_seen": 54132736 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004967602808425276, + "loss": 2.7846, + "theoretical_loss": 5.242855608496102, + "tokens_seen": 54198272 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004967502507522568, + "loss": 2.7416, + "theoretical_loss": 5.2419064791804475, + "tokens_seen": 54263808 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004967402206619859, + "loss": 2.787, + "theoretical_loss": 5.240958815984901, + "tokens_seen": 54329344 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004967301905717151, + "loss": 2.7364, + "theoretical_loss": 5.2400126148803, + "tokens_seen": 54394880 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004967201604814444, + "loss": 2.6862, + "theoretical_loss": 5.23906787185339, + "tokens_seen": 54460416 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004967101303911735, + "loss": 2.7457, + "theoretical_loss": 5.238124582906748, + "tokens_seen": 54525952 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004967001003009028, + "loss": 3.0001, + "theoretical_loss": 5.237182744058696, + "tokens_seen": 54591488 + }, + { + "epoch": 0.02, + "learning_rate": 0.000496690070210632, + "loss": 2.5721, + "theoretical_loss": 5.236242351343224, + "tokens_seen": 54657024 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004966800401203611, + "loss": 2.6815, + "theoretical_loss": 5.235303400809904, + "tokens_seen": 54722560 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004966700100300903, + "loss": 2.7991, + "theoretical_loss": 5.2343658885238185, + "tokens_seen": 54788096 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004966599799398195, + "loss": 2.6999, + "theoretical_loss": 5.2334298105654735, + "tokens_seen": 54853632 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004966499498495487, + "loss": 2.7422, + "theoretical_loss": 5.232495163030723, + "tokens_seen": 54919168 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004966399197592779, + "loss": 2.7164, + "theoretical_loss": 5.231561942030694, + "tokens_seen": 54984704 + }, + { + "epoch": 0.02, + "learning_rate": 0.000496629889669007, + "loss": 2.7593, + "theoretical_loss": 5.230630143691698, + "tokens_seen": 55050240 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004966198595787362, + "loss": 2.7121, + "theoretical_loss": 5.22969976415517, + "tokens_seen": 55115776 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004966098294884654, + "loss": 2.7679, + "theoretical_loss": 5.228770799577576, + "tokens_seen": 55181312 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004965997993981946, + "loss": 2.6129, + "theoretical_loss": 5.227843246130343, + "tokens_seen": 55246848 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004965897693079238, + "loss": 2.6336, + "theoretical_loss": 5.2269170999997865, + "tokens_seen": 55312384 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004965797392176529, + "loss": 2.6945, + "theoretical_loss": 5.225992357387031, + "tokens_seen": 55377920 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004965697091273821, + "loss": 2.7467, + "theoretical_loss": 5.225069014507933, + "tokens_seen": 55443456 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004965596790371113, + "loss": 2.7161, + "theoretical_loss": 5.224147067593012, + "tokens_seen": 55508992 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004965496489468405, + "loss": 2.7509, + "theoretical_loss": 5.223226512887375, + "tokens_seen": 55574528 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004965396188565698, + "loss": 2.6674, + "theoretical_loss": 5.222307346650636, + "tokens_seen": 55640064 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 28266, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.421344757080078, + "objective/train/theoretical_loss": 5.221389565156855, + "objective/train/tokens_used": 55705600, + "theoretical_loss": 5.221389565156855, + "tokens_seen": 55705600 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004965295887662988, + "loss": 2.6125, + "theoretical_loss": 5.221389565156855, + "tokens_seen": 55705600 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004965195586760281, + "loss": 2.7385, + "theoretical_loss": 5.2204731646944555, + "tokens_seen": 55771136 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004965095285857573, + "loss": 2.6009, + "theoretical_loss": 5.219558141566156, + "tokens_seen": 55836672 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004964994984954865, + "loss": 2.7117, + "theoretical_loss": 5.218644492088899, + "tokens_seen": 55902208 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004964894684052157, + "loss": 2.6692, + "theoretical_loss": 5.217732212593778, + "tokens_seen": 55967744 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004964794383149449, + "loss": 2.617, + "theoretical_loss": 5.216821299425968, + "tokens_seen": 56033280 + }, + { + "epoch": 0.02, + "learning_rate": 0.000496469408224674, + "loss": 2.7736, + "theoretical_loss": 5.215911748944652, + "tokens_seen": 56098816 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004964593781344032, + "loss": 2.747, + "theoretical_loss": 5.215003557522959, + "tokens_seen": 56164352 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004964493480441324, + "loss": 2.634, + "theoretical_loss": 5.21409672154788, + "tokens_seen": 56229888 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004964393179538616, + "loss": 2.7464, + "theoretical_loss": 5.213191237420215, + "tokens_seen": 56295424 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004964292878635908, + "loss": 2.7409, + "theoretical_loss": 5.212287101554496, + "tokens_seen": 56360960 + }, + { + "epoch": 0.02, + "learning_rate": 0.00049641925777332, + "loss": 2.6864, + "theoretical_loss": 5.211384310378916, + "tokens_seen": 56426496 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004964092276830491, + "loss": 2.6999, + "theoretical_loss": 5.210482860335269, + "tokens_seen": 56492032 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004963991975927783, + "loss": 2.7319, + "theoretical_loss": 5.2095827478788745, + "tokens_seen": 56557568 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004963891675025075, + "loss": 2.5708, + "theoretical_loss": 5.208683969478518, + "tokens_seen": 56623104 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004963791374122367, + "loss": 2.6001, + "theoretical_loss": 5.2077865216163755, + "tokens_seen": 56688640 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004963691073219659, + "loss": 2.6403, + "theoretical_loss": 5.206890400787959, + "tokens_seen": 56754176 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004963590772316951, + "loss": 2.58, + "theoretical_loss": 5.205995603502039, + "tokens_seen": 56819712 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004963490471414242, + "loss": 2.6272, + "theoretical_loss": 5.205102126280586, + "tokens_seen": 56885248 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004963390170511535, + "loss": 2.7034, + "theoretical_loss": 5.204209965658704, + "tokens_seen": 56950784 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004963289869608827, + "loss": 2.6516, + "theoretical_loss": 5.203319118184565, + "tokens_seen": 57016320 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004963189568706119, + "loss": 2.7487, + "theoretical_loss": 5.202429580419343, + "tokens_seen": 57081856 + }, + { + "epoch": 0.02, + "learning_rate": 0.000496308926780341, + "loss": 2.5931, + "theoretical_loss": 5.201541348937157, + "tokens_seen": 57147392 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004962988966900702, + "loss": 2.6747, + "theoretical_loss": 5.200654420324998, + "tokens_seen": 57212928 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004962888665997994, + "loss": 2.5713, + "theoretical_loss": 5.199768791182677, + "tokens_seen": 57278464 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004962788365095286, + "loss": 2.589, + "theoretical_loss": 5.198884458122748, + "tokens_seen": 57344000 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004962688064192578, + "loss": 2.5456, + "theoretical_loss": 5.198001417770461, + "tokens_seen": 57409536 + }, + { + "epoch": 0.02, + "learning_rate": 0.000496258776328987, + "loss": 2.6555, + "theoretical_loss": 5.1971196667636885, + "tokens_seen": 57475072 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004962487462387161, + "loss": 2.7297, + "theoretical_loss": 5.196239201752871, + "tokens_seen": 57540608 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004962387161484453, + "loss": 2.5553, + "theoretical_loss": 5.19536001940095, + "tokens_seen": 57606144 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004962286860581746, + "loss": 2.6298, + "theoretical_loss": 5.1944821163833135, + "tokens_seen": 57671680 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004962186559679037, + "loss": 2.5579, + "theoretical_loss": 5.19360548938773, + "tokens_seen": 57737216 + }, + { + "epoch": 0.02, + "learning_rate": 0.000496208625877633, + "loss": 2.6071, + "theoretical_loss": 5.192730135114292, + "tokens_seen": 57802752 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004961985957873621, + "loss": 2.5932, + "theoretical_loss": 5.191856050275353, + "tokens_seen": 57868288 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004961885656970913, + "loss": 2.6078, + "theoretical_loss": 5.190983231595473, + "tokens_seen": 57933824 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004961785356068205, + "loss": 2.5666, + "theoretical_loss": 5.1901116758113535, + "tokens_seen": 57999360 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004961685055165497, + "loss": 2.5328, + "theoretical_loss": 5.189241379671782, + "tokens_seen": 58064896 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004961584754262789, + "loss": 2.7239, + "theoretical_loss": 5.188372339937576, + "tokens_seen": 58130432 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004961484453360081, + "loss": 2.5654, + "theoretical_loss": 5.187504553381521, + "tokens_seen": 58195968 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004961384152457372, + "loss": 2.7473, + "theoretical_loss": 5.1866380167883115, + "tokens_seen": 58261504 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004961283851554664, + "loss": 2.7238, + "theoretical_loss": 5.1857727269545, + "tokens_seen": 58327040 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004961183550651956, + "loss": 2.5483, + "theoretical_loss": 5.184908680688437, + "tokens_seen": 58392576 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004961083249749248, + "loss": 2.5616, + "theoretical_loss": 5.18404587481021, + "tokens_seen": 58458112 + }, + { + "epoch": 0.02, + "learning_rate": 0.000496098294884654, + "loss": 2.5237, + "theoretical_loss": 5.183184306151595, + "tokens_seen": 58523648 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004960882647943831, + "loss": 2.5267, + "theoretical_loss": 5.182323971555995, + "tokens_seen": 58589184 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004960782347041123, + "loss": 2.6205, + "theoretical_loss": 5.181464867878384, + "tokens_seen": 58654720 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004960682046138415, + "loss": 2.7083, + "theoretical_loss": 5.1806069919852575, + "tokens_seen": 58720256 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004960581745235707, + "loss": 2.6558, + "theoretical_loss": 5.17975034075457, + "tokens_seen": 58785792 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004960481444333, + "loss": 2.4854, + "theoretical_loss": 5.1788949110756874, + "tokens_seen": 58851328 + }, + { + "epoch": 0.02, + "learning_rate": 0.000496038114343029, + "loss": 2.573, + "theoretical_loss": 5.178040699849327, + "tokens_seen": 58916864 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 29491, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.728625774383545, + "objective/train/theoretical_loss": 5.1771877039875065, + "objective/train/tokens_used": 58982400, + "theoretical_loss": 5.1771877039875065, + "tokens_seen": 58982400 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004960280842527583, + "loss": 2.5808, + "theoretical_loss": 5.1771877039875065, + "tokens_seen": 58982400 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004960180541624875, + "loss": 2.5483, + "theoretical_loss": 5.176335920413489, + "tokens_seen": 59047936 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004960080240722167, + "loss": 2.7546, + "theoretical_loss": 5.175485346061732, + "tokens_seen": 59113472 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004959979939819459, + "loss": 2.6282, + "theoretical_loss": 5.174635977877831, + "tokens_seen": 59179008 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004959879638916751, + "loss": 2.4998, + "theoretical_loss": 5.173787812818468, + "tokens_seen": 59244544 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004959779338014042, + "loss": 2.3818, + "theoretical_loss": 5.172940847851359, + "tokens_seen": 59310080 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004959679037111334, + "loss": 2.5603, + "theoretical_loss": 5.172095079955204, + "tokens_seen": 59375616 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004959578736208626, + "loss": 2.5791, + "theoretical_loss": 5.171250506119634, + "tokens_seen": 59441152 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004959478435305918, + "loss": 2.5457, + "theoretical_loss": 5.170407123345155, + "tokens_seen": 59506688 + }, + { + "epoch": 0.02, + "learning_rate": 0.000495937813440321, + "loss": 2.573, + "theoretical_loss": 5.169564928643103, + "tokens_seen": 59572224 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004959277833500501, + "loss": 2.5242, + "theoretical_loss": 5.168723919035591, + "tokens_seen": 59637760 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004959177532597793, + "loss": 2.5713, + "theoretical_loss": 5.167884091555456, + "tokens_seen": 59703296 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004959077231695085, + "loss": 2.5304, + "theoretical_loss": 5.167045443246213, + "tokens_seen": 59768832 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004958976930792377, + "loss": 2.5679, + "theoretical_loss": 5.166207971162001, + "tokens_seen": 59834368 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004958876629889669, + "loss": 2.5295, + "theoretical_loss": 5.1653716723675345, + "tokens_seen": 59899904 + }, + { + "epoch": 0.02, + "learning_rate": 0.000495877632898696, + "loss": 2.5758, + "theoretical_loss": 5.164536543938056, + "tokens_seen": 59965440 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004958676028084253, + "loss": 2.5171, + "theoretical_loss": 5.163702582959283, + "tokens_seen": 60030976 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004958575727181544, + "loss": 2.5181, + "theoretical_loss": 5.162869786527363, + "tokens_seen": 60096512 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004958475426278837, + "loss": 2.5304, + "theoretical_loss": 5.1620381517488205, + "tokens_seen": 60162048 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004958375125376129, + "loss": 2.6151, + "theoretical_loss": 5.161207675740515, + "tokens_seen": 60227584 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004958274824473421, + "loss": 2.4964, + "theoretical_loss": 5.160378355629584, + "tokens_seen": 60293120 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004958174523570712, + "loss": 2.4244, + "theoretical_loss": 5.159550188553405, + "tokens_seen": 60358656 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004958074222668004, + "loss": 2.5702, + "theoretical_loss": 5.15872317165954, + "tokens_seen": 60424192 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004957973921765296, + "loss": 2.4117, + "theoretical_loss": 5.1578973021056935, + "tokens_seen": 60489728 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004957873620862588, + "loss": 2.4653, + "theoretical_loss": 5.157072577059662, + "tokens_seen": 60555264 + }, + { + "epoch": 0.02, + "learning_rate": 0.000495777331995988, + "loss": 2.5175, + "theoretical_loss": 5.15624899369929, + "tokens_seen": 60620800 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004957673019057172, + "loss": 2.5641, + "theoretical_loss": 5.155426549212422, + "tokens_seen": 60686336 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004957572718154463, + "loss": 2.2979, + "theoretical_loss": 5.154605240796855, + "tokens_seen": 60751872 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004957472417251755, + "loss": 2.4524, + "theoretical_loss": 5.153785065660296, + "tokens_seen": 60817408 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004957372116349047, + "loss": 2.5274, + "theoretical_loss": 5.152966021020315, + "tokens_seen": 60882944 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004957271815446339, + "loss": 2.5021, + "theoretical_loss": 5.152148104104297, + "tokens_seen": 60948480 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004957171514543631, + "loss": 2.3868, + "theoretical_loss": 5.1513313121494, + "tokens_seen": 61014016 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004957071213640923, + "loss": 2.4331, + "theoretical_loss": 5.15051564240251, + "tokens_seen": 61079552 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004956970912738214, + "loss": 2.4219, + "theoretical_loss": 5.149701092120194, + "tokens_seen": 61145088 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004956870611835507, + "loss": 2.4445, + "theoretical_loss": 5.148887658568658, + "tokens_seen": 61210624 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004956770310932798, + "loss": 2.509, + "theoretical_loss": 5.148075339023699, + "tokens_seen": 61276160 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004956670010030091, + "loss": 2.4302, + "theoretical_loss": 5.147264130770671, + "tokens_seen": 61341696 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004956569709127383, + "loss": 2.4102, + "theoretical_loss": 5.146454031104426, + "tokens_seen": 61407232 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004956469408224674, + "loss": 2.5423, + "theoretical_loss": 5.1456450373292855, + "tokens_seen": 61472768 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004956369107321966, + "loss": 2.5201, + "theoretical_loss": 5.144837146758987, + "tokens_seen": 61538304 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004956268806419258, + "loss": 2.5425, + "theoretical_loss": 5.144030356716646, + "tokens_seen": 61603840 + }, + { + "epoch": 0.02, + "learning_rate": 0.000495616850551655, + "loss": 2.6262, + "theoretical_loss": 5.1432246645347135, + "tokens_seen": 61669376 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004956068204613842, + "loss": 2.4367, + "theoretical_loss": 5.14242006755493, + "tokens_seen": 61734912 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004955967903711133, + "loss": 2.5528, + "theoretical_loss": 5.141616563128288, + "tokens_seen": 61800448 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004955867602808425, + "loss": 2.4114, + "theoretical_loss": 5.140814148614985, + "tokens_seen": 61865984 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004955767301905717, + "loss": 2.5797, + "theoretical_loss": 5.1400128213843885, + "tokens_seen": 61931520 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004955667001003009, + "loss": 2.4277, + "theoretical_loss": 5.139212578814986, + "tokens_seen": 61997056 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004955566700100301, + "loss": 2.6051, + "theoretical_loss": 5.13841341829435, + "tokens_seen": 62062592 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004955466399197592, + "loss": 2.5286, + "theoretical_loss": 5.137615337219098, + "tokens_seen": 62128128 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004955366098294884, + "loss": 2.4757, + "theoretical_loss": 5.136818332994844, + "tokens_seen": 62193664 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 31037, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.411979913711548, + "objective/train/theoretical_loss": 5.136022403036167, + "objective/train/tokens_used": 62259200, + "theoretical_loss": 5.136022403036167, + "tokens_seen": 62259200 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004955265797392177, + "loss": 2.5195, + "theoretical_loss": 5.136022403036167, + "tokens_seen": 62259200 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004955165496489468, + "loss": 2.4498, + "theoretical_loss": 5.135227544766565, + "tokens_seen": 62324736 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004955065195586761, + "loss": 2.3778, + "theoretical_loss": 5.134433755618414, + "tokens_seen": 62390272 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004954964894684052, + "loss": 2.4696, + "theoretical_loss": 5.1336410330329345, + "tokens_seen": 62455808 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004954864593781344, + "loss": 2.5592, + "theoretical_loss": 5.132849374460145, + "tokens_seen": 62521344 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004954764292878636, + "loss": 2.3491, + "theoretical_loss": 5.132058777358827, + "tokens_seen": 62586880 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004954663991975928, + "loss": 2.543, + "theoretical_loss": 5.1312692391964845, + "tokens_seen": 62652416 + }, + { + "epoch": 0.02, + "learning_rate": 0.000495456369107322, + "loss": 2.4567, + "theoretical_loss": 5.1304807574493, + "tokens_seen": 62717952 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004954463390170512, + "loss": 2.4295, + "theoretical_loss": 5.129693329602108, + "tokens_seen": 62783488 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004954363089267803, + "loss": 2.4474, + "theoretical_loss": 5.128906953148342, + "tokens_seen": 62849024 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004954262788365095, + "loss": 2.5911, + "theoretical_loss": 5.128121625590007, + "tokens_seen": 62914560 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004954162487462387, + "loss": 2.4996, + "theoretical_loss": 5.127337344437633, + "tokens_seen": 62980096 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004954062186559679, + "loss": 2.5509, + "theoretical_loss": 5.126554107210245, + "tokens_seen": 63045632 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004953961885656971, + "loss": 2.3786, + "theoretical_loss": 5.125771911435321, + "tokens_seen": 63111168 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004953861584754263, + "loss": 2.4014, + "theoretical_loss": 5.124990754648754, + "tokens_seen": 63176704 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004953761283851555, + "loss": 2.464, + "theoretical_loss": 5.124210634394814, + "tokens_seen": 63242240 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004953660982948846, + "loss": 2.4067, + "theoretical_loss": 5.1234315482261135, + "tokens_seen": 63307776 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004953560682046139, + "loss": 2.4224, + "theoretical_loss": 5.122653493703571, + "tokens_seen": 63373312 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004953460381143431, + "loss": 2.4154, + "theoretical_loss": 5.121876468396373, + "tokens_seen": 63438848 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004953360080240723, + "loss": 2.5665, + "theoretical_loss": 5.121100469881933, + "tokens_seen": 63504384 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004953259779338014, + "loss": 2.3294, + "theoretical_loss": 5.120325495745863, + "tokens_seen": 63569920 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004953159478435306, + "loss": 2.5157, + "theoretical_loss": 5.119551543581933, + "tokens_seen": 63635456 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004953059177532598, + "loss": 2.3962, + "theoretical_loss": 5.118778610992038, + "tokens_seen": 63700992 + }, + { + "epoch": 0.02, + "learning_rate": 0.000495295887662989, + "loss": 2.3481, + "theoretical_loss": 5.118006695586154, + "tokens_seen": 63766528 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004952858575727182, + "loss": 2.4302, + "theoretical_loss": 5.1172357949823155, + "tokens_seen": 63832064 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004952758274824474, + "loss": 2.399, + "theoretical_loss": 5.116465906806567, + "tokens_seen": 63897600 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004952657973921765, + "loss": 2.4182, + "theoretical_loss": 5.11569702869294, + "tokens_seen": 63963136 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004952557673019057, + "loss": 2.3922, + "theoretical_loss": 5.114929158283407, + "tokens_seen": 64028672 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004952457372116349, + "loss": 2.4446, + "theoretical_loss": 5.114162293227853, + "tokens_seen": 64094208 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004952357071213641, + "loss": 2.403, + "theoretical_loss": 5.113396431184043, + "tokens_seen": 64159744 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004952256770310933, + "loss": 2.4638, + "theoretical_loss": 5.112631569817578, + "tokens_seen": 64225280 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004952156469408225, + "loss": 2.4406, + "theoretical_loss": 5.111867706801872, + "tokens_seen": 64290816 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004952056168505516, + "loss": 2.5326, + "theoretical_loss": 5.111104839818113, + "tokens_seen": 64356352 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004951955867602809, + "loss": 2.503, + "theoretical_loss": 5.1103429665552245, + "tokens_seen": 64421888 + }, + { + "epoch": 0.02, + "learning_rate": 0.00049518555667001, + "loss": 2.3959, + "theoretical_loss": 5.109582084709841, + "tokens_seen": 64487424 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004951755265797393, + "loss": 2.4168, + "theoretical_loss": 5.108822191986269, + "tokens_seen": 64552960 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004951654964894685, + "loss": 2.4273, + "theoretical_loss": 5.108063286096453, + "tokens_seen": 64618496 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004951554663991976, + "loss": 2.3698, + "theoretical_loss": 5.107305364759943, + "tokens_seen": 64684032 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004951454363089268, + "loss": 2.3073, + "theoretical_loss": 5.106548425703869, + "tokens_seen": 64749568 + }, + { + "epoch": 0.02, + "learning_rate": 0.000495135406218656, + "loss": 2.3978, + "theoretical_loss": 5.1057924666628915, + "tokens_seen": 64815104 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004951253761283852, + "loss": 2.3449, + "theoretical_loss": 5.105037485379189, + "tokens_seen": 64880640 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004951153460381144, + "loss": 2.2846, + "theoretical_loss": 5.104283479602407, + "tokens_seen": 64946176 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004951053159478435, + "loss": 2.3944, + "theoretical_loss": 5.103530447089642, + "tokens_seen": 65011712 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004950952858575727, + "loss": 2.3636, + "theoretical_loss": 5.102778385605397, + "tokens_seen": 65077248 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004950852557673019, + "loss": 2.3902, + "theoretical_loss": 5.102027292921557, + "tokens_seen": 65142784 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004950752256770311, + "loss": 2.3745, + "theoretical_loss": 5.101277166817351, + "tokens_seen": 65208320 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004950651955867603, + "loss": 2.5413, + "theoretical_loss": 5.100528005079328, + "tokens_seen": 65273856 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004950551654964894, + "loss": 2.435, + "theoretical_loss": 5.099779805501318, + "tokens_seen": 65339392 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004950451354062186, + "loss": 2.5376, + "theoretical_loss": 5.099032565884409, + "tokens_seen": 65404928 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004950351053159479, + "loss": 2.3042, + "theoretical_loss": 5.098286284036908, + "tokens_seen": 65470464 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 32172, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.53538179397583, + "objective/train/theoretical_loss": 5.097540957774312, + "objective/train/tokens_used": 65536000, + "theoretical_loss": 5.097540957774312, + "tokens_seen": 65536000 + }, + { + "epoch": 0.02, + "learning_rate": 0.000495025075225677, + "loss": 2.3831, + "theoretical_loss": 5.097540957774312, + "tokens_seen": 65536000 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004950150451354063, + "loss": 2.234, + "theoretical_loss": 5.096796584919282, + "tokens_seen": 65601536 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004950050150451354, + "loss": 2.3207, + "theoretical_loss": 5.096053163301608, + "tokens_seen": 65667072 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004949949849548646, + "loss": 2.3293, + "theoretical_loss": 5.095310690758177, + "tokens_seen": 65732608 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004949849548645938, + "loss": 2.3265, + "theoretical_loss": 5.094569165132949, + "tokens_seen": 65798144 + }, + { + "epoch": 0.02, + "learning_rate": 0.000494974924774323, + "loss": 2.2828, + "theoretical_loss": 5.09382858427692, + "tokens_seen": 65863680 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004949648946840522, + "loss": 2.4238, + "theoretical_loss": 5.0930889460481, + "tokens_seen": 65929216 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004949548645937814, + "loss": 2.4197, + "theoretical_loss": 5.092350248311471, + "tokens_seen": 65994752 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004949448345035105, + "loss": 2.3136, + "theoretical_loss": 5.091612488938971, + "tokens_seen": 66060288 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004949348044132397, + "loss": 2.451, + "theoretical_loss": 5.090875665809454, + "tokens_seen": 66125824 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004949247743229689, + "loss": 2.3827, + "theoretical_loss": 5.09013977680867, + "tokens_seen": 66191360 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004949147442326981, + "loss": 2.4782, + "theoretical_loss": 5.089404819829225, + "tokens_seen": 66256896 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004949047141424273, + "loss": 2.4297, + "theoretical_loss": 5.088670792770563, + "tokens_seen": 66322432 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004948946840521565, + "loss": 2.4005, + "theoretical_loss": 5.087937693538927, + "tokens_seen": 66387968 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004948846539618856, + "loss": 2.4637, + "theoretical_loss": 5.0872055200473385, + "tokens_seen": 66453504 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004948746238716148, + "loss": 2.484, + "theoretical_loss": 5.0864742702155645, + "tokens_seen": 66519040 + }, + { + "epoch": 0.02, + "learning_rate": 0.000494864593781344, + "loss": 2.2647, + "theoretical_loss": 5.0857439419700885, + "tokens_seen": 66584576 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004948545636910733, + "loss": 2.3034, + "theoretical_loss": 5.085014533244088, + "tokens_seen": 66650112 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004948445336008024, + "loss": 2.5841, + "theoretical_loss": 5.084286041977396, + "tokens_seen": 66715648 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004948345035105316, + "loss": 2.4104, + "theoretical_loss": 5.083558466116487, + "tokens_seen": 66781184 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004948244734202607, + "loss": 2.2718, + "theoretical_loss": 5.082831803614433, + "tokens_seen": 66846720 + }, + { + "epoch": 0.02, + "learning_rate": 0.00049481444332999, + "loss": 2.341, + "theoretical_loss": 5.082106052430891, + "tokens_seen": 66912256 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004948044132397192, + "loss": 2.3932, + "theoretical_loss": 5.081381210532063, + "tokens_seen": 66977792 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004947943831494484, + "loss": 2.4903, + "theoretical_loss": 5.08065727589068, + "tokens_seen": 67043328 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004947843530591776, + "loss": 2.4876, + "theoretical_loss": 5.079934246485963, + "tokens_seen": 67108864 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004947743229689067, + "loss": 2.255, + "theoretical_loss": 5.079212120303606, + "tokens_seen": 67174400 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004947642928786359, + "loss": 2.4504, + "theoretical_loss": 5.078490895335744, + "tokens_seen": 67239936 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004947542627883651, + "loss": 2.2195, + "theoretical_loss": 5.077770569580924, + "tokens_seen": 67305472 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004947442326980943, + "loss": 2.3913, + "theoretical_loss": 5.077051141044086, + "tokens_seen": 67371008 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004947342026078235, + "loss": 2.3738, + "theoretical_loss": 5.076332607736527, + "tokens_seen": 67436544 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004947241725175527, + "loss": 2.4094, + "theoretical_loss": 5.0756149676758815, + "tokens_seen": 67502080 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004947141424272818, + "loss": 2.3699, + "theoretical_loss": 5.0748982188860925, + "tokens_seen": 67567616 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004947041123370111, + "loss": 2.3402, + "theoretical_loss": 5.074182359397387, + "tokens_seen": 67633152 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004946940822467402, + "loss": 2.3899, + "theoretical_loss": 5.073467387246247, + "tokens_seen": 67698688 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004946840521564695, + "loss": 2.3823, + "theoretical_loss": 5.072753300475385, + "tokens_seen": 67764224 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004946740220661987, + "loss": 2.389, + "theoretical_loss": 5.072040097133718, + "tokens_seen": 67829760 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004946639919759278, + "loss": 2.4827, + "theoretical_loss": 5.0713277752763455, + "tokens_seen": 67895296 + }, + { + "epoch": 0.02, + "learning_rate": 0.000494653961885657, + "loss": 2.2772, + "theoretical_loss": 5.070616332964518, + "tokens_seen": 67960832 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004946439317953862, + "loss": 2.3601, + "theoretical_loss": 5.069905768265615, + "tokens_seen": 68026368 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004946339017051154, + "loss": 2.3281, + "theoretical_loss": 5.069196079253119, + "tokens_seen": 68091904 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004946238716148446, + "loss": 2.3664, + "theoretical_loss": 5.068487264006592, + "tokens_seen": 68157440 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004946138415245737, + "loss": 2.3252, + "theoretical_loss": 5.0677793206116455, + "tokens_seen": 68222976 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004946038114343029, + "loss": 2.4293, + "theoretical_loss": 5.067072247159923, + "tokens_seen": 68288512 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004945937813440321, + "loss": 2.3559, + "theoretical_loss": 5.06636604174907, + "tokens_seen": 68354048 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004945837512537613, + "loss": 2.3201, + "theoretical_loss": 5.065660702482711, + "tokens_seen": 68419584 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004945737211634905, + "loss": 2.4379, + "theoretical_loss": 5.064956227470425, + "tokens_seen": 68485120 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004945636910732196, + "loss": 2.3219, + "theoretical_loss": 5.06425261482772, + "tokens_seen": 68550656 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004945536609829488, + "loss": 2.3047, + "theoretical_loss": 5.063549862676011, + "tokens_seen": 68616192 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004945436308926781, + "loss": 2.2173, + "theoretical_loss": 5.062847969142591, + "tokens_seen": 68681728 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004945336008024072, + "loss": 2.3611, + "theoretical_loss": 5.062146932360616, + "tokens_seen": 68747264 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 33398, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.4574713706970215, + "objective/train/theoretical_loss": 5.061446750469072, + "objective/train/tokens_used": 68812800, + "theoretical_loss": 5.061446750469072, + "tokens_seen": 68812800 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004945235707121365, + "loss": 2.359, + "theoretical_loss": 5.061446750469072, + "tokens_seen": 68812800 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004945135406218656, + "loss": 2.4512, + "theoretical_loss": 5.060747421612758, + "tokens_seen": 68878336 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004945035105315948, + "loss": 2.2395, + "theoretical_loss": 5.060048943942254, + "tokens_seen": 68943872 + }, + { + "epoch": 0.02, + "learning_rate": 0.000494493480441324, + "loss": 2.2485, + "theoretical_loss": 5.059351315613908, + "tokens_seen": 69009408 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004944834503510532, + "loss": 2.3839, + "theoretical_loss": 5.058654534789804, + "tokens_seen": 69074944 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004944734202607824, + "loss": 2.3775, + "theoretical_loss": 5.057958599637746, + "tokens_seen": 69140480 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004944633901705116, + "loss": 2.2785, + "theoretical_loss": 5.057263508331225, + "tokens_seen": 69206016 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004944533600802407, + "loss": 2.3098, + "theoretical_loss": 5.0565692590494065, + "tokens_seen": 69271552 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004944433299899699, + "loss": 2.2444, + "theoretical_loss": 5.055875849977101, + "tokens_seen": 69337088 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004944332998996991, + "loss": 2.3373, + "theoretical_loss": 5.055183279304744, + "tokens_seen": 69402624 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004944232698094283, + "loss": 2.2511, + "theoretical_loss": 5.0544915452283705, + "tokens_seen": 69468160 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004944132397191575, + "loss": 2.3655, + "theoretical_loss": 5.053800645949597, + "tokens_seen": 69533696 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004944032096288867, + "loss": 2.3638, + "theoretical_loss": 5.053110579675593, + "tokens_seen": 69599232 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004943931795386158, + "loss": 2.341, + "theoretical_loss": 5.052421344619064, + "tokens_seen": 69664768 + }, + { + "epoch": 0.02, + "learning_rate": 0.000494383149448345, + "loss": 2.3183, + "theoretical_loss": 5.051732938998227, + "tokens_seen": 69730304 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004943731193580742, + "loss": 2.3337, + "theoretical_loss": 5.051045361036787, + "tokens_seen": 69795840 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004943630892678035, + "loss": 2.262, + "theoretical_loss": 5.0503586089639185, + "tokens_seen": 69861376 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004943530591775326, + "loss": 2.3009, + "theoretical_loss": 5.04967268101424, + "tokens_seen": 69926912 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004943430290872618, + "loss": 2.394, + "theoretical_loss": 5.048987575427793, + "tokens_seen": 69992448 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004943329989969909, + "loss": 2.3826, + "theoretical_loss": 5.048303290450023, + "tokens_seen": 70057984 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004943229689067202, + "loss": 2.4062, + "theoretical_loss": 5.047619824331753, + "tokens_seen": 70123520 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004943129388164494, + "loss": 2.2456, + "theoretical_loss": 5.046937175329167, + "tokens_seen": 70189056 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004943029087261786, + "loss": 2.4284, + "theoretical_loss": 5.046255341703786, + "tokens_seen": 70254592 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004942928786359078, + "loss": 2.2161, + "theoretical_loss": 5.045574321722445, + "tokens_seen": 70320128 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004942828485456369, + "loss": 2.3796, + "theoretical_loss": 5.0448941136572785, + "tokens_seen": 70385664 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004942728184553661, + "loss": 2.2709, + "theoretical_loss": 5.044214715785689, + "tokens_seen": 70451200 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004942627883650953, + "loss": 2.3823, + "theoretical_loss": 5.043536126390334, + "tokens_seen": 70516736 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004942527582748245, + "loss": 2.338, + "theoretical_loss": 5.042858343759107, + "tokens_seen": 70582272 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004942427281845537, + "loss": 2.3282, + "theoretical_loss": 5.0421813661851065, + "tokens_seen": 70647808 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004942326980942828, + "loss": 2.2828, + "theoretical_loss": 5.041505191966625, + "tokens_seen": 70713344 + }, + { + "epoch": 0.02, + "learning_rate": 0.000494222668004012, + "loss": 2.2437, + "theoretical_loss": 5.040829819407123, + "tokens_seen": 70778880 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004942126379137412, + "loss": 2.339, + "theoretical_loss": 5.040155246815212, + "tokens_seen": 70844416 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004942026078234704, + "loss": 2.3891, + "theoretical_loss": 5.039481472504633, + "tokens_seen": 70909952 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004941925777331996, + "loss": 2.3234, + "theoretical_loss": 5.038808494794232, + "tokens_seen": 70975488 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004941825476429289, + "loss": 2.3063, + "theoretical_loss": 5.038136312007948, + "tokens_seen": 71041024 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004941725175526579, + "loss": 2.3105, + "theoretical_loss": 5.037464922474786, + "tokens_seen": 71106560 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004941624874623872, + "loss": 2.2882, + "theoretical_loss": 5.036794324528802, + "tokens_seen": 71172096 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004941524573721163, + "loss": 2.3312, + "theoretical_loss": 5.036124516509076, + "tokens_seen": 71237632 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004941424272818456, + "loss": 2.2387, + "theoretical_loss": 5.035455496759701, + "tokens_seen": 71303168 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004941323971915748, + "loss": 2.2868, + "theoretical_loss": 5.034787263629758, + "tokens_seen": 71368704 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004941223671013039, + "loss": 2.3221, + "theoretical_loss": 5.0341198154732965, + "tokens_seen": 71434240 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004941123370110331, + "loss": 2.3147, + "theoretical_loss": 5.033453150649318, + "tokens_seen": 71499776 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004941023069207623, + "loss": 2.3737, + "theoretical_loss": 5.032787267521752, + "tokens_seen": 71565312 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004940922768304915, + "loss": 2.2538, + "theoretical_loss": 5.032122164459443, + "tokens_seen": 71630848 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004940822467402207, + "loss": 2.2253, + "theoretical_loss": 5.031457839836122, + "tokens_seen": 71696384 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004940722166499498, + "loss": 2.3065, + "theoretical_loss": 5.0307942920304, + "tokens_seen": 71761920 + }, + { + "epoch": 0.02, + "learning_rate": 0.000494062186559679, + "loss": 2.2222, + "theoretical_loss": 5.030131519425737, + "tokens_seen": 71827456 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004940521564694082, + "loss": 2.2243, + "theoretical_loss": 5.02946952041043, + "tokens_seen": 71892992 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004940421263791374, + "loss": 2.5455, + "theoretical_loss": 5.028808293377589, + "tokens_seen": 71958528 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004940320962888666, + "loss": 2.3095, + "theoretical_loss": 5.028147836725127, + "tokens_seen": 72024064 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 34719, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.1904053688049316, + "objective/train/theoretical_loss": 5.0274881488557295, + "objective/train/tokens_used": 72089600, + "theoretical_loss": 5.0274881488557295, + "tokens_seen": 72089600 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004940220661985958, + "loss": 2.2471, + "theoretical_loss": 5.0274881488557295, + "tokens_seen": 72089600 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004940120361083249, + "loss": 2.371, + "theoretical_loss": 5.026829228176848, + "tokens_seen": 72155136 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004940020060180542, + "loss": 2.3128, + "theoretical_loss": 5.026171073100672, + "tokens_seen": 72220672 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004939919759277834, + "loss": 2.2263, + "theoretical_loss": 5.025513682044114, + "tokens_seen": 72286208 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004939819458375126, + "loss": 2.2687, + "theoretical_loss": 5.024857053428793, + "tokens_seen": 72351744 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004939719157472418, + "loss": 2.242, + "theoretical_loss": 5.024201185681015, + "tokens_seen": 72417280 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493961885656971, + "loss": 2.2848, + "theoretical_loss": 5.023546077231752, + "tokens_seen": 72482816 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004939518555667001, + "loss": 2.2538, + "theoretical_loss": 5.0228917265166295, + "tokens_seen": 72548352 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004939418254764293, + "loss": 2.2621, + "theoretical_loss": 5.022238131975905, + "tokens_seen": 72613888 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004939317953861585, + "loss": 2.2766, + "theoretical_loss": 5.021585292054446, + "tokens_seen": 72679424 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004939217652958877, + "loss": 2.1511, + "theoretical_loss": 5.020933205201725, + "tokens_seen": 72744960 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004939117352056169, + "loss": 2.1983, + "theoretical_loss": 5.020281869871787, + "tokens_seen": 72810496 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493901705115346, + "loss": 2.3942, + "theoretical_loss": 5.019631284523241, + "tokens_seen": 72876032 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004938916750250752, + "loss": 2.3212, + "theoretical_loss": 5.01898144761924, + "tokens_seen": 72941568 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004938816449348044, + "loss": 2.27, + "theoretical_loss": 5.018332357627462, + "tokens_seen": 73007104 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004938716148445337, + "loss": 2.2111, + "theoretical_loss": 5.017684013020096, + "tokens_seen": 73072640 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004938615847542628, + "loss": 2.3017, + "theoretical_loss": 5.017036412273823, + "tokens_seen": 73138176 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493851554663992, + "loss": 2.327, + "theoretical_loss": 5.016389553869796, + "tokens_seen": 73203712 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004938415245737211, + "loss": 2.3257, + "theoretical_loss": 5.0157434362936275, + "tokens_seen": 73269248 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004938314944834504, + "loss": 2.3457, + "theoretical_loss": 5.01509805803537, + "tokens_seen": 73334784 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004938214643931796, + "loss": 2.3353, + "theoretical_loss": 5.0144534175894995, + "tokens_seen": 73400320 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004938114343029088, + "loss": 2.2615, + "theoretical_loss": 5.013809513454898, + "tokens_seen": 73465856 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493801404212638, + "loss": 2.3114, + "theoretical_loss": 5.013166344134838, + "tokens_seen": 73531392 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004937913741223671, + "loss": 2.263, + "theoretical_loss": 5.012523908136963, + "tokens_seen": 73596928 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004937813440320963, + "loss": 2.3388, + "theoretical_loss": 5.011882203973277, + "tokens_seen": 73662464 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004937713139418255, + "loss": 2.237, + "theoretical_loss": 5.011241230160119, + "tokens_seen": 73728000 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004937612838515547, + "loss": 2.3059, + "theoretical_loss": 5.0106009852181534, + "tokens_seen": 73793536 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004937512537612839, + "loss": 2.2569, + "theoretical_loss": 5.009961467672353, + "tokens_seen": 73859072 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493741223671013, + "loss": 2.212, + "theoretical_loss": 5.009322676051976, + "tokens_seen": 73924608 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004937311935807422, + "loss": 2.3597, + "theoretical_loss": 5.008684608890562, + "tokens_seen": 73990144 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004937211634904714, + "loss": 2.2744, + "theoretical_loss": 5.008047264725901, + "tokens_seen": 74055680 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004937111334002006, + "loss": 2.3329, + "theoretical_loss": 5.007410642100032, + "tokens_seen": 74121216 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004937011033099298, + "loss": 2.1329, + "theoretical_loss": 5.006774739559216, + "tokens_seen": 74186752 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004936910732196591, + "loss": 2.2341, + "theoretical_loss": 5.00613955565392, + "tokens_seen": 74252288 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004936810431293881, + "loss": 2.1303, + "theoretical_loss": 5.005505088938815, + "tokens_seen": 74317824 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004936710130391174, + "loss": 2.2476, + "theoretical_loss": 5.004871337972742, + "tokens_seen": 74383360 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004936609829488465, + "loss": 2.1802, + "theoretical_loss": 5.004238301318706, + "tokens_seen": 74448896 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004936509528585758, + "loss": 2.2953, + "theoretical_loss": 5.003605977543861, + "tokens_seen": 74514432 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493640922768305, + "loss": 2.2395, + "theoretical_loss": 5.00297436521949, + "tokens_seen": 74579968 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004936308926780341, + "loss": 2.2854, + "theoretical_loss": 5.002343462920992, + "tokens_seen": 74645504 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004936208625877633, + "loss": 2.3308, + "theoretical_loss": 5.001713269227867, + "tokens_seen": 74711040 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004936108324974925, + "loss": 2.262, + "theoretical_loss": 5.0010837827237005, + "tokens_seen": 74776576 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004936008024072217, + "loss": 2.2907, + "theoretical_loss": 5.000455001996143, + "tokens_seen": 74842112 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004935907723169509, + "loss": 2.3631, + "theoretical_loss": 4.999826925636905, + "tokens_seen": 74907648 + }, + { + "epoch": 0.02, + "learning_rate": 0.00049358074222668, + "loss": 2.3292, + "theoretical_loss": 4.999199552241732, + "tokens_seen": 74973184 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004935707121364092, + "loss": 2.3675, + "theoretical_loss": 4.998572880410396, + "tokens_seen": 75038720 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004935606820461384, + "loss": 2.1894, + "theoretical_loss": 4.997946908746677, + "tokens_seen": 75104256 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004935506519558676, + "loss": 2.3133, + "theoretical_loss": 4.997321635858347, + "tokens_seen": 75169792 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004935406218655968, + "loss": 2.261, + "theoretical_loss": 4.9966970603571585, + "tokens_seen": 75235328 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493530591775326, + "loss": 2.3353, + "theoretical_loss": 4.996073180858831, + "tokens_seen": 75300864 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 35865, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.2504303455352783, + "objective/train/theoretical_loss": 4.995449995983028, + "objective/train/tokens_used": 75366400, + "theoretical_loss": 4.995449995983028, + "tokens_seen": 75366400 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004935205616850551, + "loss": 2.2556, + "theoretical_loss": 4.995449995983028, + "tokens_seen": 75366400 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004935105315947844, + "loss": 2.271, + "theoretical_loss": 4.994827504353353, + "tokens_seen": 75431936 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004935005015045135, + "loss": 2.2495, + "theoretical_loss": 4.994205704597325, + "tokens_seen": 75497472 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004934904714142428, + "loss": 2.2533, + "theoretical_loss": 4.993584595346373, + "tokens_seen": 75563008 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004934804413239719, + "loss": 2.3012, + "theoretical_loss": 4.992964175235814, + "tokens_seen": 75628544 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004934704112337011, + "loss": 2.2547, + "theoretical_loss": 4.992344442904843, + "tokens_seen": 75694080 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004934603811434303, + "loss": 2.1872, + "theoretical_loss": 4.991725396996515, + "tokens_seen": 75759616 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004934503510531595, + "loss": 2.0976, + "theoretical_loss": 4.991107036157738, + "tokens_seen": 75825152 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004934403209628887, + "loss": 2.1728, + "theoretical_loss": 4.990489359039248, + "tokens_seen": 75890688 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004934302908726179, + "loss": 2.2403, + "theoretical_loss": 4.989872364295605, + "tokens_seen": 75956224 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493420260782347, + "loss": 2.247, + "theoretical_loss": 4.989256050585174, + "tokens_seen": 76021760 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004934102306920762, + "loss": 2.3243, + "theoretical_loss": 4.9886404165701075, + "tokens_seen": 76087296 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004934002006018054, + "loss": 2.2353, + "theoretical_loss": 4.98802546091634, + "tokens_seen": 76152832 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004933901705115346, + "loss": 2.3191, + "theoretical_loss": 4.9874111822935685, + "tokens_seen": 76218368 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004933801404212638, + "loss": 2.1415, + "theoretical_loss": 4.986797579375239, + "tokens_seen": 76283904 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493370110330993, + "loss": 2.2633, + "theoretical_loss": 4.986184650838532, + "tokens_seen": 76349440 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004933600802407221, + "loss": 2.2384, + "theoretical_loss": 4.985572395364354, + "tokens_seen": 76414976 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004933500501504513, + "loss": 2.1311, + "theoretical_loss": 4.98496081163732, + "tokens_seen": 76480512 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004933400200601805, + "loss": 2.1382, + "theoretical_loss": 4.984349898345735, + "tokens_seen": 76546048 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004933299899699098, + "loss": 2.2198, + "theoretical_loss": 4.98373965418159, + "tokens_seen": 76611584 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004933199598796389, + "loss": 2.3184, + "theoretical_loss": 4.983130077840543, + "tokens_seen": 76677120 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004933099297893682, + "loss": 2.2892, + "theoretical_loss": 4.982521168021906, + "tokens_seen": 76742656 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004932998996990972, + "loss": 2.0826, + "theoretical_loss": 4.981912923428634, + "tokens_seen": 76808192 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004932898696088265, + "loss": 2.3, + "theoretical_loss": 4.981305342767307, + "tokens_seen": 76873728 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004932798395185557, + "loss": 2.1494, + "theoretical_loss": 4.980698424748122, + "tokens_seen": 76939264 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004932698094282849, + "loss": 2.3238, + "theoretical_loss": 4.980092168084878, + "tokens_seen": 77004800 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004932597793380141, + "loss": 2.2109, + "theoretical_loss": 4.97948657149496, + "tokens_seen": 77070336 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004932497492477432, + "loss": 2.1849, + "theoretical_loss": 4.978881633699331, + "tokens_seen": 77135872 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004932397191574724, + "loss": 2.307, + "theoretical_loss": 4.9782773534225155, + "tokens_seen": 77201408 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004932296890672016, + "loss": 2.1627, + "theoretical_loss": 4.977673729392586, + "tokens_seen": 77266944 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004932196589769308, + "loss": 2.237, + "theoretical_loss": 4.977070760341153, + "tokens_seen": 77332480 + }, + { + "epoch": 0.02, + "learning_rate": 0.00049320962888666, + "loss": 2.0822, + "theoretical_loss": 4.976468445003352, + "tokens_seen": 77398016 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004931995987963893, + "loss": 2.1438, + "theoretical_loss": 4.975866782117828, + "tokens_seen": 77463552 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004931895687061183, + "loss": 2.2996, + "theoretical_loss": 4.975265770426724, + "tokens_seen": 77529088 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004931795386158476, + "loss": 2.2511, + "theoretical_loss": 4.974665408675667, + "tokens_seen": 77594624 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004931695085255767, + "loss": 2.3474, + "theoretical_loss": 4.974065695613763, + "tokens_seen": 77660160 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493159478435306, + "loss": 2.1883, + "theoretical_loss": 4.973466629993574, + "tokens_seen": 77725696 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004931494483450352, + "loss": 2.2856, + "theoretical_loss": 4.972868210571109, + "tokens_seen": 77791232 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004931394182547643, + "loss": 2.3194, + "theoretical_loss": 4.972270436105815, + "tokens_seen": 77856768 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004931293881644935, + "loss": 2.2664, + "theoretical_loss": 4.971673305360563, + "tokens_seen": 77922304 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004931193580742227, + "loss": 2.0957, + "theoretical_loss": 4.971076817101633, + "tokens_seen": 77987840 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004931093279839519, + "loss": 2.192, + "theoretical_loss": 4.970480970098703, + "tokens_seen": 78053376 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004930992978936811, + "loss": 2.2034, + "theoretical_loss": 4.96988576312484, + "tokens_seen": 78118912 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004930892678034102, + "loss": 2.2028, + "theoretical_loss": 4.969291194956485, + "tokens_seen": 78184448 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004930792377131394, + "loss": 2.2794, + "theoretical_loss": 4.968697264373438, + "tokens_seen": 78249984 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004930692076228686, + "loss": 2.2974, + "theoretical_loss": 4.968103970158854, + "tokens_seen": 78315520 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004930591775325978, + "loss": 2.1833, + "theoretical_loss": 4.967511311099223, + "tokens_seen": 78381056 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493049147442327, + "loss": 2.2794, + "theoretical_loss": 4.96691928598436, + "tokens_seen": 78446592 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004930391173520562, + "loss": 2.2013, + "theoretical_loss": 4.966327893607398, + "tokens_seen": 78512128 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004930290872617853, + "loss": 2.2991, + "theoretical_loss": 4.96573713276477, + "tokens_seen": 78577664 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 36940, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9949949979782104, + "objective/train/theoretical_loss": 4.965147002256201, + "objective/train/tokens_used": 78643200, + "theoretical_loss": 4.965147002256201, + "tokens_seen": 78643200 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004930190571715146, + "loss": 2.1442, + "theoretical_loss": 4.965147002256201, + "tokens_seen": 78643200 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004930090270812437, + "loss": 2.1125, + "theoretical_loss": 4.964557500884695, + "tokens_seen": 78708736 + }, + { + "epoch": 0.02, + "learning_rate": 0.000492998996990973, + "loss": 2.2485, + "theoretical_loss": 4.963968627456523, + "tokens_seen": 78774272 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004929889669007021, + "loss": 2.2695, + "theoretical_loss": 4.96338038078121, + "tokens_seen": 78839808 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004929789368104313, + "loss": 2.0974, + "theoretical_loss": 4.962792759671528, + "tokens_seen": 78905344 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004929689067201605, + "loss": 2.1936, + "theoretical_loss": 4.962205762943478, + "tokens_seen": 78970880 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004929588766298897, + "loss": 2.2689, + "theoretical_loss": 4.961619389416288, + "tokens_seen": 79036416 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004929488465396189, + "loss": 2.3084, + "theoretical_loss": 4.96103363791239, + "tokens_seen": 79101952 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004929388164493481, + "loss": 2.2398, + "theoretical_loss": 4.960448507257416, + "tokens_seen": 79167488 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004929287863590773, + "loss": 2.2062, + "theoretical_loss": 4.959863996280188, + "tokens_seen": 79233024 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004929187562688064, + "loss": 2.1251, + "theoretical_loss": 4.959280103812697, + "tokens_seen": 79298560 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004929087261785356, + "loss": 2.083, + "theoretical_loss": 4.958696828690105, + "tokens_seen": 79364096 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004928986960882648, + "loss": 2.1283, + "theoretical_loss": 4.958114169750726, + "tokens_seen": 79429632 + }, + { + "epoch": 0.02, + "learning_rate": 0.000492888665997994, + "loss": 2.2702, + "theoretical_loss": 4.957532125836012, + "tokens_seen": 79495168 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004928786359077232, + "loss": 2.1475, + "theoretical_loss": 4.956950695790551, + "tokens_seen": 79560704 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004928686058174523, + "loss": 2.2003, + "theoretical_loss": 4.95636987846205, + "tokens_seen": 79626240 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004928585757271815, + "loss": 2.1842, + "theoretical_loss": 4.955789672701323, + "tokens_seen": 79691776 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004928485456369107, + "loss": 2.2253, + "theoretical_loss": 4.955210077362285, + "tokens_seen": 79757312 + }, + { + "epoch": 0.02, + "learning_rate": 0.00049283851554664, + "loss": 2.2648, + "theoretical_loss": 4.954631091301934, + "tokens_seen": 79822848 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004928284854563691, + "loss": 2.2793, + "theoretical_loss": 4.954052713380349, + "tokens_seen": 79888384 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004928184553660984, + "loss": 2.2674, + "theoretical_loss": 4.953474942460673, + "tokens_seen": 79953920 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004928084252758274, + "loss": 2.3665, + "theoretical_loss": 4.952897777409104, + "tokens_seen": 80019456 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004927983951855567, + "loss": 2.3205, + "theoretical_loss": 4.952321217094882, + "tokens_seen": 80084992 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004927883650952859, + "loss": 2.1552, + "theoretical_loss": 4.951745260390284, + "tokens_seen": 80150528 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004927783350050151, + "loss": 2.2706, + "theoretical_loss": 4.951169906170607, + "tokens_seen": 80216064 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004927683049147443, + "loss": 2.2171, + "theoretical_loss": 4.950595153314161, + "tokens_seen": 80281600 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004927582748244734, + "loss": 2.2139, + "theoretical_loss": 4.950021000702261, + "tokens_seen": 80347136 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004927482447342026, + "loss": 2.2714, + "theoretical_loss": 4.949447447219208, + "tokens_seen": 80412672 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004927382146439318, + "loss": 2.2835, + "theoretical_loss": 4.948874491752287, + "tokens_seen": 80478208 + }, + { + "epoch": 0.02, + "learning_rate": 0.000492728184553661, + "loss": 2.2547, + "theoretical_loss": 4.9483021331917545, + "tokens_seen": 80543744 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004927181544633902, + "loss": 2.2009, + "theoretical_loss": 4.9477303704308255, + "tokens_seen": 80609280 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004927081243731193, + "loss": 2.2262, + "theoretical_loss": 4.947159202365664, + "tokens_seen": 80674816 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004926980942828485, + "loss": 2.2214, + "theoretical_loss": 4.946588627895375, + "tokens_seen": 80740352 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004926880641925777, + "loss": 2.134, + "theoretical_loss": 4.946018645921994, + "tokens_seen": 80805888 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004926780341023069, + "loss": 2.1625, + "theoretical_loss": 4.945449255350473, + "tokens_seen": 80871424 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004926680040120361, + "loss": 2.1859, + "theoretical_loss": 4.944880455088677, + "tokens_seen": 80936960 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004926579739217654, + "loss": 2.1861, + "theoretical_loss": 4.944312244047365, + "tokens_seen": 81002496 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004926479438314944, + "loss": 2.1955, + "theoretical_loss": 4.943744621140189, + "tokens_seen": 81068032 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004926379137412237, + "loss": 2.2435, + "theoretical_loss": 4.94317758528368, + "tokens_seen": 81133568 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004926278836509528, + "loss": 2.1718, + "theoretical_loss": 4.942611135397234, + "tokens_seen": 81199104 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004926178535606821, + "loss": 2.2062, + "theoretical_loss": 4.942045270403112, + "tokens_seen": 81264640 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004926078234704113, + "loss": 2.302, + "theoretical_loss": 4.941479989226421, + "tokens_seen": 81330176 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004925977933801404, + "loss": 2.2698, + "theoretical_loss": 4.940915290795106, + "tokens_seen": 81395712 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004925877632898696, + "loss": 2.2096, + "theoretical_loss": 4.940351174039947, + "tokens_seen": 81461248 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004925777331995988, + "loss": 2.2175, + "theoretical_loss": 4.939787637894537, + "tokens_seen": 81526784 + }, + { + "epoch": 0.02, + "learning_rate": 0.000492567703109328, + "loss": 2.171, + "theoretical_loss": 4.939224681295286, + "tokens_seen": 81592320 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004925576730190572, + "loss": 2.1318, + "theoretical_loss": 4.938662303181399, + "tokens_seen": 81657856 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004925476429287864, + "loss": 2.1405, + "theoretical_loss": 4.938100502494878, + "tokens_seen": 81723392 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004925376128385155, + "loss": 2.2237, + "theoretical_loss": 4.9375392781805, + "tokens_seen": 81788928 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004925275827482447, + "loss": 2.2024, + "theoretical_loss": 4.936978629185818, + "tokens_seen": 81854464 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 38057, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.202394485473633, + "objective/train/theoretical_loss": 4.936418554461147, + "objective/train/tokens_used": 81920000, + "theoretical_loss": 4.936418554461147, + "tokens_seen": 81920000 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004925175526579739, + "loss": 2.2192, + "theoretical_loss": 4.936418554461147, + "tokens_seen": 81920000 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004925075225677031, + "loss": 2.0992, + "theoretical_loss": 4.935859052959554, + "tokens_seen": 81985536 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004924974924774323, + "loss": 2.2714, + "theoretical_loss": 4.935300123636851, + "tokens_seen": 82051072 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004924874623871615, + "loss": 2.2555, + "theoretical_loss": 4.93474176545158, + "tokens_seen": 82116608 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004924774322968907, + "loss": 2.227, + "theoretical_loss": 4.934183977365015, + "tokens_seen": 82182144 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004924674022066199, + "loss": 2.1679, + "theoretical_loss": 4.9336267583411395, + "tokens_seen": 82247680 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004924573721163491, + "loss": 2.1307, + "theoretical_loss": 4.933070107346644, + "tokens_seen": 82313216 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004924473420260783, + "loss": 2.1567, + "theoretical_loss": 4.932514023350921, + "tokens_seen": 82378752 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004924373119358075, + "loss": 2.2098, + "theoretical_loss": 4.931958505326042, + "tokens_seen": 82444288 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004924272818455366, + "loss": 2.0739, + "theoretical_loss": 4.931403552246769, + "tokens_seen": 82509824 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004924172517552658, + "loss": 2.1842, + "theoretical_loss": 4.930849163090521, + "tokens_seen": 82575360 + }, + { + "epoch": 0.03, + "learning_rate": 0.000492407221664995, + "loss": 2.1828, + "theoretical_loss": 4.930295336837387, + "tokens_seen": 82640896 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004923971915747242, + "loss": 2.1548, + "theoretical_loss": 4.9297420724701055, + "tokens_seen": 82706432 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004923871614844534, + "loss": 2.2072, + "theoretical_loss": 4.929189368974054, + "tokens_seen": 82771968 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004923771313941825, + "loss": 2.1819, + "theoretical_loss": 4.928637225337246, + "tokens_seen": 82837504 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004923671013039117, + "loss": 2.1379, + "theoretical_loss": 4.9280856405503215, + "tokens_seen": 82903040 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004923570712136409, + "loss": 2.083, + "theoretical_loss": 4.927534613606534, + "tokens_seen": 82968576 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004923470411233702, + "loss": 2.1922, + "theoretical_loss": 4.926984143501745, + "tokens_seen": 83034112 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004923370110330993, + "loss": 2.2751, + "theoretical_loss": 4.926434229234411, + "tokens_seen": 83099648 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004923269809428286, + "loss": 2.2322, + "theoretical_loss": 4.925884869805586, + "tokens_seen": 83165184 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004923169508525576, + "loss": 2.206, + "theoretical_loss": 4.925336064218893, + "tokens_seen": 83230720 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004923069207622869, + "loss": 2.2255, + "theoretical_loss": 4.924787811480538, + "tokens_seen": 83296256 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004922968906720161, + "loss": 2.1578, + "theoretical_loss": 4.9242401105992855, + "tokens_seen": 83361792 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004922868605817453, + "loss": 2.0497, + "theoretical_loss": 4.923692960586452, + "tokens_seen": 83427328 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004922768304914745, + "loss": 2.0019, + "theoretical_loss": 4.923146360455906, + "tokens_seen": 83492864 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004922668004012036, + "loss": 2.1109, + "theoretical_loss": 4.92260030922405, + "tokens_seen": 83558400 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004922567703109328, + "loss": 2.2309, + "theoretical_loss": 4.922054805909816, + "tokens_seen": 83623936 + }, + { + "epoch": 0.03, + "learning_rate": 0.000492246740220662, + "loss": 2.052, + "theoretical_loss": 4.921509849534658, + "tokens_seen": 83689472 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004922367101303912, + "loss": 2.1835, + "theoretical_loss": 4.920965439122538, + "tokens_seen": 83755008 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004922266800401204, + "loss": 2.2854, + "theoretical_loss": 4.9204215736999295, + "tokens_seen": 83820544 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004922166499498495, + "loss": 2.2712, + "theoretical_loss": 4.919878252295793, + "tokens_seen": 83886080 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004922066198595787, + "loss": 2.0612, + "theoretical_loss": 4.919335473941581, + "tokens_seen": 83951616 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004921965897693079, + "loss": 2.1494, + "theoretical_loss": 4.918793237671224, + "tokens_seen": 84017152 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004921865596790371, + "loss": 2.2522, + "theoretical_loss": 4.918251542521121, + "tokens_seen": 84082688 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004921765295887663, + "loss": 2.215, + "theoretical_loss": 4.9177103875301365, + "tokens_seen": 84148224 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004921664994984956, + "loss": 2.0619, + "theoretical_loss": 4.917169771739586, + "tokens_seen": 84213760 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004921564694082246, + "loss": 2.0233, + "theoretical_loss": 4.916629694193233, + "tokens_seen": 84279296 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004921464393179539, + "loss": 2.192, + "theoretical_loss": 4.916090153937276, + "tokens_seen": 84344832 + }, + { + "epoch": 0.03, + "learning_rate": 0.000492136409227683, + "loss": 2.2705, + "theoretical_loss": 4.915551150020346, + "tokens_seen": 84410368 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004921263791374123, + "loss": 2.1463, + "theoretical_loss": 4.915012681493495, + "tokens_seen": 84475904 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004921163490471415, + "loss": 2.1398, + "theoretical_loss": 4.914474747410187, + "tokens_seen": 84541440 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004921063189568706, + "loss": 2.1065, + "theoretical_loss": 4.913937346826293, + "tokens_seen": 84606976 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004920962888665998, + "loss": 2.0674, + "theoretical_loss": 4.91340047880008, + "tokens_seen": 84672512 + }, + { + "epoch": 0.03, + "learning_rate": 0.000492086258776329, + "loss": 2.091, + "theoretical_loss": 4.9128641423922055, + "tokens_seen": 84738048 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004920762286860582, + "loss": 2.0549, + "theoretical_loss": 4.912328336665711, + "tokens_seen": 84803584 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004920661985957874, + "loss": 2.1916, + "theoretical_loss": 4.911793060686007, + "tokens_seen": 84869120 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004920561685055166, + "loss": 2.2318, + "theoretical_loss": 4.9112583135208725, + "tokens_seen": 84934656 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004920461384152457, + "loss": 2.2887, + "theoretical_loss": 4.910724094240447, + "tokens_seen": 85000192 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004920361083249749, + "loss": 2.1492, + "theoretical_loss": 4.910190401917214, + "tokens_seen": 85065728 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004920260782347041, + "loss": 2.0682, + "theoretical_loss": 4.909657235626009, + "tokens_seen": 85131264 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 39240, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.163038969039917, + "objective/train/theoretical_loss": 4.909124594443993, + "objective/train/tokens_used": 85196800, + "theoretical_loss": 4.909124594443993, + "tokens_seen": 85196800 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004920160481444333, + "loss": 2.1263, + "theoretical_loss": 4.909124594443993, + "tokens_seen": 85196800 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004920060180541625, + "loss": 2.1707, + "theoretical_loss": 4.908592477450659, + "tokens_seen": 85262336 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004919959879638916, + "loss": 2.0864, + "theoretical_loss": 4.908060883727821, + "tokens_seen": 85327872 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004919859578736209, + "loss": 2.1818, + "theoretical_loss": 4.9075298123596, + "tokens_seen": 85393408 + }, + { + "epoch": 0.03, + "learning_rate": 0.00049197592778335, + "loss": 2.1696, + "theoretical_loss": 4.906999262432427, + "tokens_seen": 85458944 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004919658976930793, + "loss": 2.2655, + "theoretical_loss": 4.906469233035025, + "tokens_seen": 85524480 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004919558676028084, + "loss": 2.2326, + "theoretical_loss": 4.905939723258408, + "tokens_seen": 85590016 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004919458375125377, + "loss": 2.2682, + "theoretical_loss": 4.905410732195875, + "tokens_seen": 85655552 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004919358074222668, + "loss": 2.3244, + "theoretical_loss": 4.904882258942994, + "tokens_seen": 85721088 + }, + { + "epoch": 0.03, + "learning_rate": 0.000491925777331996, + "loss": 2.1515, + "theoretical_loss": 4.904354302597603, + "tokens_seen": 85786624 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004919157472417252, + "loss": 2.0532, + "theoretical_loss": 4.903826862259798, + "tokens_seen": 85852160 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004919057171514544, + "loss": 1.9279, + "theoretical_loss": 4.90329993703193, + "tokens_seen": 85917696 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004918956870611836, + "loss": 2.0392, + "theoretical_loss": 4.902773526018591, + "tokens_seen": 85983232 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004918856569709127, + "loss": 2.2609, + "theoretical_loss": 4.902247628326613, + "tokens_seen": 86048768 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004918756268806419, + "loss": 2.1832, + "theoretical_loss": 4.9017222430650556, + "tokens_seen": 86114304 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004918655967903711, + "loss": 2.1833, + "theoretical_loss": 4.901197369345205, + "tokens_seen": 86179840 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004918555667001003, + "loss": 2.116, + "theoretical_loss": 4.9006730062805595, + "tokens_seen": 86245376 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004918455366098295, + "loss": 2.1492, + "theoretical_loss": 4.9001491529868275, + "tokens_seen": 86310912 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004918355065195586, + "loss": 2.2519, + "theoretical_loss": 4.899625808581918, + "tokens_seen": 86376448 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004918254764292878, + "loss": 2.0854, + "theoretical_loss": 4.899102972185936, + "tokens_seen": 86441984 + }, + { + "epoch": 0.03, + "learning_rate": 0.000491815446339017, + "loss": 2.1091, + "theoretical_loss": 4.898580642921171, + "tokens_seen": 86507520 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004918054162487463, + "loss": 2.1494, + "theoretical_loss": 4.898058819912093, + "tokens_seen": 86573056 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004917953861584754, + "loss": 2.1229, + "theoretical_loss": 4.897537502285347, + "tokens_seen": 86638592 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004917853560682047, + "loss": 2.1611, + "theoretical_loss": 4.897016689169741, + "tokens_seen": 86704128 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004917753259779337, + "loss": 2.1895, + "theoretical_loss": 4.896496379696243, + "tokens_seen": 86769664 + }, + { + "epoch": 0.03, + "learning_rate": 0.000491765295887663, + "loss": 2.1702, + "theoretical_loss": 4.895976572997973, + "tokens_seen": 86835200 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004917552657973922, + "loss": 2.2811, + "theoretical_loss": 4.895457268210196, + "tokens_seen": 86900736 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004917452357071214, + "loss": 2.1183, + "theoretical_loss": 4.8949384644703136, + "tokens_seen": 86966272 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004917352056168506, + "loss": 2.1894, + "theoretical_loss": 4.894420160917861, + "tokens_seen": 87031808 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004917251755265797, + "loss": 2.1255, + "theoretical_loss": 4.893902356694495, + "tokens_seen": 87097344 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004917151454363089, + "loss": 2.2124, + "theoretical_loss": 4.8933850509439925, + "tokens_seen": 87162880 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004917051153460381, + "loss": 2.1481, + "theoretical_loss": 4.89286824281224, + "tokens_seen": 87228416 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004916950852557673, + "loss": 2.2657, + "theoretical_loss": 4.892351931447224, + "tokens_seen": 87293952 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004916850551654965, + "loss": 2.0912, + "theoretical_loss": 4.891836115999034, + "tokens_seen": 87359488 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004916750250752258, + "loss": 2.2401, + "theoretical_loss": 4.891320795619847, + "tokens_seen": 87425024 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004916649949849548, + "loss": 2.1282, + "theoretical_loss": 4.890805969463923, + "tokens_seen": 87490560 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004916549648946841, + "loss": 2.1413, + "theoretical_loss": 4.8902916366876, + "tokens_seen": 87556096 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004916449348044132, + "loss": 2.1419, + "theoretical_loss": 4.889777796449285, + "tokens_seen": 87621632 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004916349047141425, + "loss": 2.0941, + "theoretical_loss": 4.889264447909451, + "tokens_seen": 87687168 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004916248746238717, + "loss": 2.1657, + "theoretical_loss": 4.888751590230626, + "tokens_seen": 87752704 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004916148445336008, + "loss": 2.2084, + "theoretical_loss": 4.888239222577389, + "tokens_seen": 87818240 + }, + { + "epoch": 0.03, + "learning_rate": 0.00049160481444333, + "loss": 2.1992, + "theoretical_loss": 4.887727344116364, + "tokens_seen": 87883776 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004915947843530592, + "loss": 2.2652, + "theoretical_loss": 4.8872159540162095, + "tokens_seen": 87949312 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004915847542627884, + "loss": 2.0996, + "theoretical_loss": 4.88670505144762, + "tokens_seen": 88014848 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004915747241725176, + "loss": 2.1603, + "theoretical_loss": 4.8861946355833075, + "tokens_seen": 88080384 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004915646940822468, + "loss": 2.1644, + "theoretical_loss": 4.885684705598009, + "tokens_seen": 88145920 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004915546639919759, + "loss": 2.088, + "theoretical_loss": 4.885175260668468, + "tokens_seen": 88211456 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004915446339017051, + "loss": 2.192, + "theoretical_loss": 4.884666299973436, + "tokens_seen": 88276992 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004915346038114343, + "loss": 2.0311, + "theoretical_loss": 4.884157822693663, + "tokens_seen": 88342528 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004915245737211635, + "loss": 2.1695, + "theoretical_loss": 4.88364982801189, + "tokens_seen": 88408064 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 40420, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.2097666263580322, + "objective/train/theoretical_loss": 4.8831423151128455, + "objective/train/tokens_used": 88473600, + "theoretical_loss": 4.8831423151128455, + "tokens_seen": 88473600 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004915145436308927, + "loss": 2.1896, + "theoretical_loss": 4.8831423151128455, + "tokens_seen": 88473600 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004915045135406218, + "loss": 2.0422, + "theoretical_loss": 4.8826352831832365, + "tokens_seen": 88539136 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004914944834503511, + "loss": 2.0988, + "theoretical_loss": 4.882128731411745, + "tokens_seen": 88604672 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004914844533600802, + "loss": 2.0341, + "theoretical_loss": 4.881622658989022, + "tokens_seen": 88670208 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004914744232698095, + "loss": 2.1243, + "theoretical_loss": 4.881117065107675, + "tokens_seen": 88735744 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004914643931795386, + "loss": 2.1129, + "theoretical_loss": 4.880611948962271, + "tokens_seen": 88801280 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004914543630892679, + "loss": 2.2901, + "theoretical_loss": 4.880107309749322, + "tokens_seen": 88866816 + }, + { + "epoch": 0.03, + "learning_rate": 0.000491444332998997, + "loss": 2.1839, + "theoretical_loss": 4.879603146667284, + "tokens_seen": 88932352 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004914343029087262, + "loss": 2.2298, + "theoretical_loss": 4.879099458916553, + "tokens_seen": 88997888 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004914242728184554, + "loss": 2.1732, + "theoretical_loss": 4.878596245699449, + "tokens_seen": 89063424 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004914142427281846, + "loss": 2.1533, + "theoretical_loss": 4.878093506220221, + "tokens_seen": 89128960 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004914042126379138, + "loss": 2.1162, + "theoretical_loss": 4.877591239685035, + "tokens_seen": 89194496 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004913941825476429, + "loss": 2.1416, + "theoretical_loss": 4.877089445301969, + "tokens_seen": 89260032 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004913841524573721, + "loss": 2.1434, + "theoretical_loss": 4.876588122281007, + "tokens_seen": 89325568 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004913741223671013, + "loss": 2.0938, + "theoretical_loss": 4.876087269834034, + "tokens_seen": 89391104 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004913640922768305, + "loss": 2.2523, + "theoretical_loss": 4.875586887174828, + "tokens_seen": 89456640 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004913540621865597, + "loss": 2.0968, + "theoretical_loss": 4.875086973519058, + "tokens_seen": 89522176 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004913440320962888, + "loss": 2.1436, + "theoretical_loss": 4.874587528084271, + "tokens_seen": 89587712 + }, + { + "epoch": 0.03, + "learning_rate": 0.000491334002006018, + "loss": 2.1251, + "theoretical_loss": 4.874088550089896, + "tokens_seen": 89653248 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004913239719157472, + "loss": 2.0605, + "theoretical_loss": 4.873590038757229, + "tokens_seen": 89718784 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004913139418254765, + "loss": 2.0302, + "theoretical_loss": 4.87309199330943, + "tokens_seen": 89784320 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004913039117352056, + "loss": 2.2222, + "theoretical_loss": 4.872594412971521, + "tokens_seen": 89849856 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004912938816449349, + "loss": 2.1099, + "theoretical_loss": 4.872097296970376, + "tokens_seen": 89915392 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004912838515546639, + "loss": 2.0196, + "theoretical_loss": 4.871600644534717, + "tokens_seen": 89980928 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004912738214643932, + "loss": 2.253, + "theoretical_loss": 4.871104454895106, + "tokens_seen": 90046464 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004912637913741224, + "loss": 2.179, + "theoretical_loss": 4.870608727283942, + "tokens_seen": 90112000 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004912537612838516, + "loss": 2.2497, + "theoretical_loss": 4.870113460935455, + "tokens_seen": 90177536 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004912437311935808, + "loss": 2.1783, + "theoretical_loss": 4.869618655085697, + "tokens_seen": 90243072 + }, + { + "epoch": 0.03, + "learning_rate": 0.00049123370110331, + "loss": 2.1701, + "theoretical_loss": 4.869124308972543, + "tokens_seen": 90308608 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004912236710130391, + "loss": 2.1516, + "theoretical_loss": 4.868630421835677, + "tokens_seen": 90374144 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004912136409227683, + "loss": 2.0638, + "theoretical_loss": 4.868136992916593, + "tokens_seen": 90439680 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004912036108324975, + "loss": 2.2692, + "theoretical_loss": 4.867644021458585, + "tokens_seen": 90505216 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004911935807422267, + "loss": 2.0886, + "theoretical_loss": 4.8671515067067475, + "tokens_seen": 90570752 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004911835506519559, + "loss": 2.0704, + "theoretical_loss": 4.866659447907962, + "tokens_seen": 90636288 + }, + { + "epoch": 0.03, + "learning_rate": 0.000491173520561685, + "loss": 2.0847, + "theoretical_loss": 4.866167844310896, + "tokens_seen": 90701824 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004911634904714142, + "loss": 2.1189, + "theoretical_loss": 4.865676695165996, + "tokens_seen": 90767360 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004911534603811434, + "loss": 2.0816, + "theoretical_loss": 4.865185999725487, + "tokens_seen": 90832896 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004911434302908726, + "loss": 2.1414, + "theoretical_loss": 4.864695757243359, + "tokens_seen": 90898432 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004911334002006019, + "loss": 2.0815, + "theoretical_loss": 4.864205966975364, + "tokens_seen": 90963968 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004911233701103309, + "loss": 2.1518, + "theoretical_loss": 4.8637166281790165, + "tokens_seen": 91029504 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004911133400200602, + "loss": 2.0639, + "theoretical_loss": 4.86322774011358, + "tokens_seen": 91095040 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004911033099297893, + "loss": 2.2088, + "theoretical_loss": 4.862739302040067, + "tokens_seen": 91160576 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004910932798395186, + "loss": 2.2344, + "theoretical_loss": 4.862251313221231, + "tokens_seen": 91226112 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004910832497492478, + "loss": 2.0753, + "theoretical_loss": 4.861763772921561, + "tokens_seen": 91291648 + }, + { + "epoch": 0.03, + "learning_rate": 0.000491073219658977, + "loss": 2.1829, + "theoretical_loss": 4.861276680407279, + "tokens_seen": 91357184 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004910631895687061, + "loss": 2.0843, + "theoretical_loss": 4.860790034946331, + "tokens_seen": 91422720 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004910531594784353, + "loss": 2.1609, + "theoretical_loss": 4.860303835808384, + "tokens_seen": 91488256 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004910431293881645, + "loss": 2.0808, + "theoretical_loss": 4.859818082264822, + "tokens_seen": 91553792 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004910330992978937, + "loss": 2.1477, + "theoretical_loss": 4.859332773588736, + "tokens_seen": 91619328 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004910230692076229, + "loss": 2.1218, + "theoretical_loss": 4.858847909054923, + "tokens_seen": 91684864 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 41581, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.0970096588134766, + "objective/train/theoretical_loss": 4.858363487939879, + "objective/train/tokens_used": 91750400, + "theoretical_loss": 4.858363487939879, + "tokens_seen": 91750400 + }, + { + "epoch": 0.03, + "learning_rate": 0.000491013039117352, + "loss": 2.0835, + "theoretical_loss": 4.858363487939879, + "tokens_seen": 91750400 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004910030090270812, + "loss": 2.1217, + "theoretical_loss": 4.857879509521796, + "tokens_seen": 91815936 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004909929789368104, + "loss": 2.0398, + "theoretical_loss": 4.857395973080552, + "tokens_seen": 91881472 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004909829488465397, + "loss": 2.1699, + "theoretical_loss": 4.856912877897713, + "tokens_seen": 91947008 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004909729187562688, + "loss": 2.1213, + "theoretical_loss": 4.85643022325652, + "tokens_seen": 92012544 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004909628886659981, + "loss": 2.173, + "theoretical_loss": 4.855948008441891, + "tokens_seen": 92078080 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004909528585757272, + "loss": 2.1174, + "theoretical_loss": 4.85546623274041, + "tokens_seen": 92143616 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004909428284854564, + "loss": 2.0028, + "theoretical_loss": 4.854984895440325, + "tokens_seen": 92209152 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004909327983951856, + "loss": 2.1224, + "theoretical_loss": 4.854503995831545, + "tokens_seen": 92274688 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004909227683049148, + "loss": 2.1763, + "theoretical_loss": 4.854023533205627, + "tokens_seen": 92340224 + }, + { + "epoch": 0.03, + "learning_rate": 0.000490912738214644, + "loss": 1.9732, + "theoretical_loss": 4.853543506855784, + "tokens_seen": 92405760 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004909027081243731, + "loss": 2.0671, + "theoretical_loss": 4.853063916076865, + "tokens_seen": 92471296 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004908926780341023, + "loss": 2.0852, + "theoretical_loss": 4.852584760165362, + "tokens_seen": 92536832 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004908826479438315, + "loss": 2.1347, + "theoretical_loss": 4.8521060384194, + "tokens_seen": 92602368 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004908726178535607, + "loss": 2.1489, + "theoretical_loss": 4.851627750138731, + "tokens_seen": 92667904 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004908625877632899, + "loss": 2.0957, + "theoretical_loss": 4.851149894624731, + "tokens_seen": 92733440 + }, + { + "epoch": 0.03, + "learning_rate": 0.000490852557673019, + "loss": 2.1201, + "theoretical_loss": 4.8506724711803955, + "tokens_seen": 92798976 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004908425275827482, + "loss": 2.0898, + "theoretical_loss": 4.850195479110333, + "tokens_seen": 92864512 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004908324974924774, + "loss": 2.0865, + "theoretical_loss": 4.849718917720762, + "tokens_seen": 92930048 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004908224674022067, + "loss": 2.1342, + "theoretical_loss": 4.849242786319503, + "tokens_seen": 92995584 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004908124373119358, + "loss": 2.209, + "theoretical_loss": 4.848767084215979, + "tokens_seen": 93061120 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004908024072216651, + "loss": 2.0819, + "theoretical_loss": 4.848291810721204, + "tokens_seen": 93126656 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004907923771313941, + "loss": 2.0173, + "theoretical_loss": 4.847816965147783, + "tokens_seen": 93192192 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004907823470411234, + "loss": 2.1506, + "theoretical_loss": 4.847342546809907, + "tokens_seen": 93257728 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004907723169508526, + "loss": 2.083, + "theoretical_loss": 4.846868555023345, + "tokens_seen": 93323264 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004907622868605818, + "loss": 2.091, + "theoretical_loss": 4.846394989105444, + "tokens_seen": 93388800 + }, + { + "epoch": 0.03, + "learning_rate": 0.000490752256770311, + "loss": 2.1831, + "theoretical_loss": 4.845921848375118, + "tokens_seen": 93454336 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004907422266800401, + "loss": 2.0908, + "theoretical_loss": 4.845449132152851, + "tokens_seen": 93519872 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004907321965897693, + "loss": 2.1894, + "theoretical_loss": 4.844976839760683, + "tokens_seen": 93585408 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004907221664994985, + "loss": 2.1162, + "theoretical_loss": 4.844504970522217, + "tokens_seen": 93650944 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004907121364092277, + "loss": 2.1915, + "theoretical_loss": 4.844033523762603, + "tokens_seen": 93716480 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004907021063189569, + "loss": 2.1348, + "theoretical_loss": 4.843562498808538, + "tokens_seen": 93782016 + }, + { + "epoch": 0.03, + "learning_rate": 0.000490692076228686, + "loss": 2.028, + "theoretical_loss": 4.843091894988265, + "tokens_seen": 93847552 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004906820461384152, + "loss": 2.1654, + "theoretical_loss": 4.842621711631564, + "tokens_seen": 93913088 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004906720160481444, + "loss": 2.0216, + "theoretical_loss": 4.842151948069745, + "tokens_seen": 93978624 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004906619859578736, + "loss": 2.154, + "theoretical_loss": 4.84168260363565, + "tokens_seen": 94044160 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004906519558676028, + "loss": 2.0338, + "theoretical_loss": 4.841213677663648, + "tokens_seen": 94109696 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004906419257773321, + "loss": 2.1501, + "theoretical_loss": 4.840745169489621, + "tokens_seen": 94175232 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004906318956870611, + "loss": 2.0931, + "theoretical_loss": 4.840277078450969, + "tokens_seen": 94240768 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004906218655967904, + "loss": 1.9382, + "theoretical_loss": 4.839809403886606, + "tokens_seen": 94306304 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004906118355065195, + "loss": 2.1272, + "theoretical_loss": 4.839342145136947, + "tokens_seen": 94371840 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004906018054162488, + "loss": 2.0277, + "theoretical_loss": 4.838875301543913, + "tokens_seen": 94437376 + }, + { + "epoch": 0.03, + "learning_rate": 0.000490591775325978, + "loss": 2.0796, + "theoretical_loss": 4.838408872450919, + "tokens_seen": 94502912 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004905817452357072, + "loss": 2.0314, + "theoretical_loss": 4.837942857202876, + "tokens_seen": 94568448 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004905717151454363, + "loss": 2.1907, + "theoretical_loss": 4.837477255146178, + "tokens_seen": 94633984 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004905616850551655, + "loss": 2.1119, + "theoretical_loss": 4.837012065628711, + "tokens_seen": 94699520 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004905516549648947, + "loss": 2.1177, + "theoretical_loss": 4.836547287999833, + "tokens_seen": 94765056 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004905416248746239, + "loss": 2.1192, + "theoretical_loss": 4.836082921610381, + "tokens_seen": 94830592 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004905315947843531, + "loss": 2.0829, + "theoretical_loss": 4.835618965812664, + "tokens_seen": 94896128 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004905215646940822, + "loss": 2.0491, + "theoretical_loss": 4.835155419960453, + "tokens_seen": 94961664 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 42677, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9773741960525513, + "objective/train/theoretical_loss": 4.8346922834089865, + "objective/train/tokens_used": 95027200, + "theoretical_loss": 4.8346922834089865, + "tokens_seen": 95027200 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004905115346038114, + "loss": 2.0068, + "theoretical_loss": 4.8346922834089865, + "tokens_seen": 95027200 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004905015045135406, + "loss": 2.0873, + "theoretical_loss": 4.834229555514956, + "tokens_seen": 95092736 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004904914744232698, + "loss": 2.0542, + "theoretical_loss": 4.833767235636511, + "tokens_seen": 95158272 + }, + { + "epoch": 0.03, + "learning_rate": 0.000490481444332999, + "loss": 2.0188, + "theoretical_loss": 4.833305323133246, + "tokens_seen": 95223808 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004904714142427281, + "loss": 2.1815, + "theoretical_loss": 4.832843817366204, + "tokens_seen": 95289344 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004904613841524574, + "loss": 2.0665, + "theoretical_loss": 4.832382717697866, + "tokens_seen": 95354880 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004904513540621865, + "loss": 2.1686, + "theoretical_loss": 4.831922023492153, + "tokens_seen": 95420416 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004904413239719158, + "loss": 2.167, + "theoretical_loss": 4.831461734114415, + "tokens_seen": 95485952 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004904312938816449, + "loss": 2.0256, + "theoretical_loss": 4.831001848931429, + "tokens_seen": 95551488 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004904212637913742, + "loss": 2.0389, + "theoretical_loss": 4.830542367311402, + "tokens_seen": 95617024 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004904112337011033, + "loss": 2.1306, + "theoretical_loss": 4.830083288623955, + "tokens_seen": 95682560 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004904012036108325, + "loss": 2.0434, + "theoretical_loss": 4.829624612240126, + "tokens_seen": 95748096 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004903911735205617, + "loss": 2.0951, + "theoretical_loss": 4.829166337532367, + "tokens_seen": 95813632 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004903811434302909, + "loss": 2.0097, + "theoretical_loss": 4.828708463874532, + "tokens_seen": 95879168 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004903711133400201, + "loss": 2.0718, + "theoretical_loss": 4.828250990641886, + "tokens_seen": 95944704 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004903610832497492, + "loss": 2.1477, + "theoretical_loss": 4.827793917211083, + "tokens_seen": 96010240 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004903510531594784, + "loss": 2.0752, + "theoretical_loss": 4.827337242960181, + "tokens_seen": 96075776 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004903410230692076, + "loss": 2.045, + "theoretical_loss": 4.8268809672686235, + "tokens_seen": 96141312 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004903309929789368, + "loss": 2.1776, + "theoretical_loss": 4.826425089517245, + "tokens_seen": 96206848 + }, + { + "epoch": 0.03, + "learning_rate": 0.000490320962888666, + "loss": 1.9944, + "theoretical_loss": 4.825969609088258, + "tokens_seen": 96272384 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004903109327983952, + "loss": 2.0146, + "theoretical_loss": 4.825514525365258, + "tokens_seen": 96337920 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004903009027081243, + "loss": 2.0592, + "theoretical_loss": 4.825059837733214, + "tokens_seen": 96403456 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004902908726178535, + "loss": 2.0679, + "theoretical_loss": 4.824605545578463, + "tokens_seen": 96468992 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004902808425275828, + "loss": 1.9782, + "theoretical_loss": 4.824151648288714, + "tokens_seen": 96534528 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004902708124373119, + "loss": 2.1722, + "theoretical_loss": 4.823698145253035, + "tokens_seen": 96600064 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004902607823470412, + "loss": 2.0851, + "theoretical_loss": 4.823245035861852, + "tokens_seen": 96665600 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004902507522567703, + "loss": 2.0499, + "theoretical_loss": 4.822792319506952, + "tokens_seen": 96731136 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004902407221664995, + "loss": 2.0976, + "theoretical_loss": 4.822339995581466, + "tokens_seen": 96796672 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004902306920762287, + "loss": 2.055, + "theoretical_loss": 4.821888063479874, + "tokens_seen": 96862208 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004902206619859579, + "loss": 2.1186, + "theoretical_loss": 4.8214365225980025, + "tokens_seen": 96927744 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004902106318956871, + "loss": 1.9946, + "theoretical_loss": 4.8209853723330145, + "tokens_seen": 96993280 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004902006018054163, + "loss": 2.053, + "theoretical_loss": 4.820534612083408, + "tokens_seen": 97058816 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004901905717151454, + "loss": 2.151, + "theoretical_loss": 4.820084241249013, + "tokens_seen": 97124352 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004901805416248746, + "loss": 2.0912, + "theoretical_loss": 4.819634259230991, + "tokens_seen": 97189888 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004901705115346038, + "loss": 2.0306, + "theoretical_loss": 4.819184665431821, + "tokens_seen": 97255424 + }, + { + "epoch": 0.03, + "learning_rate": 0.000490160481444333, + "loss": 2.1189, + "theoretical_loss": 4.818735459255306, + "tokens_seen": 97320960 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004901504513540623, + "loss": 1.9427, + "theoretical_loss": 4.818286640106564, + "tokens_seen": 97386496 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004901404212637913, + "loss": 2.0626, + "theoretical_loss": 4.817838207392027, + "tokens_seen": 97452032 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004901303911735206, + "loss": 2.153, + "theoretical_loss": 4.817390160519435, + "tokens_seen": 97517568 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004901203610832497, + "loss": 2.0585, + "theoretical_loss": 4.816942498897831, + "tokens_seen": 97583104 + }, + { + "epoch": 0.03, + "learning_rate": 0.000490110330992979, + "loss": 2.024, + "theoretical_loss": 4.816495221937563, + "tokens_seen": 97648640 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004901003009027082, + "loss": 1.9827, + "theoretical_loss": 4.816048329050273, + "tokens_seen": 97714176 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004900902708124374, + "loss": 2.1243, + "theoretical_loss": 4.815601819648897, + "tokens_seen": 97779712 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004900802407221665, + "loss": 2.1736, + "theoretical_loss": 4.815155693147666, + "tokens_seen": 97845248 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004900702106318957, + "loss": 2.1715, + "theoretical_loss": 4.81470994896209, + "tokens_seen": 97910784 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004900601805416249, + "loss": 2.112, + "theoretical_loss": 4.8142645865089655, + "tokens_seen": 97976320 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004900501504513541, + "loss": 1.9955, + "theoretical_loss": 4.813819605206369, + "tokens_seen": 98041856 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004900401203610833, + "loss": 2.0937, + "theoretical_loss": 4.813375004473651, + "tokens_seen": 98107392 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004900300902708124, + "loss": 2.1264, + "theoretical_loss": 4.812930783731433, + "tokens_seen": 98172928 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004900200601805416, + "loss": 2.0444, + "theoretical_loss": 4.8124869424016055, + "tokens_seen": 98238464 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 43774, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.1019463539123535, + "objective/train/theoretical_loss": 4.812043479907323, + "objective/train/tokens_used": 98304000, + "theoretical_loss": 4.812043479907323, + "tokens_seen": 98304000 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004900100300902708, + "loss": 2.1098, + "theoretical_loss": 4.812043479907323, + "tokens_seen": 98304000 + }, + { + "epoch": 0.03, + "learning_rate": 0.00049, + "loss": 2.0861, + "theoretical_loss": 4.8116003956730005, + "tokens_seen": 98369536 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004899899699097292, + "loss": 2.0724, + "theoretical_loss": 4.811157689124311, + "tokens_seen": 98435072 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004899799398194583, + "loss": 2.0576, + "theoretical_loss": 4.8107153596881815, + "tokens_seen": 98500608 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004899699097291876, + "loss": 2.0329, + "theoretical_loss": 4.810273406792788, + "tokens_seen": 98566144 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004899598796389167, + "loss": 2.0608, + "theoretical_loss": 4.809831829867555, + "tokens_seen": 98631680 + }, + { + "epoch": 0.03, + "learning_rate": 0.000489949849548646, + "loss": 2.098, + "theoretical_loss": 4.809390628343146, + "tokens_seen": 98697216 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004899398194583751, + "loss": 2.048, + "theoretical_loss": 4.808949801651468, + "tokens_seen": 98762752 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004899297893681044, + "loss": 2.0827, + "theoretical_loss": 4.8085093492256625, + "tokens_seen": 98828288 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004899197592778335, + "loss": 2.1243, + "theoretical_loss": 4.808069270500102, + "tokens_seen": 98893824 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004899097291875627, + "loss": 2.0994, + "theoretical_loss": 4.80762956491039, + "tokens_seen": 98959360 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004898996990972919, + "loss": 1.9871, + "theoretical_loss": 4.807190231893355, + "tokens_seen": 99024896 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004898896690070211, + "loss": 2.0382, + "theoretical_loss": 4.806751270887048, + "tokens_seen": 99090432 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004898796389167503, + "loss": 2.1254, + "theoretical_loss": 4.806312681330736, + "tokens_seen": 99155968 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004898696088264794, + "loss": 2.0685, + "theoretical_loss": 4.805874462664903, + "tokens_seen": 99221504 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004898595787362086, + "loss": 2.0966, + "theoretical_loss": 4.805436614331247, + "tokens_seen": 99287040 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004898495486459378, + "loss": 1.9404, + "theoretical_loss": 4.804999135772668, + "tokens_seen": 99352576 + }, + { + "epoch": 0.03, + "learning_rate": 0.000489839518555667, + "loss": 2.1967, + "theoretical_loss": 4.804562026433278, + "tokens_seen": 99418112 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004898294884653962, + "loss": 2.0461, + "theoretical_loss": 4.804125285758383, + "tokens_seen": 99483648 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004898194583751254, + "loss": 2.0633, + "theoretical_loss": 4.803688913194495, + "tokens_seen": 99549184 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004898094282848545, + "loss": 2.1497, + "theoretical_loss": 4.8032529081893145, + "tokens_seen": 99614720 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004897993981945837, + "loss": 2.0278, + "theoretical_loss": 4.802817270191736, + "tokens_seen": 99680256 + }, + { + "epoch": 0.03, + "learning_rate": 0.000489789368104313, + "loss": 1.9698, + "theoretical_loss": 4.802381998651841, + "tokens_seen": 99745792 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004897793380140421, + "loss": 2.0454, + "theoretical_loss": 4.801947093020896, + "tokens_seen": 99811328 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004897693079237714, + "loss": 2.118, + "theoretical_loss": 4.801512552751349, + "tokens_seen": 99876864 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004897592778335005, + "loss": 2.0551, + "theoretical_loss": 4.801078377296827, + "tokens_seen": 99942400 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004897492477432297, + "loss": 1.9065, + "theoretical_loss": 4.800644566112128, + "tokens_seen": 100007936 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004897392176529589, + "loss": 1.9779, + "theoretical_loss": 4.800211118653225, + "tokens_seen": 100073472 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004897291875626881, + "loss": 2.0569, + "theoretical_loss": 4.799778034377258, + "tokens_seen": 100139008 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004897191574724173, + "loss": 2.0307, + "theoretical_loss": 4.799345312742531, + "tokens_seen": 100204544 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004897091273821465, + "loss": 2.0272, + "theoretical_loss": 4.798912953208511, + "tokens_seen": 100270080 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004896990972918756, + "loss": 2.191, + "theoretical_loss": 4.798480955235822, + "tokens_seen": 100335616 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004896890672016048, + "loss": 2.1439, + "theoretical_loss": 4.798049318286243, + "tokens_seen": 100401152 + }, + { + "epoch": 0.03, + "learning_rate": 0.000489679037111334, + "loss": 1.9883, + "theoretical_loss": 4.797618041822707, + "tokens_seen": 100466688 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004896690070210632, + "loss": 2.0386, + "theoretical_loss": 4.797187125309293, + "tokens_seen": 100532224 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004896589769307924, + "loss": 2.1061, + "theoretical_loss": 4.796756568211227, + "tokens_seen": 100597760 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004896489468405215, + "loss": 2.021, + "theoretical_loss": 4.796326369994876, + "tokens_seen": 100663296 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004896389167502507, + "loss": 2.0023, + "theoretical_loss": 4.795896530127747, + "tokens_seen": 100728832 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004896288866599799, + "loss": 2.0345, + "theoretical_loss": 4.795467048078484, + "tokens_seen": 100794368 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004896188565697091, + "loss": 2.122, + "theoretical_loss": 4.79503792331686, + "tokens_seen": 100859904 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004896088264794384, + "loss": 2.0401, + "theoretical_loss": 4.794609155313782, + "tokens_seen": 100925440 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004895987963891674, + "loss": 2.1504, + "theoretical_loss": 4.79418074354128, + "tokens_seen": 100990976 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004895887662988967, + "loss": 2.126, + "theoretical_loss": 4.79375268747251, + "tokens_seen": 101056512 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004895787362086259, + "loss": 1.8961, + "theoretical_loss": 4.793324986581746, + "tokens_seen": 101122048 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004895687061183551, + "loss": 2.0467, + "theoretical_loss": 4.79289764034438, + "tokens_seen": 101187584 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004895586760280843, + "loss": 2.1346, + "theoretical_loss": 4.792470648236918, + "tokens_seen": 101253120 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004895486459378135, + "loss": 2.0733, + "theoretical_loss": 4.792044009736978, + "tokens_seen": 101318656 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004895386158475426, + "loss": 1.9318, + "theoretical_loss": 4.791617724323283, + "tokens_seen": 101384192 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004895285857572718, + "loss": 2.2007, + "theoretical_loss": 4.791191791475663, + "tokens_seen": 101449728 + }, + { + "epoch": 0.03, + "learning_rate": 0.000489518555667001, + "loss": 2.0408, + "theoretical_loss": 4.79076621067505, + "tokens_seen": 101515264 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 45302, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9639794826507568, + "objective/train/theoretical_loss": 4.790340981403474, + "objective/train/tokens_used": 101580800, + "theoretical_loss": 4.790340981403474, + "tokens_seen": 101580800 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004895085255767302, + "loss": 1.9571, + "theoretical_loss": 4.790340981403474, + "tokens_seen": 101580800 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004894984954864594, + "loss": 2.0976, + "theoretical_loss": 4.7899161031440585, + "tokens_seen": 101646336 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004894884653961885, + "loss": 2.0402, + "theoretical_loss": 4.7894915753810245, + "tokens_seen": 101711872 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004894784353059178, + "loss": 2.1011, + "theoretical_loss": 4.789067397599677, + "tokens_seen": 101777408 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004894684052156469, + "loss": 2.1775, + "theoretical_loss": 4.788643569286411, + "tokens_seen": 101842944 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004894583751253762, + "loss": 2.0265, + "theoretical_loss": 4.788220089928705, + "tokens_seen": 101908480 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004894483450351053, + "loss": 2.0071, + "theoretical_loss": 4.787796959015118, + "tokens_seen": 101974016 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004894383149448346, + "loss": 2.077, + "theoretical_loss": 4.787374176035287, + "tokens_seen": 102039552 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004894282848545637, + "loss": 2.0856, + "theoretical_loss": 4.78695174047992, + "tokens_seen": 102105088 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004894182547642929, + "loss": 2.0837, + "theoretical_loss": 4.786529651840802, + "tokens_seen": 102170624 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004894082246740221, + "loss": 1.9325, + "theoretical_loss": 4.786107909610784, + "tokens_seen": 102236160 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004893981945837513, + "loss": 2.0647, + "theoretical_loss": 4.785686513283783, + "tokens_seen": 102301696 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004893881644934805, + "loss": 2.134, + "theoretical_loss": 4.785265462354779, + "tokens_seen": 102367232 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004893781344032096, + "loss": 2.1121, + "theoretical_loss": 4.7848447563198135, + "tokens_seen": 102432768 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004893681043129388, + "loss": 2.0406, + "theoretical_loss": 4.784424394675983, + "tokens_seen": 102498304 + }, + { + "epoch": 0.03, + "learning_rate": 0.000489358074222668, + "loss": 2.1261, + "theoretical_loss": 4.784004376921439, + "tokens_seen": 102563840 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004893480441323972, + "loss": 2.0262, + "theoretical_loss": 4.783584702555387, + "tokens_seen": 102629376 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004893380140421264, + "loss": 2.1245, + "theoretical_loss": 4.783165371078075, + "tokens_seen": 102694912 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004893279839518556, + "loss": 2.101, + "theoretical_loss": 4.782746381990803, + "tokens_seen": 102760448 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004893179538615847, + "loss": 2.0994, + "theoretical_loss": 4.78232773479591, + "tokens_seen": 102825984 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004893079237713139, + "loss": 2.1236, + "theoretical_loss": 4.7819094289967765, + "tokens_seen": 102891520 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004892978936810432, + "loss": 2.0508, + "theoretical_loss": 4.781491464097819, + "tokens_seen": 102957056 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004892878635907723, + "loss": 2.0715, + "theoretical_loss": 4.781073839604488, + "tokens_seen": 103022592 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004892778335005016, + "loss": 1.9398, + "theoretical_loss": 4.780656555023268, + "tokens_seen": 103088128 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004892678034102307, + "loss": 2.1241, + "theoretical_loss": 4.780239609861669, + "tokens_seen": 103153664 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004892577733199599, + "loss": 2.0938, + "theoretical_loss": 4.77982300362823, + "tokens_seen": 103219200 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004892477432296891, + "loss": 1.9826, + "theoretical_loss": 4.779406735832509, + "tokens_seen": 103284736 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004892377131394183, + "loss": 2.068, + "theoretical_loss": 4.778990805985089, + "tokens_seen": 103350272 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004892276830491475, + "loss": 2.0313, + "theoretical_loss": 4.778575213597568, + "tokens_seen": 103415808 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004892176529588767, + "loss": 1.9988, + "theoretical_loss": 4.778159958182557, + "tokens_seen": 103481344 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004892076228686058, + "loss": 1.9433, + "theoretical_loss": 4.777745039253682, + "tokens_seen": 103546880 + }, + { + "epoch": 0.03, + "learning_rate": 0.000489197592778335, + "loss": 2.217, + "theoretical_loss": 4.777330456325579, + "tokens_seen": 103612416 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004891875626880642, + "loss": 1.9752, + "theoretical_loss": 4.776916208913887, + "tokens_seen": 103677952 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004891775325977934, + "loss": 2.021, + "theoretical_loss": 4.776502296535252, + "tokens_seen": 103743488 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004891675025075226, + "loss": 2.1847, + "theoretical_loss": 4.776088718707317, + "tokens_seen": 103809024 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004891574724172517, + "loss": 2.002, + "theoretical_loss": 4.77567547494873, + "tokens_seen": 103874560 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004891474423269809, + "loss": 2.0096, + "theoretical_loss": 4.775262564779128, + "tokens_seen": 103940096 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004891374122367101, + "loss": 2.0337, + "theoretical_loss": 4.774849987719145, + "tokens_seen": 104005632 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004891273821464393, + "loss": 1.8809, + "theoretical_loss": 4.774437743290404, + "tokens_seen": 104071168 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004891173520561686, + "loss": 1.9558, + "theoretical_loss": 4.774025831015517, + "tokens_seen": 104136704 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004891073219658976, + "loss": 2.0966, + "theoretical_loss": 4.7736142504180785, + "tokens_seen": 104202240 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004890972918756269, + "loss": 2.0454, + "theoretical_loss": 4.773203001022668, + "tokens_seen": 104267776 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004890872617853561, + "loss": 2.066, + "theoretical_loss": 4.772792082354843, + "tokens_seen": 104333312 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004890772316950853, + "loss": 1.9944, + "theoretical_loss": 4.77238149394114, + "tokens_seen": 104398848 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004890672016048145, + "loss": 2.1457, + "theoretical_loss": 4.771971235309065, + "tokens_seen": 104464384 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004890571715145437, + "loss": 2.0304, + "theoretical_loss": 4.771561305987104, + "tokens_seen": 104529920 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004890471414242728, + "loss": 1.9851, + "theoretical_loss": 4.771151705504707, + "tokens_seen": 104595456 + }, + { + "epoch": 0.03, + "learning_rate": 0.000489037111334002, + "loss": 2.0363, + "theoretical_loss": 4.770742433392289, + "tokens_seen": 104660992 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004890270812437312, + "loss": 2.0583, + "theoretical_loss": 4.770333489181233, + "tokens_seen": 104726528 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004890170511534604, + "loss": 2.0439, + "theoretical_loss": 4.769924872403883, + "tokens_seen": 104792064 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 46458, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9346015453338623, + "objective/train/theoretical_loss": 4.769516582593539, + "objective/train/tokens_used": 104857600, + "theoretical_loss": 4.769516582593539, + "tokens_seen": 104857600 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004890070210631896, + "loss": 1.9328, + "theoretical_loss": 4.769516582593539, + "tokens_seen": 104857600 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004889969909729187, + "loss": 1.9266, + "theoretical_loss": 4.769108619284461, + "tokens_seen": 104923136 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004889869608826479, + "loss": 2.0159, + "theoretical_loss": 4.768700982011859, + "tokens_seen": 104988672 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004889769307923771, + "loss": 1.922, + "theoretical_loss": 4.768293670311899, + "tokens_seen": 105054208 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004889669007021063, + "loss": 2.0531, + "theoretical_loss": 4.7678866837216916, + "tokens_seen": 105119744 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004889568706118355, + "loss": 2.0119, + "theoretical_loss": 4.767480021779294, + "tokens_seen": 105185280 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004889468405215647, + "loss": 2.1521, + "theoretical_loss": 4.76707368402371, + "tokens_seen": 105250816 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004889368104312939, + "loss": 2.0455, + "theoretical_loss": 4.766667669994881, + "tokens_seen": 105316352 + }, + { + "epoch": 0.03, + "learning_rate": 0.000488926780341023, + "loss": 2.0834, + "theoretical_loss": 4.76626197923369, + "tokens_seen": 105381888 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004889167502507523, + "loss": 2.1201, + "theoretical_loss": 4.765856611281953, + "tokens_seen": 105447424 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004889067201604815, + "loss": 2.0121, + "theoretical_loss": 4.765451565682423, + "tokens_seen": 105512960 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004888966900702107, + "loss": 1.9973, + "theoretical_loss": 4.765046841978783, + "tokens_seen": 105578496 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004888866599799398, + "loss": 2.12, + "theoretical_loss": 4.764642439715642, + "tokens_seen": 105644032 + }, + { + "epoch": 0.03, + "learning_rate": 0.000488876629889669, + "loss": 2.14, + "theoretical_loss": 4.764238358438539, + "tokens_seen": 105709568 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004888665997993982, + "loss": 1.9977, + "theoretical_loss": 4.763834597693935, + "tokens_seen": 105775104 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004888565697091274, + "loss": 2.0386, + "theoretical_loss": 4.763431157029213, + "tokens_seen": 105840640 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004888465396188566, + "loss": 2.0114, + "theoretical_loss": 4.763028035992675, + "tokens_seen": 105906176 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004888365095285858, + "loss": 2.167, + "theoretical_loss": 4.762625234133537, + "tokens_seen": 105971712 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004888264794383149, + "loss": 2.0469, + "theoretical_loss": 4.762222751001933, + "tokens_seen": 106037248 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004888164493480441, + "loss": 1.9822, + "theoretical_loss": 4.761820586148906, + "tokens_seen": 106102784 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004888064192577733, + "loss": 1.9865, + "theoretical_loss": 4.761418739126407, + "tokens_seen": 106168320 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004887963891675025, + "loss": 1.9245, + "theoretical_loss": 4.761017209487298, + "tokens_seen": 106233856 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004887863590772317, + "loss": 1.9159, + "theoretical_loss": 4.760615996785342, + "tokens_seen": 106299392 + }, + { + "epoch": 0.03, + "learning_rate": 0.000488776328986961, + "loss": 2.1141, + "theoretical_loss": 4.760215100575204, + "tokens_seen": 106364928 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048876629889669, + "loss": 1.9859, + "theoretical_loss": 4.759814520412451, + "tokens_seen": 106430464 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004887562688064193, + "loss": 2.1147, + "theoretical_loss": 4.759414255853547, + "tokens_seen": 106496000 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004887462387161484, + "loss": 2.0738, + "theoretical_loss": 4.759014306455848, + "tokens_seen": 106561536 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004887362086258777, + "loss": 1.9591, + "theoretical_loss": 4.758614671777604, + "tokens_seen": 106627072 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004887261785356069, + "loss": 2.058, + "theoretical_loss": 4.758215351377956, + "tokens_seen": 106692608 + }, + { + "epoch": 0.03, + "learning_rate": 0.000488716148445336, + "loss": 2.1138, + "theoretical_loss": 4.757816344816933, + "tokens_seen": 106758144 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004887061183550652, + "loss": 1.9638, + "theoretical_loss": 4.75741765165545, + "tokens_seen": 106823680 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004886960882647944, + "loss": 1.9828, + "theoretical_loss": 4.7570192714553015, + "tokens_seen": 106889216 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004886860581745236, + "loss": 1.8841, + "theoretical_loss": 4.756621203779169, + "tokens_seen": 106954752 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004886760280842528, + "loss": 2.0968, + "theoretical_loss": 4.756223448190605, + "tokens_seen": 107020288 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004886659979939819, + "loss": 2.0304, + "theoretical_loss": 4.755826004254046, + "tokens_seen": 107085824 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004886559679037111, + "loss": 2.0026, + "theoretical_loss": 4.755428871534798, + "tokens_seen": 107151360 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004886459378134403, + "loss": 2.0144, + "theoretical_loss": 4.755032049599039, + "tokens_seen": 107216896 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004886359077231695, + "loss": 2.0271, + "theoretical_loss": 4.754635538013817, + "tokens_seen": 107282432 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004886258776328988, + "loss": 1.9333, + "theoretical_loss": 4.754239336347048, + "tokens_seen": 107347968 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004886158475426278, + "loss": 2.0751, + "theoretical_loss": 4.753843444167513, + "tokens_seen": 107413504 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004886058174523571, + "loss": 2.0111, + "theoretical_loss": 4.753447861044853, + "tokens_seen": 107479040 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004885957873620863, + "loss": 1.9335, + "theoretical_loss": 4.753052586549572, + "tokens_seen": 107544576 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004885857572718155, + "loss": 2.064, + "theoretical_loss": 4.75265762025303, + "tokens_seen": 107610112 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004885757271815447, + "loss": 2.127, + "theoretical_loss": 4.752262961727446, + "tokens_seen": 107675648 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004885656970912739, + "loss": 1.9333, + "theoretical_loss": 4.75186861054589, + "tokens_seen": 107741184 + }, + { + "epoch": 0.03, + "learning_rate": 0.000488555667001003, + "loss": 1.9635, + "theoretical_loss": 4.751474566282285, + "tokens_seen": 107806720 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004885456369107322, + "loss": 2.0054, + "theoretical_loss": 4.751080828511402, + "tokens_seen": 107872256 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004885356068204614, + "loss": 2.0062, + "theoretical_loss": 4.750687396808859, + "tokens_seen": 107937792 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004885255767301906, + "loss": 2.1271, + "theoretical_loss": 4.7502942707511195, + "tokens_seen": 108003328 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004885155466399198, + "loss": 2.0457, + "theoretical_loss": 4.749901449915489, + "tokens_seen": 108068864 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 47636, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.946031093597412, + "objective/train/theoretical_loss": 4.749508933880117, + "objective/train/tokens_used": 108134400, + "theoretical_loss": 4.749508933880117, + "tokens_seen": 108134400 + }, + { + "epoch": 0.03, + "learning_rate": 0.000488505516549649, + "loss": 2.0022, + "theoretical_loss": 4.749508933880117, + "tokens_seen": 108134400 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004884954864593781, + "loss": 2.0412, + "theoretical_loss": 4.7491167222239845, + "tokens_seen": 108199936 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004884854563691073, + "loss": 1.9925, + "theoretical_loss": 4.7487248145269145, + "tokens_seen": 108265472 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004884754262788365, + "loss": 2.0008, + "theoretical_loss": 4.7483332103695615, + "tokens_seen": 108331008 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004884653961885657, + "loss": 1.9676, + "theoretical_loss": 4.747941909333411, + "tokens_seen": 108396544 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004884553660982949, + "loss": 1.9569, + "theoretical_loss": 4.747550911000781, + "tokens_seen": 108462080 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004884453360080241, + "loss": 2.013, + "theoretical_loss": 4.747160214954813, + "tokens_seen": 108527616 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004884353059177532, + "loss": 2.0776, + "theoretical_loss": 4.746769820779477, + "tokens_seen": 108593152 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004884252758274825, + "loss": 1.9755, + "theoretical_loss": 4.746379728059566, + "tokens_seen": 108658688 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004884152457372117, + "loss": 2.0054, + "theoretical_loss": 4.745989936380692, + "tokens_seen": 108724224 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004884052156469409, + "loss": 1.9479, + "theoretical_loss": 4.745600445329288, + "tokens_seen": 108789760 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048839518555667, + "loss": 2.0916, + "theoretical_loss": 4.745211254492602, + "tokens_seen": 108855296 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004883851554663992, + "loss": 2.0613, + "theoretical_loss": 4.744822363458699, + "tokens_seen": 108920832 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004883751253761284, + "loss": 1.9014, + "theoretical_loss": 4.744433771816455, + "tokens_seen": 108986368 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004883650952858576, + "loss": 1.948, + "theoretical_loss": 4.744045479155557, + "tokens_seen": 109051904 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004883550651955868, + "loss": 1.8969, + "theoretical_loss": 4.7436574850665005, + "tokens_seen": 109117440 + }, + { + "epoch": 0.03, + "learning_rate": 0.000488345035105316, + "loss": 2.1114, + "theoretical_loss": 4.743269789140587, + "tokens_seen": 109182976 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004883350050150451, + "loss": 1.8744, + "theoretical_loss": 4.742882390969925, + "tokens_seen": 109248512 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004883249749247743, + "loss": 2.0422, + "theoretical_loss": 4.742495290147421, + "tokens_seen": 109314048 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004883149448345035, + "loss": 1.9831, + "theoretical_loss": 4.742108486266783, + "tokens_seen": 109379584 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004883049147442327, + "loss": 2.1017, + "theoretical_loss": 4.74172197892252, + "tokens_seen": 109445120 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004882948846539619, + "loss": 1.9557, + "theoretical_loss": 4.741335767709934, + "tokens_seen": 109510656 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004882848545636911, + "loss": 2.0173, + "theoretical_loss": 4.740949852225122, + "tokens_seen": 109576192 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004882748244734203, + "loss": 1.9511, + "theoretical_loss": 4.740564232064973, + "tokens_seen": 109641728 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048826479438314946, + "loss": 2.0965, + "theoretical_loss": 4.7401789068271665, + "tokens_seen": 109707264 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048825476429287864, + "loss": 2.045, + "theoretical_loss": 4.73979387611017, + "tokens_seen": 109772800 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004882447342026078, + "loss": 1.8994, + "theoretical_loss": 4.739409139513235, + "tokens_seen": 109838336 + }, + { + "epoch": 0.03, + "learning_rate": 0.000488234704112337, + "loss": 2.0311, + "theoretical_loss": 4.739024696636401, + "tokens_seen": 109903872 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048822467402206624, + "loss": 2.0434, + "theoretical_loss": 4.738640547080485, + "tokens_seen": 109969408 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048821464393179536, + "loss": 1.9664, + "theoretical_loss": 4.738256690447088, + "tokens_seen": 110034944 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004882046138415246, + "loss": 1.9871, + "theoretical_loss": 4.7378731263385845, + "tokens_seen": 110100480 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004881945837512537, + "loss": 2.0606, + "theoretical_loss": 4.737489854358129, + "tokens_seen": 110166016 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048818455366098296, + "loss": 2.176, + "theoretical_loss": 4.737106874109649, + "tokens_seen": 110231552 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048817452357071214, + "loss": 2.0162, + "theoretical_loss": 4.736724185197844, + "tokens_seen": 110297088 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004881644934804413, + "loss": 1.9145, + "theoretical_loss": 4.7363417872281826, + "tokens_seen": 110362624 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004881544633901705, + "loss": 1.9693, + "theoretical_loss": 4.735959679806903, + "tokens_seen": 110428160 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048814443329989974, + "loss": 1.9105, + "theoretical_loss": 4.735577862541007, + "tokens_seen": 110493696 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048813440320962887, + "loss": 1.9061, + "theoretical_loss": 4.735196335038264, + "tokens_seen": 110559232 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004881243731193581, + "loss": 1.9451, + "theoretical_loss": 4.734815096907203, + "tokens_seen": 110624768 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048811434302908723, + "loss": 1.9437, + "theoretical_loss": 4.734434147757117, + "tokens_seen": 110690304 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048810431293881646, + "loss": 2.0108, + "theoretical_loss": 4.734053487198054, + "tokens_seen": 110755840 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048809428284854564, + "loss": 1.8542, + "theoretical_loss": 4.733673114840817, + "tokens_seen": 110821376 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004880842527582748, + "loss": 1.9158, + "theoretical_loss": 4.73329303029697, + "tokens_seen": 110886912 + }, + { + "epoch": 0.03, + "learning_rate": 0.000488074222668004, + "loss": 1.9522, + "theoretical_loss": 4.732913233178825, + "tokens_seen": 110952448 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004880641925777332, + "loss": 2.0889, + "theoretical_loss": 4.732533723099444, + "tokens_seen": 111017984 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048805416248746237, + "loss": 2.0213, + "theoretical_loss": 4.732154499672642, + "tokens_seen": 111083520 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004880441323971916, + "loss": 2.0035, + "theoretical_loss": 4.731775562512977, + "tokens_seen": 111149056 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048803410230692073, + "loss": 2.0706, + "theoretical_loss": 4.7313969112357555, + "tokens_seen": 111214592 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048802407221664997, + "loss": 2.0833, + "theoretical_loss": 4.731018545457024, + "tokens_seen": 111280128 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048801404212637915, + "loss": 1.9794, + "theoretical_loss": 4.730640464793574, + "tokens_seen": 111345664 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 48676, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.0878026485443115, + "objective/train/theoretical_loss": 4.730262668862931, + "objective/train/tokens_used": 111411200, + "theoretical_loss": 4.730262668862931, + "tokens_seen": 111411200 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048800401203610833, + "loss": 2.0542, + "theoretical_loss": 4.730262668862931, + "tokens_seen": 111411200 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004879939819458375, + "loss": 1.9923, + "theoretical_loss": 4.729885157283366, + "tokens_seen": 111476736 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004879839518555667, + "loss": 1.9879, + "theoretical_loss": 4.729507929673879, + "tokens_seen": 111542272 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004879739217652959, + "loss": 2.0555, + "theoretical_loss": 4.729130985654208, + "tokens_seen": 111607808 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004879638916750251, + "loss": 1.9474, + "theoretical_loss": 4.72875432484482, + "tokens_seen": 111673344 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004879538615847543, + "loss": 2.0365, + "theoretical_loss": 4.728377946866915, + "tokens_seen": 111738880 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048794383149448347, + "loss": 2.0074, + "theoretical_loss": 4.72800185134242, + "tokens_seen": 111804416 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048793380140421265, + "loss": 2.001, + "theoretical_loss": 4.727626037893989, + "tokens_seen": 111869952 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048792377131394183, + "loss": 2.0327, + "theoretical_loss": 4.727250506145001, + "tokens_seen": 111935488 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048791374122367107, + "loss": 2.0094, + "theoretical_loss": 4.726875255719557, + "tokens_seen": 112001024 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004879037111334002, + "loss": 1.8166, + "theoretical_loss": 4.726500286242482, + "tokens_seen": 112066560 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048789368104312943, + "loss": 1.9677, + "theoretical_loss": 4.7261255973393155, + "tokens_seen": 112132096 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048788365095285856, + "loss": 2.0293, + "theoretical_loss": 4.725751188636319, + "tokens_seen": 112197632 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004878736208625878, + "loss": 1.9929, + "theoretical_loss": 4.725377059760468, + "tokens_seen": 112263168 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048786359077231697, + "loss": 1.9674, + "theoretical_loss": 4.725003210339452, + "tokens_seen": 112328704 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048785356068204615, + "loss": 1.9573, + "theoretical_loss": 4.724629640001672, + "tokens_seen": 112394240 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048784353059177533, + "loss": 2.011, + "theoretical_loss": 4.724256348376242, + "tokens_seen": 112459776 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048783350050150457, + "loss": 1.962, + "theoretical_loss": 4.723883335092982, + "tokens_seen": 112525312 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004878234704112337, + "loss": 1.9383, + "theoretical_loss": 4.723510599782419, + "tokens_seen": 112590848 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048781344032096293, + "loss": 1.9819, + "theoretical_loss": 4.723138142075788, + "tokens_seen": 112656384 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048780341023069206, + "loss": 2.0332, + "theoretical_loss": 4.722765961605024, + "tokens_seen": 112721920 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004877933801404213, + "loss": 1.9097, + "theoretical_loss": 4.722394058002765, + "tokens_seen": 112787456 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004877833500501505, + "loss": 1.966, + "theoretical_loss": 4.7220224309023475, + "tokens_seen": 112852992 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048777331995987966, + "loss": 2.0126, + "theoretical_loss": 4.721651079937811, + "tokens_seen": 112918528 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048776328986960884, + "loss": 2.0137, + "theoretical_loss": 4.7212800047438845, + "tokens_seen": 112984064 + }, + { + "epoch": 0.03, + "learning_rate": 0.000487753259779338, + "loss": 2.0949, + "theoretical_loss": 4.720909204955996, + "tokens_seen": 113049600 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004877432296890672, + "loss": 2.1029, + "theoretical_loss": 4.720538680210268, + "tokens_seen": 113115136 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048773319959879644, + "loss": 2.0512, + "theoretical_loss": 4.720168430143508, + "tokens_seen": 113180672 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048772316950852556, + "loss": 1.9723, + "theoretical_loss": 4.719798454393219, + "tokens_seen": 113246208 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004877131394182548, + "loss": 2.0, + "theoretical_loss": 4.719428752597588, + "tokens_seen": 113311744 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004877031093279839, + "loss": 1.9745, + "theoretical_loss": 4.719059324395493, + "tokens_seen": 113377280 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048769307923771316, + "loss": 1.9602, + "theoretical_loss": 4.71869016942649, + "tokens_seen": 113442816 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048768304914744234, + "loss": 2.0744, + "theoretical_loss": 4.718321287330823, + "tokens_seen": 113508352 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004876730190571715, + "loss": 2.076, + "theoretical_loss": 4.7179526777494125, + "tokens_seen": 113573888 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004876629889669007, + "loss": 1.969, + "theoretical_loss": 4.717584340323864, + "tokens_seen": 113639424 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048765295887662994, + "loss": 2.001, + "theoretical_loss": 4.7172162746964545, + "tokens_seen": 113704960 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048764292878635907, + "loss": 2.1638, + "theoretical_loss": 4.716848480510142, + "tokens_seen": 113770496 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004876328986960883, + "loss": 1.933, + "theoretical_loss": 4.716480957408556, + "tokens_seen": 113836032 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048762286860581743, + "loss": 1.9948, + "theoretical_loss": 4.716113705035999, + "tokens_seen": 113901568 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048761283851554666, + "loss": 1.9815, + "theoretical_loss": 4.715746723037448, + "tokens_seen": 113967104 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048760280842527584, + "loss": 1.9997, + "theoretical_loss": 4.715380011058542, + "tokens_seen": 114032640 + }, + { + "epoch": 0.03, + "learning_rate": 0.000487592778335005, + "loss": 2.0318, + "theoretical_loss": 4.715013568745597, + "tokens_seen": 114098176 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004875827482447342, + "loss": 1.9958, + "theoretical_loss": 4.714647395745587, + "tokens_seen": 114163712 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004875727181544634, + "loss": 2.0038, + "theoretical_loss": 4.714281491706154, + "tokens_seen": 114229248 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048756268806419257, + "loss": 2.1302, + "theoretical_loss": 4.713915856275603, + "tokens_seen": 114294784 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004875526579739218, + "loss": 1.8952, + "theoretical_loss": 4.713550489102901, + "tokens_seen": 114360320 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048754262788365093, + "loss": 1.8787, + "theoretical_loss": 4.7131853898376725, + "tokens_seen": 114425856 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048753259779338017, + "loss": 1.8883, + "theoretical_loss": 4.7128205581302005, + "tokens_seen": 114491392 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048752256770310935, + "loss": 1.8831, + "theoretical_loss": 4.712455993631425, + "tokens_seen": 114556928 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048751253761283853, + "loss": 2.0418, + "theoretical_loss": 4.712091695992941, + "tokens_seen": 114622464 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 49853, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.0042688846588135, + "objective/train/theoretical_loss": 4.711727664866995, + "objective/train/tokens_used": 114688000, + "theoretical_loss": 4.711727664866995, + "tokens_seen": 114688000 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004875025075225677, + "loss": 1.9854, + "theoretical_loss": 4.711727664866995, + "tokens_seen": 114688000 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004874924774322969, + "loss": 2.0059, + "theoretical_loss": 4.7113638999064875, + "tokens_seen": 114753536 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048748244734202607, + "loss": 2.0743, + "theoretical_loss": 4.711000400764967, + "tokens_seen": 114819072 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004874724172517553, + "loss": 2.0879, + "theoretical_loss": 4.71063716709663, + "tokens_seen": 114884608 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048746238716148443, + "loss": 2.1133, + "theoretical_loss": 4.7102741985563235, + "tokens_seen": 114950144 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048745235707121367, + "loss": 2.0205, + "theoretical_loss": 4.709911494799533, + "tokens_seen": 115015680 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004874423269809428, + "loss": 1.9817, + "theoretical_loss": 4.709549055482395, + "tokens_seen": 115081216 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048743229689067203, + "loss": 1.8978, + "theoretical_loss": 4.709186880261683, + "tokens_seen": 115146752 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004874222668004012, + "loss": 1.8766, + "theoretical_loss": 4.708824968794813, + "tokens_seen": 115212288 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004874122367101304, + "loss": 1.9197, + "theoretical_loss": 4.70846332073984, + "tokens_seen": 115277824 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004874022066198596, + "loss": 2.0514, + "theoretical_loss": 4.708101935755455, + "tokens_seen": 115343360 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048739217652958876, + "loss": 1.9902, + "theoretical_loss": 4.707740813500988, + "tokens_seen": 115408896 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048738214643931794, + "loss": 1.9994, + "theoretical_loss": 4.707379953636399, + "tokens_seen": 115474432 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048737211634904717, + "loss": 1.9816, + "theoretical_loss": 4.7070193558222835, + "tokens_seen": 115539968 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004873620862587763, + "loss": 1.9199, + "theoretical_loss": 4.7066590197198686, + "tokens_seen": 115605504 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048735205616850553, + "loss": 2.0137, + "theoretical_loss": 4.706298944991008, + "tokens_seen": 115671040 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004873420260782347, + "loss": 2.0398, + "theoretical_loss": 4.705939131298189, + "tokens_seen": 115736576 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004873319959879639, + "loss": 1.9798, + "theoretical_loss": 4.705579578304519, + "tokens_seen": 115802112 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004873219658976931, + "loss": 2.0628, + "theoretical_loss": 4.705220285673736, + "tokens_seen": 115867648 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048731193580742226, + "loss": 2.0168, + "theoretical_loss": 4.704861253070198, + "tokens_seen": 115933184 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048730190571715144, + "loss": 1.9366, + "theoretical_loss": 4.704502480158887, + "tokens_seen": 115998720 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004872918756268807, + "loss": 1.8443, + "theoretical_loss": 4.704143966605406, + "tokens_seen": 116064256 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004872818455366098, + "loss": 2.0006, + "theoretical_loss": 4.703785712075973, + "tokens_seen": 116129792 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048727181544633904, + "loss": 1.9869, + "theoretical_loss": 4.703427716237432, + "tokens_seen": 116195328 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048726178535606816, + "loss": 1.9807, + "theoretical_loss": 4.703069978757233, + "tokens_seen": 116260864 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004872517552657974, + "loss": 1.8592, + "theoretical_loss": 4.702712499303448, + "tokens_seen": 116326400 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004872417251755266, + "loss": 2.0154, + "theoretical_loss": 4.7023552775447595, + "tokens_seen": 116391936 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048723169508525576, + "loss": 1.9726, + "theoretical_loss": 4.701998313150462, + "tokens_seen": 116457472 + }, + { + "epoch": 0.04, + "learning_rate": 0.000487221664994985, + "loss": 2.1186, + "theoretical_loss": 4.701641605790459, + "tokens_seen": 116523008 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004872116349047141, + "loss": 1.8103, + "theoretical_loss": 4.701285155135265, + "tokens_seen": 116588544 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048720160481444336, + "loss": 2.0743, + "theoretical_loss": 4.700928960856, + "tokens_seen": 116654080 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048719157472417254, + "loss": 2.0605, + "theoretical_loss": 4.70057302262439, + "tokens_seen": 116719616 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004871815446339017, + "loss": 1.9836, + "theoretical_loss": 4.700217340112767, + "tokens_seen": 116785152 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004871715145436309, + "loss": 1.9692, + "theoretical_loss": 4.699861912994063, + "tokens_seen": 116850688 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048716148445336014, + "loss": 1.8192, + "theoretical_loss": 4.699506740941813, + "tokens_seen": 116916224 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048715145436308927, + "loss": 2.1421, + "theoretical_loss": 4.699151823630155, + "tokens_seen": 116981760 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004871414242728185, + "loss": 1.9415, + "theoretical_loss": 4.698797160733818, + "tokens_seen": 117047296 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048713139418254763, + "loss": 2.0214, + "theoretical_loss": 4.6984427519281375, + "tokens_seen": 117112832 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048712136409227686, + "loss": 1.9676, + "theoretical_loss": 4.698088596889037, + "tokens_seen": 117178368 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048711133400200604, + "loss": 2.085, + "theoretical_loss": 4.697734695293039, + "tokens_seen": 117243904 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004871013039117352, + "loss": 1.9363, + "theoretical_loss": 4.697381046817256, + "tokens_seen": 117309440 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004870912738214644, + "loss": 1.914, + "theoretical_loss": 4.697027651139393, + "tokens_seen": 117374976 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004870812437311936, + "loss": 2.1255, + "theoretical_loss": 4.696674507937746, + "tokens_seen": 117440512 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048707121364092277, + "loss": 1.9092, + "theoretical_loss": 4.696321616891199, + "tokens_seen": 117506048 + }, + { + "epoch": 0.04, + "learning_rate": 0.000487061183550652, + "loss": 1.9909, + "theoretical_loss": 4.695968977679221, + "tokens_seen": 117571584 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048705115346038113, + "loss": 1.9843, + "theoretical_loss": 4.695616589981871, + "tokens_seen": 117637120 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048704112337011037, + "loss": 2.0604, + "theoretical_loss": 4.695264453479789, + "tokens_seen": 117702656 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048703109327983955, + "loss": 1.9788, + "theoretical_loss": 4.6949125678542, + "tokens_seen": 117768192 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048702106318956873, + "loss": 1.9714, + "theoretical_loss": 4.69456093278691, + "tokens_seen": 117833728 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004870110330992979, + "loss": 2.0216, + "theoretical_loss": 4.694209547960304, + "tokens_seen": 117899264 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 50996, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.982438564300537, + "objective/train/theoretical_loss": 4.693858413057347, + "objective/train/tokens_used": 117964800, + "theoretical_loss": 4.693858413057347, + "tokens_seen": 117964800 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004870010030090271, + "loss": 2.038, + "theoretical_loss": 4.693858413057347, + "tokens_seen": 117964800 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048699097291875627, + "loss": 1.8832, + "theoretical_loss": 4.693507527761584, + "tokens_seen": 118030336 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004869809428284855, + "loss": 1.9547, + "theoretical_loss": 4.6931568917571305, + "tokens_seen": 118095872 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048697091273821463, + "loss": 2.056, + "theoretical_loss": 4.692806504728683, + "tokens_seen": 118161408 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048696088264794387, + "loss": 2.0869, + "theoretical_loss": 4.692456366361506, + "tokens_seen": 118226944 + }, + { + "epoch": 0.04, + "learning_rate": 0.000486950852557673, + "loss": 1.8917, + "theoretical_loss": 4.692106476341438, + "tokens_seen": 118292480 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048694082246740223, + "loss": 2.0569, + "theoretical_loss": 4.691756834354891, + "tokens_seen": 118358016 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004869307923771314, + "loss": 1.9595, + "theoretical_loss": 4.691407440088843, + "tokens_seen": 118423552 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004869207622868606, + "loss": 1.9308, + "theoretical_loss": 4.69105829323084, + "tokens_seen": 118489088 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004869107321965898, + "loss": 2.0626, + "theoretical_loss": 4.690709393468995, + "tokens_seen": 118554624 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048690070210631896, + "loss": 2.0201, + "theoretical_loss": 4.690360740491989, + "tokens_seen": 118620160 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048689067201604814, + "loss": 2.0316, + "theoretical_loss": 4.690012333989063, + "tokens_seen": 118685696 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048688064192577737, + "loss": 2.1201, + "theoretical_loss": 4.689664173650023, + "tokens_seen": 118751232 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004868706118355065, + "loss": 1.8783, + "theoretical_loss": 4.689316259165235, + "tokens_seen": 118816768 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048686058174523573, + "loss": 1.949, + "theoretical_loss": 4.6889685902256275, + "tokens_seen": 118882304 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004868505516549649, + "loss": 1.9981, + "theoretical_loss": 4.688621166522684, + "tokens_seen": 118947840 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004868405215646941, + "loss": 2.1063, + "theoretical_loss": 4.688273987748448, + "tokens_seen": 119013376 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004868304914744233, + "loss": 2.0209, + "theoretical_loss": 4.687927053595518, + "tokens_seen": 119078912 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048682046138415246, + "loss": 1.9411, + "theoretical_loss": 4.687580363757048, + "tokens_seen": 119144448 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048681043129388164, + "loss": 1.9227, + "theoretical_loss": 4.6872339179267435, + "tokens_seen": 119209984 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004868004012036109, + "loss": 1.9659, + "theoretical_loss": 4.686887715798864, + "tokens_seen": 119275520 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048679037111334, + "loss": 2.1, + "theoretical_loss": 4.6865417570682215, + "tokens_seen": 119341056 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048678034102306924, + "loss": 2.0249, + "theoretical_loss": 4.686196041430172, + "tokens_seen": 119406592 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048677031093279836, + "loss": 1.9004, + "theoretical_loss": 4.685850568580624, + "tokens_seen": 119472128 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004867602808425276, + "loss": 1.9391, + "theoretical_loss": 4.685505338216032, + "tokens_seen": 119537664 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004867502507522568, + "loss": 2.132, + "theoretical_loss": 4.685160350033398, + "tokens_seen": 119603200 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048674022066198596, + "loss": 2.0349, + "theoretical_loss": 4.684815603730263, + "tokens_seen": 119668736 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048673019057171514, + "loss": 2.0118, + "theoretical_loss": 4.684471099004718, + "tokens_seen": 119734272 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004867201604814443, + "loss": 2.0517, + "theoretical_loss": 4.684126835555389, + "tokens_seen": 119799808 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004867101303911735, + "loss": 2.0236, + "theoretical_loss": 4.683782813081448, + "tokens_seen": 119865344 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048670010030090274, + "loss": 1.9552, + "theoretical_loss": 4.683439031282606, + "tokens_seen": 119930880 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048669007021063187, + "loss": 1.9045, + "theoretical_loss": 4.683095489859107, + "tokens_seen": 119996416 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004866800401203611, + "loss": 1.9687, + "theoretical_loss": 4.682752188511737, + "tokens_seen": 120061952 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004866700100300903, + "loss": 1.9969, + "theoretical_loss": 4.682409126941815, + "tokens_seen": 120127488 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048665997993981947, + "loss": 2.0643, + "theoretical_loss": 4.682066304851195, + "tokens_seen": 120193024 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048664994984954865, + "loss": 1.9536, + "theoretical_loss": 4.681723721942264, + "tokens_seen": 120258560 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048663991975927783, + "loss": 1.9796, + "theoretical_loss": 4.6813813779179405, + "tokens_seen": 120324096 + }, + { + "epoch": 0.04, + "learning_rate": 0.000486629889669007, + "loss": 2.0763, + "theoretical_loss": 4.681039272481674, + "tokens_seen": 120389632 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048661985957873624, + "loss": 2.1269, + "theoretical_loss": 4.680697405337443, + "tokens_seen": 120455168 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048660982948846537, + "loss": 1.9512, + "theoretical_loss": 4.680355776189755, + "tokens_seen": 120520704 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004865997993981946, + "loss": 1.8279, + "theoretical_loss": 4.680014384743641, + "tokens_seen": 120586240 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048658976930792373, + "loss": 1.9701, + "theoretical_loss": 4.679673230704664, + "tokens_seen": 120651776 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048657973921765297, + "loss": 1.8816, + "theoretical_loss": 4.679332313778904, + "tokens_seen": 120717312 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048656970912738215, + "loss": 2.054, + "theoretical_loss": 4.678991633672968, + "tokens_seen": 120782848 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048655967903711133, + "loss": 1.8939, + "theoretical_loss": 4.678651190093988, + "tokens_seen": 120848384 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004865496489468405, + "loss": 1.9776, + "theoretical_loss": 4.67831098274961, + "tokens_seen": 120913920 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048653961885656975, + "loss": 1.9649, + "theoretical_loss": 4.677971011348005, + "tokens_seen": 120979456 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004865295887662989, + "loss": 1.9983, + "theoretical_loss": 4.677631275597859, + "tokens_seen": 121044992 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004865195586760281, + "loss": 1.9257, + "theoretical_loss": 4.677291775208378, + "tokens_seen": 121110528 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048650952858575724, + "loss": 1.9454, + "theoretical_loss": 4.676952509889281, + "tokens_seen": 121176064 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 52104, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9716572761535645, + "objective/train/theoretical_loss": 4.676613479350804, + "objective/train/tokens_used": 121241600, + "theoretical_loss": 4.676613479350804, + "tokens_seen": 121241600 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048649949849548647, + "loss": 1.977, + "theoretical_loss": 4.676613479350804, + "tokens_seen": 121241600 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048648946840521565, + "loss": 1.8962, + "theoretical_loss": 4.6762746833036966, + "tokens_seen": 121307136 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048647943831494483, + "loss": 1.9531, + "theoretical_loss": 4.675936121459218, + "tokens_seen": 121372672 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048646940822467407, + "loss": 1.8941, + "theoretical_loss": 4.675597793529143, + "tokens_seen": 121438208 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004864593781344032, + "loss": 1.9296, + "theoretical_loss": 4.675259699225752, + "tokens_seen": 121503744 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048644934804413243, + "loss": 1.8856, + "theoretical_loss": 4.674921838261836, + "tokens_seen": 121569280 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004864393179538616, + "loss": 1.9628, + "theoretical_loss": 4.674584210350695, + "tokens_seen": 121634816 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004864292878635908, + "loss": 1.8942, + "theoretical_loss": 4.674246815206136, + "tokens_seen": 121700352 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048641925777332, + "loss": 1.8336, + "theoretical_loss": 4.673909652542466, + "tokens_seen": 121765888 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048640922768304916, + "loss": 2.0095, + "theoretical_loss": 4.673572722074502, + "tokens_seen": 121831424 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048639919759277834, + "loss": 2.0737, + "theoretical_loss": 4.673236023517562, + "tokens_seen": 121896960 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048638916750250757, + "loss": 1.9092, + "theoretical_loss": 4.672899556587465, + "tokens_seen": 121962496 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004863791374122367, + "loss": 1.9355, + "theoretical_loss": 4.672563321000533, + "tokens_seen": 122028032 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048636910732196593, + "loss": 2.0114, + "theoretical_loss": 4.672227316473584, + "tokens_seen": 122093568 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004863590772316951, + "loss": 1.9386, + "theoretical_loss": 4.671891542723937, + "tokens_seen": 122159104 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004863490471414243, + "loss": 2.0975, + "theoretical_loss": 4.671555999469408, + "tokens_seen": 122224640 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004863390170511535, + "loss": 2.0106, + "theoretical_loss": 4.671220686428308, + "tokens_seen": 122290176 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048632898696088266, + "loss": 1.9436, + "theoretical_loss": 4.670885603319445, + "tokens_seen": 122355712 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048631895687061184, + "loss": 1.9429, + "theoretical_loss": 4.670550749862119, + "tokens_seen": 122421248 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004863089267803411, + "loss": 1.8952, + "theoretical_loss": 4.670216125776122, + "tokens_seen": 122486784 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004862988966900702, + "loss": 1.8632, + "theoretical_loss": 4.669881730781741, + "tokens_seen": 122552320 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048628886659979944, + "loss": 1.9671, + "theoretical_loss": 4.66954756459975, + "tokens_seen": 122617856 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048627883650952857, + "loss": 1.985, + "theoretical_loss": 4.669213626951413, + "tokens_seen": 122683392 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004862688064192578, + "loss": 2.1293, + "theoretical_loss": 4.668879917558484, + "tokens_seen": 122748928 + }, + { + "epoch": 0.04, + "learning_rate": 0.000486258776328987, + "loss": 1.9738, + "theoretical_loss": 4.6685464361432025, + "tokens_seen": 122814464 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048624874623871616, + "loss": 2.0038, + "theoretical_loss": 4.668213182428295, + "tokens_seen": 122880000 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048623871614844534, + "loss": 1.9044, + "theoretical_loss": 4.667880156136971, + "tokens_seen": 122945536 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004862286860581745, + "loss": 1.9351, + "theoretical_loss": 4.667547356992927, + "tokens_seen": 123011072 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004862186559679037, + "loss": 1.9478, + "theoretical_loss": 4.667214784720338, + "tokens_seen": 123076608 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048620862587763294, + "loss": 1.9815, + "theoretical_loss": 4.666882439043865, + "tokens_seen": 123142144 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048619859578736207, + "loss": 2.0171, + "theoretical_loss": 4.666550319688646, + "tokens_seen": 123207680 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004861885656970913, + "loss": 1.9268, + "theoretical_loss": 4.6662184263803, + "tokens_seen": 123273216 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004861785356068205, + "loss": 2.0015, + "theoretical_loss": 4.665886758844923, + "tokens_seen": 123338752 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048616850551654967, + "loss": 1.9936, + "theoretical_loss": 4.66555531680909, + "tokens_seen": 123404288 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048615847542627885, + "loss": 1.9185, + "theoretical_loss": 4.665224099999852, + "tokens_seen": 123469824 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048614844533600803, + "loss": 1.9321, + "theoretical_loss": 4.664893108144735, + "tokens_seen": 123535360 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004861384152457372, + "loss": 1.8148, + "theoretical_loss": 4.664562340971735, + "tokens_seen": 123600896 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048612838515546644, + "loss": 2.1466, + "theoretical_loss": 4.664231798209327, + "tokens_seen": 123666432 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048611835506519557, + "loss": 1.7999, + "theoretical_loss": 4.663901479586454, + "tokens_seen": 123731968 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004861083249749248, + "loss": 2.0488, + "theoretical_loss": 4.663571384832531, + "tokens_seen": 123797504 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048609829488465393, + "loss": 1.9731, + "theoretical_loss": 4.663241513677443, + "tokens_seen": 123863040 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048608826479438317, + "loss": 1.9273, + "theoretical_loss": 4.662911865851541, + "tokens_seen": 123928576 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048607823470411235, + "loss": 2.0331, + "theoretical_loss": 4.662582441085648, + "tokens_seen": 123994112 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048606820461384153, + "loss": 1.9859, + "theoretical_loss": 4.662253239111049, + "tokens_seen": 124059648 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004860581745235707, + "loss": 2.0166, + "theoretical_loss": 4.661924259659497, + "tokens_seen": 124125184 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048604814443329995, + "loss": 1.9894, + "theoretical_loss": 4.661595502463211, + "tokens_seen": 124190720 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004860381143430291, + "loss": 1.9532, + "theoretical_loss": 4.661266967254868, + "tokens_seen": 124256256 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004860280842527583, + "loss": 2.0134, + "theoretical_loss": 4.660938653767614, + "tokens_seen": 124321792 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048601805416248744, + "loss": 1.8968, + "theoretical_loss": 4.660610561735049, + "tokens_seen": 124387328 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048600802407221667, + "loss": 2.1135, + "theoretical_loss": 4.66028269089124, + "tokens_seen": 124452864 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 53273, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.0666775703430176, + "objective/train/theoretical_loss": 4.659955040970709, + "objective/train/tokens_used": 124518400, + "theoretical_loss": 4.659955040970709, + "tokens_seen": 124518400 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048599799398194585, + "loss": 2.0576, + "theoretical_loss": 4.659955040970709, + "tokens_seen": 124518400 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048598796389167503, + "loss": 1.8728, + "theoretical_loss": 4.659627611708437, + "tokens_seen": 124583936 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004859779338014042, + "loss": 1.9877, + "theoretical_loss": 4.659300402839863, + "tokens_seen": 124649472 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004859679037111334, + "loss": 1.9228, + "theoretical_loss": 4.658973414100881, + "tokens_seen": 124715008 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004859578736208626, + "loss": 2.0716, + "theoretical_loss": 4.658646645227842, + "tokens_seen": 124780544 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004859478435305918, + "loss": 1.9186, + "theoretical_loss": 4.6583200959575475, + "tokens_seen": 124846080 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048593781344032094, + "loss": 2.05, + "theoretical_loss": 4.657993766027255, + "tokens_seen": 124911616 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004859277833500502, + "loss": 1.998, + "theoretical_loss": 4.657667655174674, + "tokens_seen": 124977152 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004859177532597793, + "loss": 1.8412, + "theoretical_loss": 4.657341763137963, + "tokens_seen": 125042688 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048590772316950854, + "loss": 1.9765, + "theoretical_loss": 4.657016089655732, + "tokens_seen": 125108224 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004858976930792377, + "loss": 1.9402, + "theoretical_loss": 4.65669063446704, + "tokens_seen": 125173760 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004858876629889669, + "loss": 1.8269, + "theoretical_loss": 4.656365397311393, + "tokens_seen": 125239296 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004858776328986961, + "loss": 1.8564, + "theoretical_loss": 4.656040377928743, + "tokens_seen": 125304832 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004858676028084253, + "loss": 1.745, + "theoretical_loss": 4.655715576059493, + "tokens_seen": 125370368 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048585757271815444, + "loss": 1.9644, + "theoretical_loss": 4.655390991444484, + "tokens_seen": 125435904 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004858475426278837, + "loss": 2.0008, + "theoretical_loss": 4.6550666238250065, + "tokens_seen": 125501440 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004858375125376128, + "loss": 1.9227, + "theoretical_loss": 4.65474247294279, + "tokens_seen": 125566976 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048582748244734204, + "loss": 1.9642, + "theoretical_loss": 4.654418538540009, + "tokens_seen": 125632512 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004858174523570712, + "loss": 2.0326, + "theoretical_loss": 4.654094820359278, + "tokens_seen": 125698048 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004858074222668004, + "loss": 2.029, + "theoretical_loss": 4.653771318143651, + "tokens_seen": 125763584 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004857973921765296, + "loss": 1.8177, + "theoretical_loss": 4.65344803163662, + "tokens_seen": 125829120 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048578736208625877, + "loss": 1.8894, + "theoretical_loss": 4.653124960582118, + "tokens_seen": 125894656 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048577733199598795, + "loss": 2.0904, + "theoretical_loss": 4.652802104724513, + "tokens_seen": 125960192 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004857673019057172, + "loss": 1.995, + "theoretical_loss": 4.652479463808609, + "tokens_seen": 126025728 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004857572718154463, + "loss": 1.9387, + "theoretical_loss": 4.652157037579647, + "tokens_seen": 126091264 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048574724172517554, + "loss": 1.9774, + "theoretical_loss": 4.6518348257833, + "tokens_seen": 126156800 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048573721163490467, + "loss": 1.991, + "theoretical_loss": 4.651512828165675, + "tokens_seen": 126222336 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004857271815446339, + "loss": 1.9543, + "theoretical_loss": 4.651191044473311, + "tokens_seen": 126287872 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048571715145436314, + "loss": 2.0596, + "theoretical_loss": 4.65086947445318, + "tokens_seen": 126353408 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048570712136409227, + "loss": 1.9861, + "theoretical_loss": 4.650548117852681, + "tokens_seen": 126418944 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004856970912738215, + "loss": 1.8915, + "theoretical_loss": 4.650226974419646, + "tokens_seen": 126484480 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004856870611835507, + "loss": 2.0125, + "theoretical_loss": 4.649906043902332, + "tokens_seen": 126550016 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048567703109327987, + "loss": 2.0406, + "theoretical_loss": 4.6495853260494275, + "tokens_seen": 126615552 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048566700100300905, + "loss": 1.9785, + "theoretical_loss": 4.649264820610043, + "tokens_seen": 126681088 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048565697091273823, + "loss": 2.045, + "theoretical_loss": 4.648944527333717, + "tokens_seen": 126746624 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004856469408224674, + "loss": 1.8343, + "theoretical_loss": 4.648624445970414, + "tokens_seen": 126812160 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048563691073219664, + "loss": 1.9808, + "theoretical_loss": 4.648304576270519, + "tokens_seen": 126877696 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048562688064192577, + "loss": 1.9484, + "theoretical_loss": 4.647984917984841, + "tokens_seen": 126943232 + }, + { + "epoch": 0.04, + "learning_rate": 0.000485616850551655, + "loss": 1.9859, + "theoretical_loss": 4.647665470864613, + "tokens_seen": 127008768 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048560682046138413, + "loss": 1.941, + "theoretical_loss": 4.6473462346614856, + "tokens_seen": 127074304 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048559679037111337, + "loss": 1.9272, + "theoretical_loss": 4.647027209127531, + "tokens_seen": 127139840 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048558676028084255, + "loss": 1.9835, + "theoretical_loss": 4.646708394015242, + "tokens_seen": 127205376 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048557673019057173, + "loss": 2.0247, + "theoretical_loss": 4.646389789077524, + "tokens_seen": 127270912 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004855667001003009, + "loss": 2.0364, + "theoretical_loss": 4.646071394067707, + "tokens_seen": 127336448 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048555667001003015, + "loss": 1.9808, + "theoretical_loss": 4.645753208739533, + "tokens_seen": 127401984 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004855466399197593, + "loss": 1.9593, + "theoretical_loss": 4.645435232847159, + "tokens_seen": 127467520 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004855366098294885, + "loss": 1.8694, + "theoretical_loss": 4.645117466145157, + "tokens_seen": 127533056 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048552657973921764, + "loss": 1.762, + "theoretical_loss": 4.6447999083885145, + "tokens_seen": 127598592 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048551654964894687, + "loss": 1.7661, + "theoretical_loss": 4.644482559332629, + "tokens_seen": 127664128 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048550651955867605, + "loss": 1.9376, + "theoretical_loss": 4.644165418733312, + "tokens_seen": 127729664 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 54520, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9285078048706055, + "objective/train/theoretical_loss": 4.6438484863467835, + "objective/train/tokens_used": 127795200, + "theoretical_loss": 4.6438484863467835, + "tokens_seen": 127795200 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048549648946840523, + "loss": 1.8555, + "theoretical_loss": 4.6438484863467835, + "tokens_seen": 127795200 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004854864593781344, + "loss": 2.0183, + "theoretical_loss": 4.643531761929674, + "tokens_seen": 127860736 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004854764292878636, + "loss": 1.8378, + "theoretical_loss": 4.643215245239025, + "tokens_seen": 127926272 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004854663991975928, + "loss": 1.9205, + "theoretical_loss": 4.642898936032285, + "tokens_seen": 127991808 + }, + { + "epoch": 0.04, + "learning_rate": 0.000485456369107322, + "loss": 1.8459, + "theoretical_loss": 4.6425828340673085, + "tokens_seen": 128057344 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048544633901705114, + "loss": 2.0277, + "theoretical_loss": 4.642266939102357, + "tokens_seen": 128122880 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004854363089267804, + "loss": 1.94, + "theoretical_loss": 4.641951250896098, + "tokens_seen": 128188416 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004854262788365095, + "loss": 2.13, + "theoretical_loss": 4.641635769207603, + "tokens_seen": 128253952 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048541624874623874, + "loss": 1.8952, + "theoretical_loss": 4.641320493796349, + "tokens_seen": 128319488 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004854062186559679, + "loss": 2.0047, + "theoretical_loss": 4.641005424422213, + "tokens_seen": 128385024 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004853961885656971, + "loss": 1.9127, + "theoretical_loss": 4.640690560845474, + "tokens_seen": 128450560 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004853861584754263, + "loss": 2.0324, + "theoretical_loss": 4.640375902826815, + "tokens_seen": 128516096 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004853761283851555, + "loss": 2.007, + "theoretical_loss": 4.640061450127316, + "tokens_seen": 128581632 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048536609829488464, + "loss": 1.969, + "theoretical_loss": 4.639747202508459, + "tokens_seen": 128647168 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004853560682046139, + "loss": 2.0053, + "theoretical_loss": 4.639433159732121, + "tokens_seen": 128712704 + }, + { + "epoch": 0.04, + "learning_rate": 0.000485346038114343, + "loss": 1.9676, + "theoretical_loss": 4.639119321560581, + "tokens_seen": 128778240 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048533600802407224, + "loss": 1.8795, + "theoretical_loss": 4.638805687756511, + "tokens_seen": 128843776 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004853259779338014, + "loss": 2.0351, + "theoretical_loss": 4.63849225808298, + "tokens_seen": 128909312 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004853159478435306, + "loss": 1.875, + "theoretical_loss": 4.6381790323034515, + "tokens_seen": 128974848 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004853059177532598, + "loss": 1.9196, + "theoretical_loss": 4.637866010181784, + "tokens_seen": 129040384 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048529588766298897, + "loss": 1.9634, + "theoretical_loss": 4.63755319148223, + "tokens_seen": 129105920 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048528585757271815, + "loss": 1.9545, + "theoretical_loss": 4.637240575969432, + "tokens_seen": 129171456 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004852758274824474, + "loss": 1.9029, + "theoretical_loss": 4.636928163408426, + "tokens_seen": 129236992 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004852657973921765, + "loss": 1.9985, + "theoretical_loss": 4.636615953564637, + "tokens_seen": 129302528 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048525576730190574, + "loss": 1.9192, + "theoretical_loss": 4.636303946203881, + "tokens_seen": 129368064 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048524573721163487, + "loss": 1.9506, + "theoretical_loss": 4.635992141092364, + "tokens_seen": 129433600 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004852357071213641, + "loss": 1.9235, + "theoretical_loss": 4.635680537996679, + "tokens_seen": 129499136 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004852256770310933, + "loss": 1.9834, + "theoretical_loss": 4.635369136683805, + "tokens_seen": 129564672 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048521564694082247, + "loss": 1.9563, + "theoretical_loss": 4.6350579369211085, + "tokens_seen": 129630208 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048520561685055165, + "loss": 1.905, + "theoretical_loss": 4.634746938476344, + "tokens_seen": 129695744 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004851955867602809, + "loss": 1.9578, + "theoretical_loss": 4.634436141117645, + "tokens_seen": 129761280 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048518555667001, + "loss": 2.0146, + "theoretical_loss": 4.634125544613536, + "tokens_seen": 129826816 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048517552657973925, + "loss": 1.9966, + "theoretical_loss": 4.63381514873292, + "tokens_seen": 129892352 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004851654964894684, + "loss": 1.912, + "theoretical_loss": 4.633504953245085, + "tokens_seen": 129957888 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004851554663991976, + "loss": 1.9069, + "theoretical_loss": 4.633194957919697, + "tokens_seen": 130023424 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004851454363089268, + "loss": 1.9715, + "theoretical_loss": 4.632885162526806, + "tokens_seen": 130088960 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048513540621865597, + "loss": 1.8884, + "theoretical_loss": 4.63257556683684, + "tokens_seen": 130154496 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048512537612838515, + "loss": 1.8518, + "theoretical_loss": 4.632266170620607, + "tokens_seen": 130220032 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048511534603811433, + "loss": 1.9325, + "theoretical_loss": 4.631956973649293, + "tokens_seen": 130285568 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004851053159478435, + "loss": 1.9781, + "theoretical_loss": 4.631647975694461, + "tokens_seen": 130351104 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048509528585757275, + "loss": 1.9977, + "theoretical_loss": 4.631339176528051, + "tokens_seen": 130416640 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004850852557673019, + "loss": 1.945, + "theoretical_loss": 4.631030575922379, + "tokens_seen": 130482176 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004850752256770311, + "loss": 2.0648, + "theoretical_loss": 4.630722173650135, + "tokens_seen": 130547712 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048506519558676024, + "loss": 1.8705, + "theoretical_loss": 4.630413969484383, + "tokens_seen": 130613248 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004850551654964895, + "loss": 2.0333, + "theoretical_loss": 4.630105963198563, + "tokens_seen": 130678784 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048504513540621866, + "loss": 2.0196, + "theoretical_loss": 4.629798154566485, + "tokens_seen": 130744320 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048503510531594784, + "loss": 1.9621, + "theoretical_loss": 4.62949054336233, + "tokens_seen": 130809856 + }, + { + "epoch": 0.04, + "learning_rate": 0.000485025075225677, + "loss": 1.9137, + "theoretical_loss": 4.629183129360653, + "tokens_seen": 130875392 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048501504513540625, + "loss": 1.9377, + "theoretical_loss": 4.628875912336378, + "tokens_seen": 130940928 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004850050150451354, + "loss": 1.989, + "theoretical_loss": 4.628568892064797, + "tokens_seen": 131006464 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 55660, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8139686584472656, + "objective/train/theoretical_loss": 4.62826206832157, + "objective/train/tokens_used": 131072000, + "theoretical_loss": 4.62826206832157, + "tokens_seen": 131072000 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004849949849548646, + "loss": 1.9759, + "theoretical_loss": 4.62826206832157, + "tokens_seen": 131072000 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004849849548645938, + "loss": 1.9828, + "theoretical_loss": 4.6279554408827295, + "tokens_seen": 131137536 + }, + { + "epoch": 0.04, + "learning_rate": 0.000484974924774323, + "loss": 1.925, + "theoretical_loss": 4.62764900952467, + "tokens_seen": 131203072 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004849648946840522, + "loss": 1.9347, + "theoretical_loss": 4.627342774024155, + "tokens_seen": 131268608 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048495486459378134, + "loss": 2.0488, + "theoretical_loss": 4.6270367341583105, + "tokens_seen": 131334144 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004849448345035106, + "loss": 2.0093, + "theoretical_loss": 4.626730889704632, + "tokens_seen": 131399680 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004849348044132397, + "loss": 1.9417, + "theoretical_loss": 4.626425240440973, + "tokens_seen": 131465216 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048492477432296894, + "loss": 1.9827, + "theoretical_loss": 4.626119786145557, + "tokens_seen": 131530752 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004849147442326981, + "loss": 1.9908, + "theoretical_loss": 4.625814526596962, + "tokens_seen": 131596288 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004849047141424273, + "loss": 1.9381, + "theoretical_loss": 4.625509461574135, + "tokens_seen": 131661824 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004848946840521565, + "loss": 1.893, + "theoretical_loss": 4.625204590856377, + "tokens_seen": 131727360 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004848846539618857, + "loss": 1.9782, + "theoretical_loss": 4.624899914223355, + "tokens_seen": 131792896 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048487462387161484, + "loss": 1.8265, + "theoretical_loss": 4.624595431455091, + "tokens_seen": 131858432 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004848645937813441, + "loss": 1.9925, + "theoretical_loss": 4.62429114233197, + "tokens_seen": 131923968 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004848545636910732, + "loss": 2.0754, + "theoretical_loss": 4.623987046634729, + "tokens_seen": 131989504 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048484453360080244, + "loss": 1.9093, + "theoretical_loss": 4.623683144144465, + "tokens_seen": 132055040 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004848345035105316, + "loss": 2.0108, + "theoretical_loss": 4.623379434642632, + "tokens_seen": 132120576 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004848244734202608, + "loss": 1.911, + "theoretical_loss": 4.623075917911039, + "tokens_seen": 132186112 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048481444332999, + "loss": 1.8546, + "theoretical_loss": 4.622772593731848, + "tokens_seen": 132251648 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048480441323971917, + "loss": 1.8616, + "theoretical_loss": 4.622469461887577, + "tokens_seen": 132317184 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048479438314944835, + "loss": 2.018, + "theoretical_loss": 4.622166522161096, + "tokens_seen": 132382720 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004847843530591776, + "loss": 2.0425, + "theoretical_loss": 4.621863774335628, + "tokens_seen": 132448256 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004847743229689067, + "loss": 1.8536, + "theoretical_loss": 4.621561218194747, + "tokens_seen": 132513792 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048476429287863594, + "loss": 1.9847, + "theoretical_loss": 4.62125885352238, + "tokens_seen": 132579328 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048475426278836507, + "loss": 1.9652, + "theoretical_loss": 4.620956680102801, + "tokens_seen": 132644864 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004847442326980943, + "loss": 1.9284, + "theoretical_loss": 4.620654697720637, + "tokens_seen": 132710400 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004847342026078235, + "loss": 1.8797, + "theoretical_loss": 4.620352906160861, + "tokens_seen": 132775936 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048472417251755267, + "loss": 1.9047, + "theoretical_loss": 4.6200513052087935, + "tokens_seen": 132841472 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048471414242728185, + "loss": 1.9609, + "theoretical_loss": 4.6197498946501065, + "tokens_seen": 132907008 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004847041123370111, + "loss": 1.9019, + "theoretical_loss": 4.6194486742708145, + "tokens_seen": 132972544 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004846940822467402, + "loss": 2.0541, + "theoretical_loss": 4.6191476438572785, + "tokens_seen": 133038080 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048468405215646945, + "loss": 1.9483, + "theoretical_loss": 4.618846803196206, + "tokens_seen": 133103616 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004846740220661986, + "loss": 1.8489, + "theoretical_loss": 4.618546152074648, + "tokens_seen": 133169152 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004846639919759278, + "loss": 1.8667, + "theoretical_loss": 4.61824569028, + "tokens_seen": 133234688 + }, + { + "epoch": 0.04, + "learning_rate": 0.000484653961885657, + "loss": 1.9174, + "theoretical_loss": 4.617945417599998, + "tokens_seen": 133300224 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048464393179538617, + "loss": 1.9296, + "theoretical_loss": 4.617645333822722, + "tokens_seen": 133365760 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048463390170511535, + "loss": 1.9152, + "theoretical_loss": 4.617345438736594, + "tokens_seen": 133431296 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048462387161484453, + "loss": 2.0336, + "theoretical_loss": 4.617045732130375, + "tokens_seen": 133496832 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004846138415245737, + "loss": 2.2147, + "theoretical_loss": 4.616746213793167, + "tokens_seen": 133562368 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048460381143430295, + "loss": 1.9809, + "theoretical_loss": 4.616446883514412, + "tokens_seen": 133627904 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004845937813440321, + "loss": 1.8973, + "theoretical_loss": 4.616147741083891, + "tokens_seen": 133693440 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004845837512537613, + "loss": 1.9036, + "theoretical_loss": 4.6158487862917195, + "tokens_seen": 133758976 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048457372116349044, + "loss": 1.9479, + "theoretical_loss": 4.615550018928353, + "tokens_seen": 133824512 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004845636910732197, + "loss": 1.8944, + "theoretical_loss": 4.615251438784584, + "tokens_seen": 133890048 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048455366098294886, + "loss": 1.914, + "theoretical_loss": 4.61495304565154, + "tokens_seen": 133955584 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048454363089267804, + "loss": 1.921, + "theoretical_loss": 4.614654839320682, + "tokens_seen": 134021120 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004845336008024072, + "loss": 2.0494, + "theoretical_loss": 4.614356819583808, + "tokens_seen": 134086656 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048452357071213645, + "loss": 1.9364, + "theoretical_loss": 4.614058986233047, + "tokens_seen": 134152192 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004845135406218656, + "loss": 1.996, + "theoretical_loss": 4.613761339060863, + "tokens_seen": 134217728 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004845035105315948, + "loss": 1.9385, + "theoretical_loss": 4.613463877860051, + "tokens_seen": 134283264 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 57372, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.1588454246520996, + "objective/train/theoretical_loss": 4.61316660242374, + "objective/train/tokens_used": 134348800, + "theoretical_loss": 4.61316660242374, + "tokens_seen": 134348800 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048449348044132394, + "loss": 2.0275, + "theoretical_loss": 4.61316660242374, + "tokens_seen": 134348800 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004844834503510532, + "loss": 1.9812, + "theoretical_loss": 4.6128695125453865, + "tokens_seen": 134414336 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048447342026078236, + "loss": 1.9473, + "theoretical_loss": 4.612572608018779, + "tokens_seen": 134479872 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048446339017051154, + "loss": 2.1076, + "theoretical_loss": 4.612275888638036, + "tokens_seen": 134545408 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004844533600802407, + "loss": 2.091, + "theoretical_loss": 4.611979354197603, + "tokens_seen": 134610944 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004844433299899699, + "loss": 1.9263, + "theoretical_loss": 4.6116830044922565, + "tokens_seen": 134676480 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004844332998996991, + "loss": 2.2892, + "theoretical_loss": 4.611386839317097, + "tokens_seen": 134742016 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004844232698094283, + "loss": 1.9761, + "theoretical_loss": 4.611090858467554, + "tokens_seen": 134807552 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048441323971915745, + "loss": 2.2484, + "theoretical_loss": 4.610795061739384, + "tokens_seen": 134873088 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004844032096288867, + "loss": 1.9418, + "theoretical_loss": 4.610499448928666, + "tokens_seen": 134938624 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048439317953861586, + "loss": 2.0212, + "theoretical_loss": 4.610204019831805, + "tokens_seen": 135004160 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048438314944834504, + "loss": 2.0047, + "theoretical_loss": 4.609908774245532, + "tokens_seen": 135069696 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004843731193580742, + "loss": 2.0945, + "theoretical_loss": 4.609613711966898, + "tokens_seen": 135135232 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004843630892678034, + "loss": 2.0294, + "theoretical_loss": 4.60931883279328, + "tokens_seen": 135200768 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004843530591775326, + "loss": 1.9846, + "theoretical_loss": 4.6090241365223745, + "tokens_seen": 135266304 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004843430290872618, + "loss": 2.025, + "theoretical_loss": 4.6087296229522, + "tokens_seen": 135331840 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048433299899699095, + "loss": 2.0609, + "theoretical_loss": 4.608435291881099, + "tokens_seen": 135397376 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004843229689067202, + "loss": 1.9405, + "theoretical_loss": 4.608141143107728, + "tokens_seen": 135462912 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004843129388164493, + "loss": 1.9703, + "theoretical_loss": 4.607847176431067, + "tokens_seen": 135528448 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048430290872617855, + "loss": 2.0571, + "theoretical_loss": 4.6075533916504146, + "tokens_seen": 135593984 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048429287863590773, + "loss": 1.9809, + "theoretical_loss": 4.607259788565385, + "tokens_seen": 135659520 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004842828485456369, + "loss": 1.8207, + "theoretical_loss": 4.606966366975914, + "tokens_seen": 135725056 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004842728184553661, + "loss": 1.9558, + "theoretical_loss": 4.606673126682249, + "tokens_seen": 135790592 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048426278836509527, + "loss": 1.9064, + "theoretical_loss": 4.6063800674849595, + "tokens_seen": 135856128 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048425275827482445, + "loss": 2.0567, + "theoretical_loss": 4.606087189184924, + "tokens_seen": 135921664 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004842427281845537, + "loss": 1.8861, + "theoretical_loss": 4.6057944915833415, + "tokens_seen": 135987200 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048423269809428287, + "loss": 1.8697, + "theoretical_loss": 4.6055019744817205, + "tokens_seen": 136052736 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048422266800401205, + "loss": 1.9669, + "theoretical_loss": 4.605209637681886, + "tokens_seen": 136118272 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004842126379137413, + "loss": 1.8449, + "theoretical_loss": 4.604917480985975, + "tokens_seen": 136183808 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004842026078234704, + "loss": 1.8791, + "theoretical_loss": 4.604625504196438, + "tokens_seen": 136249344 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048419257773319965, + "loss": 1.914, + "theoretical_loss": 4.604333707116033, + "tokens_seen": 136314880 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004841825476429288, + "loss": 1.8305, + "theoretical_loss": 4.604042089547834, + "tokens_seen": 136380416 + }, + { + "epoch": 0.04, + "learning_rate": 0.000484172517552658, + "loss": 2.035, + "theoretical_loss": 4.603750651295222, + "tokens_seen": 136445952 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004841624874623872, + "loss": 2.0025, + "theoretical_loss": 4.603459392161888, + "tokens_seen": 136511488 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048415245737211637, + "loss": 1.8691, + "theoretical_loss": 4.603168311951835, + "tokens_seen": 136577024 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048414242728184555, + "loss": 1.8669, + "theoretical_loss": 4.60287741046937, + "tokens_seen": 136642560 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048413239719157473, + "loss": 1.9829, + "theoretical_loss": 4.602586687519111, + "tokens_seen": 136708096 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004841223671013039, + "loss": 1.9853, + "theoretical_loss": 4.602296142905981, + "tokens_seen": 136773632 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048411233701103315, + "loss": 1.9079, + "theoretical_loss": 4.602005776435211, + "tokens_seen": 136839168 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004841023069207623, + "loss": 1.9444, + "theoretical_loss": 4.601715587912338, + "tokens_seen": 136904704 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004840922768304915, + "loss": 2.0372, + "theoretical_loss": 4.601425577143203, + "tokens_seen": 136970240 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048408224674022064, + "loss": 2.0213, + "theoretical_loss": 4.601135743933952, + "tokens_seen": 137035776 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004840722166499499, + "loss": 2.092, + "theoretical_loss": 4.600846088091036, + "tokens_seen": 137101312 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048406218655967906, + "loss": 1.8936, + "theoretical_loss": 4.600556609421206, + "tokens_seen": 137166848 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048405215646940824, + "loss": 2.0878, + "theoretical_loss": 4.600267307731521, + "tokens_seen": 137232384 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004840421263791374, + "loss": 2.027, + "theoretical_loss": 4.5999781828293385, + "tokens_seen": 137297920 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048403209628886665, + "loss": 1.9098, + "theoretical_loss": 4.5996892345223195, + "tokens_seen": 137363456 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004840220661985958, + "loss": 2.0104, + "theoretical_loss": 4.599400462618423, + "tokens_seen": 137428992 + }, + { + "epoch": 0.04, + "learning_rate": 0.000484012036108325, + "loss": 1.9239, + "theoretical_loss": 4.59911186692591, + "tokens_seen": 137494528 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048400200601805414, + "loss": 1.8549, + "theoretical_loss": 4.598823447253343, + "tokens_seen": 137560064 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 58515, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.0190987586975098, + "objective/train/theoretical_loss": 4.598535203409582, + "objective/train/tokens_used": 137625600, + "theoretical_loss": 4.598535203409582, + "tokens_seen": 137625600 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004839919759277834, + "loss": 1.9452, + "theoretical_loss": 4.598535203409582, + "tokens_seen": 137625600 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048398194583751256, + "loss": 1.9955, + "theoretical_loss": 4.598247135203783, + "tokens_seen": 137691136 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048397191574724174, + "loss": 2.0024, + "theoretical_loss": 4.5979592424454045, + "tokens_seen": 137756672 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004839618856569709, + "loss": 1.9964, + "theoretical_loss": 4.597671524944198, + "tokens_seen": 137822208 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004839518555667001, + "loss": 2.0265, + "theoretical_loss": 4.597383982510216, + "tokens_seen": 137887744 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004839418254764293, + "loss": 1.8103, + "theoretical_loss": 4.597096614953802, + "tokens_seen": 137953280 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004839317953861585, + "loss": 1.9574, + "theoretical_loss": 4.596809422085599, + "tokens_seen": 138018816 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048392176529588765, + "loss": 1.789, + "theoretical_loss": 4.596522403716542, + "tokens_seen": 138084352 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004839117352056169, + "loss": 1.8943, + "theoretical_loss": 4.596235559657863, + "tokens_seen": 138149888 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048390170511534606, + "loss": 1.9381, + "theoretical_loss": 4.595948889721086, + "tokens_seen": 138215424 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048389167502507524, + "loss": 1.9409, + "theoretical_loss": 4.595662393718026, + "tokens_seen": 138280960 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004838816449348044, + "loss": 1.9916, + "theoretical_loss": 4.595376071460796, + "tokens_seen": 138346496 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004838716148445336, + "loss": 1.9587, + "theoretical_loss": 4.595089922761796, + "tokens_seen": 138412032 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004838615847542628, + "loss": 2.0796, + "theoretical_loss": 4.594803947433717, + "tokens_seen": 138477568 + }, + { + "epoch": 0.04, + "learning_rate": 0.000483851554663992, + "loss": 1.8934, + "theoretical_loss": 4.594518145289545, + "tokens_seen": 138543104 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048384152457372115, + "loss": 1.9935, + "theoretical_loss": 4.594232516142553, + "tokens_seen": 138608640 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004838314944834504, + "loss": 1.8861, + "theoretical_loss": 4.593947059806302, + "tokens_seen": 138674176 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004838214643931795, + "loss": 2.1611, + "theoretical_loss": 4.593661776094647, + "tokens_seen": 138739712 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048381143430290875, + "loss": 1.9821, + "theoretical_loss": 4.593376664821726, + "tokens_seen": 138805248 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048380140421263793, + "loss": 1.9509, + "theoretical_loss": 4.593091725801967, + "tokens_seen": 138870784 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004837913741223671, + "loss": 1.8744, + "theoretical_loss": 4.592806958850087, + "tokens_seen": 138936320 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004837813440320963, + "loss": 1.9989, + "theoretical_loss": 4.592522363781086, + "tokens_seen": 139001856 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048377131394182547, + "loss": 1.7862, + "theoretical_loss": 4.592237940410252, + "tokens_seen": 139067392 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048376128385155465, + "loss": 1.9901, + "theoretical_loss": 4.591953688553159, + "tokens_seen": 139132928 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004837512537612839, + "loss": 2.0565, + "theoretical_loss": 4.591669608025663, + "tokens_seen": 139198464 + }, + { + "epoch": 0.04, + "learning_rate": 0.000483741223671013, + "loss": 2.0259, + "theoretical_loss": 4.591385698643908, + "tokens_seen": 139264000 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048373119358074225, + "loss": 1.9337, + "theoretical_loss": 4.591101960224321, + "tokens_seen": 139329536 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048372116349047143, + "loss": 1.9248, + "theoretical_loss": 4.590818392583608, + "tokens_seen": 139395072 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004837111334002006, + "loss": 1.9743, + "theoretical_loss": 4.590534995538762, + "tokens_seen": 139460608 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004837011033099298, + "loss": 1.9242, + "theoretical_loss": 4.590251768907057, + "tokens_seen": 139526144 + }, + { + "epoch": 0.04, + "learning_rate": 0.000483691073219659, + "loss": 1.8895, + "theoretical_loss": 4.5899687125060495, + "tokens_seen": 139591680 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048368104312938816, + "loss": 2.0289, + "theoretical_loss": 4.589685826153573, + "tokens_seen": 139657216 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004836710130391174, + "loss": 1.8852, + "theoretical_loss": 4.589403109667746, + "tokens_seen": 139722752 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004836609829488465, + "loss": 1.8805, + "theoretical_loss": 4.589120562866961, + "tokens_seen": 139788288 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048365095285857575, + "loss": 1.9913, + "theoretical_loss": 4.588838185569897, + "tokens_seen": 139853824 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004836409227683049, + "loss": 1.9751, + "theoretical_loss": 4.588555977595506, + "tokens_seen": 139919360 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004836308926780341, + "loss": 2.0349, + "theoretical_loss": 4.58827393876302, + "tokens_seen": 139984896 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004836208625877633, + "loss": 1.8944, + "theoretical_loss": 4.587992068891948, + "tokens_seen": 140050432 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004836108324974925, + "loss": 1.9102, + "theoretical_loss": 4.587710367802076, + "tokens_seen": 140115968 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048360080240722166, + "loss": 1.9981, + "theoretical_loss": 4.587428835313469, + "tokens_seen": 140181504 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048359077231695084, + "loss": 2.0375, + "theoretical_loss": 4.587147471246463, + "tokens_seen": 140247040 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048358074222668, + "loss": 1.9063, + "theoretical_loss": 4.586866275421673, + "tokens_seen": 140312576 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048357071213640926, + "loss": 1.961, + "theoretical_loss": 4.586585247659986, + "tokens_seen": 140378112 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004835606820461384, + "loss": 1.8945, + "theoretical_loss": 4.586304387782567, + "tokens_seen": 140443648 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004835506519558676, + "loss": 1.9595, + "theoretical_loss": 4.5860236956108515, + "tokens_seen": 140509184 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004835406218655968, + "loss": 1.8776, + "theoretical_loss": 4.585743170966548, + "tokens_seen": 140574720 + }, + { + "epoch": 0.04, + "learning_rate": 0.000483530591775326, + "loss": 1.9955, + "theoretical_loss": 4.58546281367164, + "tokens_seen": 140640256 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048352056168505516, + "loss": 1.9528, + "theoretical_loss": 4.585182623548381, + "tokens_seen": 140705792 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048351053159478434, + "loss": 2.0253, + "theoretical_loss": 4.584902600419295, + "tokens_seen": 140771328 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004835005015045135, + "loss": 1.8449, + "theoretical_loss": 4.584622744107181, + "tokens_seen": 140836864 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 59715, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9665565490722656, + "objective/train/theoretical_loss": 4.584343054435103, + "objective/train/tokens_used": 140902400, + "theoretical_loss": 4.584343054435103, + "tokens_seen": 140902400 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048349047141424276, + "loss": 1.9612, + "theoretical_loss": 4.584343054435103, + "tokens_seen": 140902400 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048348044132397194, + "loss": 1.9396, + "theoretical_loss": 4.5840635312264, + "tokens_seen": 140967936 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004834704112337011, + "loss": 1.9409, + "theoretical_loss": 4.583784174304677, + "tokens_seen": 141033472 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004834603811434303, + "loss": 1.8454, + "theoretical_loss": 4.583504983493809, + "tokens_seen": 141099008 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004834503510531595, + "loss": 1.8002, + "theoretical_loss": 4.583225958617938, + "tokens_seen": 141164544 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004834403209628887, + "loss": 1.8609, + "theoretical_loss": 4.582947099501475, + "tokens_seen": 141230080 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048343029087261785, + "loss": 1.9303, + "theoretical_loss": 4.5826684059691, + "tokens_seen": 141295616 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004834202607823471, + "loss": 1.974, + "theoretical_loss": 4.582389877845754, + "tokens_seen": 141361152 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048341023069207626, + "loss": 1.9605, + "theoretical_loss": 4.58211151495665, + "tokens_seen": 141426688 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048340020060180544, + "loss": 1.6998, + "theoretical_loss": 4.581833317127263, + "tokens_seen": 141492224 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004833901705115346, + "loss": 1.9118, + "theoretical_loss": 4.581555284183335, + "tokens_seen": 141557760 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004833801404212638, + "loss": 1.9537, + "theoretical_loss": 4.581277415950872, + "tokens_seen": 141623296 + }, + { + "epoch": 0.04, + "learning_rate": 0.000483370110330993, + "loss": 2.155, + "theoretical_loss": 4.580999712256142, + "tokens_seen": 141688832 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004833600802407222, + "loss": 1.8769, + "theoretical_loss": 4.5807221729256815, + "tokens_seen": 141754368 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048335005015045135, + "loss": 1.8758, + "theoretical_loss": 4.580444797786283, + "tokens_seen": 141819904 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004833400200601806, + "loss": 1.9161, + "theoretical_loss": 4.5801675866650085, + "tokens_seen": 141885440 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004833299899699097, + "loss": 1.993, + "theoretical_loss": 4.579890539389177, + "tokens_seen": 141950976 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048331995987963895, + "loss": 2.0895, + "theoretical_loss": 4.579613655786369, + "tokens_seen": 142016512 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048330992978936813, + "loss": 1.8633, + "theoretical_loss": 4.579336935684431, + "tokens_seen": 142082048 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004832998996990973, + "loss": 2.0321, + "theoretical_loss": 4.579060378911464, + "tokens_seen": 142147584 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004832898696088265, + "loss": 1.9348, + "theoretical_loss": 4.578783985295832, + "tokens_seen": 142213120 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048327983951855567, + "loss": 2.0021, + "theoretical_loss": 4.578507754666157, + "tokens_seen": 142278656 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048326980942828485, + "loss": 1.9287, + "theoretical_loss": 4.57823168685132, + "tokens_seen": 142344192 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004832597793380141, + "loss": 1.9027, + "theoretical_loss": 4.577955781680462, + "tokens_seen": 142409728 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004832497492477432, + "loss": 1.9524, + "theoretical_loss": 4.57768003898298, + "tokens_seen": 142475264 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048323971915747245, + "loss": 2.1208, + "theoretical_loss": 4.577404458588529, + "tokens_seen": 142540800 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048322968906720163, + "loss": 1.9644, + "theoretical_loss": 4.577129040327021, + "tokens_seen": 142606336 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004832196589769308, + "loss": 1.8463, + "theoretical_loss": 4.576853784028622, + "tokens_seen": 142671872 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048320962888666, + "loss": 1.8897, + "theoretical_loss": 4.576578689523759, + "tokens_seen": 142737408 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004831995987963892, + "loss": 1.9843, + "theoretical_loss": 4.576303756643108, + "tokens_seen": 142802944 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048318956870611836, + "loss": 1.7977, + "theoretical_loss": 4.576028985217606, + "tokens_seen": 142868480 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004831795386158476, + "loss": 2.0478, + "theoretical_loss": 4.5757543750784375, + "tokens_seen": 142934016 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004831695085255767, + "loss": 1.9699, + "theoretical_loss": 4.5754799260570485, + "tokens_seen": 142999552 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048315947843530595, + "loss": 1.9781, + "theoretical_loss": 4.575205637985132, + "tokens_seen": 143065088 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004831494483450351, + "loss": 1.9391, + "theoretical_loss": 4.574931510694635, + "tokens_seen": 143130624 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004831394182547643, + "loss": 1.9165, + "theoretical_loss": 4.574657544017761, + "tokens_seen": 143196160 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004831293881644935, + "loss": 2.0497, + "theoretical_loss": 4.57438373778696, + "tokens_seen": 143261696 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004831193580742227, + "loss": 1.8759, + "theoretical_loss": 4.574110091834936, + "tokens_seen": 143327232 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048310932798395186, + "loss": 2.0278, + "theoretical_loss": 4.573836605994645, + "tokens_seen": 143392768 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048309929789368104, + "loss": 2.0334, + "theoretical_loss": 4.573563280099288, + "tokens_seen": 143458304 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004830892678034102, + "loss": 1.883, + "theoretical_loss": 4.573290113982324, + "tokens_seen": 143523840 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048307923771313946, + "loss": 1.9767, + "theoretical_loss": 4.573017107477455, + "tokens_seen": 143589376 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004830692076228686, + "loss": 1.8966, + "theoretical_loss": 4.572744260418633, + "tokens_seen": 143654912 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004830591775325978, + "loss": 1.9019, + "theoretical_loss": 4.572471572640059, + "tokens_seen": 143720448 + }, + { + "epoch": 0.04, + "learning_rate": 0.000483049147442327, + "loss": 2.0065, + "theoretical_loss": 4.572199043976185, + "tokens_seen": 143785984 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004830391173520562, + "loss": 1.8285, + "theoretical_loss": 4.571926674261705, + "tokens_seen": 143851520 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048302908726178536, + "loss": 1.9178, + "theoretical_loss": 4.571654463331563, + "tokens_seen": 143917056 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048301905717151454, + "loss": 1.9844, + "theoretical_loss": 4.571382411020949, + "tokens_seen": 143982592 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004830090270812437, + "loss": 1.8647, + "theoretical_loss": 4.5711105171653, + "tokens_seen": 144048128 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048299899699097296, + "loss": 1.8027, + "theoretical_loss": 4.570838781600296, + "tokens_seen": 144113664 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 60961, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8796693086624146, + "objective/train/theoretical_loss": 4.570567204161865, + "objective/train/tokens_used": 144179200, + "theoretical_loss": 4.570567204161865, + "tokens_seen": 144179200 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004829889669007021, + "loss": 1.9151, + "theoretical_loss": 4.570567204161865, + "tokens_seen": 144179200 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004829789368104313, + "loss": 1.8232, + "theoretical_loss": 4.570295784686176, + "tokens_seen": 144244736 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048296890672016045, + "loss": 1.9931, + "theoretical_loss": 4.570024523009646, + "tokens_seen": 144310272 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004829588766298897, + "loss": 1.9579, + "theoretical_loss": 4.569753418968933, + "tokens_seen": 144375808 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048294884653961887, + "loss": 1.9824, + "theoretical_loss": 4.569482472400939, + "tokens_seen": 144441344 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048293881644934805, + "loss": 1.8761, + "theoretical_loss": 4.5692116831428065, + "tokens_seen": 144506880 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048292878635907723, + "loss": 2.085, + "theoretical_loss": 4.568941051031925, + "tokens_seen": 144572416 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048291875626880646, + "loss": 1.9584, + "theoretical_loss": 4.568670575905921, + "tokens_seen": 144637952 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004829087261785356, + "loss": 1.8864, + "theoretical_loss": 4.568400257602666, + "tokens_seen": 144703488 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004828986960882648, + "loss": 1.9501, + "theoretical_loss": 4.568130095960268, + "tokens_seen": 144769024 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048288866599799395, + "loss": 1.9352, + "theoretical_loss": 4.56786009081708, + "tokens_seen": 144834560 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004828786359077232, + "loss": 1.8794, + "theoretical_loss": 4.567590242011692, + "tokens_seen": 144900096 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048286860581745237, + "loss": 1.9511, + "theoretical_loss": 4.567320549382931, + "tokens_seen": 144965632 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048285857572718155, + "loss": 1.9116, + "theoretical_loss": 4.56705101276987, + "tokens_seen": 145031168 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048284854563691073, + "loss": 1.9385, + "theoretical_loss": 4.566781632011814, + "tokens_seen": 145096704 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004828385155466399, + "loss": 1.9864, + "theoretical_loss": 4.566512406948309, + "tokens_seen": 145162240 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004828284854563691, + "loss": 1.7871, + "theoretical_loss": 4.566243337419137, + "tokens_seen": 145227776 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048281845536609833, + "loss": 1.9912, + "theoretical_loss": 4.5659744232643185, + "tokens_seen": 145293312 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048280842527582746, + "loss": 2.0185, + "theoretical_loss": 4.565705664324111, + "tokens_seen": 145358848 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004827983951855567, + "loss": 1.9257, + "theoretical_loss": 4.565437060439006, + "tokens_seen": 145424384 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004827883650952858, + "loss": 1.9116, + "theoretical_loss": 4.565168611449732, + "tokens_seen": 145489920 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048277833500501505, + "loss": 1.8944, + "theoretical_loss": 4.564900317197251, + "tokens_seen": 145555456 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048276830491474423, + "loss": 1.8114, + "theoretical_loss": 4.564632177522766, + "tokens_seen": 145620992 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004827582748244734, + "loss": 1.9365, + "theoretical_loss": 4.564364192267705, + "tokens_seen": 145686528 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004827482447342026, + "loss": 1.8697, + "theoretical_loss": 4.564096361273736, + "tokens_seen": 145752064 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048273821464393183, + "loss": 1.9974, + "theoretical_loss": 4.563828684382761, + "tokens_seen": 145817600 + }, + { + "epoch": 0.04, + "learning_rate": 0.000482728184553661, + "loss": 1.9142, + "theoretical_loss": 4.563561161436912, + "tokens_seen": 145883136 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004827181544633902, + "loss": 1.9102, + "theoretical_loss": 4.563293792278554, + "tokens_seen": 145948672 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004827081243731194, + "loss": 2.0365, + "theoretical_loss": 4.563026576750286, + "tokens_seen": 146014208 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048269809428284856, + "loss": 1.8996, + "theoretical_loss": 4.562759514694935, + "tokens_seen": 146079744 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004826880641925778, + "loss": 1.9387, + "theoretical_loss": 4.5624926059555655, + "tokens_seen": 146145280 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004826780341023069, + "loss": 1.8498, + "theoretical_loss": 4.562225850375466, + "tokens_seen": 146210816 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048266800401203615, + "loss": 2.0166, + "theoretical_loss": 4.561959247798159, + "tokens_seen": 146276352 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004826579739217653, + "loss": 1.8781, + "theoretical_loss": 4.561692798067396, + "tokens_seen": 146341888 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004826479438314945, + "loss": 1.9669, + "theoretical_loss": 4.561426501027158, + "tokens_seen": 146407424 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004826379137412237, + "loss": 1.9973, + "theoretical_loss": 4.561160356521656, + "tokens_seen": 146472960 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004826278836509529, + "loss": 1.9241, + "theoretical_loss": 4.5608943643953275, + "tokens_seen": 146538496 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048261785356068206, + "loss": 1.952, + "theoretical_loss": 4.56062852449284, + "tokens_seen": 146604032 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048260782347041124, + "loss": 1.8821, + "theoretical_loss": 4.5603628366590865, + "tokens_seen": 146669568 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004825977933801404, + "loss": 1.8405, + "theoretical_loss": 4.560097300739191, + "tokens_seen": 146735104 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048258776328986966, + "loss": 2.0965, + "theoretical_loss": 4.5598319165785, + "tokens_seen": 146800640 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004825777331995988, + "loss": 1.7838, + "theoretical_loss": 4.55956668402259, + "tokens_seen": 146866176 + }, + { + "epoch": 0.04, + "learning_rate": 0.000482567703109328, + "loss": 1.9149, + "theoretical_loss": 4.559301602917261, + "tokens_seen": 146931712 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004825576730190572, + "loss": 1.9638, + "theoretical_loss": 4.55903667310854, + "tokens_seen": 146997248 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004825476429287864, + "loss": 1.8508, + "theoretical_loss": 4.558771894442677, + "tokens_seen": 147062784 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048253761283851556, + "loss": 1.9026, + "theoretical_loss": 4.558507266766151, + "tokens_seen": 147128320 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048252758274824474, + "loss": 1.9706, + "theoretical_loss": 4.558242789925659, + "tokens_seen": 147193856 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004825175526579739, + "loss": 1.877, + "theoretical_loss": 4.557978463768127, + "tokens_seen": 147259392 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048250752256770316, + "loss": 1.8545, + "theoretical_loss": 4.557714288140703, + "tokens_seen": 147324928 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004824974924774323, + "loss": 1.8493, + "theoretical_loss": 4.557450262890757, + "tokens_seen": 147390464 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 62112, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7899954319000244, + "objective/train/theoretical_loss": 4.557186387865883, + "objective/train/tokens_used": 147456000, + "theoretical_loss": 4.557186387865883, + "tokens_seen": 147456000 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004824874623871615, + "loss": 1.8217, + "theoretical_loss": 4.557186387865883, + "tokens_seen": 147456000 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048247743229689065, + "loss": 1.925, + "theoretical_loss": 4.556922662913895, + "tokens_seen": 147521536 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004824674022066199, + "loss": 1.857, + "theoretical_loss": 4.55665908788283, + "tokens_seen": 147587072 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048245737211634907, + "loss": 1.8476, + "theoretical_loss": 4.556395662620945, + "tokens_seen": 147652608 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048244734202607825, + "loss": 1.7914, + "theoretical_loss": 4.5561323869767225, + "tokens_seen": 147718144 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048243731193580743, + "loss": 2.109, + "theoretical_loss": 4.555869260798859, + "tokens_seen": 147783680 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048242728184553666, + "loss": 2.0133, + "theoretical_loss": 4.5556062839362745, + "tokens_seen": 147849216 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004824172517552658, + "loss": 1.9214, + "theoretical_loss": 4.555343456238109, + "tokens_seen": 147914752 + }, + { + "epoch": 0.04, + "learning_rate": 0.000482407221664995, + "loss": 1.9126, + "theoretical_loss": 4.55508077755372, + "tokens_seen": 147980288 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048239719157472415, + "loss": 1.9944, + "theoretical_loss": 4.554818247732685, + "tokens_seen": 148045824 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004823871614844534, + "loss": 1.8649, + "theoretical_loss": 4.554555866624798, + "tokens_seen": 148111360 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048237713139418257, + "loss": 1.9124, + "theoretical_loss": 4.554293634080074, + "tokens_seen": 148176896 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048236710130391175, + "loss": 1.8927, + "theoretical_loss": 4.554031549948743, + "tokens_seen": 148242432 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048235707121364093, + "loss": 1.8653, + "theoretical_loss": 4.5537696140812525, + "tokens_seen": 148307968 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004823470411233701, + "loss": 1.9041, + "theoretical_loss": 4.553507826328268, + "tokens_seen": 148373504 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004823370110330993, + "loss": 2.0012, + "theoretical_loss": 4.553246186540671, + "tokens_seen": 148439040 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048232698094282853, + "loss": 2.0774, + "theoretical_loss": 4.552984694569555, + "tokens_seen": 148504576 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048231695085255766, + "loss": 1.844, + "theoretical_loss": 4.552723350266236, + "tokens_seen": 148570112 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004823069207622869, + "loss": 1.8785, + "theoretical_loss": 4.552462153482239, + "tokens_seen": 148635648 + }, + { + "epoch": 0.05, + "learning_rate": 0.000482296890672016, + "loss": 1.9292, + "theoretical_loss": 4.552201104069307, + "tokens_seen": 148701184 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048228686058174525, + "loss": 1.9367, + "theoretical_loss": 4.551940201879396, + "tokens_seen": 148766720 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048227683049147443, + "loss": 1.8841, + "theoretical_loss": 4.551679446764677, + "tokens_seen": 148832256 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004822668004012036, + "loss": 1.9449, + "theoretical_loss": 4.551418838577531, + "tokens_seen": 148897792 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004822567703109328, + "loss": 1.8894, + "theoretical_loss": 4.551158377170557, + "tokens_seen": 148963328 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048224674022066203, + "loss": 1.9383, + "theoretical_loss": 4.5508980623965645, + "tokens_seen": 149028864 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048223671013039116, + "loss": 1.8651, + "theoretical_loss": 4.550637894108572, + "tokens_seen": 149094400 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004822266800401204, + "loss": 1.8567, + "theoretical_loss": 4.5503778721598165, + "tokens_seen": 149159936 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004822166499498495, + "loss": 2.1398, + "theoretical_loss": 4.55011799640374, + "tokens_seen": 149225472 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048220661985957876, + "loss": 1.9277, + "theoretical_loss": 4.549858266694001, + "tokens_seen": 149291008 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048219658976930794, + "loss": 1.8661, + "theoretical_loss": 4.549598682884463, + "tokens_seen": 149356544 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004821865596790371, + "loss": 1.9282, + "theoretical_loss": 4.549339244829204, + "tokens_seen": 149422080 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004821765295887663, + "loss": 1.9147, + "theoretical_loss": 4.549079952382511, + "tokens_seen": 149487616 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004821664994984955, + "loss": 1.8857, + "theoretical_loss": 4.548820805398881, + "tokens_seen": 149553152 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048215646940822466, + "loss": 1.875, + "theoretical_loss": 4.548561803733017, + "tokens_seen": 149618688 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004821464393179539, + "loss": 1.9125, + "theoretical_loss": 4.548302947239834, + "tokens_seen": 149684224 + }, + { + "epoch": 0.05, + "learning_rate": 0.000482136409227683, + "loss": 1.8883, + "theoretical_loss": 4.548044235774454, + "tokens_seen": 149749760 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048212637913741226, + "loss": 1.9913, + "theoretical_loss": 4.547785669192207, + "tokens_seen": 149815296 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004821163490471414, + "loss": 1.8231, + "theoretical_loss": 4.547527247348631, + "tokens_seen": 149880832 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004821063189568706, + "loss": 1.8009, + "theoretical_loss": 4.547268970099471, + "tokens_seen": 149946368 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004820962888665998, + "loss": 1.8717, + "theoretical_loss": 4.547010837300677, + "tokens_seen": 150011904 + }, + { + "epoch": 0.05, + "learning_rate": 0.000482086258776329, + "loss": 1.9013, + "theoretical_loss": 4.546752848808408, + "tokens_seen": 150077440 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048207622868605816, + "loss": 1.993, + "theoretical_loss": 4.546495004479027, + "tokens_seen": 150142976 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004820661985957874, + "loss": 1.9689, + "theoretical_loss": 4.546237304169104, + "tokens_seen": 150208512 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048205616850551653, + "loss": 1.9376, + "theoretical_loss": 4.545979747735413, + "tokens_seen": 150274048 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048204613841524576, + "loss": 1.8142, + "theoretical_loss": 4.5457223350349345, + "tokens_seen": 150339584 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004820361083249749, + "loss": 1.8846, + "theoretical_loss": 4.545465065924851, + "tokens_seen": 150405120 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004820260782347041, + "loss": 1.8657, + "theoretical_loss": 4.54520794026255, + "tokens_seen": 150470656 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004820160481444333, + "loss": 1.9199, + "theoretical_loss": 4.544950957905625, + "tokens_seen": 150536192 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004820060180541625, + "loss": 1.9142, + "theoretical_loss": 4.54469411871187, + "tokens_seen": 150601728 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048199598796389167, + "loss": 1.9506, + "theoretical_loss": 4.544437422539282, + "tokens_seen": 150667264 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 63294, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8111944198608398, + "objective/train/theoretical_loss": 4.544180869246063, + "objective/train/tokens_used": 150732800, + "theoretical_loss": 4.544180869246063, + "tokens_seen": 150732800 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048198595787362085, + "loss": 1.7763, + "theoretical_loss": 4.544180869246063, + "tokens_seen": 150732800 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004819759277833501, + "loss": 1.9394, + "theoretical_loss": 4.543924458690612, + "tokens_seen": 150798336 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048196589769307927, + "loss": 1.9353, + "theoretical_loss": 4.5436681907315375, + "tokens_seen": 150863872 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048195586760280845, + "loss": 1.8317, + "theoretical_loss": 4.543412065227642, + "tokens_seen": 150929408 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048194583751253763, + "loss": 1.8921, + "theoretical_loss": 4.5431560820379335, + "tokens_seen": 150994944 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048193580742226686, + "loss": 1.9087, + "theoretical_loss": 4.54290024102162, + "tokens_seen": 151060480 + }, + { + "epoch": 0.05, + "learning_rate": 0.000481925777331996, + "loss": 1.7975, + "theoretical_loss": 4.542644542038108, + "tokens_seen": 151126016 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004819157472417252, + "loss": 1.8234, + "theoretical_loss": 4.542388984947005, + "tokens_seen": 151191552 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048190571715145435, + "loss": 1.9103, + "theoretical_loss": 4.5421335696081195, + "tokens_seen": 151257088 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004818956870611836, + "loss": 1.876, + "theoretical_loss": 4.541878295881456, + "tokens_seen": 151322624 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048188565697091277, + "loss": 1.9183, + "theoretical_loss": 4.541623163627221, + "tokens_seen": 151388160 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048187562688064195, + "loss": 1.9519, + "theoretical_loss": 4.541368172705817, + "tokens_seen": 151453696 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048186559679037113, + "loss": 1.9017, + "theoretical_loss": 4.541113322977846, + "tokens_seen": 151519232 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004818555667001003, + "loss": 1.9711, + "theoretical_loss": 4.540858614304108, + "tokens_seen": 151584768 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004818455366098295, + "loss": 1.8077, + "theoretical_loss": 4.5406040465455995, + "tokens_seen": 151650304 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048183550651955873, + "loss": 1.883, + "theoretical_loss": 4.5403496195635125, + "tokens_seen": 151715840 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048182547642928786, + "loss": 1.9822, + "theoretical_loss": 4.54009533321924, + "tokens_seen": 151781376 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004818154463390171, + "loss": 1.8812, + "theoretical_loss": 4.539841187374366, + "tokens_seen": 151846912 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004818054162487462, + "loss": 1.8571, + "theoretical_loss": 4.539587181890674, + "tokens_seen": 151912448 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048179538615847545, + "loss": 1.9318, + "theoretical_loss": 4.539333316630143, + "tokens_seen": 151977984 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048178535606820463, + "loss": 1.8283, + "theoretical_loss": 4.539079591454943, + "tokens_seen": 152043520 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004817753259779338, + "loss": 1.8234, + "theoretical_loss": 4.5388260062274455, + "tokens_seen": 152109056 + }, + { + "epoch": 0.05, + "learning_rate": 0.000481765295887663, + "loss": 1.8944, + "theoretical_loss": 4.538572560810211, + "tokens_seen": 152174592 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048175526579739223, + "loss": 1.815, + "theoretical_loss": 4.538319255065995, + "tokens_seen": 152240128 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048174523570712136, + "loss": 1.9054, + "theoretical_loss": 4.53806608885775, + "tokens_seen": 152305664 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004817352056168506, + "loss": 1.8051, + "theoretical_loss": 4.537813062048618, + "tokens_seen": 152371200 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004817251755265797, + "loss": 1.9054, + "theoretical_loss": 4.537560174501936, + "tokens_seen": 152436736 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048171514543630896, + "loss": 1.8507, + "theoretical_loss": 4.537307426081232, + "tokens_seen": 152502272 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048170511534603814, + "loss": 1.9375, + "theoretical_loss": 4.53705481665023, + "tokens_seen": 152567808 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004816950852557673, + "loss": 1.8465, + "theoretical_loss": 4.536802346072841, + "tokens_seen": 152633344 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004816850551654965, + "loss": 1.8206, + "theoretical_loss": 4.536550014213171, + "tokens_seen": 152698880 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004816750250752257, + "loss": 1.9292, + "theoretical_loss": 4.536297820935517, + "tokens_seen": 152764416 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048166499498495486, + "loss": 2.0147, + "theoretical_loss": 4.536045766104366, + "tokens_seen": 152829952 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004816549648946841, + "loss": 1.9022, + "theoretical_loss": 4.535793849584396, + "tokens_seen": 152895488 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004816449348044132, + "loss": 1.9183, + "theoretical_loss": 4.535542071240473, + "tokens_seen": 152961024 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048163490471414246, + "loss": 1.9172, + "theoretical_loss": 4.535290430937657, + "tokens_seen": 153026560 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004816248746238716, + "loss": 1.992, + "theoretical_loss": 4.535038928541194, + "tokens_seen": 153092096 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004816148445336008, + "loss": 1.816, + "theoretical_loss": 4.534787563916522, + "tokens_seen": 153157632 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048160481444333, + "loss": 1.9888, + "theoretical_loss": 4.534536336929264, + "tokens_seen": 153223168 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004815947843530592, + "loss": 1.8792, + "theoretical_loss": 4.534285247445236, + "tokens_seen": 153288704 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048158475426278837, + "loss": 1.836, + "theoretical_loss": 4.534034295330439, + "tokens_seen": 153354240 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004815747241725176, + "loss": 1.81, + "theoretical_loss": 4.5337834804510635, + "tokens_seen": 153419776 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048156469408224673, + "loss": 1.8421, + "theoretical_loss": 4.533532802673484, + "tokens_seen": 153485312 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048155466399197596, + "loss": 1.7624, + "theoretical_loss": 4.533282261864267, + "tokens_seen": 153550848 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004815446339017051, + "loss": 1.7916, + "theoretical_loss": 4.533031857890164, + "tokens_seen": 153616384 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004815346038114343, + "loss": 1.9005, + "theoretical_loss": 4.532781590618109, + "tokens_seen": 153681920 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004815245737211635, + "loss": 1.8373, + "theoretical_loss": 4.532531459915229, + "tokens_seen": 153747456 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004815145436308927, + "loss": 1.9814, + "theoretical_loss": 4.532281465648831, + "tokens_seen": 153812992 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048150451354062187, + "loss": 1.8066, + "theoretical_loss": 4.532031607686411, + "tokens_seen": 153878528 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048149448345035105, + "loss": 1.8928, + "theoretical_loss": 4.531781885895647, + "tokens_seen": 153944064 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 64627, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7535953521728516, + "objective/train/theoretical_loss": 4.531532300144404, + "objective/train/tokens_used": 154009600, + "theoretical_loss": 4.531532300144404, + "tokens_seen": 154009600 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048148445336008023, + "loss": 1.8486, + "theoretical_loss": 4.531532300144404, + "tokens_seen": 154009600 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048147442326980947, + "loss": 1.899, + "theoretical_loss": 4.531282850300731, + "tokens_seen": 154075136 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004814643931795386, + "loss": 1.8367, + "theoretical_loss": 4.53103353623286, + "tokens_seen": 154140672 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048145436308926783, + "loss": 1.8792, + "theoretical_loss": 4.530784357809209, + "tokens_seen": 154206208 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048144433299899696, + "loss": 1.8146, + "theoretical_loss": 4.5305353148983745, + "tokens_seen": 154271744 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004814343029087262, + "loss": 1.7997, + "theoretical_loss": 4.530286407369143, + "tokens_seen": 154337280 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048142427281845537, + "loss": 1.806, + "theoretical_loss": 4.530037635090476, + "tokens_seen": 154402816 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048141424272818455, + "loss": 1.7984, + "theoretical_loss": 4.529788997931526, + "tokens_seen": 154468352 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048140421263791373, + "loss": 1.8205, + "theoretical_loss": 4.52954049576162, + "tokens_seen": 154533888 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048139418254764297, + "loss": 1.917, + "theoretical_loss": 4.529292128450271, + "tokens_seen": 154599424 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004813841524573721, + "loss": 1.9199, + "theoretical_loss": 4.52904389586717, + "tokens_seen": 154664960 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048137412236710133, + "loss": 1.8787, + "theoretical_loss": 4.528795797882193, + "tokens_seen": 154730496 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048136409227683046, + "loss": 1.916, + "theoretical_loss": 4.528547834365394, + "tokens_seen": 154796032 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004813540621865597, + "loss": 1.828, + "theoretical_loss": 4.52830000518701, + "tokens_seen": 154861568 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004813440320962889, + "loss": 1.7833, + "theoretical_loss": 4.528052310217453, + "tokens_seen": 154927104 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048133400200601806, + "loss": 1.9079, + "theoretical_loss": 4.527804749327321, + "tokens_seen": 154992640 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048132397191574724, + "loss": 1.8039, + "theoretical_loss": 4.527557322387387, + "tokens_seen": 155058176 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004813139418254764, + "loss": 1.8613, + "theoretical_loss": 4.527310029268604, + "tokens_seen": 155123712 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004813039117352056, + "loss": 2.031, + "theoretical_loss": 4.527062869842107, + "tokens_seen": 155189248 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048129388164493483, + "loss": 1.8818, + "theoretical_loss": 4.526815843979204, + "tokens_seen": 155254784 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048128385155466396, + "loss": 1.764, + "theoretical_loss": 4.5265689515513845, + "tokens_seen": 155320320 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004812738214643932, + "loss": 1.7745, + "theoretical_loss": 4.526322192430316, + "tokens_seen": 155385856 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004812637913741223, + "loss": 1.8288, + "theoretical_loss": 4.526075566487842, + "tokens_seen": 155451392 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048125376128385156, + "loss": 1.923, + "theoretical_loss": 4.525829073595984, + "tokens_seen": 155516928 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048124373119358074, + "loss": 1.8225, + "theoretical_loss": 4.5255827136269415, + "tokens_seen": 155582464 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004812337011033099, + "loss": 1.7678, + "theoretical_loss": 4.525336486453088, + "tokens_seen": 155648000 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048122367101303916, + "loss": 1.7769, + "theoretical_loss": 4.525090391946974, + "tokens_seen": 155713536 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048121364092276834, + "loss": 1.8517, + "theoretical_loss": 4.524844429981327, + "tokens_seen": 155779072 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004812036108324975, + "loss": 1.8757, + "theoretical_loss": 4.52459860042905, + "tokens_seen": 155844608 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004811935807422267, + "loss": 1.9301, + "theoretical_loss": 4.524352903163219, + "tokens_seen": 155910144 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004811835506519559, + "loss": 1.9354, + "theoretical_loss": 4.524107338057089, + "tokens_seen": 155975680 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048117352056168506, + "loss": 1.8005, + "theoretical_loss": 4.523861904984086, + "tokens_seen": 156041216 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004811634904714143, + "loss": 1.7698, + "theoretical_loss": 4.5236166038178105, + "tokens_seen": 156106752 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004811534603811434, + "loss": 1.9425, + "theoretical_loss": 4.52337143443204, + "tokens_seen": 156172288 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048114343029087266, + "loss": 1.8734, + "theoretical_loss": 4.523126396700722, + "tokens_seen": 156237824 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004811334002006018, + "loss": 2.0152, + "theoretical_loss": 4.522881490497981, + "tokens_seen": 156303360 + }, + { + "epoch": 0.05, + "learning_rate": 0.000481123370110331, + "loss": 1.8864, + "theoretical_loss": 4.522636715698111, + "tokens_seen": 156368896 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004811133400200602, + "loss": 1.9455, + "theoretical_loss": 4.522392072175582, + "tokens_seen": 156434432 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004811033099297894, + "loss": 1.8527, + "theoretical_loss": 4.522147559805034, + "tokens_seen": 156499968 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048109327983951857, + "loss": 1.8936, + "theoretical_loss": 4.5219031784612795, + "tokens_seen": 156565504 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004810832497492478, + "loss": 1.8704, + "theoretical_loss": 4.521658928019305, + "tokens_seen": 156631040 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048107321965897693, + "loss": 1.7697, + "theoretical_loss": 4.521414808354266, + "tokens_seen": 156696576 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048106318956870616, + "loss": 1.9562, + "theoretical_loss": 4.52117081934149, + "tokens_seen": 156762112 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004810531594784353, + "loss": 1.855, + "theoretical_loss": 4.520926960856476, + "tokens_seen": 156827648 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004810431293881645, + "loss": 1.9667, + "theoretical_loss": 4.520683232774893, + "tokens_seen": 156893184 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004810330992978937, + "loss": 1.9148, + "theoretical_loss": 4.520439634972579, + "tokens_seen": 156958720 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004810230692076229, + "loss": 1.9235, + "theoretical_loss": 4.520196167325546, + "tokens_seen": 157024256 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048101303911735207, + "loss": 1.8956, + "theoretical_loss": 4.519952829709972, + "tokens_seen": 157089792 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048100300902708125, + "loss": 1.7984, + "theoretical_loss": 4.519709622002205, + "tokens_seen": 157155328 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048099297893681043, + "loss": 1.8182, + "theoretical_loss": 4.519466544078762, + "tokens_seen": 157220864 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 65757, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8998664617538452, + "objective/train/theoretical_loss": 4.519223595816332, + "objective/train/tokens_used": 157286400, + "theoretical_loss": 4.519223595816332, + "tokens_seen": 157286400 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048098294884653967, + "loss": 1.9051, + "theoretical_loss": 4.519223595816332, + "tokens_seen": 157286400 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004809729187562688, + "loss": 1.9033, + "theoretical_loss": 4.518980777091768, + "tokens_seen": 157351936 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048096288866599803, + "loss": 1.8386, + "theoretical_loss": 4.518738087782093, + "tokens_seen": 157417472 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048095285857572716, + "loss": 1.8895, + "theoretical_loss": 4.5184955277644985, + "tokens_seen": 157483008 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004809428284854564, + "loss": 2.0323, + "theoretical_loss": 4.518253096916341, + "tokens_seen": 157548544 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048093279839518557, + "loss": 1.8835, + "theoretical_loss": 4.518010795115151, + "tokens_seen": 157614080 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048092276830491475, + "loss": 1.8612, + "theoretical_loss": 4.517768622238616, + "tokens_seen": 157679616 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048091273821464393, + "loss": 1.7087, + "theoretical_loss": 4.5175265781646, + "tokens_seen": 157745152 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048090270812437317, + "loss": 1.8653, + "theoretical_loss": 4.517284662771125, + "tokens_seen": 157810688 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004808926780341023, + "loss": 1.9005, + "theoretical_loss": 4.5170428759363865, + "tokens_seen": 157876224 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048088264794383153, + "loss": 2.041, + "theoretical_loss": 4.516801217538739, + "tokens_seen": 157941760 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048087261785356066, + "loss": 2.0099, + "theoretical_loss": 4.516559687456708, + "tokens_seen": 158007296 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004808625877632899, + "loss": 1.8628, + "theoretical_loss": 4.516318285568982, + "tokens_seen": 158072832 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004808525576730191, + "loss": 1.8063, + "theoretical_loss": 4.516077011754414, + "tokens_seen": 158138368 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048084252758274826, + "loss": 2.0319, + "theoretical_loss": 4.51583586589202, + "tokens_seen": 158203904 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048083249749247744, + "loss": 1.8786, + "theoretical_loss": 4.515594847860985, + "tokens_seen": 158269440 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004808224674022066, + "loss": 1.9404, + "theoretical_loss": 4.515353957540656, + "tokens_seen": 158334976 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004808124373119358, + "loss": 1.9961, + "theoretical_loss": 4.515113194810541, + "tokens_seen": 158400512 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048080240722166503, + "loss": 1.9312, + "theoretical_loss": 4.514872559550316, + "tokens_seen": 158466048 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048079237713139416, + "loss": 1.8581, + "theoretical_loss": 4.514632051639815, + "tokens_seen": 158531584 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004807823470411234, + "loss": 1.8961, + "theoretical_loss": 4.51439167095904, + "tokens_seen": 158597120 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004807723169508525, + "loss": 1.8483, + "theoretical_loss": 4.5141514173881525, + "tokens_seen": 158662656 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048076228686058176, + "loss": 1.9004, + "theoretical_loss": 4.513911290807478, + "tokens_seen": 158728192 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048075225677031094, + "loss": 1.7617, + "theoretical_loss": 4.513671291097502, + "tokens_seen": 158793728 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004807422266800401, + "loss": 1.9387, + "theoretical_loss": 4.513431418138874, + "tokens_seen": 158859264 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004807321965897693, + "loss": 1.9882, + "theoretical_loss": 4.513191671812404, + "tokens_seen": 158924800 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048072216649949854, + "loss": 1.9221, + "theoretical_loss": 4.512952051999061, + "tokens_seen": 158990336 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048071213640922766, + "loss": 1.8418, + "theoretical_loss": 4.512712558579979, + "tokens_seen": 159055872 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004807021063189569, + "loss": 1.8975, + "theoretical_loss": 4.512473191436451, + "tokens_seen": 159121408 + }, + { + "epoch": 0.05, + "learning_rate": 0.000480692076228686, + "loss": 1.9629, + "theoretical_loss": 4.512233950449927, + "tokens_seen": 159186944 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048068204613841526, + "loss": 1.7667, + "theoretical_loss": 4.511994835502023, + "tokens_seen": 159252480 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048067201604814444, + "loss": 1.8919, + "theoretical_loss": 4.511755846474511, + "tokens_seen": 159318016 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004806619859578736, + "loss": 2.0707, + "theoretical_loss": 4.511516983249322, + "tokens_seen": 159383552 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004806519558676028, + "loss": 1.8853, + "theoretical_loss": 4.511278245708548, + "tokens_seen": 159449088 + }, + { + "epoch": 0.05, + "learning_rate": 0.000480641925777332, + "loss": 1.9224, + "theoretical_loss": 4.511039633734439, + "tokens_seen": 159514624 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048063189568706117, + "loss": 1.8359, + "theoretical_loss": 4.5108011472094045, + "tokens_seen": 159580160 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004806218655967904, + "loss": 1.8872, + "theoretical_loss": 4.510562786016012, + "tokens_seen": 159645696 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048061183550651953, + "loss": 1.8675, + "theoretical_loss": 4.5103245500369855, + "tokens_seen": 159711232 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048060180541624877, + "loss": 1.8736, + "theoretical_loss": 4.510086439155209, + "tokens_seen": 159776768 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048059177532597795, + "loss": 1.9262, + "theoretical_loss": 4.509848453253722, + "tokens_seen": 159842304 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048058174523570713, + "loss": 2.001, + "theoretical_loss": 4.509610592215726, + "tokens_seen": 159907840 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004805717151454363, + "loss": 1.9164, + "theoretical_loss": 4.509372855924571, + "tokens_seen": 159973376 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004805616850551655, + "loss": 1.7763, + "theoretical_loss": 4.5091352442637715, + "tokens_seen": 160038912 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048055165496489467, + "loss": 1.8798, + "theoretical_loss": 4.508897757116995, + "tokens_seen": 160104448 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004805416248746239, + "loss": 1.907, + "theoretical_loss": 4.508660394368066, + "tokens_seen": 160169984 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048053159478435303, + "loss": 1.9742, + "theoretical_loss": 4.508423155900964, + "tokens_seen": 160235520 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048052156469408227, + "loss": 1.8149, + "theoretical_loss": 4.508186041599823, + "tokens_seen": 160301056 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004805115346038114, + "loss": 1.7339, + "theoretical_loss": 4.5079490513489375, + "tokens_seen": 160366592 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048050150451354063, + "loss": 1.9188, + "theoretical_loss": 4.50771218503275, + "tokens_seen": 160432128 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004804914744232698, + "loss": 2.0351, + "theoretical_loss": 4.507475442535863, + "tokens_seen": 160497664 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 66913, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.09786319732666, + "objective/train/theoretical_loss": 4.507238823743033, + "objective/train/tokens_used": 160563200, + "theoretical_loss": 4.507238823743033, + "tokens_seen": 160563200 + }, + { + "epoch": 0.05, + "learning_rate": 0.000480481444332999, + "loss": 1.9574, + "theoretical_loss": 4.507238823743033, + "tokens_seen": 160563200 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048047141424272823, + "loss": 1.8228, + "theoretical_loss": 4.507002328539167, + "tokens_seen": 160628736 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048046138415245736, + "loss": 1.7859, + "theoretical_loss": 4.506765956809331, + "tokens_seen": 160694272 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004804513540621866, + "loss": 1.8968, + "theoretical_loss": 4.50652970843874, + "tokens_seen": 160759808 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048044132397191577, + "loss": 1.8651, + "theoretical_loss": 4.506293583312766, + "tokens_seen": 160825344 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048043129388164495, + "loss": 1.6559, + "theoretical_loss": 4.506057581316933, + "tokens_seen": 160890880 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048042126379137413, + "loss": 1.8197, + "theoretical_loss": 4.505821702336917, + "tokens_seen": 160956416 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048041123370110337, + "loss": 1.8893, + "theoretical_loss": 4.505585946258547, + "tokens_seen": 161021952 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004804012036108325, + "loss": 1.8618, + "theoretical_loss": 4.505350312967805, + "tokens_seen": 161087488 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048039117352056173, + "loss": 1.7682, + "theoretical_loss": 4.5051148023508265, + "tokens_seen": 161153024 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048038114343029086, + "loss": 1.9401, + "theoretical_loss": 4.504879414293896, + "tokens_seen": 161218560 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004803711133400201, + "loss": 1.9618, + "theoretical_loss": 4.50464414868345, + "tokens_seen": 161284096 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004803610832497493, + "loss": 1.8745, + "theoretical_loss": 4.504409005406078, + "tokens_seen": 161349632 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048035105315947846, + "loss": 1.9029, + "theoretical_loss": 4.504173984348521, + "tokens_seen": 161415168 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048034102306920764, + "loss": 1.9127, + "theoretical_loss": 4.503939085397667, + "tokens_seen": 161480704 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004803309929789368, + "loss": 1.9016, + "theoretical_loss": 4.5037043084405575, + "tokens_seen": 161546240 + }, + { + "epoch": 0.05, + "learning_rate": 0.000480320962888666, + "loss": 1.8548, + "theoretical_loss": 4.503469653364386, + "tokens_seen": 161611776 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048031093279839523, + "loss": 1.941, + "theoretical_loss": 4.503235120056491, + "tokens_seen": 161677312 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048030090270812436, + "loss": 1.907, + "theoretical_loss": 4.503000708404365, + "tokens_seen": 161742848 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004802908726178536, + "loss": 1.7849, + "theoretical_loss": 4.502766418295647, + "tokens_seen": 161808384 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004802808425275827, + "loss": 1.8813, + "theoretical_loss": 4.502532249618129, + "tokens_seen": 161873920 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048027081243731196, + "loss": 1.6863, + "theoretical_loss": 4.502298202259748, + "tokens_seen": 161939456 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048026078234704114, + "loss": 1.844, + "theoretical_loss": 4.502064276108593, + "tokens_seen": 162004992 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004802507522567703, + "loss": 1.8705, + "theoretical_loss": 4.501830471052898, + "tokens_seen": 162070528 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004802407221664995, + "loss": 1.8797, + "theoretical_loss": 4.501596786981047, + "tokens_seen": 162136064 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048023069207622874, + "loss": 1.8275, + "theoretical_loss": 4.501363223781574, + "tokens_seen": 162201600 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048022066198595786, + "loss": 1.8441, + "theoretical_loss": 4.501129781343157, + "tokens_seen": 162267136 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004802106318956871, + "loss": 1.8469, + "theoretical_loss": 4.500896459554624, + "tokens_seen": 162332672 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048020060180541623, + "loss": 1.9162, + "theoretical_loss": 4.500663258304948, + "tokens_seen": 162398208 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048019057171514546, + "loss": 1.8629, + "theoretical_loss": 4.500430177483253, + "tokens_seen": 162463744 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048018054162487464, + "loss": 1.7615, + "theoretical_loss": 4.500197216978804, + "tokens_seen": 162529280 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004801705115346038, + "loss": 1.7546, + "theoretical_loss": 4.499964376681016, + "tokens_seen": 162594816 + }, + { + "epoch": 0.05, + "learning_rate": 0.000480160481444333, + "loss": 1.8942, + "theoretical_loss": 4.49973165647945, + "tokens_seen": 162660352 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004801504513540622, + "loss": 1.9138, + "theoretical_loss": 4.4994990562638115, + "tokens_seen": 162725888 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048014042126379137, + "loss": 1.9891, + "theoretical_loss": 4.499266575923954, + "tokens_seen": 162791424 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004801303911735206, + "loss": 1.8261, + "theoretical_loss": 4.499034215349875, + "tokens_seen": 162856960 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048012036108324973, + "loss": 1.7725, + "theoretical_loss": 4.498801974431714, + "tokens_seen": 162922496 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048011033099297897, + "loss": 1.8221, + "theoretical_loss": 4.498569853059762, + "tokens_seen": 162988032 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048010030090270815, + "loss": 1.8698, + "theoretical_loss": 4.49833785112445, + "tokens_seen": 163053568 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048009027081243733, + "loss": 1.8533, + "theoretical_loss": 4.498105968516354, + "tokens_seen": 163119104 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004800802407221665, + "loss": 1.8899, + "theoretical_loss": 4.497874205126195, + "tokens_seen": 163184640 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004800702106318957, + "loss": 1.9078, + "theoretical_loss": 4.497642560844838, + "tokens_seen": 163250176 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048006018054162487, + "loss": 2.0557, + "theoretical_loss": 4.497411035563291, + "tokens_seen": 163315712 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004800501504513541, + "loss": 1.8185, + "theoretical_loss": 4.497179629172704, + "tokens_seen": 163381248 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048004012036108323, + "loss": 1.8714, + "theoretical_loss": 4.4969483415643765, + "tokens_seen": 163446784 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048003009027081247, + "loss": 1.8753, + "theoretical_loss": 4.496717172629742, + "tokens_seen": 163512320 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004800200601805416, + "loss": 1.9686, + "theoretical_loss": 4.496486122260382, + "tokens_seen": 163577856 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048001003009027083, + "loss": 1.7575, + "theoretical_loss": 4.4962551903480215, + "tokens_seen": 163643392 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048, + "loss": 1.8742, + "theoretical_loss": 4.496024376784522, + "tokens_seen": 163708928 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004799899699097292, + "loss": 1.8151, + "theoretical_loss": 4.495793681461894, + "tokens_seen": 163774464 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 68034, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8759820461273193, + "objective/train/theoretical_loss": 4.495563104272286, + "objective/train/tokens_used": 163840000, + "theoretical_loss": 4.495563104272286, + "tokens_seen": 163840000 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004799799398194584, + "loss": 1.8935, + "theoretical_loss": 4.495563104272286, + "tokens_seen": 163840000 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047996990972918756, + "loss": 1.7944, + "theoretical_loss": 4.495332645107985, + "tokens_seen": 163905536 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047995987963891674, + "loss": 1.8803, + "theoretical_loss": 4.495102303861426, + "tokens_seen": 163971072 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047994984954864597, + "loss": 1.8419, + "theoretical_loss": 4.4948720804251785, + "tokens_seen": 164036608 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004799398194583751, + "loss": 1.7363, + "theoretical_loss": 4.494641974691959, + "tokens_seen": 164102144 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047992978936810433, + "loss": 1.8228, + "theoretical_loss": 4.494411986554619, + "tokens_seen": 164167680 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004799197592778335, + "loss": 1.8407, + "theoretical_loss": 4.494182115906153, + "tokens_seen": 164233216 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004799097291875627, + "loss": 1.7393, + "theoretical_loss": 4.493952362639694, + "tokens_seen": 164298752 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004798996990972919, + "loss": 1.9264, + "theoretical_loss": 4.493722726648519, + "tokens_seen": 164364288 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047988966900702106, + "loss": 1.9542, + "theoretical_loss": 4.493493207826037, + "tokens_seen": 164429824 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047987963891675024, + "loss": 1.7552, + "theoretical_loss": 4.493263806065803, + "tokens_seen": 164495360 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004798696088264795, + "loss": 1.7857, + "theoretical_loss": 4.493034521261509, + "tokens_seen": 164560896 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004798595787362086, + "loss": 1.9283, + "theoretical_loss": 4.492805353306984, + "tokens_seen": 164626432 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047984954864593784, + "loss": 1.8938, + "theoretical_loss": 4.492576302096199, + "tokens_seen": 164691968 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047983951855566696, + "loss": 1.8091, + "theoretical_loss": 4.49234736752326, + "tokens_seen": 164757504 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004798294884653962, + "loss": 1.8954, + "theoretical_loss": 4.492118549482413, + "tokens_seen": 164823040 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004798194583751254, + "loss": 1.8802, + "theoretical_loss": 4.491889847868041, + "tokens_seen": 164888576 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047980942828485456, + "loss": 1.8934, + "theoretical_loss": 4.491661262574665, + "tokens_seen": 164954112 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047979939819458374, + "loss": 1.9162, + "theoretical_loss": 4.491432793496945, + "tokens_seen": 165019648 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004797893681043129, + "loss": 1.8467, + "theoretical_loss": 4.4912044405296765, + "tokens_seen": 165085184 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004797793380140421, + "loss": 1.8671, + "theoretical_loss": 4.49097620356779, + "tokens_seen": 165150720 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047976930792377134, + "loss": 1.8632, + "theoretical_loss": 4.490748082506356, + "tokens_seen": 165216256 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047975927783350047, + "loss": 1.8333, + "theoretical_loss": 4.490520077240582, + "tokens_seen": 165281792 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004797492477432297, + "loss": 1.7771, + "theoretical_loss": 4.490292187665808, + "tokens_seen": 165347328 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004797392176529589, + "loss": 1.8731, + "theoretical_loss": 4.490064413677514, + "tokens_seen": 165412864 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047972918756268807, + "loss": 1.9695, + "theoretical_loss": 4.4898367551713125, + "tokens_seen": 165478400 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004797191574724173, + "loss": 1.874, + "theoretical_loss": 4.489609212042954, + "tokens_seen": 165543936 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047970912738214643, + "loss": 1.9639, + "theoretical_loss": 4.489381784188322, + "tokens_seen": 165609472 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047969909729187566, + "loss": 1.9809, + "theoretical_loss": 4.489154471503438, + "tokens_seen": 165675008 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047968906720160484, + "loss": 1.9281, + "theoretical_loss": 4.488927273884456, + "tokens_seen": 165740544 + }, + { + "epoch": 0.05, + "learning_rate": 0.000479679037111334, + "loss": 1.8148, + "theoretical_loss": 4.488700191227665, + "tokens_seen": 165806080 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004796690070210632, + "loss": 1.7855, + "theoretical_loss": 4.488473223429491, + "tokens_seen": 165871616 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004796589769307924, + "loss": 1.7946, + "theoretical_loss": 4.48824637038649, + "tokens_seen": 165937152 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047964894684052157, + "loss": 1.7789, + "theoretical_loss": 4.488019631995354, + "tokens_seen": 166002688 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004796389167502508, + "loss": 1.9848, + "theoretical_loss": 4.48779300815291, + "tokens_seen": 166068224 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047962888665997993, + "loss": 1.8128, + "theoretical_loss": 4.4875664987561175, + "tokens_seen": 166133760 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047961885656970917, + "loss": 1.962, + "theoretical_loss": 4.487340103702068, + "tokens_seen": 166199296 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047960882647943835, + "loss": 1.7685, + "theoretical_loss": 4.487113822887988, + "tokens_seen": 166264832 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047959879638916753, + "loss": 1.8175, + "theoretical_loss": 4.486887656211234, + "tokens_seen": 166330368 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004795887662988967, + "loss": 1.7765, + "theoretical_loss": 4.486661603569301, + "tokens_seen": 166395904 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004795787362086259, + "loss": 1.8213, + "theoretical_loss": 4.486435664859809, + "tokens_seen": 166461440 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047956870611835507, + "loss": 1.8348, + "theoretical_loss": 4.486209839980516, + "tokens_seen": 166526976 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004795586760280843, + "loss": 1.8333, + "theoretical_loss": 4.485984128829307, + "tokens_seen": 166592512 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047954864593781343, + "loss": 1.8999, + "theoretical_loss": 4.485758531304205, + "tokens_seen": 166658048 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047953861584754267, + "loss": 1.9694, + "theoretical_loss": 4.485533047303358, + "tokens_seen": 166723584 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004795285857572718, + "loss": 1.9473, + "theoretical_loss": 4.485307676725048, + "tokens_seen": 166789120 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047951855566700103, + "loss": 1.8574, + "theoretical_loss": 4.485082419467691, + "tokens_seen": 166854656 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004795085255767302, + "loss": 1.8264, + "theoretical_loss": 4.484857275429828, + "tokens_seen": 166920192 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004794984954864594, + "loss": 1.8212, + "theoretical_loss": 4.484632244510136, + "tokens_seen": 166985728 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004794884653961886, + "loss": 1.8254, + "theoretical_loss": 4.4844073266074185, + "tokens_seen": 167051264 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 69574, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9637398719787598, + "objective/train/theoretical_loss": 4.484182521620611, + "objective/train/tokens_used": 167116800, + "theoretical_loss": 4.484182521620611, + "tokens_seen": 167116800 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047947843530591776, + "loss": 1.8428, + "theoretical_loss": 4.484182521620611, + "tokens_seen": 167116800 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047946840521564694, + "loss": 1.7845, + "theoretical_loss": 4.48395782944878, + "tokens_seen": 167182336 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047945837512537617, + "loss": 1.8533, + "theoretical_loss": 4.483733249991118, + "tokens_seen": 167247872 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004794483450351053, + "loss": 1.8598, + "theoretical_loss": 4.483508783146952, + "tokens_seen": 167313408 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047943831494483453, + "loss": 1.8805, + "theoretical_loss": 4.483284428815735, + "tokens_seen": 167378944 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004794282848545637, + "loss": 1.8257, + "theoretical_loss": 4.483060186897049, + "tokens_seen": 167444480 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004794182547642929, + "loss": 1.8869, + "theoretical_loss": 4.482836057290607, + "tokens_seen": 167510016 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004794082246740221, + "loss": 1.9014, + "theoretical_loss": 4.4826120398962495, + "tokens_seen": 167575552 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047939819458375126, + "loss": 1.8137, + "theoretical_loss": 4.482388134613943, + "tokens_seen": 167641088 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047938816449348044, + "loss": 1.8814, + "theoretical_loss": 4.482164341343788, + "tokens_seen": 167706624 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004793781344032097, + "loss": 1.7821, + "theoretical_loss": 4.481940659986008, + "tokens_seen": 167772160 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004793681043129388, + "loss": 1.7868, + "theoretical_loss": 4.481717090440956, + "tokens_seen": 167837696 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047935807422266804, + "loss": 1.7669, + "theoretical_loss": 4.481493632609112, + "tokens_seen": 167903232 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047934804413239716, + "loss": 1.844, + "theoretical_loss": 4.481270286391084, + "tokens_seen": 167968768 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004793380140421264, + "loss": 1.7384, + "theoretical_loss": 4.481047051687609, + "tokens_seen": 168034304 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004793279839518556, + "loss": 1.9285, + "theoretical_loss": 4.480823928399547, + "tokens_seen": 168099840 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047931795386158476, + "loss": 1.8336, + "theoretical_loss": 4.480600916427887, + "tokens_seen": 168165376 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047930792377131394, + "loss": 1.8853, + "theoretical_loss": 4.480378015673745, + "tokens_seen": 168230912 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004792978936810431, + "loss": 1.9212, + "theoretical_loss": 4.480155226038362, + "tokens_seen": 168296448 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004792878635907723, + "loss": 1.7997, + "theoretical_loss": 4.479932547423107, + "tokens_seen": 168361984 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047927783350050154, + "loss": 1.8857, + "theoretical_loss": 4.479709979729472, + "tokens_seen": 168427520 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047926780341023067, + "loss": 1.8425, + "theoretical_loss": 4.479487522859078, + "tokens_seen": 168493056 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004792577733199599, + "loss": 1.7901, + "theoretical_loss": 4.4792651767136675, + "tokens_seen": 168558592 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004792477432296891, + "loss": 1.8226, + "theoretical_loss": 4.479042941195113, + "tokens_seen": 168624128 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047923771313941827, + "loss": 1.8587, + "theoretical_loss": 4.478820816205408, + "tokens_seen": 168689664 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047922768304914745, + "loss": 1.79, + "theoretical_loss": 4.4785988016466725, + "tokens_seen": 168755200 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047921765295887663, + "loss": 1.8538, + "theoretical_loss": 4.4783768974211515, + "tokens_seen": 168820736 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004792076228686058, + "loss": 1.9558, + "theoretical_loss": 4.478155103431215, + "tokens_seen": 168886272 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047919759277833504, + "loss": 1.8175, + "theoretical_loss": 4.477933419579356, + "tokens_seen": 168951808 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047918756268806417, + "loss": 1.8759, + "theoretical_loss": 4.477711845768191, + "tokens_seen": 169017344 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004791775325977934, + "loss": 1.8803, + "theoretical_loss": 4.477490381900461, + "tokens_seen": 169082880 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047916750250752253, + "loss": 1.86, + "theoretical_loss": 4.477269027879032, + "tokens_seen": 169148416 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047915747241725177, + "loss": 1.8632, + "theoretical_loss": 4.477047783606892, + "tokens_seen": 169213952 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047914744232698095, + "loss": 1.9413, + "theoretical_loss": 4.476826648987151, + "tokens_seen": 169279488 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047913741223671013, + "loss": 1.8372, + "theoretical_loss": 4.476605623923044, + "tokens_seen": 169345024 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004791273821464393, + "loss": 1.7951, + "theoretical_loss": 4.47638470831793, + "tokens_seen": 169410560 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047911735205616855, + "loss": 1.9596, + "theoretical_loss": 4.476163902075286, + "tokens_seen": 169476096 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004791073219658977, + "loss": 1.8814, + "theoretical_loss": 4.475943205098716, + "tokens_seen": 169541632 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004790972918756269, + "loss": 1.8457, + "theoretical_loss": 4.475722617291944, + "tokens_seen": 169607168 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047908726178535604, + "loss": 2.022, + "theoretical_loss": 4.4755021385588165, + "tokens_seen": 169672704 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047907723169508527, + "loss": 1.9155, + "theoretical_loss": 4.475281768803301, + "tokens_seen": 169738240 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047906720160481445, + "loss": 1.7643, + "theoretical_loss": 4.475061507929487, + "tokens_seen": 169803776 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047905717151454363, + "loss": 1.9975, + "theoretical_loss": 4.474841355841587, + "tokens_seen": 169869312 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004790471414242728, + "loss": 1.8395, + "theoretical_loss": 4.474621312443932, + "tokens_seen": 169934848 + }, + { + "epoch": 0.05, + "learning_rate": 0.000479037111334002, + "loss": 1.9898, + "theoretical_loss": 4.474401377640977, + "tokens_seen": 170000384 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004790270812437312, + "loss": 1.8421, + "theoretical_loss": 4.474181551337294, + "tokens_seen": 170065920 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004790170511534604, + "loss": 1.8998, + "theoretical_loss": 4.473961833437579, + "tokens_seen": 170131456 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047900702106318954, + "loss": 1.9838, + "theoretical_loss": 4.473742223846646, + "tokens_seen": 170196992 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004789969909729188, + "loss": 1.822, + "theoretical_loss": 4.473522722469431, + "tokens_seen": 170262528 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004789869608826479, + "loss": 1.9278, + "theoretical_loss": 4.47330332921099, + "tokens_seen": 170328064 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 70789, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9570930004119873, + "objective/train/theoretical_loss": 4.473084043976495, + "objective/train/tokens_used": 170393600, + "theoretical_loss": 4.473084043976495, + "tokens_seen": 170393600 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047897693079237714, + "loss": 1.9023, + "theoretical_loss": 4.473084043976495, + "tokens_seen": 170393600 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047896690070210637, + "loss": 1.9553, + "theoretical_loss": 4.472864866671245, + "tokens_seen": 170459136 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004789568706118355, + "loss": 1.884, + "theoretical_loss": 4.472645797200651, + "tokens_seen": 170524672 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047894684052156473, + "loss": 1.8748, + "theoretical_loss": 4.472426835470247, + "tokens_seen": 170590208 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004789368104312939, + "loss": 1.9037, + "theoretical_loss": 4.472207981385686, + "tokens_seen": 170655744 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004789267803410231, + "loss": 1.8762, + "theoretical_loss": 4.471989234852737, + "tokens_seen": 170721280 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004789167502507523, + "loss": 1.8569, + "theoretical_loss": 4.471770595777292, + "tokens_seen": 170786816 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047890672016048146, + "loss": 1.8308, + "theoretical_loss": 4.4715520640653565, + "tokens_seen": 170852352 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047889669007021064, + "loss": 1.8257, + "theoretical_loss": 4.471333639623059, + "tokens_seen": 170917888 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004788866599799399, + "loss": 1.9129, + "theoretical_loss": 4.471115322356643, + "tokens_seen": 170983424 + }, + { + "epoch": 0.05, + "learning_rate": 0.000478876629889669, + "loss": 1.7896, + "theoretical_loss": 4.470897112172471, + "tokens_seen": 171048960 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047886659979939824, + "loss": 2.0778, + "theoretical_loss": 4.470679008977022, + "tokens_seen": 171114496 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047885656970912736, + "loss": 1.8114, + "theoretical_loss": 4.470461012676894, + "tokens_seen": 171180032 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004788465396188566, + "loss": 1.6854, + "theoretical_loss": 4.470243123178802, + "tokens_seen": 171245568 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004788365095285858, + "loss": 1.8587, + "theoretical_loss": 4.4700253403895776, + "tokens_seen": 171311104 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047882647943831496, + "loss": 1.895, + "theoretical_loss": 4.469807664216168, + "tokens_seen": 171376640 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047881644934804414, + "loss": 1.9044, + "theoretical_loss": 4.469590094565638, + "tokens_seen": 171442176 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004788064192577733, + "loss": 1.9978, + "theoretical_loss": 4.469372631345172, + "tokens_seen": 171507712 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004787963891675025, + "loss": 1.7586, + "theoretical_loss": 4.469155274462064, + "tokens_seen": 171573248 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047878635907723174, + "loss": 1.8075, + "theoretical_loss": 4.468938023823732, + "tokens_seen": 171638784 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047877632898696087, + "loss": 1.8921, + "theoretical_loss": 4.468720879337704, + "tokens_seen": 171704320 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004787662988966901, + "loss": 1.8943, + "theoretical_loss": 4.468503840911627, + "tokens_seen": 171769856 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004787562688064193, + "loss": 1.8887, + "theoretical_loss": 4.46828690845326, + "tokens_seen": 171835392 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047874623871614847, + "loss": 1.936, + "theoretical_loss": 4.468070081870483, + "tokens_seen": 171900928 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047873620862587765, + "loss": 1.7717, + "theoretical_loss": 4.467853361071284, + "tokens_seen": 171966464 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047872617853560683, + "loss": 1.8779, + "theoretical_loss": 4.467636745963772, + "tokens_seen": 172032000 + }, + { + "epoch": 0.05, + "learning_rate": 0.000478716148445336, + "loss": 1.8006, + "theoretical_loss": 4.467420236456169, + "tokens_seen": 172097536 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047870611835506524, + "loss": 1.6922, + "theoretical_loss": 4.46720383245681, + "tokens_seen": 172163072 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047869608826479437, + "loss": 1.9822, + "theoretical_loss": 4.466987533874145, + "tokens_seen": 172228608 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004786860581745236, + "loss": 1.9252, + "theoretical_loss": 4.4667713406167415, + "tokens_seen": 172294144 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047867602808425273, + "loss": 1.8408, + "theoretical_loss": 4.466555252593277, + "tokens_seen": 172359680 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047866599799398197, + "loss": 1.9607, + "theoretical_loss": 4.4663392697125435, + "tokens_seen": 172425216 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047865596790371115, + "loss": 1.7748, + "theoretical_loss": 4.466123391883448, + "tokens_seen": 172490752 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047864593781344033, + "loss": 1.8863, + "theoretical_loss": 4.4659076190150095, + "tokens_seen": 172556288 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004786359077231695, + "loss": 1.928, + "theoretical_loss": 4.465691951016364, + "tokens_seen": 172621824 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047862587763289875, + "loss": 1.7897, + "theoretical_loss": 4.465476387796755, + "tokens_seen": 172687360 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004786158475426279, + "loss": 1.7607, + "theoretical_loss": 4.465260929265543, + "tokens_seen": 172752896 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004786058174523571, + "loss": 1.8875, + "theoretical_loss": 4.4650455753322005, + "tokens_seen": 172818432 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047859578736208624, + "loss": 1.7212, + "theoretical_loss": 4.464830325906312, + "tokens_seen": 172883968 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047858575727181547, + "loss": 1.8209, + "theoretical_loss": 4.464615180897574, + "tokens_seen": 172949504 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047857572718154465, + "loss": 1.6843, + "theoretical_loss": 4.464400140215796, + "tokens_seen": 173015040 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047856569709127383, + "loss": 1.7962, + "theoretical_loss": 4.4641852037708984, + "tokens_seen": 173080576 + }, + { + "epoch": 0.05, + "learning_rate": 0.000478555667001003, + "loss": 1.7699, + "theoretical_loss": 4.463970371472916, + "tokens_seen": 173146112 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004785456369107322, + "loss": 1.8306, + "theoretical_loss": 4.463755643231992, + "tokens_seen": 173211648 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004785356068204614, + "loss": 1.846, + "theoretical_loss": 4.463541018958384, + "tokens_seen": 173277184 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004785255767301906, + "loss": 1.9998, + "theoretical_loss": 4.463326498562459, + "tokens_seen": 173342720 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047851554663991974, + "loss": 1.8235, + "theoretical_loss": 4.463112081954696, + "tokens_seen": 173408256 + }, + { + "epoch": 0.05, + "learning_rate": 0.000478505516549649, + "loss": 1.8905, + "theoretical_loss": 4.462897769045684, + "tokens_seen": 173473792 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004784954864593781, + "loss": 1.7852, + "theoretical_loss": 4.462683559746123, + "tokens_seen": 173539328 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047848545636910734, + "loss": 1.9151, + "theoretical_loss": 4.4624694539668255, + "tokens_seen": 173604864 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 71990, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7329236268997192, + "objective/train/theoretical_loss": 4.462255451618713, + "objective/train/tokens_used": 173670400, + "theoretical_loss": 4.462255451618713, + "tokens_seen": 173670400 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004784754262788365, + "loss": 1.8058, + "theoretical_loss": 4.462255451618713, + "tokens_seen": 173670400 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004784653961885657, + "loss": 1.7413, + "theoretical_loss": 4.462041552612815, + "tokens_seen": 173735936 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004784553660982949, + "loss": 1.828, + "theoretical_loss": 4.461827756860275, + "tokens_seen": 173801472 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004784453360080241, + "loss": 1.845, + "theoretical_loss": 4.461614064272345, + "tokens_seen": 173867008 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047843530591775324, + "loss": 1.8338, + "theoretical_loss": 4.461400474760383, + "tokens_seen": 173932544 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004784252758274825, + "loss": 1.7585, + "theoretical_loss": 4.461186988235864, + "tokens_seen": 173998080 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004784152457372116, + "loss": 1.7789, + "theoretical_loss": 4.460973604610366, + "tokens_seen": 174063616 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047840521564694084, + "loss": 1.8831, + "theoretical_loss": 4.4607603237955775, + "tokens_seen": 174129152 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047839518555667, + "loss": 1.7797, + "theoretical_loss": 4.460547145703298, + "tokens_seen": 174194688 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004783851554663992, + "loss": 1.8606, + "theoretical_loss": 4.460334070245436, + "tokens_seen": 174260224 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004783751253761284, + "loss": 1.9206, + "theoretical_loss": 4.460121097334004, + "tokens_seen": 174325760 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047836509528585756, + "loss": 1.8674, + "theoretical_loss": 4.4599082268811285, + "tokens_seen": 174391296 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047835506519558675, + "loss": 1.8316, + "theoretical_loss": 4.459695458799041, + "tokens_seen": 174456832 + }, + { + "epoch": 0.05, + "learning_rate": 0.000478345035105316, + "loss": 1.9675, + "theoretical_loss": 4.459482793000083, + "tokens_seen": 174522368 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004783350050150451, + "loss": 1.8524, + "theoretical_loss": 4.459270229396701, + "tokens_seen": 174587904 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047832497492477434, + "loss": 1.8676, + "theoretical_loss": 4.459057767901454, + "tokens_seen": 174653440 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047831494483450347, + "loss": 1.8692, + "theoretical_loss": 4.458845408427004, + "tokens_seen": 174718976 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004783049147442327, + "loss": 1.7387, + "theoretical_loss": 4.458633150886124, + "tokens_seen": 174784512 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004782948846539619, + "loss": 1.7887, + "theoretical_loss": 4.458420995191691, + "tokens_seen": 174850048 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047828485456369107, + "loss": 1.7138, + "theoretical_loss": 4.458208941256691, + "tokens_seen": 174915584 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047827482447342025, + "loss": 1.8316, + "theoretical_loss": 4.457996988994218, + "tokens_seen": 174981120 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004782647943831495, + "loss": 1.7814, + "theoretical_loss": 4.457785138317469, + "tokens_seen": 175046656 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004782547642928786, + "loss": 1.8261, + "theoretical_loss": 4.4575733891397515, + "tokens_seen": 175112192 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047824473420260785, + "loss": 1.7809, + "theoretical_loss": 4.457361741374477, + "tokens_seen": 175177728 + }, + { + "epoch": 0.05, + "learning_rate": 0.000478234704112337, + "loss": 1.7504, + "theoretical_loss": 4.457150194935165, + "tokens_seen": 175243264 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004782246740220662, + "loss": 1.9297, + "theoretical_loss": 4.456938749735439, + "tokens_seen": 175308800 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047821464393179544, + "loss": 1.8589, + "theoretical_loss": 4.45672740568903, + "tokens_seen": 175374336 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047820461384152457, + "loss": 1.8781, + "theoretical_loss": 4.456516162709774, + "tokens_seen": 175439872 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004781945837512538, + "loss": 1.852, + "theoretical_loss": 4.456305020711613, + "tokens_seen": 175505408 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047818455366098293, + "loss": 1.8335, + "theoretical_loss": 4.456093979608594, + "tokens_seen": 175570944 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047817452357071217, + "loss": 1.7862, + "theoretical_loss": 4.455883039314869, + "tokens_seen": 175636480 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047816449348044135, + "loss": 1.8388, + "theoretical_loss": 4.455672199744695, + "tokens_seen": 175702016 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047815446339017053, + "loss": 1.8903, + "theoretical_loss": 4.455461460812435, + "tokens_seen": 175767552 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004781444332998997, + "loss": 1.8927, + "theoretical_loss": 4.455250822432556, + "tokens_seen": 175833088 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047813440320962895, + "loss": 1.8587, + "theoretical_loss": 4.4550402845196295, + "tokens_seen": 175898624 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004781243731193581, + "loss": 1.7638, + "theoretical_loss": 4.45482984698833, + "tokens_seen": 175964160 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004781143430290873, + "loss": 1.8055, + "theoretical_loss": 4.454619509753439, + "tokens_seen": 176029696 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047810431293881644, + "loss": 1.8083, + "theoretical_loss": 4.45440927272984, + "tokens_seen": 176095232 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047809428284854567, + "loss": 1.7916, + "theoretical_loss": 4.454199135832521, + "tokens_seen": 176160768 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047808425275827485, + "loss": 1.8189, + "theoretical_loss": 4.4539890989765745, + "tokens_seen": 176226304 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047807422266800403, + "loss": 1.761, + "theoretical_loss": 4.453779162077193, + "tokens_seen": 176291840 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004780641925777332, + "loss": 1.745, + "theoretical_loss": 4.453569325049679, + "tokens_seen": 176357376 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004780541624874624, + "loss": 1.9617, + "theoretical_loss": 4.453359587809431, + "tokens_seen": 176422912 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004780441323971916, + "loss": 1.8346, + "theoretical_loss": 4.453149950271956, + "tokens_seen": 176488448 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004780341023069208, + "loss": 1.87, + "theoretical_loss": 4.452940412352861, + "tokens_seen": 176553984 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047802407221664994, + "loss": 1.8097, + "theoretical_loss": 4.452730973967856, + "tokens_seen": 176619520 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004780140421263792, + "loss": 1.9451, + "theoretical_loss": 4.452521635032754, + "tokens_seen": 176685056 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004780040120361083, + "loss": 1.8851, + "theoretical_loss": 4.452312395463473, + "tokens_seen": 176750592 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047799398194583754, + "loss": 1.6717, + "theoretical_loss": 4.452103255176027, + "tokens_seen": 176816128 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004779839518555667, + "loss": 1.8679, + "theoretical_loss": 4.451894214086537, + "tokens_seen": 176881664 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 73013, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8558073043823242, + "objective/train/theoretical_loss": 4.4516852721112254, + "objective/train/tokens_used": 176947200, + "theoretical_loss": 4.4516852721112254, + "tokens_seen": 176947200 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004779739217652959, + "loss": 1.8276, + "theoretical_loss": 4.4516852721112254, + "tokens_seen": 176947200 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004779638916750251, + "loss": 1.6993, + "theoretical_loss": 4.451476429166416, + "tokens_seen": 177012736 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004779538615847543, + "loss": 1.826, + "theoretical_loss": 4.451267685168532, + "tokens_seen": 177078272 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047794383149448344, + "loss": 1.9454, + "theoretical_loss": 4.451059040034102, + "tokens_seen": 177143808 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004779338014042127, + "loss": 1.864, + "theoretical_loss": 4.450850493679752, + "tokens_seen": 177209344 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004779237713139418, + "loss": 1.742, + "theoretical_loss": 4.450642046022211, + "tokens_seen": 177274880 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047791374122367104, + "loss": 1.8723, + "theoretical_loss": 4.45043369697831, + "tokens_seen": 177340416 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004779037111334002, + "loss": 1.7697, + "theoretical_loss": 4.450225446464978, + "tokens_seen": 177405952 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004778936810431294, + "loss": 1.8755, + "theoretical_loss": 4.450017294399247, + "tokens_seen": 177471488 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004778836509528586, + "loss": 1.9198, + "theoretical_loss": 4.449809240698249, + "tokens_seen": 177537024 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047787362086258776, + "loss": 1.8773, + "theoretical_loss": 4.449601285279215, + "tokens_seen": 177602560 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047786359077231695, + "loss": 1.7724, + "theoretical_loss": 4.449393428059478, + "tokens_seen": 177668096 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004778535606820462, + "loss": 1.8094, + "theoretical_loss": 4.449185668956469, + "tokens_seen": 177733632 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004778435305917753, + "loss": 1.7762, + "theoretical_loss": 4.448978007887721, + "tokens_seen": 177799168 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047783350050150454, + "loss": 1.8401, + "theoretical_loss": 4.448770444770865, + "tokens_seen": 177864704 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047782347041123367, + "loss": 1.8183, + "theoretical_loss": 4.448562979523633, + "tokens_seen": 177930240 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004778134403209629, + "loss": 1.8403, + "theoretical_loss": 4.448355612063855, + "tokens_seen": 177995776 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004778034102306921, + "loss": 1.7026, + "theoretical_loss": 4.4481483423094605, + "tokens_seen": 178061312 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047779338014042127, + "loss": 1.8815, + "theoretical_loss": 4.447941170178478, + "tokens_seen": 178126848 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047778335005015045, + "loss": 1.7765, + "theoretical_loss": 4.447734095589036, + "tokens_seen": 178192384 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004777733199598797, + "loss": 1.8267, + "theoretical_loss": 4.4475271184593606, + "tokens_seen": 178257920 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004777632898696088, + "loss": 1.776, + "theoretical_loss": 4.447320238707777, + "tokens_seen": 178323456 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047775325977933805, + "loss": 1.9414, + "theoretical_loss": 4.447113456252707, + "tokens_seen": 178388992 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004777432296890672, + "loss": 1.7726, + "theoretical_loss": 4.446906771012674, + "tokens_seen": 178454528 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004777331995987964, + "loss": 1.8388, + "theoretical_loss": 4.446700182906298, + "tokens_seen": 178520064 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004777231695085256, + "loss": 1.7529, + "theoretical_loss": 4.446493691852297, + "tokens_seen": 178585600 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047771313941825477, + "loss": 1.8891, + "theoretical_loss": 4.4462872977694845, + "tokens_seen": 178651136 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047770310932798395, + "loss": 1.767, + "theoretical_loss": 4.4460810005767755, + "tokens_seen": 178716672 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047769307923771313, + "loss": 1.8358, + "theoretical_loss": 4.445874800193181, + "tokens_seen": 178782208 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004776830491474423, + "loss": 1.8577, + "theoretical_loss": 4.445668696537808, + "tokens_seen": 178847744 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047767301905717155, + "loss": 1.8441, + "theoretical_loss": 4.4454626895298635, + "tokens_seen": 178913280 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004776629889669007, + "loss": 1.7651, + "theoretical_loss": 4.4452567790886475, + "tokens_seen": 178978816 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004776529588766299, + "loss": 1.7746, + "theoretical_loss": 4.445050965133561, + "tokens_seen": 179044352 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047764292878635904, + "loss": 1.718, + "theoretical_loss": 4.4448452475841, + "tokens_seen": 179109888 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004776328986960883, + "loss": 1.7676, + "theoretical_loss": 4.444639626359856, + "tokens_seen": 179175424 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047762286860581746, + "loss": 1.8974, + "theoretical_loss": 4.444434101380519, + "tokens_seen": 179240960 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047761283851554664, + "loss": 1.8785, + "theoretical_loss": 4.444228672565874, + "tokens_seen": 179306496 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004776028084252758, + "loss": 1.9031, + "theoretical_loss": 4.444023339835803, + "tokens_seen": 179372032 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047759277833500505, + "loss": 1.8133, + "theoretical_loss": 4.443818103110283, + "tokens_seen": 179437568 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004775827482447342, + "loss": 1.8272, + "theoretical_loss": 4.443612962309387, + "tokens_seen": 179503104 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004775727181544634, + "loss": 1.8224, + "theoretical_loss": 4.4434079173532846, + "tokens_seen": 179568640 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047756268806419254, + "loss": 1.832, + "theoretical_loss": 4.4432029681622405, + "tokens_seen": 179634176 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004775526579739218, + "loss": 1.7976, + "theoretical_loss": 4.442998114656614, + "tokens_seen": 179699712 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047754262788365096, + "loss": 1.6862, + "theoretical_loss": 4.442793356756861, + "tokens_seen": 179765248 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047753259779338014, + "loss": 1.8277, + "theoretical_loss": 4.442588694383532, + "tokens_seen": 179830784 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004775225677031093, + "loss": 1.7795, + "theoretical_loss": 4.44238412745727, + "tokens_seen": 179896320 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004775125376128385, + "loss": 1.7944, + "theoretical_loss": 4.442179655898819, + "tokens_seen": 179961856 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004775025075225677, + "loss": 1.8766, + "theoretical_loss": 4.44197527962901, + "tokens_seen": 180027392 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004774924774322969, + "loss": 1.7942, + "theoretical_loss": 4.441770998568773, + "tokens_seen": 180092928 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047748244734202605, + "loss": 1.7673, + "theoretical_loss": 4.441566812639133, + "tokens_seen": 180158464 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 74146, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7549896240234375, + "objective/train/theoretical_loss": 4.441362721761207, + "objective/train/tokens_used": 180224000, + "theoretical_loss": 4.441362721761207, + "tokens_seen": 180224000 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004774724172517553, + "loss": 1.7526, + "theoretical_loss": 4.441362721761207, + "tokens_seen": 180224000 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004774623871614845, + "loss": 1.8397, + "theoretical_loss": 4.441158725856207, + "tokens_seen": 180289536 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047745235707121364, + "loss": 1.8651, + "theoretical_loss": 4.440954824845438, + "tokens_seen": 180355072 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004774423269809429, + "loss": 1.9575, + "theoretical_loss": 4.4407510186503, + "tokens_seen": 180420608 + }, + { + "epoch": 0.05, + "learning_rate": 0.000477432296890672, + "loss": 1.8298, + "theoretical_loss": 4.440547307192286, + "tokens_seen": 180486144 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047742226680040124, + "loss": 1.9459, + "theoretical_loss": 4.440343690392984, + "tokens_seen": 180551680 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004774122367101304, + "loss": 1.8788, + "theoretical_loss": 4.4401401681740715, + "tokens_seen": 180617216 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004774022066198596, + "loss": 1.7434, + "theoretical_loss": 4.439936740457323, + "tokens_seen": 180682752 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004773921765295888, + "loss": 1.8112, + "theoretical_loss": 4.439733407164605, + "tokens_seen": 180748288 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047738214643931797, + "loss": 1.9476, + "theoretical_loss": 4.439530168217875, + "tokens_seen": 180813824 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047737211634904715, + "loss": 1.9098, + "theoretical_loss": 4.439327023539187, + "tokens_seen": 180879360 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004773620862587764, + "loss": 1.8219, + "theoretical_loss": 4.439123973050684, + "tokens_seen": 180944896 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004773520561685055, + "loss": 1.814, + "theoretical_loss": 4.438921016674603, + "tokens_seen": 181010432 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047734202607823474, + "loss": 1.7813, + "theoretical_loss": 4.438718154333274, + "tokens_seen": 181075968 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047733199598796387, + "loss": 1.7999, + "theoretical_loss": 4.438515385949118, + "tokens_seen": 181141504 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004773219658976931, + "loss": 1.7753, + "theoretical_loss": 4.438312711444647, + "tokens_seen": 181207040 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004773119358074223, + "loss": 1.8436, + "theoretical_loss": 4.438110130742469, + "tokens_seen": 181272576 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047730190571715147, + "loss": 1.7952, + "theoretical_loss": 4.437907643765281, + "tokens_seen": 181338112 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047729187562688065, + "loss": 1.7743, + "theoretical_loss": 4.43770525043587, + "tokens_seen": 181403648 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004772818455366099, + "loss": 1.882, + "theoretical_loss": 4.437502950677116, + "tokens_seen": 181469184 + }, + { + "epoch": 0.06, + "learning_rate": 0.000477271815446339, + "loss": 1.7737, + "theoretical_loss": 4.437300744411992, + "tokens_seen": 181534720 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047726178535606825, + "loss": 1.681, + "theoretical_loss": 4.43709863156356, + "tokens_seen": 181600256 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004772517552657974, + "loss": 1.8414, + "theoretical_loss": 4.436896612054975, + "tokens_seen": 181665792 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004772417251755266, + "loss": 1.8781, + "theoretical_loss": 4.43669468580948, + "tokens_seen": 181731328 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004772316950852558, + "loss": 1.8442, + "theoretical_loss": 4.43649285275041, + "tokens_seen": 181796864 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047722166499498497, + "loss": 1.9343, + "theoretical_loss": 4.436291112801193, + "tokens_seen": 181862400 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047721163490471415, + "loss": 1.8826, + "theoretical_loss": 4.436089465885345, + "tokens_seen": 181927936 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047720160481444333, + "loss": 1.7489, + "theoretical_loss": 4.435887911926472, + "tokens_seen": 181993472 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004771915747241725, + "loss": 1.6771, + "theoretical_loss": 4.435686450848273, + "tokens_seen": 182059008 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047718154463390175, + "loss": 1.7286, + "theoretical_loss": 4.435485082574533, + "tokens_seen": 182124544 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004771715145436309, + "loss": 1.7231, + "theoretical_loss": 4.43528380702913, + "tokens_seen": 182190080 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004771614844533601, + "loss": 1.7531, + "theoretical_loss": 4.435082624136031, + "tokens_seen": 182255616 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047715145436308924, + "loss": 1.7635, + "theoretical_loss": 4.4348815338192935, + "tokens_seen": 182321152 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004771414242728185, + "loss": 1.8331, + "theoretical_loss": 4.434680536003062, + "tokens_seen": 182386688 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047713139418254766, + "loss": 1.806, + "theoretical_loss": 4.434479630611573, + "tokens_seen": 182452224 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047712136409227684, + "loss": 1.8298, + "theoretical_loss": 4.43427881756915, + "tokens_seen": 182517760 + }, + { + "epoch": 0.06, + "learning_rate": 0.000477111334002006, + "loss": 1.8436, + "theoretical_loss": 4.434078096800209, + "tokens_seen": 182583296 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047710130391173525, + "loss": 1.7748, + "theoretical_loss": 4.433877468229251, + "tokens_seen": 182648832 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004770912738214644, + "loss": 1.7875, + "theoretical_loss": 4.433676931780868, + "tokens_seen": 182714368 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004770812437311936, + "loss": 1.6981, + "theoretical_loss": 4.43347648737974, + "tokens_seen": 182779904 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047707121364092274, + "loss": 1.8335, + "theoretical_loss": 4.433276134950637, + "tokens_seen": 182845440 + }, + { + "epoch": 0.06, + "learning_rate": 0.000477061183550652, + "loss": 1.7855, + "theoretical_loss": 4.4330758744184156, + "tokens_seen": 182910976 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047705115346038116, + "loss": 1.9753, + "theoretical_loss": 4.432875705708021, + "tokens_seen": 182976512 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047704112337011034, + "loss": 1.7554, + "theoretical_loss": 4.432675628744488, + "tokens_seen": 183042048 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004770310932798395, + "loss": 1.8607, + "theoretical_loss": 4.432475643452937, + "tokens_seen": 183107584 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004770210631895687, + "loss": 1.759, + "theoretical_loss": 4.432275749758579, + "tokens_seen": 183173120 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004770110330992979, + "loss": 1.8441, + "theoretical_loss": 4.432075947586711, + "tokens_seen": 183238656 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004770010030090271, + "loss": 1.8455, + "theoretical_loss": 4.431876236862717, + "tokens_seen": 183304192 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047699097291875625, + "loss": 1.7014, + "theoretical_loss": 4.43167661751207, + "tokens_seen": 183369728 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004769809428284855, + "loss": 1.928, + "theoretical_loss": 4.431477089460331, + "tokens_seen": 183435264 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 75209, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8564612865447998, + "objective/train/theoretical_loss": 4.431277652633145, + "objective/train/tokens_used": 183500800, + "theoretical_loss": 4.431277652633145, + "tokens_seen": 183500800 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047697091273821466, + "loss": 1.7744, + "theoretical_loss": 4.431277652633145, + "tokens_seen": 183500800 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047696088264794384, + "loss": 1.9069, + "theoretical_loss": 4.431078306956248, + "tokens_seen": 183566336 + }, + { + "epoch": 0.06, + "learning_rate": 0.000476950852557673, + "loss": 1.9241, + "theoretical_loss": 4.43087905235546, + "tokens_seen": 183631872 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004769408224674022, + "loss": 1.7814, + "theoretical_loss": 4.43067988875669, + "tokens_seen": 183697408 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004769307923771314, + "loss": 1.8495, + "theoretical_loss": 4.430480816085931, + "tokens_seen": 183762944 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004769207622868606, + "loss": 1.8451, + "theoretical_loss": 4.430281834269265, + "tokens_seen": 183828480 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047691073219658975, + "loss": 1.7834, + "theoretical_loss": 4.4300829432328594, + "tokens_seen": 183894016 + }, + { + "epoch": 0.06, + "learning_rate": 0.000476900702106319, + "loss": 1.8122, + "theoretical_loss": 4.429884142902968, + "tokens_seen": 183959552 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004768906720160481, + "loss": 1.7016, + "theoretical_loss": 4.429685433205931, + "tokens_seen": 184025088 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047688064192577735, + "loss": 1.8084, + "theoretical_loss": 4.429486814068174, + "tokens_seen": 184090624 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047687061183550653, + "loss": 1.8902, + "theoretical_loss": 4.4292882854162094, + "tokens_seen": 184156160 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004768605817452357, + "loss": 1.7954, + "theoretical_loss": 4.429089847176634, + "tokens_seen": 184221696 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004768505516549649, + "loss": 1.7949, + "theoretical_loss": 4.428891499276132, + "tokens_seen": 184287232 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047684052156469407, + "loss": 1.9035, + "theoretical_loss": 4.42869324164147, + "tokens_seen": 184352768 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047683049147442325, + "loss": 1.7285, + "theoretical_loss": 4.428495074199505, + "tokens_seen": 184418304 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004768204613841525, + "loss": 1.7177, + "theoretical_loss": 4.428296996877176, + "tokens_seen": 184483840 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004768104312938816, + "loss": 1.7721, + "theoretical_loss": 4.428099009601506, + "tokens_seen": 184549376 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047680040120361085, + "loss": 1.8879, + "theoretical_loss": 4.4279011122996055, + "tokens_seen": 184614912 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047679037111334003, + "loss": 1.7302, + "theoretical_loss": 4.427703304898668, + "tokens_seen": 184680448 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004767803410230692, + "loss": 1.7743, + "theoretical_loss": 4.427505587325974, + "tokens_seen": 184745984 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004767703109327984, + "loss": 1.7922, + "theoretical_loss": 4.427307959508885, + "tokens_seen": 184811520 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004767602808425276, + "loss": 1.8638, + "theoretical_loss": 4.4271104213748504, + "tokens_seen": 184877056 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047675025075225676, + "loss": 1.7951, + "theoretical_loss": 4.426912972851404, + "tokens_seen": 184942592 + }, + { + "epoch": 0.06, + "learning_rate": 0.000476740220661986, + "loss": 1.9447, + "theoretical_loss": 4.426715613866159, + "tokens_seen": 185008128 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004767301905717151, + "loss": 1.8062, + "theoretical_loss": 4.426518344346819, + "tokens_seen": 185073664 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047672016048144435, + "loss": 1.7944, + "theoretical_loss": 4.426321164221168, + "tokens_seen": 185139200 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047671013039117353, + "loss": 1.7433, + "theoretical_loss": 4.426124073417074, + "tokens_seen": 185204736 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004767001003009027, + "loss": 1.8309, + "theoretical_loss": 4.42592707186249, + "tokens_seen": 185270272 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047669007021063195, + "loss": 1.6871, + "theoretical_loss": 4.42573015948545, + "tokens_seen": 185335808 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004766800401203611, + "loss": 1.8429, + "theoretical_loss": 4.425533336214076, + "tokens_seen": 185401344 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004766700100300903, + "loss": 1.7728, + "theoretical_loss": 4.425336601976568, + "tokens_seen": 185466880 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047665997993981944, + "loss": 1.7935, + "theoretical_loss": 4.425139956701214, + "tokens_seen": 185532416 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004766499498495487, + "loss": 1.8143, + "theoretical_loss": 4.42494340031638, + "tokens_seen": 185597952 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047663991975927786, + "loss": 1.7342, + "theoretical_loss": 4.424746932750521, + "tokens_seen": 185663488 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047662988966900704, + "loss": 1.9316, + "theoretical_loss": 4.4245505539321694, + "tokens_seen": 185729024 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004766198595787362, + "loss": 1.9217, + "theoretical_loss": 4.424354263789943, + "tokens_seen": 185794560 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047660982948846545, + "loss": 1.8363, + "theoretical_loss": 4.424158062252541, + "tokens_seen": 185860096 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004765997993981946, + "loss": 1.8556, + "theoretical_loss": 4.423961949248748, + "tokens_seen": 185925632 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004765897693079238, + "loss": 1.8439, + "theoretical_loss": 4.423765924707427, + "tokens_seen": 185991168 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047657973921765294, + "loss": 1.7671, + "theoretical_loss": 4.4235699885575235, + "tokens_seen": 186056704 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004765697091273822, + "loss": 1.7064, + "theoretical_loss": 4.423374140728068, + "tokens_seen": 186122240 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047655967903711136, + "loss": 1.7648, + "theoretical_loss": 4.423178381148172, + "tokens_seen": 186187776 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047654964894684054, + "loss": 1.8559, + "theoretical_loss": 4.422982709747026, + "tokens_seen": 186253312 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004765396188565697, + "loss": 1.7719, + "theoretical_loss": 4.4227871264539065, + "tokens_seen": 186318848 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004765295887662989, + "loss": 1.6946, + "theoretical_loss": 4.422591631198167, + "tokens_seen": 186384384 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004765195586760281, + "loss": 1.7326, + "theoretical_loss": 4.422396223909248, + "tokens_seen": 186449920 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004765095285857573, + "loss": 1.8727, + "theoretical_loss": 4.422200904516666, + "tokens_seen": 186515456 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047649949849548645, + "loss": 1.7308, + "theoretical_loss": 4.4220056729500214, + "tokens_seen": 186580992 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004764894684052157, + "loss": 1.7988, + "theoretical_loss": 4.4218105291389955, + "tokens_seen": 186646528 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047647943831494486, + "loss": 1.7091, + "theoretical_loss": 4.4216154730133495, + "tokens_seen": 186712064 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 76351, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.060258388519287, + "objective/train/theoretical_loss": 4.421420504502928, + "objective/train/tokens_used": 186777600, + "theoretical_loss": 4.421420504502928, + "tokens_seen": 186777600 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047646940822467404, + "loss": 1.9284, + "theoretical_loss": 4.421420504502928, + "tokens_seen": 186777600 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004764593781344032, + "loss": 1.799, + "theoretical_loss": 4.421225623537652, + "tokens_seen": 186843136 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004764493480441324, + "loss": 1.7183, + "theoretical_loss": 4.421030830047528, + "tokens_seen": 186908672 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004764393179538616, + "loss": 1.8273, + "theoretical_loss": 4.420836123962639, + "tokens_seen": 186974208 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004764292878635908, + "loss": 1.7885, + "theoretical_loss": 4.420641505213151, + "tokens_seen": 187039744 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047641925777331995, + "loss": 1.8903, + "theoretical_loss": 4.42044697372931, + "tokens_seen": 187105280 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004764092276830492, + "loss": 1.7128, + "theoretical_loss": 4.420252529441439, + "tokens_seen": 187170816 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004763991975927783, + "loss": 1.7955, + "theoretical_loss": 4.420058172279945, + "tokens_seen": 187236352 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047638916750250755, + "loss": 1.7214, + "theoretical_loss": 4.419863902175313, + "tokens_seen": 187301888 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047637913741223673, + "loss": 1.844, + "theoretical_loss": 4.419669719058108, + "tokens_seen": 187367424 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004763691073219659, + "loss": 1.7422, + "theoretical_loss": 4.419475622858974, + "tokens_seen": 187432960 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004763590772316951, + "loss": 1.7768, + "theoretical_loss": 4.419281613508637, + "tokens_seen": 187498496 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047634904714142427, + "loss": 1.8007, + "theoretical_loss": 4.419087690937898, + "tokens_seen": 187564032 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047633901705115345, + "loss": 1.7868, + "theoretical_loss": 4.418893855077641, + "tokens_seen": 187629568 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004763289869608827, + "loss": 1.8132, + "theoretical_loss": 4.4187001058588296, + "tokens_seen": 187695104 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004763189568706118, + "loss": 1.7417, + "theoretical_loss": 4.418506443212503, + "tokens_seen": 187760640 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047630892678034105, + "loss": 1.7494, + "theoretical_loss": 4.4183128670697815, + "tokens_seen": 187826176 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047629889669007023, + "loss": 1.7467, + "theoretical_loss": 4.418119377361863, + "tokens_seen": 187891712 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004762888665997994, + "loss": 1.7924, + "theoretical_loss": 4.417925974020028, + "tokens_seen": 187957248 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004762788365095286, + "loss": 1.9139, + "theoretical_loss": 4.4177326569756294, + "tokens_seen": 188022784 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004762688064192578, + "loss": 1.8531, + "theoretical_loss": 4.417539426160104, + "tokens_seen": 188088320 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047625877632898696, + "loss": 1.7441, + "theoretical_loss": 4.417346281504963, + "tokens_seen": 188153856 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004762487462387162, + "loss": 1.8382, + "theoretical_loss": 4.417153222941798, + "tokens_seen": 188219392 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004762387161484453, + "loss": 1.9315, + "theoretical_loss": 4.416960250402279, + "tokens_seen": 188284928 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047622868605817455, + "loss": 1.7635, + "theoretical_loss": 4.416767363818153, + "tokens_seen": 188350464 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004762186559679037, + "loss": 1.8468, + "theoretical_loss": 4.4165745631212445, + "tokens_seen": 188416000 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004762086258776329, + "loss": 1.7607, + "theoretical_loss": 4.416381848243456, + "tokens_seen": 188481536 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004761985957873621, + "loss": 1.8874, + "theoretical_loss": 4.41618921911677, + "tokens_seen": 188547072 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004761885656970913, + "loss": 1.7372, + "theoretical_loss": 4.415996675673242, + "tokens_seen": 188612608 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047617853560682046, + "loss": 1.8811, + "theoretical_loss": 4.415804217845009, + "tokens_seen": 188678144 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047616850551654964, + "loss": 1.8768, + "theoretical_loss": 4.415611845564284, + "tokens_seen": 188743680 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004761584754262788, + "loss": 1.7971, + "theoretical_loss": 4.415419558763356, + "tokens_seen": 188809216 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047614844533600806, + "loss": 1.7964, + "theoretical_loss": 4.415227357374592, + "tokens_seen": 188874752 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004761384152457372, + "loss": 1.8487, + "theoretical_loss": 4.415035241330436, + "tokens_seen": 188940288 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004761283851554664, + "loss": 1.8112, + "theoretical_loss": 4.41484321056341, + "tokens_seen": 189005824 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004761183550651956, + "loss": 1.749, + "theoretical_loss": 4.4146512650061105, + "tokens_seen": 189071360 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004761083249749248, + "loss": 1.7317, + "theoretical_loss": 4.414459404591211, + "tokens_seen": 189136896 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047609829488465396, + "loss": 1.8013, + "theoretical_loss": 4.414267629251463, + "tokens_seen": 189202432 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047608826479438314, + "loss": 1.7576, + "theoretical_loss": 4.414075938919694, + "tokens_seen": 189267968 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004760782347041123, + "loss": 1.5949, + "theoretical_loss": 4.413884333528807, + "tokens_seen": 189333504 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047606820461384156, + "loss": 1.783, + "theoretical_loss": 4.413692813011781, + "tokens_seen": 189399040 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004760581745235707, + "loss": 1.8018, + "theoretical_loss": 4.413501377301671, + "tokens_seen": 189464576 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004760481444332999, + "loss": 1.8555, + "theoretical_loss": 4.41331002633161, + "tokens_seen": 189530112 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047603811434302905, + "loss": 1.8091, + "theoretical_loss": 4.413118760034804, + "tokens_seen": 189595648 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004760280842527583, + "loss": 1.9037, + "theoretical_loss": 4.4129275783445365, + "tokens_seen": 189661184 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047601805416248746, + "loss": 1.7337, + "theoretical_loss": 4.412736481194166, + "tokens_seen": 189726720 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047600802407221665, + "loss": 1.8267, + "theoretical_loss": 4.412545468517127, + "tokens_seen": 189792256 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004759979939819458, + "loss": 1.8315, + "theoretical_loss": 4.412354540246929, + "tokens_seen": 189857792 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047598796389167506, + "loss": 1.7881, + "theoretical_loss": 4.4121636963171555, + "tokens_seen": 189923328 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004759779338014042, + "loss": 1.7662, + "theoretical_loss": 4.411972936661467, + "tokens_seen": 189988864 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 77536, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6228785514831543, + "objective/train/theoretical_loss": 4.411782261213599, + "objective/train/tokens_used": 190054400, + "theoretical_loss": 4.411782261213599, + "tokens_seen": 190054400 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004759679037111334, + "loss": 1.7039, + "theoretical_loss": 4.411782261213599, + "tokens_seen": 190054400 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004759578736208626, + "loss": 1.8091, + "theoretical_loss": 4.41159166990736, + "tokens_seen": 190119936 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004759478435305918, + "loss": 1.784, + "theoretical_loss": 4.411401162676636, + "tokens_seen": 190185472 + }, + { + "epoch": 0.06, + "learning_rate": 0.000475937813440321, + "loss": 1.6501, + "theoretical_loss": 4.411210739455386, + "tokens_seen": 190251008 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047592778335005015, + "loss": 1.8019, + "theoretical_loss": 4.4110204001776445, + "tokens_seen": 190316544 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004759177532597794, + "loss": 1.7573, + "theoretical_loss": 4.410830144777519, + "tokens_seen": 190382080 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004759077231695085, + "loss": 1.6876, + "theoretical_loss": 4.410639973189192, + "tokens_seen": 190447616 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047589769307923775, + "loss": 1.8802, + "theoretical_loss": 4.410449885346922, + "tokens_seen": 190513152 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047588766298896693, + "loss": 1.7647, + "theoretical_loss": 4.4102598811850395, + "tokens_seen": 190578688 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004758776328986961, + "loss": 1.8023, + "theoretical_loss": 4.410069960637949, + "tokens_seen": 190644224 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004758676028084253, + "loss": 1.7415, + "theoretical_loss": 4.409880123640131, + "tokens_seen": 190709760 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047585757271815447, + "loss": 1.853, + "theoretical_loss": 4.409690370126139, + "tokens_seen": 190775296 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047584754262788365, + "loss": 1.703, + "theoretical_loss": 4.409500700030598, + "tokens_seen": 190840832 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004758375125376129, + "loss": 1.7434, + "theoretical_loss": 4.409311113288211, + "tokens_seen": 190906368 + }, + { + "epoch": 0.06, + "learning_rate": 0.000475827482447342, + "loss": 1.8396, + "theoretical_loss": 4.409121609833749, + "tokens_seen": 190971904 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047581745235707125, + "loss": 1.7634, + "theoretical_loss": 4.408932189602061, + "tokens_seen": 191037440 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047580742226680043, + "loss": 1.7709, + "theoretical_loss": 4.408742852528068, + "tokens_seen": 191102976 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004757973921765296, + "loss": 1.9023, + "theoretical_loss": 4.408553598546764, + "tokens_seen": 191168512 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004757873620862588, + "loss": 1.7562, + "theoretical_loss": 4.408364427593215, + "tokens_seen": 191234048 + }, + { + "epoch": 0.06, + "learning_rate": 0.000475777331995988, + "loss": 1.7767, + "theoretical_loss": 4.40817533960256, + "tokens_seen": 191299584 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047576730190571716, + "loss": 1.8096, + "theoretical_loss": 4.407986334510014, + "tokens_seen": 191365120 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004757572718154464, + "loss": 1.7941, + "theoretical_loss": 4.407797412250861, + "tokens_seen": 191430656 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004757472417251755, + "loss": 1.8461, + "theoretical_loss": 4.40760857276046, + "tokens_seen": 191496192 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047573721163490475, + "loss": 1.8025, + "theoretical_loss": 4.407419815974242, + "tokens_seen": 191561728 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004757271815446339, + "loss": 1.7275, + "theoretical_loss": 4.407231141827708, + "tokens_seen": 191627264 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004757171514543631, + "loss": 1.6884, + "theoretical_loss": 4.407042550256436, + "tokens_seen": 191692800 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004757071213640923, + "loss": 1.7522, + "theoretical_loss": 4.406854041196072, + "tokens_seen": 191758336 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004756970912738215, + "loss": 1.8429, + "theoretical_loss": 4.406665614582338, + "tokens_seen": 191823872 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047568706118355066, + "loss": 1.8248, + "theoretical_loss": 4.406477270351024, + "tokens_seen": 191889408 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047567703109327984, + "loss": 1.8172, + "theoretical_loss": 4.4062890084379935, + "tokens_seen": 191954944 + }, + { + "epoch": 0.06, + "learning_rate": 0.000475667001003009, + "loss": 1.7967, + "theoretical_loss": 4.406100828779183, + "tokens_seen": 192020480 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047565697091273826, + "loss": 1.8942, + "theoretical_loss": 4.4059127313106, + "tokens_seen": 192086016 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004756469408224674, + "loss": 1.7867, + "theoretical_loss": 4.405724715968323, + "tokens_seen": 192151552 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004756369107321966, + "loss": 1.8035, + "theoretical_loss": 4.405536782688502, + "tokens_seen": 192217088 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004756268806419258, + "loss": 1.7871, + "theoretical_loss": 4.405348931407359, + "tokens_seen": 192282624 + }, + { + "epoch": 0.06, + "learning_rate": 0.000475616850551655, + "loss": 1.8437, + "theoretical_loss": 4.405161162061187, + "tokens_seen": 192348160 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047560682046138416, + "loss": 1.8227, + "theoretical_loss": 4.404973474586351, + "tokens_seen": 192413696 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047559679037111334, + "loss": 1.7459, + "theoretical_loss": 4.404785868919285, + "tokens_seen": 192479232 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004755867602808425, + "loss": 1.8285, + "theoretical_loss": 4.404598344996497, + "tokens_seen": 192544768 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047557673019057176, + "loss": 1.7495, + "theoretical_loss": 4.404410902754562, + "tokens_seen": 192610304 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004755667001003009, + "loss": 1.7507, + "theoretical_loss": 4.404223542130129, + "tokens_seen": 192675840 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004755566700100301, + "loss": 1.6094, + "theoretical_loss": 4.404036263059917, + "tokens_seen": 192741376 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047554663991975925, + "loss": 1.7948, + "theoretical_loss": 4.403849065480714, + "tokens_seen": 192806912 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004755366098294885, + "loss": 1.7329, + "theoretical_loss": 4.403661949329381, + "tokens_seen": 192872448 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047552657973921766, + "loss": 1.7796, + "theoretical_loss": 4.403474914542848, + "tokens_seen": 192937984 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047551654964894685, + "loss": 1.8633, + "theoretical_loss": 4.403287961058112, + "tokens_seen": 193003520 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047550651955867603, + "loss": 1.946, + "theoretical_loss": 4.403101088812248, + "tokens_seen": 193069056 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047549648946840526, + "loss": 1.7618, + "theoretical_loss": 4.4029142977423925, + "tokens_seen": 193134592 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004754864593781344, + "loss": 1.8202, + "theoretical_loss": 4.402727587785758, + "tokens_seen": 193200128 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004754764292878636, + "loss": 1.6898, + "theoretical_loss": 4.402540958879625, + "tokens_seen": 193265664 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 78681, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7246482372283936, + "objective/train/theoretical_loss": 4.402354410961341, + "objective/train/tokens_used": 193331200, + "theoretical_loss": 4.402354410961341, + "tokens_seen": 193331200 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047546639919759275, + "loss": 1.7187, + "theoretical_loss": 4.402354410961341, + "tokens_seen": 193331200 + }, + { + "epoch": 0.06, + "learning_rate": 0.000475456369107322, + "loss": 1.8393, + "theoretical_loss": 4.402167943968328, + "tokens_seen": 193396736 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047544633901705117, + "loss": 1.8361, + "theoretical_loss": 4.4019815578380745, + "tokens_seen": 193462272 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047543630892678035, + "loss": 1.7718, + "theoretical_loss": 4.401795252508139, + "tokens_seen": 193527808 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047542627883650953, + "loss": 1.8109, + "theoretical_loss": 4.401609027916149, + "tokens_seen": 193593344 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004754162487462387, + "loss": 1.7735, + "theoretical_loss": 4.4014228839998015, + "tokens_seen": 193658880 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004754062186559679, + "loss": 1.7855, + "theoretical_loss": 4.401236820696863, + "tokens_seen": 193724416 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047539618856569713, + "loss": 1.6993, + "theoretical_loss": 4.40105083794517, + "tokens_seen": 193789952 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047538615847542625, + "loss": 1.7557, + "theoretical_loss": 4.400864935682625, + "tokens_seen": 193855488 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004753761283851555, + "loss": 1.7621, + "theoretical_loss": 4.400679113847202, + "tokens_seen": 193921024 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004753660982948846, + "loss": 1.7904, + "theoretical_loss": 4.400493372376943, + "tokens_seen": 193986560 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047535606820461385, + "loss": 1.7937, + "theoretical_loss": 4.400307711209957, + "tokens_seen": 194052096 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047534603811434303, + "loss": 1.754, + "theoretical_loss": 4.400122130284426, + "tokens_seen": 194117632 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004753360080240722, + "loss": 1.7313, + "theoretical_loss": 4.399936629538595, + "tokens_seen": 194183168 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004753259779338014, + "loss": 1.7938, + "theoretical_loss": 4.3997512089107795, + "tokens_seen": 194248704 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047531594784353063, + "loss": 1.7482, + "theoretical_loss": 4.399565868339366, + "tokens_seen": 194314240 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047530591775325976, + "loss": 1.7491, + "theoretical_loss": 4.399380607762804, + "tokens_seen": 194379776 + }, + { + "epoch": 0.06, + "learning_rate": 0.000475295887662989, + "loss": 1.6983, + "theoretical_loss": 4.399195427119616, + "tokens_seen": 194445312 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004752858575727181, + "loss": 1.7682, + "theoretical_loss": 4.399010326348388, + "tokens_seen": 194510848 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047527582748244736, + "loss": 1.8121, + "theoretical_loss": 4.398825305387778, + "tokens_seen": 194576384 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047526579739217654, + "loss": 1.8724, + "theoretical_loss": 4.39864036417651, + "tokens_seen": 194641920 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004752557673019057, + "loss": 1.8277, + "theoretical_loss": 4.398455502653373, + "tokens_seen": 194707456 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004752457372116349, + "loss": 1.59, + "theoretical_loss": 4.398270720757228, + "tokens_seen": 194772992 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004752357071213641, + "loss": 1.8022, + "theoretical_loss": 4.398086018427001, + "tokens_seen": 194838528 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047522567703109326, + "loss": 1.9117, + "theoretical_loss": 4.397901395601687, + "tokens_seen": 194904064 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004752156469408225, + "loss": 1.8677, + "theoretical_loss": 4.397716852220345, + "tokens_seen": 194969600 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004752056168505517, + "loss": 1.7885, + "theoretical_loss": 4.397532388222105, + "tokens_seen": 195035136 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047519558676028086, + "loss": 1.7231, + "theoretical_loss": 4.397348003546161, + "tokens_seen": 195100672 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047518555667001004, + "loss": 1.7511, + "theoretical_loss": 4.397163698131776, + "tokens_seen": 195166208 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004751755265797392, + "loss": 1.8558, + "theoretical_loss": 4.39697947191828, + "tokens_seen": 195231744 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047516549648946846, + "loss": 1.774, + "theoretical_loss": 4.396795324845069, + "tokens_seen": 195297280 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004751554663991976, + "loss": 1.6426, + "theoretical_loss": 4.396611256851604, + "tokens_seen": 195362816 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004751454363089268, + "loss": 1.6834, + "theoretical_loss": 4.3964272678774154, + "tokens_seen": 195428352 + }, + { + "epoch": 0.06, + "learning_rate": 0.000475135406218656, + "loss": 1.7877, + "theoretical_loss": 4.396243357862099, + "tokens_seen": 195493888 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004751253761283852, + "loss": 1.7188, + "theoretical_loss": 4.396059526745318, + "tokens_seen": 195559424 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047511534603811436, + "loss": 1.7579, + "theoretical_loss": 4.395875774466799, + "tokens_seen": 195624960 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047510531594784354, + "loss": 1.7735, + "theoretical_loss": 4.395692100966338, + "tokens_seen": 195690496 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004750952858575727, + "loss": 1.6245, + "theoretical_loss": 4.395508506183796, + "tokens_seen": 195756032 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047508525576730196, + "loss": 1.7387, + "theoretical_loss": 4.395324990059098, + "tokens_seen": 195821568 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004750752256770311, + "loss": 1.705, + "theoretical_loss": 4.395141552532239, + "tokens_seen": 195887104 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004750651955867603, + "loss": 1.9404, + "theoretical_loss": 4.394958193543277, + "tokens_seen": 195952640 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047505516549648945, + "loss": 1.7179, + "theoretical_loss": 4.394774913032337, + "tokens_seen": 196018176 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004750451354062187, + "loss": 1.8743, + "theoretical_loss": 4.394591710939608, + "tokens_seen": 196083712 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047503510531594787, + "loss": 1.7165, + "theoretical_loss": 4.394408587205344, + "tokens_seen": 196149248 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047502507522567705, + "loss": 1.7043, + "theoretical_loss": 4.39422554176987, + "tokens_seen": 196214784 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047501504513540623, + "loss": 1.7071, + "theoretical_loss": 4.39404257457357, + "tokens_seen": 196280320 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047500501504513546, + "loss": 1.9059, + "theoretical_loss": 4.393859685556895, + "tokens_seen": 196345856 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004749949849548646, + "loss": 1.796, + "theoretical_loss": 4.393676874660364, + "tokens_seen": 196411392 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004749849548645938, + "loss": 1.6938, + "theoretical_loss": 4.393494141824557, + "tokens_seen": 196476928 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047497492477432295, + "loss": 1.7275, + "theoretical_loss": 4.393311486990122, + "tokens_seen": 196542464 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 79941, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6307528018951416, + "objective/train/theoretical_loss": 4.39312891009777, + "objective/train/tokens_used": 196608000, + "theoretical_loss": 4.39312891009777, + "tokens_seen": 196608000 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004749648946840522, + "loss": 1.7307, + "theoretical_loss": 4.39312891009777, + "tokens_seen": 196608000 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047495486459378137, + "loss": 1.7506, + "theoretical_loss": 4.392946411088279, + "tokens_seen": 196673536 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047494483450351055, + "loss": 1.7897, + "theoretical_loss": 4.392763989902489, + "tokens_seen": 196739072 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047493480441323973, + "loss": 1.7444, + "theoretical_loss": 4.3925816464813074, + "tokens_seen": 196804608 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004749247743229689, + "loss": 1.8318, + "theoretical_loss": 4.392399380765704, + "tokens_seen": 196870144 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004749147442326981, + "loss": 1.7192, + "theoretical_loss": 4.392217192696712, + "tokens_seen": 196935680 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047490471414242733, + "loss": 1.5759, + "theoretical_loss": 4.392035082215434, + "tokens_seen": 197001216 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047489468405215646, + "loss": 1.7208, + "theoretical_loss": 4.391853049263032, + "tokens_seen": 197066752 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004748846539618857, + "loss": 1.8486, + "theoretical_loss": 4.391671093780733, + "tokens_seen": 197132288 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004748746238716148, + "loss": 1.8148, + "theoretical_loss": 4.39148921570983, + "tokens_seen": 197197824 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047486459378134405, + "loss": 1.796, + "theoretical_loss": 4.391307414991678, + "tokens_seen": 197263360 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047485456369107323, + "loss": 1.7478, + "theoretical_loss": 4.391125691567697, + "tokens_seen": 197328896 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004748445336008024, + "loss": 1.7406, + "theoretical_loss": 4.390944045379371, + "tokens_seen": 197394432 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004748345035105316, + "loss": 1.7532, + "theoretical_loss": 4.390762476368245, + "tokens_seen": 197459968 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047482447342026083, + "loss": 1.8653, + "theoretical_loss": 4.390580984475932, + "tokens_seen": 197525504 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047481444332998996, + "loss": 1.8253, + "theoretical_loss": 4.3903995696441065, + "tokens_seen": 197591040 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004748044132397192, + "loss": 1.7858, + "theoretical_loss": 4.390218231814504, + "tokens_seen": 197656576 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004747943831494483, + "loss": 1.6603, + "theoretical_loss": 4.390036970928929, + "tokens_seen": 197722112 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047478435305917756, + "loss": 1.6621, + "theoretical_loss": 4.3898557869292425, + "tokens_seen": 197787648 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047477432296890674, + "loss": 1.8352, + "theoretical_loss": 4.389674679757374, + "tokens_seen": 197853184 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004747642928786359, + "loss": 1.6582, + "theoretical_loss": 4.389493649355314, + "tokens_seen": 197918720 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004747542627883651, + "loss": 1.7203, + "theoretical_loss": 4.389312695665115, + "tokens_seen": 197984256 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004747442326980943, + "loss": 1.7285, + "theoretical_loss": 4.389131818628895, + "tokens_seen": 198049792 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047473420260782346, + "loss": 1.7475, + "theoretical_loss": 4.388951018188832, + "tokens_seen": 198115328 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004747241725175527, + "loss": 1.7353, + "theoretical_loss": 4.388770294287169, + "tokens_seen": 198180864 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004747141424272818, + "loss": 1.9815, + "theoretical_loss": 4.38858964686621, + "tokens_seen": 198246400 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047470411233701106, + "loss": 1.8014, + "theoretical_loss": 4.388409075868323, + "tokens_seen": 198311936 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004746940822467402, + "loss": 1.7851, + "theoretical_loss": 4.388228581235937, + "tokens_seen": 198377472 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004746840521564694, + "loss": 1.7858, + "theoretical_loss": 4.388048162911545, + "tokens_seen": 198443008 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004746740220661986, + "loss": 1.8557, + "theoretical_loss": 4.3878678208377, + "tokens_seen": 198508544 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004746639919759278, + "loss": 1.7148, + "theoretical_loss": 4.38768755495702, + "tokens_seen": 198574080 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047465396188565696, + "loss": 1.7505, + "theoretical_loss": 4.387507365212182, + "tokens_seen": 198639616 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004746439317953862, + "loss": 1.9025, + "theoretical_loss": 4.387327251545927, + "tokens_seen": 198705152 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004746339017051153, + "loss": 1.7834, + "theoretical_loss": 4.387147213901059, + "tokens_seen": 198770688 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047462387161484456, + "loss": 1.792, + "theoretical_loss": 4.386967252220441, + "tokens_seen": 198836224 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004746138415245737, + "loss": 1.6678, + "theoretical_loss": 4.386787366447, + "tokens_seen": 198901760 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004746038114343029, + "loss": 1.6898, + "theoretical_loss": 4.386607556523723, + "tokens_seen": 198967296 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004745937813440321, + "loss": 1.6961, + "theoretical_loss": 4.386427822393661, + "tokens_seen": 199032832 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004745837512537613, + "loss": 1.8633, + "theoretical_loss": 4.386248163999923, + "tokens_seen": 199098368 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047457372116349047, + "loss": 1.819, + "theoretical_loss": 4.3860685812856826, + "tokens_seen": 199163904 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047456369107321965, + "loss": 1.799, + "theoretical_loss": 4.385889074194173, + "tokens_seen": 199229440 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047455366098294883, + "loss": 1.7357, + "theoretical_loss": 4.385709642668688, + "tokens_seen": 199294976 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047454363089267807, + "loss": 1.747, + "theoretical_loss": 4.385530286652586, + "tokens_seen": 199360512 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004745336008024072, + "loss": 1.8163, + "theoretical_loss": 4.385351006089283, + "tokens_seen": 199426048 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047452357071213643, + "loss": 1.7781, + "theoretical_loss": 4.385171800922257, + "tokens_seen": 199491584 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047451354062186555, + "loss": 1.7181, + "theoretical_loss": 4.384992671095047, + "tokens_seen": 199557120 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004745035105315948, + "loss": 1.7806, + "theoretical_loss": 4.384813616551252, + "tokens_seen": 199622656 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047449348044132397, + "loss": 1.8312, + "theoretical_loss": 4.384634637234534, + "tokens_seen": 199688192 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047448345035105315, + "loss": 1.8045, + "theoretical_loss": 4.384455733088612, + "tokens_seen": 199753728 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004744734202607824, + "loss": 1.7852, + "theoretical_loss": 4.38427690405727, + "tokens_seen": 199819264 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 81028, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.657170295715332, + "objective/train/theoretical_loss": 4.384098150084349, + "objective/train/tokens_used": 199884800, + "theoretical_loss": 4.384098150084349, + "tokens_seen": 199884800 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047446339017051157, + "loss": 1.7515, + "theoretical_loss": 4.384098150084349, + "tokens_seen": 199884800 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047445336008024075, + "loss": 1.7682, + "theoretical_loss": 4.383919471113751, + "tokens_seen": 199950336 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047444332998996993, + "loss": 1.7568, + "theoretical_loss": 4.383740867089441, + "tokens_seen": 200015872 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004744332998996991, + "loss": 1.8527, + "theoretical_loss": 4.383562337955439, + "tokens_seen": 200081408 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004744232698094283, + "loss": 1.8253, + "theoretical_loss": 4.3833838836558305, + "tokens_seen": 200146944 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047441323971915753, + "loss": 1.7543, + "theoretical_loss": 4.383205504134758, + "tokens_seen": 200212480 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047440320962888666, + "loss": 1.7333, + "theoretical_loss": 4.383027199336424, + "tokens_seen": 200278016 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004743931795386159, + "loss": 1.8392, + "theoretical_loss": 4.382848969205092, + "tokens_seen": 200343552 + }, + { + "epoch": 0.06, + "learning_rate": 0.000474383149448345, + "loss": 1.7855, + "theoretical_loss": 4.3826708136850865, + "tokens_seen": 200409088 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047437311935807425, + "loss": 1.7443, + "theoretical_loss": 4.382492732720786, + "tokens_seen": 200474624 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047436308926780343, + "loss": 1.7972, + "theoretical_loss": 4.382314726256637, + "tokens_seen": 200540160 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004743530591775326, + "loss": 1.7806, + "theoretical_loss": 4.382136794237138, + "tokens_seen": 200605696 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004743430290872618, + "loss": 1.8331, + "theoretical_loss": 4.3819589366068525, + "tokens_seen": 200671232 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047433299899699103, + "loss": 1.7529, + "theoretical_loss": 4.381781153310399, + "tokens_seen": 200736768 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047432296890672016, + "loss": 1.7489, + "theoretical_loss": 4.381603444292459, + "tokens_seen": 200802304 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004743129388164494, + "loss": 1.7846, + "theoretical_loss": 4.38142580949777, + "tokens_seen": 200867840 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004743029087261785, + "loss": 1.7225, + "theoretical_loss": 4.381248248871131, + "tokens_seen": 200933376 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047429287863590776, + "loss": 1.5943, + "theoretical_loss": 4.3810707623574, + "tokens_seen": 200998912 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047428284854563694, + "loss": 1.7371, + "theoretical_loss": 4.380893349901491, + "tokens_seen": 201064448 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004742728184553661, + "loss": 1.8018, + "theoretical_loss": 4.380716011448381, + "tokens_seen": 201129984 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004742627883650953, + "loss": 1.9216, + "theoretical_loss": 4.380538746943103, + "tokens_seen": 201195520 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004742527582748245, + "loss": 1.8638, + "theoretical_loss": 4.380361556330749, + "tokens_seen": 201261056 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047424272818455366, + "loss": 1.818, + "theoretical_loss": 4.380184439556472, + "tokens_seen": 201326592 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004742326980942829, + "loss": 1.7171, + "theoretical_loss": 4.38000739656548, + "tokens_seen": 201392128 + }, + { + "epoch": 0.06, + "learning_rate": 0.000474222668004012, + "loss": 1.8047, + "theoretical_loss": 4.379830427303041, + "tokens_seen": 201457664 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047421263791374126, + "loss": 1.8714, + "theoretical_loss": 4.379653531714483, + "tokens_seen": 201523200 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004742026078234704, + "loss": 1.808, + "theoretical_loss": 4.37947670974519, + "tokens_seen": 201588736 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004741925777331996, + "loss": 1.8163, + "theoretical_loss": 4.379299961340605, + "tokens_seen": 201654272 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004741825476429288, + "loss": 1.8523, + "theoretical_loss": 4.37912328644623, + "tokens_seen": 201719808 + }, + { + "epoch": 0.06, + "learning_rate": 0.000474172517552658, + "loss": 1.6702, + "theoretical_loss": 4.378946685007623, + "tokens_seen": 201785344 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047416248746238716, + "loss": 1.8057, + "theoretical_loss": 4.378770156970403, + "tokens_seen": 201850880 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004741524573721164, + "loss": 1.8237, + "theoretical_loss": 4.378593702280243, + "tokens_seen": 201916416 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004741424272818455, + "loss": 1.7204, + "theoretical_loss": 4.378417320882879, + "tokens_seen": 201981952 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047413239719157476, + "loss": 1.7818, + "theoretical_loss": 4.378241012724099, + "tokens_seen": 202047488 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004741223671013039, + "loss": 1.7394, + "theoretical_loss": 4.378064777749753, + "tokens_seen": 202113024 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004741123370110331, + "loss": 1.7441, + "theoretical_loss": 4.377888615905746, + "tokens_seen": 202178560 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004741023069207623, + "loss": 1.8015, + "theoretical_loss": 4.377712527138042, + "tokens_seen": 202244096 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004740922768304915, + "loss": 1.7172, + "theoretical_loss": 4.3775365113926625, + "tokens_seen": 202309632 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047408224674022067, + "loss": 1.6785, + "theoretical_loss": 4.377360568615685, + "tokens_seen": 202375168 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047407221664994985, + "loss": 1.7272, + "theoretical_loss": 4.377184698753244, + "tokens_seen": 202440704 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047406218655967903, + "loss": 1.7652, + "theoretical_loss": 4.377008901751536, + "tokens_seen": 202506240 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047405215646940827, + "loss": 1.8725, + "theoretical_loss": 4.376833177556806, + "tokens_seen": 202571776 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004740421263791374, + "loss": 1.7845, + "theoretical_loss": 4.3766575261153635, + "tokens_seen": 202637312 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047403209628886663, + "loss": 1.7098, + "theoretical_loss": 4.3764819473735725, + "tokens_seen": 202702848 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047402206619859575, + "loss": 1.7717, + "theoretical_loss": 4.376306441277852, + "tokens_seen": 202768384 + }, + { + "epoch": 0.06, + "learning_rate": 0.000474012036108325, + "loss": 1.7126, + "theoretical_loss": 4.376131007774681, + "tokens_seen": 202833920 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047400200601805417, + "loss": 1.6819, + "theoretical_loss": 4.375955646810594, + "tokens_seen": 202899456 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047399197592778335, + "loss": 1.7511, + "theoretical_loss": 4.37578035833218, + "tokens_seen": 202964992 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047398194583751253, + "loss": 1.8706, + "theoretical_loss": 4.375605142286087, + "tokens_seen": 203030528 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047397191574724177, + "loss": 1.8242, + "theoretical_loss": 4.37542999861902, + "tokens_seen": 203096064 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 82204, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7427034378051758, + "objective/train/theoretical_loss": 4.375254927277737, + "objective/train/tokens_used": 203161600, + "theoretical_loss": 4.375254927277737, + "tokens_seen": 203161600 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004739618856569709, + "loss": 1.7186, + "theoretical_loss": 4.375254927277737, + "tokens_seen": 203161600 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047395185556670013, + "loss": 1.767, + "theoretical_loss": 4.375079928209058, + "tokens_seen": 203227136 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047394182547642926, + "loss": 1.8283, + "theoretical_loss": 4.374905001359851, + "tokens_seen": 203292672 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004739317953861585, + "loss": 1.7463, + "theoretical_loss": 4.374730146677049, + "tokens_seen": 203358208 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004739217652958877, + "loss": 1.7634, + "theoretical_loss": 4.374555364107635, + "tokens_seen": 203423744 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047391173520561686, + "loss": 1.7135, + "theoretical_loss": 4.37438065359865, + "tokens_seen": 203489280 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047390170511534604, + "loss": 1.6781, + "theoretical_loss": 4.374206015097193, + "tokens_seen": 203554816 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004738916750250752, + "loss": 1.7042, + "theoretical_loss": 4.374031448550414, + "tokens_seen": 203620352 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004738816449348044, + "loss": 1.7709, + "theoretical_loss": 4.373856953905523, + "tokens_seen": 203685888 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047387161484453363, + "loss": 1.7627, + "theoretical_loss": 4.373682531109785, + "tokens_seen": 203751424 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047386158475426276, + "loss": 1.7698, + "theoretical_loss": 4.37350818011052, + "tokens_seen": 203816960 + }, + { + "epoch": 0.06, + "learning_rate": 0.000473851554663992, + "loss": 1.7924, + "theoretical_loss": 4.3733339008551, + "tokens_seen": 203882496 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004738415245737211, + "loss": 1.6176, + "theoretical_loss": 4.373159693290961, + "tokens_seen": 203948032 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047383149448345036, + "loss": 1.5773, + "theoretical_loss": 4.372985557365586, + "tokens_seen": 204013568 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047382146439317954, + "loss": 1.7066, + "theoretical_loss": 4.372811493026518, + "tokens_seen": 204079104 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004738114343029087, + "loss": 1.7299, + "theoretical_loss": 4.3726375002213524, + "tokens_seen": 204144640 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004738014042126379, + "loss": 1.7836, + "theoretical_loss": 4.372463578897743, + "tokens_seen": 204210176 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047379137412236714, + "loss": 1.8589, + "theoretical_loss": 4.372289729003397, + "tokens_seen": 204275712 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047378134403209626, + "loss": 1.8164, + "theoretical_loss": 4.372115950486075, + "tokens_seen": 204341248 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004737713139418255, + "loss": 1.7726, + "theoretical_loss": 4.371942243293595, + "tokens_seen": 204406784 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004737612838515546, + "loss": 1.7794, + "theoretical_loss": 4.371768607373829, + "tokens_seen": 204472320 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047375125376128386, + "loss": 1.736, + "theoretical_loss": 4.371595042674704, + "tokens_seen": 204537856 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047374122367101304, + "loss": 1.8719, + "theoretical_loss": 4.371421549144202, + "tokens_seen": 204603392 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004737311935807422, + "loss": 1.6912, + "theoretical_loss": 4.371248126730357, + "tokens_seen": 204668928 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047372116349047146, + "loss": 1.7641, + "theoretical_loss": 4.3710747753812615, + "tokens_seen": 204734464 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004737111334002006, + "loss": 1.675, + "theoretical_loss": 4.370901495045061, + "tokens_seen": 204800000 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004737011033099298, + "loss": 1.767, + "theoretical_loss": 4.370728285669954, + "tokens_seen": 204865536 + }, + { + "epoch": 0.06, + "learning_rate": 0.000473691073219659, + "loss": 1.7801, + "theoretical_loss": 4.370555147204195, + "tokens_seen": 204931072 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004736810431293882, + "loss": 1.7707, + "theoretical_loss": 4.370382079596092, + "tokens_seen": 204996608 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047367101303911736, + "loss": 1.6679, + "theoretical_loss": 4.370209082794006, + "tokens_seen": 205062144 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004736609829488466, + "loss": 1.687, + "theoretical_loss": 4.370036156746357, + "tokens_seen": 205127680 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004736509528585757, + "loss": 1.8486, + "theoretical_loss": 4.369863301401612, + "tokens_seen": 205193216 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047364092276830496, + "loss": 1.6753, + "theoretical_loss": 4.369690516708298, + "tokens_seen": 205258752 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004736308926780341, + "loss": 1.7955, + "theoretical_loss": 4.3695178026149915, + "tokens_seen": 205324288 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004736208625877633, + "loss": 1.8218, + "theoretical_loss": 4.369345159070327, + "tokens_seen": 205389824 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004736108324974925, + "loss": 1.7908, + "theoretical_loss": 4.369172586022989, + "tokens_seen": 205455360 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004736008024072217, + "loss": 1.6694, + "theoretical_loss": 4.369000083421716, + "tokens_seen": 205520896 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047359077231695087, + "loss": 1.6487, + "theoretical_loss": 4.368827651215304, + "tokens_seen": 205586432 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047358074222668005, + "loss": 1.7703, + "theoretical_loss": 4.368655289352598, + "tokens_seen": 205651968 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047357071213640923, + "loss": 1.8056, + "theoretical_loss": 4.3684829977824995, + "tokens_seen": 205717504 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047356068204613847, + "loss": 1.759, + "theoretical_loss": 4.368310776453961, + "tokens_seen": 205783040 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004735506519558676, + "loss": 1.7157, + "theoretical_loss": 4.368138625315989, + "tokens_seen": 205848576 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047354062186559683, + "loss": 1.7627, + "theoretical_loss": 4.367966544317646, + "tokens_seen": 205914112 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047353059177532595, + "loss": 1.7625, + "theoretical_loss": 4.367794533408043, + "tokens_seen": 205979648 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004735205616850552, + "loss": 1.7723, + "theoretical_loss": 4.367622592536348, + "tokens_seen": 206045184 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047351053159478437, + "loss": 1.722, + "theoretical_loss": 4.36745072165178, + "tokens_seen": 206110720 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047350050150451355, + "loss": 1.8467, + "theoretical_loss": 4.36727892070361, + "tokens_seen": 206176256 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047349047141424273, + "loss": 1.7696, + "theoretical_loss": 4.367107189641167, + "tokens_seen": 206241792 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047348044132397197, + "loss": 1.6861, + "theoretical_loss": 4.3669355284138245, + "tokens_seen": 206307328 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004734704112337011, + "loss": 1.7294, + "theoretical_loss": 4.366763936971017, + "tokens_seen": 206372864 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 83354, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.812289834022522, + "objective/train/theoretical_loss": 4.366592415262226, + "objective/train/tokens_used": 206438400, + "theoretical_loss": 4.366592415262226, + "tokens_seen": 206438400 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047346038114343033, + "loss": 1.8261, + "theoretical_loss": 4.366592415262226, + "tokens_seen": 206438400 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047345035105315946, + "loss": 1.8405, + "theoretical_loss": 4.366420963236989, + "tokens_seen": 206503936 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004734403209628887, + "loss": 1.7694, + "theoretical_loss": 4.366249580844894, + "tokens_seen": 206569472 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004734302908726179, + "loss": 1.636, + "theoretical_loss": 4.366078268035581, + "tokens_seen": 206635008 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047342026078234706, + "loss": 1.8203, + "theoretical_loss": 4.365907024758744, + "tokens_seen": 206700544 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047341023069207624, + "loss": 1.7157, + "theoretical_loss": 4.365735850964131, + "tokens_seen": 206766080 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004734002006018054, + "loss": 1.7413, + "theoretical_loss": 4.365564746601537, + "tokens_seen": 206831616 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004733901705115346, + "loss": 1.7623, + "theoretical_loss": 4.365393711620814, + "tokens_seen": 206897152 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047338014042126383, + "loss": 1.7014, + "theoretical_loss": 4.365222745971863, + "tokens_seen": 206962688 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047337011033099296, + "loss": 1.8622, + "theoretical_loss": 4.365051849604638, + "tokens_seen": 207028224 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004733600802407222, + "loss": 1.7738, + "theoretical_loss": 4.364881022469148, + "tokens_seen": 207093760 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004733500501504513, + "loss": 1.7296, + "theoretical_loss": 4.364710264515448, + "tokens_seen": 207159296 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047334002006018056, + "loss": 1.8674, + "theoretical_loss": 4.36453957569365, + "tokens_seen": 207224832 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047332998996990974, + "loss": 1.7477, + "theoretical_loss": 4.364368955953913, + "tokens_seen": 207290368 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004733199598796389, + "loss": 1.7474, + "theoretical_loss": 4.364198405246453, + "tokens_seen": 207355904 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004733099297893681, + "loss": 1.7957, + "theoretical_loss": 4.364027923521535, + "tokens_seen": 207421440 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047329989969909734, + "loss": 1.7073, + "theoretical_loss": 4.363857510729473, + "tokens_seen": 207486976 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047328986960882646, + "loss": 1.764, + "theoretical_loss": 4.363687166820637, + "tokens_seen": 207552512 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004732798395185557, + "loss": 1.7332, + "theoretical_loss": 4.363516891745445, + "tokens_seen": 207618048 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004732698094282848, + "loss": 1.7595, + "theoretical_loss": 4.36334668545437, + "tokens_seen": 207683584 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047325977933801406, + "loss": 1.869, + "theoretical_loss": 4.363176547897931, + "tokens_seen": 207749120 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047324974924774324, + "loss": 1.6917, + "theoretical_loss": 4.363006479026702, + "tokens_seen": 207814656 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004732397191574724, + "loss": 1.7181, + "theoretical_loss": 4.3628364787913085, + "tokens_seen": 207880192 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004732296890672016, + "loss": 1.8768, + "theoretical_loss": 4.362666547142424, + "tokens_seen": 207945728 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004732196589769308, + "loss": 1.7592, + "theoretical_loss": 4.362496684030775, + "tokens_seen": 208011264 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047320962888665997, + "loss": 1.8336, + "theoretical_loss": 4.36232688940714, + "tokens_seen": 208076800 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004731995987963892, + "loss": 1.7714, + "theoretical_loss": 4.362157163222346, + "tokens_seen": 208142336 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047318956870611833, + "loss": 1.7771, + "theoretical_loss": 4.361987505427272, + "tokens_seen": 208207872 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047317953861584757, + "loss": 1.7804, + "theoretical_loss": 4.361817915972846, + "tokens_seen": 208273408 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047316950852557675, + "loss": 1.7863, + "theoretical_loss": 4.36164839481005, + "tokens_seen": 208338944 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047315947843530593, + "loss": 1.853, + "theoretical_loss": 4.361478941889915, + "tokens_seen": 208404480 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004731494483450351, + "loss": 1.7744, + "theoretical_loss": 4.36130955716352, + "tokens_seen": 208470016 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004731394182547643, + "loss": 1.717, + "theoretical_loss": 4.361140240581998, + "tokens_seen": 208535552 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047312938816449347, + "loss": 1.8272, + "theoretical_loss": 4.36097099209653, + "tokens_seen": 208601088 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004731193580742227, + "loss": 1.7842, + "theoretical_loss": 4.36080181165835, + "tokens_seen": 208666624 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047310932798395183, + "loss": 1.7088, + "theoretical_loss": 4.360632699218739, + "tokens_seen": 208732160 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047309929789368107, + "loss": 1.8339, + "theoretical_loss": 4.360463654729029, + "tokens_seen": 208797696 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004730892678034102, + "loss": 1.7561, + "theoretical_loss": 4.360294678140605, + "tokens_seen": 208863232 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047307923771313943, + "loss": 1.6648, + "theoretical_loss": 4.360125769404897, + "tokens_seen": 208928768 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004730692076228686, + "loss": 1.707, + "theoretical_loss": 4.35995692847339, + "tokens_seen": 208994304 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004730591775325978, + "loss": 1.7322, + "theoretical_loss": 4.359788155297615, + "tokens_seen": 209059840 + }, + { + "epoch": 0.06, + "learning_rate": 0.000473049147442327, + "loss": 1.7809, + "theoretical_loss": 4.359619449829156, + "tokens_seen": 209125376 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047303911735205615, + "loss": 1.8513, + "theoretical_loss": 4.359450812019643, + "tokens_seen": 209190912 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047302908726178534, + "loss": 1.6942, + "theoretical_loss": 4.359282241820761, + "tokens_seen": 209256448 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047301905717151457, + "loss": 1.8711, + "theoretical_loss": 4.359113739184238, + "tokens_seen": 209321984 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004730090270812437, + "loss": 1.8279, + "theoretical_loss": 4.358945304061858, + "tokens_seen": 209387520 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047299899699097293, + "loss": 1.8764, + "theoretical_loss": 4.35877693640545, + "tokens_seen": 209453056 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004729889669007021, + "loss": 1.7245, + "theoretical_loss": 4.358608636166894, + "tokens_seen": 209518592 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004729789368104313, + "loss": 1.8795, + "theoretical_loss": 4.358440403298119, + "tokens_seen": 209584128 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047296890672016053, + "loss": 1.6978, + "theoretical_loss": 4.3582722377511045, + "tokens_seen": 209649664 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 84957, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7861697673797607, + "objective/train/theoretical_loss": 4.358104139477878, + "objective/train/tokens_used": 209715200, + "theoretical_loss": 4.358104139477878, + "tokens_seen": 209715200 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047295887662988966, + "loss": 1.7862, + "theoretical_loss": 4.358104139477878, + "tokens_seen": 209715200 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004729488465396189, + "loss": 1.7305, + "theoretical_loss": 4.357936108430517, + "tokens_seen": 209780736 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004729388164493481, + "loss": 1.6971, + "theoretical_loss": 4.357768144561146, + "tokens_seen": 209846272 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047292878635907726, + "loss": 1.8378, + "theoretical_loss": 4.357600247821942, + "tokens_seen": 209911808 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047291875626880644, + "loss": 1.6917, + "theoretical_loss": 4.3574324181651285, + "tokens_seen": 209977344 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004729087261785356, + "loss": 1.8151, + "theoretical_loss": 4.357264655542978, + "tokens_seen": 210042880 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004728986960882648, + "loss": 1.8716, + "theoretical_loss": 4.357096959907813, + "tokens_seen": 210108416 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047288866599799403, + "loss": 1.7544, + "theoretical_loss": 4.356929331212004, + "tokens_seen": 210173952 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047287863590772316, + "loss": 1.8002, + "theoretical_loss": 4.356761769407969, + "tokens_seen": 210239488 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004728686058174524, + "loss": 1.822, + "theoretical_loss": 4.356594274448178, + "tokens_seen": 210305024 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004728585757271815, + "loss": 1.7119, + "theoretical_loss": 4.356426846285147, + "tokens_seen": 210370560 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047284854563691076, + "loss": 1.8, + "theoretical_loss": 4.3562594848714395, + "tokens_seen": 210436096 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047283851554663994, + "loss": 1.7333, + "theoretical_loss": 4.3560921901596705, + "tokens_seen": 210501632 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004728284854563691, + "loss": 1.6943, + "theoretical_loss": 4.355924962102502, + "tokens_seen": 210567168 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004728184553660983, + "loss": 1.7035, + "theoretical_loss": 4.3557578006526425, + "tokens_seen": 210632704 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047280842527582754, + "loss": 1.7918, + "theoretical_loss": 4.355590705762852, + "tokens_seen": 210698240 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047279839518555666, + "loss": 1.8061, + "theoretical_loss": 4.355423677385937, + "tokens_seen": 210763776 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004727883650952859, + "loss": 1.8395, + "theoretical_loss": 4.355256715474751, + "tokens_seen": 210829312 + }, + { + "epoch": 0.06, + "learning_rate": 0.000472778335005015, + "loss": 1.8367, + "theoretical_loss": 4.355089819982198, + "tokens_seen": 210894848 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047276830491474426, + "loss": 1.7835, + "theoretical_loss": 4.354922990861229, + "tokens_seen": 210960384 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047275827482447344, + "loss": 1.7442, + "theoretical_loss": 4.354756228064842, + "tokens_seen": 211025920 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004727482447342026, + "loss": 1.7197, + "theoretical_loss": 4.354589531546083, + "tokens_seen": 211091456 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004727382146439318, + "loss": 1.7773, + "theoretical_loss": 4.354422901258048, + "tokens_seen": 211156992 + }, + { + "epoch": 0.06, + "learning_rate": 0.000472728184553661, + "loss": 1.6554, + "theoretical_loss": 4.354256337153878, + "tokens_seen": 211222528 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047271815446339017, + "loss": 1.8183, + "theoretical_loss": 4.354089839186763, + "tokens_seen": 211288064 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004727081243731194, + "loss": 1.7126, + "theoretical_loss": 4.35392340730994, + "tokens_seen": 211353600 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047269809428284853, + "loss": 1.7094, + "theoretical_loss": 4.353757041476695, + "tokens_seen": 211419136 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047268806419257777, + "loss": 1.7717, + "theoretical_loss": 4.35359074164036, + "tokens_seen": 211484672 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047267803410230695, + "loss": 1.7511, + "theoretical_loss": 4.3534245077543146, + "tokens_seen": 211550208 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047266800401203613, + "loss": 1.6662, + "theoretical_loss": 4.353258339771986, + "tokens_seen": 211615744 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004726579739217653, + "loss": 1.6516, + "theoretical_loss": 4.353092237646849, + "tokens_seen": 211681280 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004726479438314945, + "loss": 1.7501, + "theoretical_loss": 4.3529262013324255, + "tokens_seen": 211746816 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047263791374122367, + "loss": 1.6926, + "theoretical_loss": 4.352760230782284, + "tokens_seen": 211812352 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004726278836509529, + "loss": 1.6797, + "theoretical_loss": 4.352594325950041, + "tokens_seen": 211877888 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047261785356068203, + "loss": 1.7829, + "theoretical_loss": 4.35242848678936, + "tokens_seen": 211943424 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047260782347041127, + "loss": 1.7184, + "theoretical_loss": 4.35226271325395, + "tokens_seen": 212008960 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004725977933801404, + "loss": 1.7752, + "theoretical_loss": 4.352097005297569, + "tokens_seen": 212074496 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047258776328986963, + "loss": 1.6256, + "theoretical_loss": 4.351931362874019, + "tokens_seen": 212140032 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004725777331995988, + "loss": 1.7665, + "theoretical_loss": 4.351765785937154, + "tokens_seen": 212205568 + }, + { + "epoch": 0.06, + "learning_rate": 0.000472567703109328, + "loss": 1.8141, + "theoretical_loss": 4.3516002744408695, + "tokens_seen": 212271104 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004725576730190572, + "loss": 1.7432, + "theoretical_loss": 4.351434828339109, + "tokens_seen": 212336640 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047254764292878636, + "loss": 1.8781, + "theoretical_loss": 4.351269447585864, + "tokens_seen": 212402176 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047253761283851554, + "loss": 1.8555, + "theoretical_loss": 4.3511041321351716, + "tokens_seen": 212467712 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047252758274824477, + "loss": 1.7428, + "theoretical_loss": 4.350938881941116, + "tokens_seen": 212533248 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004725175526579739, + "loss": 1.8037, + "theoretical_loss": 4.350773696957827, + "tokens_seen": 212598784 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047250752256770313, + "loss": 1.7252, + "theoretical_loss": 4.350608577139481, + "tokens_seen": 212664320 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004724974924774323, + "loss": 1.7746, + "theoretical_loss": 4.350443522440301, + "tokens_seen": 212729856 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004724874623871615, + "loss": 1.6912, + "theoretical_loss": 4.350278532814556, + "tokens_seen": 212795392 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004724774322968907, + "loss": 1.9488, + "theoretical_loss": 4.350113608216563, + "tokens_seen": 212860928 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047246740220661986, + "loss": 1.7812, + "theoretical_loss": 4.3499487486006805, + "tokens_seen": 212926464 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 86198, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8377131223678589, + "objective/train/theoretical_loss": 4.349783953921318, + "objective/train/tokens_used": 212992000, + "theoretical_loss": 4.349783953921318, + "tokens_seen": 212992000 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047245737211634904, + "loss": 1.7293, + "theoretical_loss": 4.349783953921318, + "tokens_seen": 212992000 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004724473420260783, + "loss": 1.8676, + "theoretical_loss": 4.349619224132929, + "tokens_seen": 213057536 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004724373119358074, + "loss": 1.7243, + "theoretical_loss": 4.349454559190011, + "tokens_seen": 213123072 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047242728184553664, + "loss": 1.7526, + "theoretical_loss": 4.349289959047112, + "tokens_seen": 213188608 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047241725175526576, + "loss": 1.7828, + "theoretical_loss": 4.349125423658821, + "tokens_seen": 213254144 + }, + { + "epoch": 0.06, + "learning_rate": 0.000472407221664995, + "loss": 1.7507, + "theoretical_loss": 4.348960952979776, + "tokens_seen": 213319680 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004723971915747242, + "loss": 1.8321, + "theoretical_loss": 4.3487965469646594, + "tokens_seen": 213385216 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047238716148445336, + "loss": 1.8013, + "theoretical_loss": 4.3486322055682, + "tokens_seen": 213450752 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047237713139418254, + "loss": 1.7026, + "theoretical_loss": 4.348467928745171, + "tokens_seen": 213516288 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004723671013039117, + "loss": 1.6742, + "theoretical_loss": 4.348303716450391, + "tokens_seen": 213581824 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004723570712136409, + "loss": 1.7826, + "theoretical_loss": 4.348139568638727, + "tokens_seen": 213647360 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047234704112337014, + "loss": 1.8446, + "theoretical_loss": 4.347975485265087, + "tokens_seen": 213712896 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047233701103309927, + "loss": 1.8201, + "theoretical_loss": 4.347811466284428, + "tokens_seen": 213778432 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004723269809428285, + "loss": 1.7842, + "theoretical_loss": 4.34764751165175, + "tokens_seen": 213843968 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004723169508525577, + "loss": 1.8428, + "theoretical_loss": 4.347483621322099, + "tokens_seen": 213909504 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047230692076228686, + "loss": 1.7814, + "theoretical_loss": 4.347319795250567, + "tokens_seen": 213975040 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047229689067201605, + "loss": 1.7519, + "theoretical_loss": 4.347156033392291, + "tokens_seen": 214040576 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004722868605817452, + "loss": 1.7441, + "theoretical_loss": 4.346992335702451, + "tokens_seen": 214106112 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004722768304914744, + "loss": 1.7971, + "theoretical_loss": 4.346828702136273, + "tokens_seen": 214171648 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047226680040120364, + "loss": 1.7973, + "theoretical_loss": 4.34666513264903, + "tokens_seen": 214237184 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047225677031093277, + "loss": 1.7195, + "theoretical_loss": 4.346501627196037, + "tokens_seen": 214302720 + }, + { + "epoch": 0.06, + "learning_rate": 0.000472246740220662, + "loss": 1.7618, + "theoretical_loss": 4.346338185732655, + "tokens_seen": 214368256 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047223671013039113, + "loss": 1.8123, + "theoretical_loss": 4.3461748082142915, + "tokens_seen": 214433792 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047222668004012037, + "loss": 1.8297, + "theoretical_loss": 4.346011494596395, + "tokens_seen": 214499328 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004722166499498496, + "loss": 1.7847, + "theoretical_loss": 4.34584824483446, + "tokens_seen": 214564864 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047220661985957873, + "loss": 1.8062, + "theoretical_loss": 4.345685058884029, + "tokens_seen": 214630400 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047219658976930797, + "loss": 1.7216, + "theoretical_loss": 4.345521936700684, + "tokens_seen": 214695936 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047218655967903715, + "loss": 1.7592, + "theoretical_loss": 4.345358878240053, + "tokens_seen": 214761472 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047217652958876633, + "loss": 1.8087, + "theoretical_loss": 4.345195883457811, + "tokens_seen": 214827008 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004721664994984955, + "loss": 1.6792, + "theoretical_loss": 4.3450329523096745, + "tokens_seen": 214892544 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004721564694082247, + "loss": 1.8433, + "theoretical_loss": 4.344870084751406, + "tokens_seen": 214958080 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047214643931795387, + "loss": 1.8827, + "theoretical_loss": 4.34470728073881, + "tokens_seen": 215023616 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004721364092276831, + "loss": 1.7492, + "theoretical_loss": 4.344544540227737, + "tokens_seen": 215089152 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047212637913741223, + "loss": 1.8401, + "theoretical_loss": 4.344381863174082, + "tokens_seen": 215154688 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047211634904714147, + "loss": 1.6476, + "theoretical_loss": 4.344219249533783, + "tokens_seen": 215220224 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004721063189568706, + "loss": 1.8421, + "theoretical_loss": 4.34405669926282, + "tokens_seen": 215285760 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047209628886659983, + "loss": 1.7383, + "theoretical_loss": 4.343894212317224, + "tokens_seen": 215351296 + }, + { + "epoch": 0.07, + "learning_rate": 0.000472086258776329, + "loss": 1.8349, + "theoretical_loss": 4.34373178865306, + "tokens_seen": 215416832 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004720762286860582, + "loss": 1.829, + "theoretical_loss": 4.343569428226445, + "tokens_seen": 215482368 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004720661985957874, + "loss": 1.7774, + "theoretical_loss": 4.343407130993536, + "tokens_seen": 215547904 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047205616850551656, + "loss": 1.6854, + "theoretical_loss": 4.343244896910534, + "tokens_seen": 215613440 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047204613841524574, + "loss": 1.6007, + "theoretical_loss": 4.343082725933684, + "tokens_seen": 215678976 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047203610832497497, + "loss": 1.6659, + "theoretical_loss": 4.342920618019275, + "tokens_seen": 215744512 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004720260782347041, + "loss": 1.7069, + "theoretical_loss": 4.342758573123639, + "tokens_seen": 215810048 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047201604814443333, + "loss": 1.7361, + "theoretical_loss": 4.342596591203153, + "tokens_seen": 215875584 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004720060180541625, + "loss": 1.7797, + "theoretical_loss": 4.3424346722142335, + "tokens_seen": 215941120 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004719959879638917, + "loss": 1.8018, + "theoretical_loss": 4.342272816113344, + "tokens_seen": 216006656 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004719859578736209, + "loss": 1.7144, + "theoretical_loss": 4.342111022856992, + "tokens_seen": 216072192 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047197592778335006, + "loss": 1.7713, + "theoretical_loss": 4.341949292401725, + "tokens_seen": 216137728 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047196589769307924, + "loss": 1.7497, + "theoretical_loss": 4.341787624704135, + "tokens_seen": 216203264 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 87357, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9433777332305908, + "objective/train/theoretical_loss": 4.341626019720858, + "objective/train/tokens_used": 216268800, + "theoretical_loss": 4.341626019720858, + "tokens_seen": 216268800 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004719558676028085, + "loss": 1.78, + "theoretical_loss": 4.341626019720858, + "tokens_seen": 216268800 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004719458375125376, + "loss": 1.9139, + "theoretical_loss": 4.341464477408572, + "tokens_seen": 216334336 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047193580742226684, + "loss": 1.6871, + "theoretical_loss": 4.341302997724, + "tokens_seen": 216399872 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047192577733199596, + "loss": 1.7596, + "theoretical_loss": 4.341141580623905, + "tokens_seen": 216465408 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004719157472417252, + "loss": 1.9171, + "theoretical_loss": 4.340980226065096, + "tokens_seen": 216530944 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004719057171514544, + "loss": 1.7146, + "theoretical_loss": 4.340818934004421, + "tokens_seen": 216596480 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047189568706118356, + "loss": 1.7787, + "theoretical_loss": 4.340657704398774, + "tokens_seen": 216662016 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047188565697091274, + "loss": 1.6893, + "theoretical_loss": 4.340496537205093, + "tokens_seen": 216727552 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004718756268806419, + "loss": 1.6909, + "theoretical_loss": 4.340335432380353, + "tokens_seen": 216793088 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004718655967903711, + "loss": 1.7254, + "theoretical_loss": 4.340174389881579, + "tokens_seen": 216858624 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047185556670010034, + "loss": 1.7824, + "theoretical_loss": 4.340013409665833, + "tokens_seen": 216924160 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047184553660982947, + "loss": 1.8047, + "theoretical_loss": 4.339852491690221, + "tokens_seen": 216989696 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004718355065195587, + "loss": 1.6993, + "theoretical_loss": 4.339691635911892, + "tokens_seen": 217055232 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004718254764292879, + "loss": 1.7637, + "theoretical_loss": 4.339530842288037, + "tokens_seen": 217120768 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047181544633901706, + "loss": 1.7254, + "theoretical_loss": 4.339370110775892, + "tokens_seen": 217186304 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047180541624874625, + "loss": 1.7784, + "theoretical_loss": 4.33920944133273, + "tokens_seen": 217251840 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004717953861584754, + "loss": 1.7356, + "theoretical_loss": 4.3390488339158715, + "tokens_seen": 217317376 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004717853560682046, + "loss": 1.7679, + "theoretical_loss": 4.338888288482675, + "tokens_seen": 217382912 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047177532597793384, + "loss": 1.7418, + "theoretical_loss": 4.338727804990544, + "tokens_seen": 217448448 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047176529588766297, + "loss": 1.7096, + "theoretical_loss": 4.338567383396924, + "tokens_seen": 217513984 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004717552657973922, + "loss": 1.7454, + "theoretical_loss": 4.338407023659302, + "tokens_seen": 217579520 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047174523570712133, + "loss": 1.6892, + "theoretical_loss": 4.338246725735205, + "tokens_seen": 217645056 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047173520561685057, + "loss": 1.6366, + "theoretical_loss": 4.338086489582205, + "tokens_seen": 217710592 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047172517552657975, + "loss": 1.6688, + "theoretical_loss": 4.3379263151579135, + "tokens_seen": 217776128 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047171514543630893, + "loss": 1.714, + "theoretical_loss": 4.337766202419986, + "tokens_seen": 217841664 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004717051153460381, + "loss": 1.7574, + "theoretical_loss": 4.337606151326119, + "tokens_seen": 217907200 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047169508525576735, + "loss": 1.7152, + "theoretical_loss": 4.337446161834048, + "tokens_seen": 217972736 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004716850551654965, + "loss": 1.7497, + "theoretical_loss": 4.3372862339015565, + "tokens_seen": 218038272 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004716750250752257, + "loss": 1.7789, + "theoretical_loss": 4.337126367486462, + "tokens_seen": 218103808 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047166499498495484, + "loss": 1.8792, + "theoretical_loss": 4.33696656254663, + "tokens_seen": 218169344 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047165496489468407, + "loss": 1.7586, + "theoretical_loss": 4.336806819039962, + "tokens_seen": 218234880 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047164493480441325, + "loss": 1.7617, + "theoretical_loss": 4.336647136924407, + "tokens_seen": 218300416 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047163490471414243, + "loss": 1.7389, + "theoretical_loss": 4.33648751615795, + "tokens_seen": 218365952 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004716248746238716, + "loss": 1.7308, + "theoretical_loss": 4.33632795669862, + "tokens_seen": 218431488 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004716148445336008, + "loss": 1.8515, + "theoretical_loss": 4.336168458504486, + "tokens_seen": 218497024 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047160481444333, + "loss": 1.5931, + "theoretical_loss": 4.336009021533661, + "tokens_seen": 218562560 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004715947843530592, + "loss": 1.8253, + "theoretical_loss": 4.335849645744295, + "tokens_seen": 218628096 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047158475426278834, + "loss": 1.7049, + "theoretical_loss": 4.3356903310945825, + "tokens_seen": 218693632 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004715747241725176, + "loss": 1.7473, + "theoretical_loss": 4.335531077542758, + "tokens_seen": 218759168 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004715646940822467, + "loss": 1.7906, + "theoretical_loss": 4.335371885047096, + "tokens_seen": 218824704 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047155466399197594, + "loss": 1.7439, + "theoretical_loss": 4.335212753565914, + "tokens_seen": 218890240 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004715446339017051, + "loss": 1.8022, + "theoretical_loss": 4.33505368305757, + "tokens_seen": 218955776 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004715346038114343, + "loss": 1.7383, + "theoretical_loss": 4.33489467348046, + "tokens_seen": 219021312 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004715245737211635, + "loss": 1.7166, + "theoretical_loss": 4.334735724793026, + "tokens_seen": 219086848 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004715145436308927, + "loss": 1.7755, + "theoretical_loss": 4.334576836953744, + "tokens_seen": 219152384 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047150451354062184, + "loss": 1.7666, + "theoretical_loss": 4.334418009921139, + "tokens_seen": 219217920 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004714944834503511, + "loss": 1.8513, + "theoretical_loss": 4.334259243653769, + "tokens_seen": 219283456 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004714844533600802, + "loss": 1.7605, + "theoretical_loss": 4.334100538110237, + "tokens_seen": 219348992 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047147442326980944, + "loss": 1.7996, + "theoretical_loss": 4.333941893249184, + "tokens_seen": 219414528 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004714643931795387, + "loss": 1.7103, + "theoretical_loss": 4.333783309029296, + "tokens_seen": 219480064 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 88560, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6464080810546875, + "objective/train/theoretical_loss": 4.3336247854092935, + "objective/train/tokens_used": 219545600, + "theoretical_loss": 4.3336247854092935, + "tokens_seen": 219545600 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004714543630892678, + "loss": 1.6198, + "theoretical_loss": 4.3336247854092935, + "tokens_seen": 219545600 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047144433299899704, + "loss": 1.8285, + "theoretical_loss": 4.333466322347942, + "tokens_seen": 219611136 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047143430290872616, + "loss": 1.8293, + "theoretical_loss": 4.333307919804045, + "tokens_seen": 219676672 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004714242728184554, + "loss": 1.7677, + "theoretical_loss": 4.333149577736446, + "tokens_seen": 219742208 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004714142427281846, + "loss": 1.6264, + "theoretical_loss": 4.332991296104032, + "tokens_seen": 219807744 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047140421263791376, + "loss": 1.901, + "theoretical_loss": 4.332833074865725, + "tokens_seen": 219873280 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047139418254764294, + "loss": 1.7377, + "theoretical_loss": 4.332674913980492, + "tokens_seen": 219938816 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004713841524573721, + "loss": 1.7431, + "theoretical_loss": 4.332516813407337, + "tokens_seen": 220004352 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004713741223671013, + "loss": 1.8339, + "theoretical_loss": 4.332358773105307, + "tokens_seen": 220069888 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047136409227683054, + "loss": 1.8619, + "theoretical_loss": 4.332200793033485, + "tokens_seen": 220135424 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047135406218655967, + "loss": 1.7744, + "theoretical_loss": 4.332042873150997, + "tokens_seen": 220200960 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004713440320962889, + "loss": 1.7255, + "theoretical_loss": 4.3318850134170095, + "tokens_seen": 220266496 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004713340020060181, + "loss": 1.8033, + "theoretical_loss": 4.331727213790725, + "tokens_seen": 220332032 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047132397191574726, + "loss": 1.717, + "theoretical_loss": 4.331569474231388, + "tokens_seen": 220397568 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047131394182547645, + "loss": 1.7596, + "theoretical_loss": 4.331411794698286, + "tokens_seen": 220463104 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047130391173520563, + "loss": 1.7626, + "theoretical_loss": 4.33125417515074, + "tokens_seen": 220528640 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004712938816449348, + "loss": 1.6863, + "theoretical_loss": 4.331096615548115, + "tokens_seen": 220594176 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047128385155466404, + "loss": 1.6783, + "theoretical_loss": 4.330939115849814, + "tokens_seen": 220659712 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047127382146439317, + "loss": 1.7017, + "theoretical_loss": 4.33078167601528, + "tokens_seen": 220725248 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004712637913741224, + "loss": 1.6987, + "theoretical_loss": 4.330624296003996, + "tokens_seen": 220790784 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047125376128385153, + "loss": 1.8712, + "theoretical_loss": 4.330466975775483, + "tokens_seen": 220856320 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047124373119358077, + "loss": 1.7094, + "theoretical_loss": 4.330309715289301, + "tokens_seen": 220921856 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047123370110330995, + "loss": 1.7085, + "theoretical_loss": 4.330152514505053, + "tokens_seen": 220987392 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047122367101303913, + "loss": 1.7695, + "theoretical_loss": 4.329995373382378, + "tokens_seen": 221052928 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004712136409227683, + "loss": 1.7863, + "theoretical_loss": 4.329838291880954, + "tokens_seen": 221118464 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047120361083249755, + "loss": 1.9036, + "theoretical_loss": 4.329681269960499, + "tokens_seen": 221184000 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004711935807422267, + "loss": 1.8795, + "theoretical_loss": 4.329524307580773, + "tokens_seen": 221249536 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004711835506519559, + "loss": 1.7942, + "theoretical_loss": 4.3293674047015696, + "tokens_seen": 221315072 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047117352056168504, + "loss": 1.7811, + "theoretical_loss": 4.329210561282726, + "tokens_seen": 221380608 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047116349047141427, + "loss": 1.7132, + "theoretical_loss": 4.329053777284115, + "tokens_seen": 221446144 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047115346038114345, + "loss": 1.7785, + "theoretical_loss": 4.328897052665653, + "tokens_seen": 221511680 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047114343029087263, + "loss": 1.719, + "theoretical_loss": 4.328740387387289, + "tokens_seen": 221577216 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004711334002006018, + "loss": 1.8532, + "theoretical_loss": 4.328583781409016, + "tokens_seen": 221642752 + }, + { + "epoch": 0.07, + "learning_rate": 0.000471123370110331, + "loss": 1.7324, + "theoretical_loss": 4.328427234690864, + "tokens_seen": 221708288 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004711133400200602, + "loss": 1.5996, + "theoretical_loss": 4.328270747192901, + "tokens_seen": 221773824 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004711033099297894, + "loss": 1.6405, + "theoretical_loss": 4.328114318875236, + "tokens_seen": 221839360 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047109327983951854, + "loss": 1.7078, + "theoretical_loss": 4.327957949698012, + "tokens_seen": 221904896 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004710832497492478, + "loss": 1.7112, + "theoretical_loss": 4.327801639621417, + "tokens_seen": 221970432 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004710732196589769, + "loss": 1.8038, + "theoretical_loss": 4.3276453886056725, + "tokens_seen": 222035968 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047106318956870614, + "loss": 1.6582, + "theoretical_loss": 4.32748919661104, + "tokens_seen": 222101504 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004710531594784353, + "loss": 1.7868, + "theoretical_loss": 4.327333063597821, + "tokens_seen": 222167040 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004710431293881645, + "loss": 1.7257, + "theoretical_loss": 4.327176989526352, + "tokens_seen": 222232576 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004710330992978937, + "loss": 1.7557, + "theoretical_loss": 4.327020974357012, + "tokens_seen": 222298112 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004710230692076229, + "loss": 1.8298, + "theoretical_loss": 4.326865018050215, + "tokens_seen": 222363648 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047101303911735204, + "loss": 1.7474, + "theoretical_loss": 4.326709120566415, + "tokens_seen": 222429184 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004710030090270813, + "loss": 1.8134, + "theoretical_loss": 4.326553281866103, + "tokens_seen": 222494720 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004709929789368104, + "loss": 1.713, + "theoretical_loss": 4.326397501909811, + "tokens_seen": 222560256 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047098294884653964, + "loss": 1.6891, + "theoretical_loss": 4.326241780658104, + "tokens_seen": 222625792 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004709729187562688, + "loss": 1.7225, + "theoretical_loss": 4.32608611807159, + "tokens_seen": 222691328 + }, + { + "epoch": 0.07, + "learning_rate": 0.000470962888665998, + "loss": 1.7214, + "theoretical_loss": 4.325930514110912, + "tokens_seen": 222756864 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 89737, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6674587726593018, + "objective/train/theoretical_loss": 4.325774968736753, + "objective/train/tokens_used": 222822400, + "theoretical_loss": 4.325774968736753, + "tokens_seen": 222822400 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004709528585757272, + "loss": 1.6715, + "theoretical_loss": 4.325774968736753, + "tokens_seen": 222822400 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047094282848545636, + "loss": 1.7765, + "theoretical_loss": 4.325619481909833, + "tokens_seen": 222887936 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047093279839518555, + "loss": 1.7363, + "theoretical_loss": 4.325464053590908, + "tokens_seen": 222953472 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004709227683049148, + "loss": 1.7944, + "theoretical_loss": 4.325308683740776, + "tokens_seen": 223019008 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004709127382146439, + "loss": 1.7202, + "theoretical_loss": 4.325153372320269, + "tokens_seen": 223084544 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047090270812437314, + "loss": 1.7051, + "theoretical_loss": 4.32499811929026, + "tokens_seen": 223150080 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047089267803410227, + "loss": 1.7274, + "theoretical_loss": 4.324842924611655, + "tokens_seen": 223215616 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004708826479438315, + "loss": 1.8601, + "theoretical_loss": 4.324687788245403, + "tokens_seen": 223281152 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004708726178535607, + "loss": 1.7688, + "theoretical_loss": 4.3245327101524875, + "tokens_seen": 223346688 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047086258776328987, + "loss": 1.6863, + "theoretical_loss": 4.324377690293929, + "tokens_seen": 223412224 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047085255767301905, + "loss": 1.7461, + "theoretical_loss": 4.324222728630787, + "tokens_seen": 223477760 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004708425275827483, + "loss": 1.7605, + "theoretical_loss": 4.324067825124159, + "tokens_seen": 223543296 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004708324974924774, + "loss": 1.6723, + "theoretical_loss": 4.323912979735178, + "tokens_seen": 223608832 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047082246740220665, + "loss": 1.782, + "theoretical_loss": 4.323758192425016, + "tokens_seen": 223674368 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004708124373119358, + "loss": 1.8077, + "theoretical_loss": 4.323603463154882, + "tokens_seen": 223739904 + }, + { + "epoch": 0.07, + "learning_rate": 0.000470802407221665, + "loss": 1.596, + "theoretical_loss": 4.323448791886022, + "tokens_seen": 223805440 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004707923771313942, + "loss": 1.7106, + "theoretical_loss": 4.323294178579717, + "tokens_seen": 223870976 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047078234704112337, + "loss": 1.7921, + "theoretical_loss": 4.323139623197289, + "tokens_seen": 223936512 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047077231695085255, + "loss": 1.8188, + "theoretical_loss": 4.322985125700096, + "tokens_seen": 224002048 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047076228686058173, + "loss": 1.691, + "theoretical_loss": 4.322830686049532, + "tokens_seen": 224067584 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004707522567703109, + "loss": 1.8178, + "theoretical_loss": 4.322676304207027, + "tokens_seen": 224133120 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047074222668004015, + "loss": 1.6633, + "theoretical_loss": 4.32252198013405, + "tokens_seen": 224198656 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004707321965897693, + "loss": 1.7004, + "theoretical_loss": 4.32236771379211, + "tokens_seen": 224264192 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004707221664994985, + "loss": 1.7446, + "theoretical_loss": 4.322213505142744, + "tokens_seen": 224329728 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047071213640922775, + "loss": 1.7927, + "theoretical_loss": 4.322059354147535, + "tokens_seen": 224395264 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004707021063189569, + "loss": 1.7239, + "theoretical_loss": 4.321905260768097, + "tokens_seen": 224460800 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004706920762286861, + "loss": 1.8023, + "theoretical_loss": 4.321751224966083, + "tokens_seen": 224526336 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047068204613841524, + "loss": 1.7283, + "theoretical_loss": 4.321597246703183, + "tokens_seen": 224591872 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047067201604814447, + "loss": 1.7003, + "theoretical_loss": 4.321443325941123, + "tokens_seen": 224657408 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047066198595787365, + "loss": 1.6586, + "theoretical_loss": 4.3212894626416665, + "tokens_seen": 224722944 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047065195586760283, + "loss": 1.7413, + "theoretical_loss": 4.321135656766611, + "tokens_seen": 224788480 + }, + { + "epoch": 0.07, + "learning_rate": 0.000470641925777332, + "loss": 1.7777, + "theoretical_loss": 4.320981908277793, + "tokens_seen": 224854016 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004706318956870612, + "loss": 1.6462, + "theoretical_loss": 4.320828217137086, + "tokens_seen": 224919552 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004706218655967904, + "loss": 1.6785, + "theoretical_loss": 4.320674583306396, + "tokens_seen": 224985088 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004706118355065196, + "loss": 1.6832, + "theoretical_loss": 4.32052100674767, + "tokens_seen": 225050624 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047060180541624874, + "loss": 1.8006, + "theoretical_loss": 4.320367487422891, + "tokens_seen": 225116160 + }, + { + "epoch": 0.07, + "learning_rate": 0.000470591775325978, + "loss": 1.666, + "theoretical_loss": 4.3202140252940735, + "tokens_seen": 225181696 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004705817452357071, + "loss": 1.6483, + "theoretical_loss": 4.3200606203232725, + "tokens_seen": 225247232 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047057171514543634, + "loss": 1.801, + "theoretical_loss": 4.319907272472579, + "tokens_seen": 225312768 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004705616850551655, + "loss": 1.7072, + "theoretical_loss": 4.319753981704119, + "tokens_seen": 225378304 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004705516549648947, + "loss": 1.766, + "theoretical_loss": 4.319600747980056, + "tokens_seen": 225443840 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004705416248746239, + "loss": 1.7142, + "theoretical_loss": 4.319447571262586, + "tokens_seen": 225509376 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004705315947843531, + "loss": 1.6915, + "theoretical_loss": 4.3192944515139455, + "tokens_seen": 225574912 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047052156469408224, + "loss": 1.7173, + "theoretical_loss": 4.319141388696404, + "tokens_seen": 225640448 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004705115346038115, + "loss": 1.7548, + "theoretical_loss": 4.318988382772269, + "tokens_seen": 225705984 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004705015045135406, + "loss": 1.7308, + "theoretical_loss": 4.318835433703883, + "tokens_seen": 225771520 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047049147442326984, + "loss": 1.6799, + "theoretical_loss": 4.318682541453622, + "tokens_seen": 225837056 + }, + { + "epoch": 0.07, + "learning_rate": 0.000470481444332999, + "loss": 1.6708, + "theoretical_loss": 4.318529705983903, + "tokens_seen": 225902592 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004704714142427282, + "loss": 1.7649, + "theoretical_loss": 4.318376927257174, + "tokens_seen": 225968128 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004704613841524574, + "loss": 1.7105, + "theoretical_loss": 4.3182242052359205, + "tokens_seen": 226033664 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 90912, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8519130945205688, + "objective/train/theoretical_loss": 4.318071539882665, + "objective/train/tokens_used": 226099200, + "theoretical_loss": 4.318071539882665, + "tokens_seen": 226099200 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047045135406218656, + "loss": 1.7591, + "theoretical_loss": 4.318071539882665, + "tokens_seen": 226099200 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047044132397191575, + "loss": 1.7135, + "theoretical_loss": 4.317918931159963, + "tokens_seen": 226164736 + }, + { + "epoch": 0.07, + "learning_rate": 0.000470431293881645, + "loss": 1.7454, + "theoretical_loss": 4.317766379030407, + "tokens_seen": 226230272 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004704212637913741, + "loss": 1.6553, + "theoretical_loss": 4.317613883456625, + "tokens_seen": 226295808 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047041123370110334, + "loss": 1.7506, + "theoretical_loss": 4.317461444401281, + "tokens_seen": 226361344 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047040120361083247, + "loss": 1.7758, + "theoretical_loss": 4.317309061827073, + "tokens_seen": 226426880 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004703911735205617, + "loss": 1.7284, + "theoretical_loss": 4.317156735696736, + "tokens_seen": 226492416 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004703811434302909, + "loss": 1.8049, + "theoretical_loss": 4.3170044659730396, + "tokens_seen": 226557952 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047037111334002007, + "loss": 1.7239, + "theoretical_loss": 4.316852252618788, + "tokens_seen": 226623488 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047036108324974925, + "loss": 1.5861, + "theoretical_loss": 4.316700095596822, + "tokens_seen": 226689024 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004703510531594785, + "loss": 1.7742, + "theoretical_loss": 4.3165479948700165, + "tokens_seen": 226754560 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004703410230692076, + "loss": 1.7184, + "theoretical_loss": 4.316395950401283, + "tokens_seen": 226820096 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047033099297893685, + "loss": 1.6999, + "theoretical_loss": 4.316243962153566, + "tokens_seen": 226885632 + }, + { + "epoch": 0.07, + "learning_rate": 0.000470320962888666, + "loss": 1.7357, + "theoretical_loss": 4.316092030089847, + "tokens_seen": 226951168 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004703109327983952, + "loss": 1.6417, + "theoretical_loss": 4.315940154173143, + "tokens_seen": 227016704 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004703009027081244, + "loss": 1.7559, + "theoretical_loss": 4.315788334366504, + "tokens_seen": 227082240 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047029087261785357, + "loss": 1.7229, + "theoretical_loss": 4.315636570633015, + "tokens_seen": 227147776 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047028084252758275, + "loss": 1.8085, + "theoretical_loss": 4.315484862935799, + "tokens_seen": 227213312 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047027081243731193, + "loss": 1.6269, + "theoretical_loss": 4.315333211238009, + "tokens_seen": 227278848 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004702607823470411, + "loss": 1.7543, + "theoretical_loss": 4.315181615502838, + "tokens_seen": 227344384 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047025075225677035, + "loss": 1.6108, + "theoretical_loss": 4.31503007569351, + "tokens_seen": 227409920 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004702407221664995, + "loss": 1.6216, + "theoretical_loss": 4.314878591773287, + "tokens_seen": 227475456 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004702306920762287, + "loss": 1.6661, + "theoretical_loss": 4.31472716370546, + "tokens_seen": 227540992 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047022066198595784, + "loss": 1.6578, + "theoretical_loss": 4.314575791453362, + "tokens_seen": 227606528 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004702106318956871, + "loss": 1.6841, + "theoretical_loss": 4.314424474980355, + "tokens_seen": 227672064 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047020060180541626, + "loss": 1.6013, + "theoretical_loss": 4.31427321424984, + "tokens_seen": 227737600 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047019057171514544, + "loss": 1.7559, + "theoretical_loss": 4.314122009225248, + "tokens_seen": 227803136 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004701805416248746, + "loss": 1.8134, + "theoretical_loss": 4.313970859870048, + "tokens_seen": 227868672 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047017051153460385, + "loss": 1.7695, + "theoretical_loss": 4.313819766147741, + "tokens_seen": 227934208 + }, + { + "epoch": 0.07, + "learning_rate": 0.000470160481444333, + "loss": 1.7959, + "theoretical_loss": 4.313668728021866, + "tokens_seen": 227999744 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004701504513540622, + "loss": 1.7898, + "theoretical_loss": 4.313517745455991, + "tokens_seen": 228065280 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047014042126379134, + "loss": 1.6536, + "theoretical_loss": 4.313366818413724, + "tokens_seen": 228130816 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004701303911735206, + "loss": 1.781, + "theoretical_loss": 4.313215946858702, + "tokens_seen": 228196352 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047012036108324976, + "loss": 1.7156, + "theoretical_loss": 4.313065130754601, + "tokens_seen": 228261888 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047011033099297894, + "loss": 1.708, + "theoretical_loss": 4.312914370065128, + "tokens_seen": 228327424 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004701003009027081, + "loss": 1.7704, + "theoretical_loss": 4.312763664754025, + "tokens_seen": 228392960 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004700902708124373, + "loss": 1.6844, + "theoretical_loss": 4.312613014785069, + "tokens_seen": 228458496 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004700802407221665, + "loss": 1.7918, + "theoretical_loss": 4.3124624201220705, + "tokens_seen": 228524032 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004700702106318957, + "loss": 1.6064, + "theoretical_loss": 4.312311880728873, + "tokens_seen": 228589568 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047006018054162484, + "loss": 1.7647, + "theoretical_loss": 4.312161396569356, + "tokens_seen": 228655104 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004700501504513541, + "loss": 1.7886, + "theoretical_loss": 4.312010967607431, + "tokens_seen": 228720640 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047004012036108326, + "loss": 1.7678, + "theoretical_loss": 4.311860593807045, + "tokens_seen": 228786176 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047003009027081244, + "loss": 1.7824, + "theoretical_loss": 4.311710275132178, + "tokens_seen": 228851712 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004700200601805416, + "loss": 1.7283, + "theoretical_loss": 4.3115600115468435, + "tokens_seen": 228917248 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004700100300902708, + "loss": 1.7498, + "theoretical_loss": 4.311409803015089, + "tokens_seen": 228982784 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047, + "loss": 1.7381, + "theoretical_loss": 4.3112596495009985, + "tokens_seen": 229048320 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004699899699097292, + "loss": 1.8247, + "theoretical_loss": 4.311109550968684, + "tokens_seen": 229113856 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046997993981945835, + "loss": 1.818, + "theoretical_loss": 4.310959507382296, + "tokens_seen": 229179392 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004699699097291876, + "loss": 1.6735, + "theoretical_loss": 4.310809518706017, + "tokens_seen": 229244928 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046995987963891676, + "loss": 1.7625, + "theoretical_loss": 4.310659584904062, + "tokens_seen": 229310464 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 92008, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.636131763458252, + "objective/train/theoretical_loss": 4.310509705940682, + "objective/train/tokens_used": 229376000, + "theoretical_loss": 4.310509705940682, + "tokens_seen": 229376000 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046994984954864595, + "loss": 1.6673, + "theoretical_loss": 4.310509705940682, + "tokens_seen": 229376000 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004699398194583752, + "loss": 1.7624, + "theoretical_loss": 4.310359881780159, + "tokens_seen": 229441536 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004699297893681043, + "loss": 1.7813, + "theoretical_loss": 4.310210112386811, + "tokens_seen": 229507072 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046991975927783354, + "loss": 1.7978, + "theoretical_loss": 4.310060397724987, + "tokens_seen": 229572608 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046990972918756267, + "loss": 1.7852, + "theoretical_loss": 4.30991073775907, + "tokens_seen": 229638144 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004698996990972919, + "loss": 1.7464, + "theoretical_loss": 4.3097611324534775, + "tokens_seen": 229703680 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004698896690070211, + "loss": 1.8092, + "theoretical_loss": 4.309611581772659, + "tokens_seen": 229769216 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046987963891675027, + "loss": 1.7588, + "theoretical_loss": 4.309462085681098, + "tokens_seen": 229834752 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046986960882647945, + "loss": 1.6603, + "theoretical_loss": 4.3093126441433105, + "tokens_seen": 229900288 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004698595787362087, + "loss": 1.7012, + "theoretical_loss": 4.3091632571238465, + "tokens_seen": 229965824 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004698495486459378, + "loss": 1.7958, + "theoretical_loss": 4.30901392458729, + "tokens_seen": 230031360 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046983951855566705, + "loss": 1.7455, + "theoretical_loss": 4.308864646498254, + "tokens_seen": 230096896 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004698294884653962, + "loss": 1.5212, + "theoretical_loss": 4.30871542282139, + "tokens_seen": 230162432 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004698194583751254, + "loss": 1.734, + "theoretical_loss": 4.308566253521378, + "tokens_seen": 230227968 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004698094282848546, + "loss": 1.8291, + "theoretical_loss": 4.308417138562934, + "tokens_seen": 230293504 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046979939819458377, + "loss": 1.7104, + "theoretical_loss": 4.308268077910806, + "tokens_seen": 230359040 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046978936810431295, + "loss": 1.7629, + "theoretical_loss": 4.308119071529774, + "tokens_seen": 230424576 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046977933801404213, + "loss": 1.7272, + "theoretical_loss": 4.3079701193846525, + "tokens_seen": 230490112 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004697693079237713, + "loss": 1.7455, + "theoretical_loss": 4.307821221440286, + "tokens_seen": 230555648 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046975927783350055, + "loss": 1.8077, + "theoretical_loss": 4.3076723776615555, + "tokens_seen": 230621184 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004697492477432297, + "loss": 1.8101, + "theoretical_loss": 4.307523588013371, + "tokens_seen": 230686720 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004697392176529589, + "loss": 1.7468, + "theoretical_loss": 4.307374852460679, + "tokens_seen": 230752256 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046972918756268804, + "loss": 1.6786, + "theoretical_loss": 4.307226170968456, + "tokens_seen": 230817792 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004697191574724173, + "loss": 1.7237, + "theoretical_loss": 4.307077543501713, + "tokens_seen": 230883328 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046970912738214646, + "loss": 1.6065, + "theoretical_loss": 4.3069289700254885, + "tokens_seen": 230948864 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046969909729187564, + "loss": 1.6724, + "theoretical_loss": 4.306780450504861, + "tokens_seen": 231014400 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004696890672016048, + "loss": 1.6905, + "theoretical_loss": 4.3066319849049375, + "tokens_seen": 231079936 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046967903711133405, + "loss": 1.8063, + "theoretical_loss": 4.306483573190857, + "tokens_seen": 231145472 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004696690070210632, + "loss": 1.9493, + "theoretical_loss": 4.306335215327791, + "tokens_seen": 231211008 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004696589769307924, + "loss": 1.7925, + "theoretical_loss": 4.306186911280945, + "tokens_seen": 231276544 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046964894684052154, + "loss": 1.6382, + "theoretical_loss": 4.306038661015557, + "tokens_seen": 231342080 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004696389167502508, + "loss": 1.7865, + "theoretical_loss": 4.305890464496895, + "tokens_seen": 231407616 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046962888665997996, + "loss": 1.8414, + "theoretical_loss": 4.305742321690261, + "tokens_seen": 231473152 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046961885656970914, + "loss": 1.807, + "theoretical_loss": 4.305594232560988, + "tokens_seen": 231538688 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004696088264794383, + "loss": 1.7907, + "theoretical_loss": 4.305446197074443, + "tokens_seen": 231604224 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004695987963891675, + "loss": 1.6758, + "theoretical_loss": 4.305298215196023, + "tokens_seen": 231669760 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004695887662988967, + "loss": 1.7871, + "theoretical_loss": 4.305150286891159, + "tokens_seen": 231735296 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004695787362086259, + "loss": 1.7498, + "theoretical_loss": 4.3050024121253125, + "tokens_seen": 231800832 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046956870611835505, + "loss": 1.7899, + "theoretical_loss": 4.304854590863979, + "tokens_seen": 231866368 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004695586760280843, + "loss": 1.6899, + "theoretical_loss": 4.304706823072683, + "tokens_seen": 231931904 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046954864593781346, + "loss": 1.8388, + "theoretical_loss": 4.304559108716984, + "tokens_seen": 231997440 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046953861584754264, + "loss": 1.5997, + "theoretical_loss": 4.304411447762472, + "tokens_seen": 232062976 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004695285857572718, + "loss": 1.7554, + "theoretical_loss": 4.304263840174768, + "tokens_seen": 232128512 + }, + { + "epoch": 0.07, + "learning_rate": 0.000469518555667001, + "loss": 1.8602, + "theoretical_loss": 4.304116285919527, + "tokens_seen": 232194048 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004695085255767302, + "loss": 1.807, + "theoretical_loss": 4.303968784962433, + "tokens_seen": 232259584 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004694984954864594, + "loss": 1.7912, + "theoretical_loss": 4.303821337269205, + "tokens_seen": 232325120 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046948846539618855, + "loss": 1.5735, + "theoretical_loss": 4.303673942805591, + "tokens_seen": 232390656 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004694784353059178, + "loss": 1.7663, + "theoretical_loss": 4.303526601537371, + "tokens_seen": 232456192 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004694684052156469, + "loss": 1.7424, + "theoretical_loss": 4.303379313430359, + "tokens_seen": 232521728 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046945837512537615, + "loss": 1.6462, + "theoretical_loss": 4.303232078450399, + "tokens_seen": 232587264 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 93187, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.658179759979248, + "objective/train/theoretical_loss": 4.303084896563364, + "objective/train/tokens_used": 232652800, + "theoretical_loss": 4.303084896563364, + "tokens_seen": 232652800 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004694483450351053, + "loss": 1.6189, + "theoretical_loss": 4.303084896563364, + "tokens_seen": 232652800 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004694383149448345, + "loss": 1.7253, + "theoretical_loss": 4.302937767735164, + "tokens_seen": 232718336 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004694282848545637, + "loss": 1.7441, + "theoretical_loss": 4.302790691931736, + "tokens_seen": 232783872 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046941825476429287, + "loss": 1.7285, + "theoretical_loss": 4.30264366911905, + "tokens_seen": 232849408 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046940822467402205, + "loss": 1.8286, + "theoretical_loss": 4.3024966992631075, + "tokens_seen": 232914944 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004693981945837513, + "loss": 1.7776, + "theoretical_loss": 4.302349782329942, + "tokens_seen": 232980480 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004693881644934804, + "loss": 1.7399, + "theoretical_loss": 4.302202918285616, + "tokens_seen": 233046016 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046937813440320965, + "loss": 1.665, + "theoretical_loss": 4.302056107096226, + "tokens_seen": 233111552 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046936810431293883, + "loss": 1.7162, + "theoretical_loss": 4.301909348727899, + "tokens_seen": 233177088 + }, + { + "epoch": 0.07, + "learning_rate": 0.000469358074222668, + "loss": 1.7945, + "theoretical_loss": 4.301762643146792, + "tokens_seen": 233242624 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004693480441323972, + "loss": 1.7496, + "theoretical_loss": 4.301615990319094, + "tokens_seen": 233308160 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004693380140421264, + "loss": 1.71, + "theoretical_loss": 4.301469390211025, + "tokens_seen": 233373696 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046932798395185555, + "loss": 1.6445, + "theoretical_loss": 4.301322842788839, + "tokens_seen": 233439232 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004693179538615848, + "loss": 1.6803, + "theoretical_loss": 4.301176348018814, + "tokens_seen": 233504768 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004693079237713139, + "loss": 1.782, + "theoretical_loss": 4.301029905867265, + "tokens_seen": 233570304 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046929789368104315, + "loss": 1.7307, + "theoretical_loss": 4.300883516300539, + "tokens_seen": 233635840 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004692878635907723, + "loss": 1.6654, + "theoretical_loss": 4.300737179285008, + "tokens_seen": 233701376 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004692778335005015, + "loss": 1.669, + "theoretical_loss": 4.30059089478708, + "tokens_seen": 233766912 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004692678034102307, + "loss": 1.7476, + "theoretical_loss": 4.300444662773192, + "tokens_seen": 233832448 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004692577733199599, + "loss": 1.7533, + "theoretical_loss": 4.30029848320981, + "tokens_seen": 233897984 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046924774322968906, + "loss": 1.7624, + "theoretical_loss": 4.300152356063435, + "tokens_seen": 233963520 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046923771313941824, + "loss": 1.8104, + "theoretical_loss": 4.300006281300598, + "tokens_seen": 234029056 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004692276830491474, + "loss": 1.6047, + "theoretical_loss": 4.299860258887856, + "tokens_seen": 234094592 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046921765295887666, + "loss": 1.7249, + "theoretical_loss": 4.299714288791802, + "tokens_seen": 234160128 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046920762286860584, + "loss": 1.7413, + "theoretical_loss": 4.299568370979056, + "tokens_seen": 234225664 + }, + { + "epoch": 0.07, + "learning_rate": 0.000469197592778335, + "loss": 1.703, + "theoretical_loss": 4.299422505416272, + "tokens_seen": 234291200 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046918756268806425, + "loss": 1.7413, + "theoretical_loss": 4.2992766920701335, + "tokens_seen": 234356736 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004691775325977934, + "loss": 1.6135, + "theoretical_loss": 4.299130930907352, + "tokens_seen": 234422272 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004691675025075226, + "loss": 1.7604, + "theoretical_loss": 4.2989852218946725, + "tokens_seen": 234487808 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046915747241725174, + "loss": 1.633, + "theoretical_loss": 4.298839564998869, + "tokens_seen": 234553344 + }, + { + "epoch": 0.07, + "learning_rate": 0.000469147442326981, + "loss": 1.7927, + "theoretical_loss": 4.298693960186747, + "tokens_seen": 234618880 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046913741223671016, + "loss": 1.7545, + "theoretical_loss": 4.298548407425141, + "tokens_seen": 234684416 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046912738214643934, + "loss": 1.8618, + "theoretical_loss": 4.298402906680918, + "tokens_seen": 234749952 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004691173520561685, + "loss": 1.7684, + "theoretical_loss": 4.298257457920973, + "tokens_seen": 234815488 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004691073219658977, + "loss": 1.766, + "theoretical_loss": 4.298112061112232, + "tokens_seen": 234881024 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004690972918756269, + "loss": 1.8295, + "theoretical_loss": 4.297966716221653, + "tokens_seen": 234946560 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004690872617853561, + "loss": 1.704, + "theoretical_loss": 4.297821423216222, + "tokens_seen": 235012096 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046907723169508525, + "loss": 1.8155, + "theoretical_loss": 4.297676182062955, + "tokens_seen": 235077632 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004690672016048145, + "loss": 1.7369, + "theoretical_loss": 4.297530992728901, + "tokens_seen": 235143168 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046905717151454366, + "loss": 1.689, + "theoretical_loss": 4.297385855181136, + "tokens_seen": 235208704 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046904714142427284, + "loss": 1.7395, + "theoretical_loss": 4.297240769386768, + "tokens_seen": 235274240 + }, + { + "epoch": 0.07, + "learning_rate": 0.000469037111334002, + "loss": 1.682, + "theoretical_loss": 4.297095735312934, + "tokens_seen": 235339776 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004690270812437312, + "loss": 1.6274, + "theoretical_loss": 4.296950752926801, + "tokens_seen": 235405312 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004690170511534604, + "loss": 1.7697, + "theoretical_loss": 4.296805822195568, + "tokens_seen": 235470848 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004690070210631896, + "loss": 1.7127, + "theoretical_loss": 4.29666094308646, + "tokens_seen": 235536384 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046899699097291875, + "loss": 1.8147, + "theoretical_loss": 4.296516115566736, + "tokens_seen": 235601920 + }, + { + "epoch": 0.07, + "learning_rate": 0.000468986960882648, + "loss": 1.7597, + "theoretical_loss": 4.296371339603684, + "tokens_seen": 235667456 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004689769307923771, + "loss": 1.7956, + "theoretical_loss": 4.296226615164618, + "tokens_seen": 235732992 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046896690070210635, + "loss": 1.6725, + "theoretical_loss": 4.296081942216887, + "tokens_seen": 235798528 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046895687061183553, + "loss": 1.7241, + "theoretical_loss": 4.295937320727866, + "tokens_seen": 235864064 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 94470, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6630239486694336, + "objective/train/theoretical_loss": 4.295792750664964, + "objective/train/tokens_used": 235929600, + "theoretical_loss": 4.295792750664964, + "tokens_seen": 235929600 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004689468405215647, + "loss": 1.7879, + "theoretical_loss": 4.295792750664964, + "tokens_seen": 235929600 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004689368104312939, + "loss": 1.6086, + "theoretical_loss": 4.295648231995613, + "tokens_seen": 235995136 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046892678034102307, + "loss": 1.6636, + "theoretical_loss": 4.295503764687281, + "tokens_seen": 236060672 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046891675025075225, + "loss": 1.7302, + "theoretical_loss": 4.295359348707462, + "tokens_seen": 236126208 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004689067201604815, + "loss": 1.847, + "theoretical_loss": 4.2952149840236835, + "tokens_seen": 236191744 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004688966900702106, + "loss": 1.7145, + "theoretical_loss": 4.295070670603496, + "tokens_seen": 236257280 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046888665997993985, + "loss": 1.7603, + "theoretical_loss": 4.294926408414486, + "tokens_seen": 236322816 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046887662988966903, + "loss": 1.7972, + "theoretical_loss": 4.294782197424265, + "tokens_seen": 236388352 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004688665997993982, + "loss": 1.6577, + "theoretical_loss": 4.294638037600478, + "tokens_seen": 236453888 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004688565697091274, + "loss": 1.7304, + "theoretical_loss": 4.2944939289107955, + "tokens_seen": 236519424 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004688465396188566, + "loss": 1.7665, + "theoretical_loss": 4.29434987132292, + "tokens_seen": 236584960 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046883650952858575, + "loss": 1.6579, + "theoretical_loss": 4.294205864804581, + "tokens_seen": 236650496 + }, + { + "epoch": 0.07, + "learning_rate": 0.000468826479438315, + "loss": 1.7712, + "theoretical_loss": 4.294061909323539, + "tokens_seen": 236716032 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004688164493480441, + "loss": 1.7757, + "theoretical_loss": 4.293918004847585, + "tokens_seen": 236781568 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046880641925777335, + "loss": 1.7514, + "theoretical_loss": 4.293774151344537, + "tokens_seen": 236847104 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004687963891675025, + "loss": 1.7375, + "theoretical_loss": 4.2936303487822425, + "tokens_seen": 236912640 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004687863590772317, + "loss": 1.7668, + "theoretical_loss": 4.293486597128578, + "tokens_seen": 236978176 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004687763289869609, + "loss": 1.789, + "theoretical_loss": 4.293342896351452, + "tokens_seen": 237043712 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004687662988966901, + "loss": 1.6791, + "theoretical_loss": 4.293199246418796, + "tokens_seen": 237109248 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046875626880641926, + "loss": 1.7438, + "theoretical_loss": 4.293055647298578, + "tokens_seen": 237174784 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046874623871614844, + "loss": 1.7041, + "theoretical_loss": 4.292912098958791, + "tokens_seen": 237240320 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004687362086258776, + "loss": 1.7646, + "theoretical_loss": 4.292768601367455, + "tokens_seen": 237305856 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046872617853560686, + "loss": 1.7503, + "theoretical_loss": 4.292625154492623, + "tokens_seen": 237371392 + }, + { + "epoch": 0.07, + "learning_rate": 0.000468716148445336, + "loss": 1.6933, + "theoretical_loss": 4.292481758302375, + "tokens_seen": 237436928 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004687061183550652, + "loss": 1.6932, + "theoretical_loss": 4.2923384127648205, + "tokens_seen": 237502464 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004686960882647944, + "loss": 1.593, + "theoretical_loss": 4.292195117848098, + "tokens_seen": 237568000 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004686860581745236, + "loss": 1.6681, + "theoretical_loss": 4.292051873520373, + "tokens_seen": 237633536 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046867602808425276, + "loss": 1.7876, + "theoretical_loss": 4.291908679749842, + "tokens_seen": 237699072 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046866599799398194, + "loss": 1.6345, + "theoretical_loss": 4.2917655365047285, + "tokens_seen": 237764608 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004686559679037111, + "loss": 1.6957, + "theoretical_loss": 4.291622443753287, + "tokens_seen": 237830144 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046864593781344036, + "loss": 1.6237, + "theoretical_loss": 4.291479401463799, + "tokens_seen": 237895680 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004686359077231695, + "loss": 1.7134, + "theoretical_loss": 4.291336409604574, + "tokens_seen": 237961216 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004686258776328987, + "loss": 1.7155, + "theoretical_loss": 4.291193468143952, + "tokens_seen": 238026752 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046861584754262785, + "loss": 1.6798, + "theoretical_loss": 4.2910505770503, + "tokens_seen": 238092288 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004686058174523571, + "loss": 1.738, + "theoretical_loss": 4.290907736292016, + "tokens_seen": 238157824 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046859578736208626, + "loss": 1.7638, + "theoretical_loss": 4.290764945837522, + "tokens_seen": 238223360 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046858575727181545, + "loss": 1.659, + "theoretical_loss": 4.290622205655273, + "tokens_seen": 238288896 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004685757271815446, + "loss": 1.6735, + "theoretical_loss": 4.290479515713751, + "tokens_seen": 238354432 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004685656970912738, + "loss": 1.7675, + "theoretical_loss": 4.290336875981465, + "tokens_seen": 238419968 + }, + { + "epoch": 0.07, + "learning_rate": 0.000468555667001003, + "loss": 1.6542, + "theoretical_loss": 4.290194286426955, + "tokens_seen": 238485504 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004685456369107322, + "loss": 1.7145, + "theoretical_loss": 4.290051747018785, + "tokens_seen": 238551040 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046853560682046135, + "loss": 1.656, + "theoretical_loss": 4.289909257725553, + "tokens_seen": 238616576 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004685255767301906, + "loss": 1.6839, + "theoretical_loss": 4.289766818515882, + "tokens_seen": 238682112 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046851554663991977, + "loss": 1.655, + "theoretical_loss": 4.2896244293584225, + "tokens_seen": 238747648 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046850551654964895, + "loss": 1.8208, + "theoretical_loss": 4.2894820902218544, + "tokens_seen": 238813184 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046849548645937813, + "loss": 1.7191, + "theoretical_loss": 4.289339801074888, + "tokens_seen": 238878720 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004684854563691073, + "loss": 1.7987, + "theoretical_loss": 4.289197561886258, + "tokens_seen": 238944256 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004684754262788365, + "loss": 1.7447, + "theoretical_loss": 4.289055372624729, + "tokens_seen": 239009792 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046846539618856573, + "loss": 1.6909, + "theoretical_loss": 4.288913233259092, + "tokens_seen": 239075328 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004684553660982949, + "loss": 1.652, + "theoretical_loss": 4.28877114375817, + "tokens_seen": 239140864 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 95545, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.81060791015625, + "objective/train/theoretical_loss": 4.288629104090811, + "objective/train/tokens_used": 239206400, + "theoretical_loss": 4.288629104090811, + "tokens_seen": 239206400 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004684453360080241, + "loss": 1.8308, + "theoretical_loss": 4.288629104090811, + "tokens_seen": 239206400 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046843530591775327, + "loss": 1.7265, + "theoretical_loss": 4.28848711422589, + "tokens_seen": 239271936 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046842527582748245, + "loss": 1.8166, + "theoretical_loss": 4.288345174132313, + "tokens_seen": 239337472 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004684152457372117, + "loss": 1.7244, + "theoretical_loss": 4.288203283779012, + "tokens_seen": 239403008 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004684052156469408, + "loss": 1.6195, + "theoretical_loss": 4.288061443134947, + "tokens_seen": 239468544 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046839518555667005, + "loss": 1.7998, + "theoretical_loss": 4.287919652169107, + "tokens_seen": 239534080 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046838515546639923, + "loss": 1.7595, + "theoretical_loss": 4.287777910850507, + "tokens_seen": 239599616 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004683751253761284, + "loss": 1.8619, + "theoretical_loss": 4.287636219148191, + "tokens_seen": 239665152 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004683650952858576, + "loss": 1.6978, + "theoretical_loss": 4.287494577031231, + "tokens_seen": 239730688 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004683550651955868, + "loss": 1.7168, + "theoretical_loss": 4.287352984468727, + "tokens_seen": 239796224 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046834503510531595, + "loss": 1.8227, + "theoretical_loss": 4.287211441429805, + "tokens_seen": 239861760 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004683350050150452, + "loss": 1.7731, + "theoretical_loss": 4.2870699478836185, + "tokens_seen": 239927296 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004683249749247743, + "loss": 1.7122, + "theoretical_loss": 4.286928503799352, + "tokens_seen": 239992832 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046831494483450355, + "loss": 1.8176, + "theoretical_loss": 4.286787109146213, + "tokens_seen": 240058368 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004683049147442327, + "loss": 1.6242, + "theoretical_loss": 4.286645763893442, + "tokens_seen": 240123904 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004682948846539619, + "loss": 1.6575, + "theoretical_loss": 4.2865044680103015, + "tokens_seen": 240189440 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004682848545636911, + "loss": 1.5961, + "theoretical_loss": 4.286363221466084, + "tokens_seen": 240254976 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004682748244734203, + "loss": 1.7673, + "theoretical_loss": 4.286222024230111, + "tokens_seen": 240320512 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046826479438314946, + "loss": 1.606, + "theoretical_loss": 4.286080876271729, + "tokens_seen": 240386048 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046825476429287864, + "loss": 1.7394, + "theoretical_loss": 4.285939777560312, + "tokens_seen": 240451584 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004682447342026078, + "loss": 1.7709, + "theoretical_loss": 4.285798728065264, + "tokens_seen": 240517120 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046823470411233706, + "loss": 1.8098, + "theoretical_loss": 4.285657727756012, + "tokens_seen": 240582656 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004682246740220662, + "loss": 1.6741, + "theoretical_loss": 4.285516776602016, + "tokens_seen": 240648192 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004682146439317954, + "loss": 1.678, + "theoretical_loss": 4.2853758745727575, + "tokens_seen": 240713728 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004682046138415246, + "loss": 1.7293, + "theoretical_loss": 4.285235021637748, + "tokens_seen": 240779264 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004681945837512538, + "loss": 1.7785, + "theoretical_loss": 4.285094217766527, + "tokens_seen": 240844800 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046818455366098296, + "loss": 1.7735, + "theoretical_loss": 4.2849534629286605, + "tokens_seen": 240910336 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046817452357071214, + "loss": 1.7333, + "theoretical_loss": 4.284812757093741, + "tokens_seen": 240975872 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004681644934804413, + "loss": 1.6552, + "theoretical_loss": 4.284672100231388, + "tokens_seen": 241041408 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046815446339017056, + "loss": 1.6866, + "theoretical_loss": 4.284531492311248, + "tokens_seen": 241106944 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004681444332998997, + "loss": 1.6913, + "theoretical_loss": 4.284390933302997, + "tokens_seen": 241172480 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004681344032096289, + "loss": 1.7576, + "theoretical_loss": 4.284250423176335, + "tokens_seen": 241238016 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046812437311935805, + "loss": 1.693, + "theoretical_loss": 4.28410996190099, + "tokens_seen": 241303552 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004681143430290873, + "loss": 1.5822, + "theoretical_loss": 4.283969549446717, + "tokens_seen": 241369088 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046810431293881646, + "loss": 1.6514, + "theoretical_loss": 4.2838291857833, + "tokens_seen": 241434624 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046809428284854565, + "loss": 1.7562, + "theoretical_loss": 4.283688870880545, + "tokens_seen": 241500160 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004680842527582748, + "loss": 1.7997, + "theoretical_loss": 4.28354860470829, + "tokens_seen": 241565696 + }, + { + "epoch": 0.07, + "learning_rate": 0.000468074222668004, + "loss": 1.7165, + "theoretical_loss": 4.283408387236397, + "tokens_seen": 241631232 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004680641925777332, + "loss": 1.624, + "theoretical_loss": 4.283268218434755, + "tokens_seen": 241696768 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004680541624874624, + "loss": 1.8737, + "theoretical_loss": 4.283128098273281, + "tokens_seen": 241762304 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046804413239719155, + "loss": 1.9072, + "theoretical_loss": 4.282988026721917, + "tokens_seen": 241827840 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004680341023069208, + "loss": 1.6305, + "theoretical_loss": 4.282848003750634, + "tokens_seen": 241893376 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046802407221664997, + "loss": 1.5579, + "theoretical_loss": 4.282708029329427, + "tokens_seen": 241958912 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046801404212637915, + "loss": 1.7278, + "theoretical_loss": 4.2825681034283205, + "tokens_seen": 242024448 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046800401203610833, + "loss": 1.6548, + "theoretical_loss": 4.2824282260173625, + "tokens_seen": 242089984 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004679939819458375, + "loss": 1.7205, + "theoretical_loss": 4.282288397066631, + "tokens_seen": 242155520 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004679839518555667, + "loss": 1.6879, + "theoretical_loss": 4.282148616546227, + "tokens_seen": 242221056 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046797392176529593, + "loss": 1.6664, + "theoretical_loss": 4.282008884426281, + "tokens_seen": 242286592 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046796389167502505, + "loss": 1.7874, + "theoretical_loss": 4.281869200676948, + "tokens_seen": 242352128 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004679538615847543, + "loss": 1.7613, + "theoretical_loss": 4.281729565268412, + "tokens_seen": 242417664 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 97293, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8152554035186768, + "objective/train/theoretical_loss": 4.28158997817088, + "objective/train/tokens_used": 242483200, + "theoretical_loss": 4.28158997817088, + "tokens_seen": 242483200 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004679438314944834, + "loss": 1.7366, + "theoretical_loss": 4.28158997817088, + "tokens_seen": 242483200 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046793380140421265, + "loss": 1.6954, + "theoretical_loss": 4.281450439354587, + "tokens_seen": 242548736 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046792377131394183, + "loss": 1.7155, + "theoretical_loss": 4.281310948789796, + "tokens_seen": 242614272 + }, + { + "epoch": 0.07, + "learning_rate": 0.000467913741223671, + "loss": 1.6753, + "theoretical_loss": 4.2811715064467935, + "tokens_seen": 242679808 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004679037111334002, + "loss": 1.6617, + "theoretical_loss": 4.281032112295894, + "tokens_seen": 242745344 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046789368104312943, + "loss": 1.7092, + "theoretical_loss": 4.2808927663074385, + "tokens_seen": 242810880 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046788365095285856, + "loss": 1.7408, + "theoretical_loss": 4.280753468451793, + "tokens_seen": 242876416 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004678736208625878, + "loss": 1.6914, + "theoretical_loss": 4.28061421869935, + "tokens_seen": 242941952 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004678635907723169, + "loss": 1.7117, + "theoretical_loss": 4.28047501702053, + "tokens_seen": 243007488 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046785356068204616, + "loss": 1.7334, + "theoretical_loss": 4.280335863385778, + "tokens_seen": 243073024 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046784353059177534, + "loss": 1.7076, + "theoretical_loss": 4.280196757765564, + "tokens_seen": 243138560 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004678335005015045, + "loss": 1.7734, + "theoretical_loss": 4.280057700130388, + "tokens_seen": 243204096 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004678234704112337, + "loss": 1.6348, + "theoretical_loss": 4.279918690450771, + "tokens_seen": 243269632 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004678134403209629, + "loss": 1.6184, + "theoretical_loss": 4.279779728697263, + "tokens_seen": 243335168 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046780341023069206, + "loss": 1.7207, + "theoretical_loss": 4.279640814840441, + "tokens_seen": 243400704 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004677933801404213, + "loss": 1.8434, + "theoretical_loss": 4.279501948850905, + "tokens_seen": 243466240 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004677833500501504, + "loss": 1.7237, + "theoretical_loss": 4.279363130699284, + "tokens_seen": 243531776 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046777331995987966, + "loss": 1.8167, + "theoretical_loss": 4.279224360356231, + "tokens_seen": 243597312 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004677632898696088, + "loss": 1.6899, + "theoretical_loss": 4.279085637792425, + "tokens_seen": 243662848 + }, + { + "epoch": 0.07, + "learning_rate": 0.000467753259779338, + "loss": 1.7415, + "theoretical_loss": 4.278946962978573, + "tokens_seen": 243728384 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004677432296890672, + "loss": 1.6914, + "theoretical_loss": 4.278808335885403, + "tokens_seen": 243793920 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004677331995987964, + "loss": 1.8687, + "theoretical_loss": 4.278669756483674, + "tokens_seen": 243859456 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046772316950852556, + "loss": 1.7547, + "theoretical_loss": 4.278531224744167, + "tokens_seen": 243924992 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004677131394182548, + "loss": 1.6504, + "theoretical_loss": 4.278392740637694, + "tokens_seen": 243990528 + }, + { + "epoch": 0.07, + "learning_rate": 0.000467703109327984, + "loss": 1.6931, + "theoretical_loss": 4.278254304135085, + "tokens_seen": 244056064 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046769307923771316, + "loss": 1.6579, + "theoretical_loss": 4.2781159152072, + "tokens_seen": 244121600 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046768304914744234, + "loss": 1.7124, + "theoretical_loss": 4.277977573824928, + "tokens_seen": 244187136 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004676730190571715, + "loss": 1.6795, + "theoretical_loss": 4.277839279959177, + "tokens_seen": 244252672 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046766298896690076, + "loss": 1.6029, + "theoretical_loss": 4.277701033580884, + "tokens_seen": 244318208 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004676529588766299, + "loss": 1.7564, + "theoretical_loss": 4.277562834661013, + "tokens_seen": 244383744 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004676429287863591, + "loss": 1.709, + "theoretical_loss": 4.277424683170549, + "tokens_seen": 244449280 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046763289869608825, + "loss": 1.7488, + "theoretical_loss": 4.277286579080508, + "tokens_seen": 244514816 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004676228686058175, + "loss": 1.6525, + "theoretical_loss": 4.277148522361927, + "tokens_seen": 244580352 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046761283851554666, + "loss": 1.7078, + "theoretical_loss": 4.277010512985871, + "tokens_seen": 244645888 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046760280842527585, + "loss": 1.7094, + "theoretical_loss": 4.2768725509234296, + "tokens_seen": 244711424 + }, + { + "epoch": 0.07, + "learning_rate": 0.000467592778335005, + "loss": 1.673, + "theoretical_loss": 4.276734636145717, + "tokens_seen": 244776960 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004675827482447342, + "loss": 1.6325, + "theoretical_loss": 4.276596768623875, + "tokens_seen": 244842496 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004675727181544634, + "loss": 1.725, + "theoretical_loss": 4.276458948329068, + "tokens_seen": 244908032 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004675626880641926, + "loss": 1.6884, + "theoretical_loss": 4.276321175232487, + "tokens_seen": 244973568 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046755265797392175, + "loss": 1.736, + "theoretical_loss": 4.276183449305348, + "tokens_seen": 245039104 + }, + { + "epoch": 0.07, + "learning_rate": 0.000467542627883651, + "loss": 1.7105, + "theoretical_loss": 4.276045770518895, + "tokens_seen": 245104640 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046753259779338017, + "loss": 1.6535, + "theoretical_loss": 4.275908138844392, + "tokens_seen": 245170176 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046752256770310935, + "loss": 1.7219, + "theoretical_loss": 4.275770554253132, + "tokens_seen": 245235712 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046751253761283853, + "loss": 1.6575, + "theoretical_loss": 4.275633016716432, + "tokens_seen": 245301248 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004675025075225677, + "loss": 1.7139, + "theoretical_loss": 4.275495526205635, + "tokens_seen": 245366784 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004674924774322969, + "loss": 1.7327, + "theoretical_loss": 4.275358082692106, + "tokens_seen": 245432320 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046748244734202613, + "loss": 1.6401, + "theoretical_loss": 4.275220686147241, + "tokens_seen": 245497856 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046747241725175525, + "loss": 1.8064, + "theoretical_loss": 4.275083336542455, + "tokens_seen": 245563392 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004674623871614845, + "loss": 1.7259, + "theoretical_loss": 4.274946033849189, + "tokens_seen": 245628928 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004674523570712136, + "loss": 1.6354, + "theoretical_loss": 4.2748087780389135, + "tokens_seen": 245694464 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 98436, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.710029125213623, + "objective/train/theoretical_loss": 4.27467156908312, + "objective/train/tokens_used": 245760000, + "theoretical_loss": 4.27467156908312, + "tokens_seen": 245760000 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046744232698094285, + "loss": 1.7835, + "theoretical_loss": 4.27467156908312, + "tokens_seen": 245760000 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046743229689067203, + "loss": 1.6884, + "theoretical_loss": 4.274534406953325, + "tokens_seen": 245825536 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004674222668004012, + "loss": 1.7106, + "theoretical_loss": 4.274397291621073, + "tokens_seen": 245891072 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004674122367101304, + "loss": 1.7619, + "theoretical_loss": 4.274260223057928, + "tokens_seen": 245956608 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046740220661985963, + "loss": 1.7592, + "theoretical_loss": 4.274123201235483, + "tokens_seen": 246022144 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046739217652958876, + "loss": 1.7344, + "theoretical_loss": 4.273986226125356, + "tokens_seen": 246087680 + }, + { + "epoch": 0.07, + "learning_rate": 0.000467382146439318, + "loss": 1.7139, + "theoretical_loss": 4.273849297699186, + "tokens_seen": 246153216 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004673721163490471, + "loss": 1.6644, + "theoretical_loss": 4.273712415928643, + "tokens_seen": 246218752 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046736208625877636, + "loss": 1.7082, + "theoretical_loss": 4.273575580785415, + "tokens_seen": 246284288 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046735205616850554, + "loss": 1.742, + "theoretical_loss": 4.27343879224122, + "tokens_seen": 246349824 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004673420260782347, + "loss": 1.7646, + "theoretical_loss": 4.273302050267796, + "tokens_seen": 246415360 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004673319959879639, + "loss": 1.7898, + "theoretical_loss": 4.27316535483691, + "tokens_seen": 246480896 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004673219658976931, + "loss": 1.652, + "theoretical_loss": 4.273028705920351, + "tokens_seen": 246546432 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046731193580742226, + "loss": 1.7322, + "theoretical_loss": 4.272892103489934, + "tokens_seen": 246611968 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004673019057171515, + "loss": 1.8962, + "theoretical_loss": 4.272755547517495, + "tokens_seen": 246677504 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004672918756268806, + "loss": 1.7024, + "theoretical_loss": 4.272619037974902, + "tokens_seen": 246743040 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046728184553660986, + "loss": 1.7445, + "theoretical_loss": 4.272482574834039, + "tokens_seen": 246808576 + }, + { + "epoch": 0.07, + "learning_rate": 0.000467271815446339, + "loss": 1.6916, + "theoretical_loss": 4.272346158066821, + "tokens_seen": 246874112 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004672617853560682, + "loss": 1.7167, + "theoretical_loss": 4.272209787645183, + "tokens_seen": 246939648 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004672517552657974, + "loss": 1.7275, + "theoretical_loss": 4.272073463541087, + "tokens_seen": 247005184 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004672417251755266, + "loss": 1.6625, + "theoretical_loss": 4.271937185726519, + "tokens_seen": 247070720 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046723169508525576, + "loss": 1.6707, + "theoretical_loss": 4.2718009541734885, + "tokens_seen": 247136256 + }, + { + "epoch": 0.07, + "learning_rate": 0.000467221664994985, + "loss": 1.699, + "theoretical_loss": 4.27166476885403, + "tokens_seen": 247201792 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004672116349047141, + "loss": 1.8059, + "theoretical_loss": 4.271528629740201, + "tokens_seen": 247267328 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046720160481444336, + "loss": 1.8197, + "theoretical_loss": 4.271392536804088, + "tokens_seen": 247332864 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004671915747241725, + "loss": 1.6823, + "theoretical_loss": 4.2712564900177945, + "tokens_seen": 247398400 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004671815446339017, + "loss": 1.852, + "theoretical_loss": 4.271120489353454, + "tokens_seen": 247463936 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004671715145436309, + "loss": 1.67, + "theoretical_loss": 4.270984534783221, + "tokens_seen": 247529472 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004671614844533601, + "loss": 1.7299, + "theoretical_loss": 4.270848626279275, + "tokens_seen": 247595008 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046715145436308927, + "loss": 1.6347, + "theoretical_loss": 4.270712763813822, + "tokens_seen": 247660544 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046714142427281845, + "loss": 1.7785, + "theoretical_loss": 4.270576947359088, + "tokens_seen": 247726080 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046713139418254763, + "loss": 1.6973, + "theoretical_loss": 4.270441176887327, + "tokens_seen": 247791616 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046712136409227686, + "loss": 1.7061, + "theoretical_loss": 4.270305452370814, + "tokens_seen": 247857152 + }, + { + "epoch": 0.08, + "learning_rate": 0.000467111334002006, + "loss": 1.5564, + "theoretical_loss": 4.27016977378185, + "tokens_seen": 247922688 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004671013039117352, + "loss": 1.7128, + "theoretical_loss": 4.270034141092759, + "tokens_seen": 247988224 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046709127382146435, + "loss": 1.68, + "theoretical_loss": 4.269898554275889, + "tokens_seen": 248053760 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004670812437311936, + "loss": 1.7211, + "theoretical_loss": 4.269763013303613, + "tokens_seen": 248119296 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046707121364092277, + "loss": 1.7393, + "theoretical_loss": 4.269627518148326, + "tokens_seen": 248184832 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046706118355065195, + "loss": 1.7064, + "theoretical_loss": 4.26949206878245, + "tokens_seen": 248250368 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046705115346038113, + "loss": 1.6622, + "theoretical_loss": 4.269356665178426, + "tokens_seen": 248315904 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046704112337011037, + "loss": 1.7251, + "theoretical_loss": 4.269221307308725, + "tokens_seen": 248381440 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004670310932798395, + "loss": 1.586, + "theoretical_loss": 4.269085995145837, + "tokens_seen": 248446976 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046702106318956873, + "loss": 1.7202, + "theoretical_loss": 4.268950728662278, + "tokens_seen": 248512512 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046701103309929786, + "loss": 1.6685, + "theoretical_loss": 4.268815507830587, + "tokens_seen": 248578048 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004670010030090271, + "loss": 1.7355, + "theoretical_loss": 4.268680332623328, + "tokens_seen": 248643584 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004669909729187563, + "loss": 1.7107, + "theoretical_loss": 4.268545203013085, + "tokens_seen": 248709120 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046698094282848545, + "loss": 1.7682, + "theoretical_loss": 4.26841011897247, + "tokens_seen": 248774656 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046697091273821464, + "loss": 1.6871, + "theoretical_loss": 4.268275080474118, + "tokens_seen": 248840192 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004669608826479438, + "loss": 1.753, + "theoretical_loss": 4.268140087490686, + "tokens_seen": 248905728 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046695085255767305, + "loss": 1.7233, + "theoretical_loss": 4.2680051399948535, + "tokens_seen": 248971264 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 99585, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6530349254608154, + "objective/train/theoretical_loss": 4.267870237959327, + "objective/train/tokens_used": 249036800, + "theoretical_loss": 4.267870237959327, + "tokens_seen": 249036800 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046694082246740223, + "loss": 1.5918, + "theoretical_loss": 4.267870237959327, + "tokens_seen": 249036800 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004669307923771314, + "loss": 1.8316, + "theoretical_loss": 4.267735381356835, + "tokens_seen": 249102336 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004669207622868606, + "loss": 1.8411, + "theoretical_loss": 4.26760057016013, + "tokens_seen": 249167872 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046691073219658983, + "loss": 1.7709, + "theoretical_loss": 4.2674658043419855, + "tokens_seen": 249233408 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046690070210631896, + "loss": 1.7546, + "theoretical_loss": 4.267331083875202, + "tokens_seen": 249298944 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004668906720160482, + "loss": 1.6948, + "theoretical_loss": 4.267196408732601, + "tokens_seen": 249364480 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004668806419257773, + "loss": 1.8047, + "theoretical_loss": 4.267061778887029, + "tokens_seen": 249430016 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046687061183550656, + "loss": 1.6734, + "theoretical_loss": 4.2669271943113545, + "tokens_seen": 249495552 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046686058174523574, + "loss": 1.6921, + "theoretical_loss": 4.26679265497847, + "tokens_seen": 249561088 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004668505516549649, + "loss": 1.5625, + "theoretical_loss": 4.266658160861293, + "tokens_seen": 249626624 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004668405215646941, + "loss": 1.7462, + "theoretical_loss": 4.266523711932761, + "tokens_seen": 249692160 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004668304914744233, + "loss": 1.6447, + "theoretical_loss": 4.266389308165837, + "tokens_seen": 249757696 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046682046138415246, + "loss": 1.6386, + "theoretical_loss": 4.266254949533507, + "tokens_seen": 249823232 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004668104312938817, + "loss": 1.6848, + "theoretical_loss": 4.26612063600878, + "tokens_seen": 249888768 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004668004012036108, + "loss": 1.6807, + "theoretical_loss": 4.265986367564688, + "tokens_seen": 249954304 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046679037111334006, + "loss": 1.6282, + "theoretical_loss": 4.265852144174286, + "tokens_seen": 250019840 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004667803410230692, + "loss": 1.7594, + "theoretical_loss": 4.265717965810653, + "tokens_seen": 250085376 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004667703109327984, + "loss": 1.6308, + "theoretical_loss": 4.265583832446892, + "tokens_seen": 250150912 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004667602808425276, + "loss": 1.7563, + "theoretical_loss": 4.265449744056126, + "tokens_seen": 250216448 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004667502507522568, + "loss": 1.7011, + "theoretical_loss": 4.265315700611504, + "tokens_seen": 250281984 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046674022066198596, + "loss": 1.7566, + "theoretical_loss": 4.2651817020861955, + "tokens_seen": 250347520 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004667301905717152, + "loss": 1.671, + "theoretical_loss": 4.265047748453396, + "tokens_seen": 250413056 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004667201604814443, + "loss": 1.7042, + "theoretical_loss": 4.264913839686323, + "tokens_seen": 250478592 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046671013039117356, + "loss": 1.7407, + "theoretical_loss": 4.264779975758215, + "tokens_seen": 250544128 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004667001003009027, + "loss": 1.8194, + "theoretical_loss": 4.2646461566423355, + "tokens_seen": 250609664 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004666900702106319, + "loss": 1.7065, + "theoretical_loss": 4.26451238231197, + "tokens_seen": 250675200 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004666800401203611, + "loss": 1.673, + "theoretical_loss": 4.2643786527404295, + "tokens_seen": 250740736 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004666700100300903, + "loss": 1.6372, + "theoretical_loss": 4.264244967901043, + "tokens_seen": 250806272 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046665997993981947, + "loss": 1.7074, + "theoretical_loss": 4.264111327767164, + "tokens_seen": 250871808 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046664994984954865, + "loss": 1.758, + "theoretical_loss": 4.263977732312174, + "tokens_seen": 250937344 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046663991975927783, + "loss": 1.6629, + "theoretical_loss": 4.26384418150947, + "tokens_seen": 251002880 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046662988966900706, + "loss": 1.6581, + "theoretical_loss": 4.263710675332476, + "tokens_seen": 251068416 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004666198595787362, + "loss": 1.6544, + "theoretical_loss": 4.263577213754638, + "tokens_seen": 251133952 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046660982948846543, + "loss": 1.6197, + "theoretical_loss": 4.2634437967494225, + "tokens_seen": 251199488 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046659979939819455, + "loss": 1.8185, + "theoretical_loss": 4.2633104242903235, + "tokens_seen": 251265024 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004665897693079238, + "loss": 1.737, + "theoretical_loss": 4.263177096350852, + "tokens_seen": 251330560 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046657973921765297, + "loss": 1.6505, + "theoretical_loss": 4.263043812904547, + "tokens_seen": 251396096 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046656970912738215, + "loss": 1.75, + "theoretical_loss": 4.262910573924965, + "tokens_seen": 251461632 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046655967903711133, + "loss": 1.589, + "theoretical_loss": 4.26277737938569, + "tokens_seen": 251527168 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046654964894684057, + "loss": 1.6738, + "theoretical_loss": 4.262644229260324, + "tokens_seen": 251592704 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004665396188565697, + "loss": 1.6935, + "theoretical_loss": 4.262511123522495, + "tokens_seen": 251658240 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046652958876629893, + "loss": 1.6689, + "theoretical_loss": 4.262378062145853, + "tokens_seen": 251723776 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046651955867602806, + "loss": 1.6685, + "theoretical_loss": 4.262245045104068, + "tokens_seen": 251789312 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004665095285857573, + "loss": 1.7451, + "theoretical_loss": 4.262112072370837, + "tokens_seen": 251854848 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004664994984954865, + "loss": 1.7476, + "theoretical_loss": 4.261979143919873, + "tokens_seen": 251920384 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046648946840521565, + "loss": 1.6644, + "theoretical_loss": 4.261846259724917, + "tokens_seen": 251985920 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046647943831494484, + "loss": 1.7374, + "theoretical_loss": 4.261713419759731, + "tokens_seen": 252051456 + }, + { + "epoch": 0.08, + "learning_rate": 0.000466469408224674, + "loss": 1.7318, + "theoretical_loss": 4.2615806239981, + "tokens_seen": 252116992 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004664593781344032, + "loss": 1.7289, + "theoretical_loss": 4.261447872413827, + "tokens_seen": 252182528 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046644934804413243, + "loss": 1.7206, + "theoretical_loss": 4.261315164980743, + "tokens_seen": 252248064 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 100742, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6555662155151367, + "objective/train/theoretical_loss": 4.261182501672698, + "objective/train/tokens_used": 252313600, + "theoretical_loss": 4.261182501672698, + "tokens_seen": 252313600 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046643931795386156, + "loss": 1.6862, + "theoretical_loss": 4.261182501672698, + "tokens_seen": 252313600 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004664292878635908, + "loss": 1.7439, + "theoretical_loss": 4.261049882463565, + "tokens_seen": 252379136 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004664192577733199, + "loss": 1.7267, + "theoretical_loss": 4.26091730732724, + "tokens_seen": 252444672 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046640922768304916, + "loss": 1.7576, + "theoretical_loss": 4.2607847762376405, + "tokens_seen": 252510208 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046639919759277834, + "loss": 1.6226, + "theoretical_loss": 4.260652289168706, + "tokens_seen": 252575744 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004663891675025075, + "loss": 1.6639, + "theoretical_loss": 4.260519846094399, + "tokens_seen": 252641280 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004663791374122367, + "loss": 1.6794, + "theoretical_loss": 4.2603874469887035, + "tokens_seen": 252706816 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046636910732196594, + "loss": 1.6861, + "theoretical_loss": 4.260255091825625, + "tokens_seen": 252772352 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046635907723169506, + "loss": 1.5753, + "theoretical_loss": 4.260122780579193, + "tokens_seen": 252837888 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004663490471414243, + "loss": 1.6234, + "theoretical_loss": 4.259990513223458, + "tokens_seen": 252903424 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004663390170511534, + "loss": 1.7423, + "theoretical_loss": 4.259858289732491, + "tokens_seen": 252968960 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046632898696088266, + "loss": 1.6893, + "theoretical_loss": 4.259726110080388, + "tokens_seen": 253034496 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046631895687061184, + "loss": 1.8106, + "theoretical_loss": 4.259593974241264, + "tokens_seen": 253100032 + }, + { + "epoch": 0.08, + "learning_rate": 0.000466308926780341, + "loss": 1.6929, + "theoretical_loss": 4.25946188218926, + "tokens_seen": 253165568 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004662988966900702, + "loss": 1.7523, + "theoretical_loss": 4.259329833898534, + "tokens_seen": 253231104 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004662888665997994, + "loss": 1.8477, + "theoretical_loss": 4.25919782934327, + "tokens_seen": 253296640 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046627883650952857, + "loss": 1.7539, + "theoretical_loss": 4.259065868497672, + "tokens_seen": 253362176 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004662688064192578, + "loss": 1.6898, + "theoretical_loss": 4.258933951335965, + "tokens_seen": 253427712 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046625877632898693, + "loss": 1.6771, + "theoretical_loss": 4.258802077832399, + "tokens_seen": 253493248 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046624874623871616, + "loss": 1.6106, + "theoretical_loss": 4.258670247961243, + "tokens_seen": 253558784 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046623871614844535, + "loss": 1.7893, + "theoretical_loss": 4.258538461696787, + "tokens_seen": 253624320 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004662286860581745, + "loss": 1.5957, + "theoretical_loss": 4.258406719013347, + "tokens_seen": 253689856 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004662186559679037, + "loss": 1.7369, + "theoretical_loss": 4.258275019885257, + "tokens_seen": 253755392 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004662086258776329, + "loss": 1.7892, + "theoretical_loss": 4.258143364286875, + "tokens_seen": 253820928 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004661985957873621, + "loss": 1.6849, + "theoretical_loss": 4.258011752192579, + "tokens_seen": 253886464 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004661885656970913, + "loss": 1.6448, + "theoretical_loss": 4.257880183576768, + "tokens_seen": 253952000 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004661785356068205, + "loss": 1.5739, + "theoretical_loss": 4.257748658413867, + "tokens_seen": 254017536 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046616850551654967, + "loss": 1.6557, + "theoretical_loss": 4.257617176678317, + "tokens_seen": 254083072 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046615847542627885, + "loss": 1.685, + "theoretical_loss": 4.2574857383445845, + "tokens_seen": 254148608 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046614844533600803, + "loss": 1.7413, + "theoretical_loss": 4.257354343387156, + "tokens_seen": 254214144 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046613841524573727, + "loss": 1.6793, + "theoretical_loss": 4.257222991780541, + "tokens_seen": 254279680 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004661283851554664, + "loss": 1.6938, + "theoretical_loss": 4.257091683499268, + "tokens_seen": 254345216 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046611835506519563, + "loss": 1.6747, + "theoretical_loss": 4.25696041851789, + "tokens_seen": 254410752 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046610832497492475, + "loss": 1.6786, + "theoretical_loss": 4.256829196810978, + "tokens_seen": 254476288 + }, + { + "epoch": 0.08, + "learning_rate": 0.000466098294884654, + "loss": 1.7797, + "theoretical_loss": 4.256698018353129, + "tokens_seen": 254541824 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046608826479438317, + "loss": 1.7249, + "theoretical_loss": 4.256566883118958, + "tokens_seen": 254607360 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046607823470411235, + "loss": 1.7657, + "theoretical_loss": 4.256435791083101, + "tokens_seen": 254672896 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046606820461384153, + "loss": 1.7616, + "theoretical_loss": 4.256304742220219, + "tokens_seen": 254738432 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046605817452357077, + "loss": 1.6633, + "theoretical_loss": 4.25617373650499, + "tokens_seen": 254803968 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004660481444332999, + "loss": 1.7892, + "theoretical_loss": 4.256042773912117, + "tokens_seen": 254869504 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046603811434302913, + "loss": 1.7192, + "theoretical_loss": 4.255911854416323, + "tokens_seen": 254935040 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046602808425275826, + "loss": 1.5271, + "theoretical_loss": 4.255780977992352, + "tokens_seen": 255000576 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004660180541624875, + "loss": 1.691, + "theoretical_loss": 4.255650144614968, + "tokens_seen": 255066112 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004660080240722167, + "loss": 1.6403, + "theoretical_loss": 4.255519354258959, + "tokens_seen": 255131648 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046599799398194586, + "loss": 1.6681, + "theoretical_loss": 4.255388606899133, + "tokens_seen": 255197184 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046598796389167504, + "loss": 1.5944, + "theoretical_loss": 4.255257902510319, + "tokens_seen": 255262720 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004659779338014042, + "loss": 1.7333, + "theoretical_loss": 4.255127241067368, + "tokens_seen": 255328256 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004659679037111334, + "loss": 1.6125, + "theoretical_loss": 4.25499662254515, + "tokens_seen": 255393792 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046595787362086263, + "loss": 1.61, + "theoretical_loss": 4.254866046918559, + "tokens_seen": 255459328 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046594784353059176, + "loss": 1.7383, + "theoretical_loss": 4.254735514162508, + "tokens_seen": 255524864 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 101923, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9140307903289795, + "objective/train/theoretical_loss": 4.254605024251933, + "objective/train/tokens_used": 255590400, + "theoretical_loss": 4.254605024251933, + "tokens_seen": 255590400 + }, + { + "epoch": 0.08, + "learning_rate": 0.000465937813440321, + "loss": 1.8056, + "theoretical_loss": 4.254605024251933, + "tokens_seen": 255590400 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004659277833500501, + "loss": 1.7362, + "theoretical_loss": 4.254474577161789, + "tokens_seen": 255655936 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046591775325977936, + "loss": 1.7036, + "theoretical_loss": 4.254344172867053, + "tokens_seen": 255721472 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046590772316950854, + "loss": 1.683, + "theoretical_loss": 4.254213811342723, + "tokens_seen": 255787008 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004658976930792377, + "loss": 1.7829, + "theoretical_loss": 4.254083492563819, + "tokens_seen": 255852544 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004658876629889669, + "loss": 1.6992, + "theoretical_loss": 4.253953216505381, + "tokens_seen": 255918080 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046587763289869614, + "loss": 1.6679, + "theoretical_loss": 4.253822983142468, + "tokens_seen": 255983616 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046586760280842526, + "loss": 1.71, + "theoretical_loss": 4.253692792450165, + "tokens_seen": 256049152 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004658575727181545, + "loss": 1.7124, + "theoretical_loss": 4.253562644403573, + "tokens_seen": 256114688 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004658475426278836, + "loss": 1.7023, + "theoretical_loss": 4.253432538977816, + "tokens_seen": 256180224 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046583751253761286, + "loss": 1.6105, + "theoretical_loss": 4.253302476148039, + "tokens_seen": 256245760 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046582748244734204, + "loss": 1.7953, + "theoretical_loss": 4.253172455889407, + "tokens_seen": 256311296 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004658174523570712, + "loss": 1.6734, + "theoretical_loss": 4.2530424781771075, + "tokens_seen": 256376832 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004658074222668004, + "loss": 1.7921, + "theoretical_loss": 4.252912542986345, + "tokens_seen": 256442368 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004657973921765296, + "loss": 1.6086, + "theoretical_loss": 4.2527826502923505, + "tokens_seen": 256507904 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046578736208625877, + "loss": 1.7322, + "theoretical_loss": 4.25265280007037, + "tokens_seen": 256573440 + }, + { + "epoch": 0.08, + "learning_rate": 0.000465777331995988, + "loss": 1.7023, + "theoretical_loss": 4.252522992295676, + "tokens_seen": 256638976 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046576730190571713, + "loss": 1.7916, + "theoretical_loss": 4.252393226943555, + "tokens_seen": 256704512 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046575727181544636, + "loss": 1.6622, + "theoretical_loss": 4.252263503989321, + "tokens_seen": 256770048 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046574724172517555, + "loss": 1.8871, + "theoretical_loss": 4.252133823408303, + "tokens_seen": 256835584 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004657372116349047, + "loss": 1.7644, + "theoretical_loss": 4.2520041851758545, + "tokens_seen": 256901120 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004657271815446339, + "loss": 1.6526, + "theoretical_loss": 4.251874589267347, + "tokens_seen": 256966656 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004657171514543631, + "loss": 1.8165, + "theoretical_loss": 4.251745035658176, + "tokens_seen": 257032192 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046570712136409227, + "loss": 1.6164, + "theoretical_loss": 4.251615524323752, + "tokens_seen": 257097728 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004656970912738215, + "loss": 1.6897, + "theoretical_loss": 4.251486055239512, + "tokens_seen": 257163264 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046568706118355063, + "loss": 1.7912, + "theoretical_loss": 4.251356628380911, + "tokens_seen": 257228800 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046567703109327987, + "loss": 1.7508, + "theoretical_loss": 4.251227243723424, + "tokens_seen": 257294336 + }, + { + "epoch": 0.08, + "learning_rate": 0.000465667001003009, + "loss": 1.7219, + "theoretical_loss": 4.251097901242546, + "tokens_seen": 257359872 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046565697091273823, + "loss": 1.675, + "theoretical_loss": 4.250968600913795, + "tokens_seen": 257425408 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004656469408224674, + "loss": 1.7091, + "theoretical_loss": 4.250839342712707, + "tokens_seen": 257490944 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004656369107321966, + "loss": 1.7201, + "theoretical_loss": 4.250710126614838, + "tokens_seen": 257556480 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004656268806419258, + "loss": 1.6855, + "theoretical_loss": 4.250580952595769, + "tokens_seen": 257622016 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046561685055165495, + "loss": 1.6148, + "theoretical_loss": 4.250451820631095, + "tokens_seen": 257687552 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046560682046138414, + "loss": 1.7517, + "theoretical_loss": 4.2503227306964355, + "tokens_seen": 257753088 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046559679037111337, + "loss": 1.7524, + "theoretical_loss": 4.25019368276743, + "tokens_seen": 257818624 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004655867602808425, + "loss": 1.8399, + "theoretical_loss": 4.250064676819735, + "tokens_seen": 257884160 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046557673019057173, + "loss": 1.7836, + "theoretical_loss": 4.249935712829033, + "tokens_seen": 257949696 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004655667001003009, + "loss": 1.6849, + "theoretical_loss": 4.249806790771022, + "tokens_seen": 258015232 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004655566700100301, + "loss": 1.6629, + "theoretical_loss": 4.249677910621422, + "tokens_seen": 258080768 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004655466399197593, + "loss": 1.792, + "theoretical_loss": 4.249549072355974, + "tokens_seen": 258146304 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046553660982948846, + "loss": 1.7649, + "theoretical_loss": 4.2494202759504365, + "tokens_seen": 258211840 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046552657973921764, + "loss": 1.6687, + "theoretical_loss": 4.249291521380591, + "tokens_seen": 258277376 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004655165496489469, + "loss": 1.7005, + "theoretical_loss": 4.249162808622237, + "tokens_seen": 258342912 + }, + { + "epoch": 0.08, + "learning_rate": 0.000465506519558676, + "loss": 1.7778, + "theoretical_loss": 4.249034137651197, + "tokens_seen": 258408448 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046549648946840524, + "loss": 1.7071, + "theoretical_loss": 4.248905508443311, + "tokens_seen": 258473984 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046548645937813436, + "loss": 1.7059, + "theoretical_loss": 4.248776920974439, + "tokens_seen": 258539520 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004654764292878636, + "loss": 1.6685, + "theoretical_loss": 4.248648375220464, + "tokens_seen": 258605056 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004654663991975928, + "loss": 1.7429, + "theoretical_loss": 4.248519871157285, + "tokens_seen": 258670592 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046545636910732196, + "loss": 1.7259, + "theoretical_loss": 4.248391408760824, + "tokens_seen": 258736128 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004654463390170512, + "loss": 1.6462, + "theoretical_loss": 4.248262988007022, + "tokens_seen": 258801664 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 103108, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7217928171157837, + "objective/train/theoretical_loss": 4.248134608871841, + "objective/train/tokens_used": 258867200, + "theoretical_loss": 4.248134608871841, + "tokens_seen": 258867200 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004654363089267803, + "loss": 1.7536, + "theoretical_loss": 4.248134608871841, + "tokens_seen": 258867200 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046542627883650956, + "loss": 1.6453, + "theoretical_loss": 4.24800627133126, + "tokens_seen": 258932736 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046541624874623874, + "loss": 1.7165, + "theoretical_loss": 4.247877975361281, + "tokens_seen": 258998272 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004654062186559679, + "loss": 1.7058, + "theoretical_loss": 4.247749720937925, + "tokens_seen": 259063808 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004653961885656971, + "loss": 1.7035, + "theoretical_loss": 4.247621508037232, + "tokens_seen": 259129344 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046538615847542634, + "loss": 1.7691, + "theoretical_loss": 4.247493336635264, + "tokens_seen": 259194880 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046537612838515546, + "loss": 1.6906, + "theoretical_loss": 4.2473652067081, + "tokens_seen": 259260416 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004653660982948847, + "loss": 1.6934, + "theoretical_loss": 4.247237118231841, + "tokens_seen": 259325952 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004653560682046138, + "loss": 1.6557, + "theoretical_loss": 4.247109071182607, + "tokens_seen": 259391488 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046534603811434306, + "loss": 1.6267, + "theoretical_loss": 4.246981065536538, + "tokens_seen": 259457024 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046533600802407224, + "loss": 1.5815, + "theoretical_loss": 4.246853101269793, + "tokens_seen": 259522560 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004653259779338014, + "loss": 1.7822, + "theoretical_loss": 4.2467251783585525, + "tokens_seen": 259588096 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004653159478435306, + "loss": 1.7833, + "theoretical_loss": 4.246597296779014, + "tokens_seen": 259653632 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004653059177532598, + "loss": 1.7397, + "theoretical_loss": 4.246469456507398, + "tokens_seen": 259719168 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046529588766298897, + "loss": 1.7402, + "theoretical_loss": 4.246341657519943, + "tokens_seen": 259784704 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004652858575727182, + "loss": 1.747, + "theoretical_loss": 4.246213899792905, + "tokens_seen": 259850240 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046527582748244733, + "loss": 1.6449, + "theoretical_loss": 4.246086183302564, + "tokens_seen": 259915776 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046526579739217656, + "loss": 1.62, + "theoretical_loss": 4.2459585080252165, + "tokens_seen": 259981312 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046525576730190575, + "loss": 1.6836, + "theoretical_loss": 4.245830873937179, + "tokens_seen": 260046848 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004652457372116349, + "loss": 1.7848, + "theoretical_loss": 4.24570328101479, + "tokens_seen": 260112384 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004652357071213641, + "loss": 1.7458, + "theoretical_loss": 4.245575729234403, + "tokens_seen": 260177920 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004652256770310933, + "loss": 1.6128, + "theoretical_loss": 4.245448218572396, + "tokens_seen": 260243456 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046521564694082247, + "loss": 1.7903, + "theoretical_loss": 4.245320749005163, + "tokens_seen": 260308992 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004652056168505517, + "loss": 1.648, + "theoretical_loss": 4.245193320509118, + "tokens_seen": 260374528 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046519558676028083, + "loss": 1.6644, + "theoretical_loss": 4.245065933060697, + "tokens_seen": 260440064 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046518555667001007, + "loss": 1.6481, + "theoretical_loss": 4.244938586636351, + "tokens_seen": 260505600 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004651755265797392, + "loss": 1.676, + "theoretical_loss": 4.244811281212556, + "tokens_seen": 260571136 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046516549648946843, + "loss": 1.75, + "theoretical_loss": 4.244684016765802, + "tokens_seen": 260636672 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004651554663991976, + "loss": 1.6776, + "theoretical_loss": 4.2445567932726025, + "tokens_seen": 260702208 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004651454363089268, + "loss": 1.7455, + "theoretical_loss": 4.244429610709488, + "tokens_seen": 260767744 + }, + { + "epoch": 0.08, + "learning_rate": 0.000465135406218656, + "loss": 1.5978, + "theoretical_loss": 4.244302469053007, + "tokens_seen": 260833280 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046512537612838515, + "loss": 1.6689, + "theoretical_loss": 4.244175368279733, + "tokens_seen": 260898816 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046511534603811434, + "loss": 1.7129, + "theoretical_loss": 4.244048308366253, + "tokens_seen": 260964352 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046510531594784357, + "loss": 1.6014, + "theoretical_loss": 4.243921289289175, + "tokens_seen": 261029888 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004650952858575727, + "loss": 1.6672, + "theoretical_loss": 4.243794311025128, + "tokens_seen": 261095424 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046508525576730193, + "loss": 1.7487, + "theoretical_loss": 4.243667373550757, + "tokens_seen": 261160960 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004650752256770311, + "loss": 1.6586, + "theoretical_loss": 4.24354047684273, + "tokens_seen": 261226496 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004650651955867603, + "loss": 1.6979, + "theoretical_loss": 4.243413620877732, + "tokens_seen": 261292032 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004650551654964895, + "loss": 1.6207, + "theoretical_loss": 4.243286805632467, + "tokens_seen": 261357568 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046504513540621866, + "loss": 1.707, + "theoretical_loss": 4.243160031083658, + "tokens_seen": 261423104 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046503510531594784, + "loss": 1.612, + "theoretical_loss": 4.24303329720805, + "tokens_seen": 261488640 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004650250752256771, + "loss": 1.7061, + "theoretical_loss": 4.242906603982402, + "tokens_seen": 261554176 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004650150451354062, + "loss": 1.7664, + "theoretical_loss": 4.242779951383497, + "tokens_seen": 261619712 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046500501504513544, + "loss": 1.6664, + "theoretical_loss": 4.242653339388135, + "tokens_seen": 261685248 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046499498495486456, + "loss": 1.7533, + "theoretical_loss": 4.242526767973134, + "tokens_seen": 261750784 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004649849548645938, + "loss": 1.6617, + "theoretical_loss": 4.242400237115333, + "tokens_seen": 261816320 + }, + { + "epoch": 0.08, + "learning_rate": 0.000464974924774323, + "loss": 1.6123, + "theoretical_loss": 4.242273746791589, + "tokens_seen": 261881856 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046496489468405216, + "loss": 1.7798, + "theoretical_loss": 4.242147296978778, + "tokens_seen": 261947392 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046495486459378134, + "loss": 1.7319, + "theoretical_loss": 4.242020887653795, + "tokens_seen": 262012928 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004649448345035105, + "loss": 1.697, + "theoretical_loss": 4.241894518793555, + "tokens_seen": 262078464 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 104233, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7379779815673828, + "objective/train/theoretical_loss": 4.24176819037499, + "objective/train/tokens_used": 262144000, + "theoretical_loss": 4.24176819037499, + "tokens_seen": 262144000 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004649348044132397, + "loss": 1.6834, + "theoretical_loss": 4.24176819037499, + "tokens_seen": 262144000 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046492477432296894, + "loss": 1.738, + "theoretical_loss": 4.2416419023750525, + "tokens_seen": 262209536 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046491474423269807, + "loss": 1.6673, + "theoretical_loss": 4.241515654770713, + "tokens_seen": 262275072 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004649047141424273, + "loss": 1.6562, + "theoretical_loss": 4.241389447538961, + "tokens_seen": 262340608 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004648946840521565, + "loss": 1.6838, + "theoretical_loss": 4.241263280656806, + "tokens_seen": 262406144 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046488465396188566, + "loss": 1.6802, + "theoretical_loss": 4.241137154101275, + "tokens_seen": 262471680 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046487462387161485, + "loss": 1.7131, + "theoretical_loss": 4.2410110678494135, + "tokens_seen": 262537216 + }, + { + "epoch": 0.08, + "learning_rate": 0.000464864593781344, + "loss": 1.6751, + "theoretical_loss": 4.240885021878286, + "tokens_seen": 262602752 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004648545636910732, + "loss": 1.6701, + "theoretical_loss": 4.240759016164978, + "tokens_seen": 262668288 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046484453360080244, + "loss": 1.7015, + "theoretical_loss": 4.240633050686592, + "tokens_seen": 262733824 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046483450351053157, + "loss": 1.6672, + "theoretical_loss": 4.2405071254202475, + "tokens_seen": 262799360 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004648244734202608, + "loss": 1.8094, + "theoretical_loss": 4.240381240343085, + "tokens_seen": 262864896 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046481444332998993, + "loss": 1.7274, + "theoretical_loss": 4.240255395432264, + "tokens_seen": 262930432 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046480441323971917, + "loss": 1.6625, + "theoretical_loss": 4.240129590664961, + "tokens_seen": 262995968 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046479438314944835, + "loss": 1.6395, + "theoretical_loss": 4.240003826018373, + "tokens_seen": 263061504 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046478435305917753, + "loss": 1.7288, + "theoretical_loss": 4.239878101469713, + "tokens_seen": 263127040 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004647743229689067, + "loss": 1.7392, + "theoretical_loss": 4.239752416996215, + "tokens_seen": 263192576 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046476429287863595, + "loss": 1.7359, + "theoretical_loss": 4.239626772575131, + "tokens_seen": 263258112 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004647542627883651, + "loss": 1.6072, + "theoretical_loss": 4.239501168183732, + "tokens_seen": 263323648 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004647442326980943, + "loss": 1.6014, + "theoretical_loss": 4.239375603799305, + "tokens_seen": 263389184 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046473420260782344, + "loss": 1.6769, + "theoretical_loss": 4.2392500793991585, + "tokens_seen": 263454720 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046472417251755267, + "loss": 1.5579, + "theoretical_loss": 4.239124594960619, + "tokens_seen": 263520256 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004647141424272819, + "loss": 1.6461, + "theoretical_loss": 4.23899915046103, + "tokens_seen": 263585792 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046470411233701103, + "loss": 1.7038, + "theoretical_loss": 4.238873745877756, + "tokens_seen": 263651328 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046469408224674027, + "loss": 1.6748, + "theoretical_loss": 4.238748381188176, + "tokens_seen": 263716864 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004646840521564694, + "loss": 1.8229, + "theoretical_loss": 4.238623056369692, + "tokens_seen": 263782400 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046467402206619863, + "loss": 1.6299, + "theoretical_loss": 4.238497771399721, + "tokens_seen": 263847936 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004646639919759278, + "loss": 1.7408, + "theoretical_loss": 4.2383725262556995, + "tokens_seen": 263913472 + }, + { + "epoch": 0.08, + "learning_rate": 0.000464653961885657, + "loss": 1.6695, + "theoretical_loss": 4.238247320915083, + "tokens_seen": 263979008 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004646439317953862, + "loss": 1.5775, + "theoretical_loss": 4.238122155355345, + "tokens_seen": 264044544 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046463390170511535, + "loss": 1.7264, + "theoretical_loss": 4.237997029553977, + "tokens_seen": 264110080 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046462387161484454, + "loss": 1.6611, + "theoretical_loss": 4.237871943488488, + "tokens_seen": 264175616 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046461384152457377, + "loss": 1.7228, + "theoretical_loss": 4.237746897136408, + "tokens_seen": 264241152 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004646038114343029, + "loss": 1.635, + "theoretical_loss": 4.237621890475284, + "tokens_seen": 264306688 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046459378134403213, + "loss": 1.7143, + "theoretical_loss": 4.237496923482678, + "tokens_seen": 264372224 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004645837512537613, + "loss": 1.69, + "theoretical_loss": 4.237371996136175, + "tokens_seen": 264437760 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004645737211634905, + "loss": 1.6853, + "theoretical_loss": 4.237247108413376, + "tokens_seen": 264503296 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004645636910732197, + "loss": 1.584, + "theoretical_loss": 4.237122260291901, + "tokens_seen": 264568832 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046455366098294886, + "loss": 1.632, + "theoretical_loss": 4.236997451749387, + "tokens_seen": 264634368 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046454363089267804, + "loss": 1.7635, + "theoretical_loss": 4.236872682763489, + "tokens_seen": 264699904 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004645336008024073, + "loss": 1.6925, + "theoretical_loss": 4.236747953311884, + "tokens_seen": 264765440 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004645235707121364, + "loss": 1.6228, + "theoretical_loss": 4.236623263372261, + "tokens_seen": 264830976 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046451354062186564, + "loss": 1.6598, + "theoretical_loss": 4.236498612922331, + "tokens_seen": 264896512 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046450351053159476, + "loss": 1.6286, + "theoretical_loss": 4.236374001939822, + "tokens_seen": 264962048 + }, + { + "epoch": 0.08, + "learning_rate": 0.000464493480441324, + "loss": 1.6654, + "theoretical_loss": 4.2362494304024825, + "tokens_seen": 265027584 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004644834503510532, + "loss": 1.7653, + "theoretical_loss": 4.236124898288074, + "tokens_seen": 265093120 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046447342026078236, + "loss": 1.7093, + "theoretical_loss": 4.236000405574382, + "tokens_seen": 265158656 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046446339017051154, + "loss": 1.7484, + "theoretical_loss": 4.235875952239204, + "tokens_seen": 265224192 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004644533600802407, + "loss": 1.6917, + "theoretical_loss": 4.235751538260359, + "tokens_seen": 265289728 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004644433299899699, + "loss": 1.705, + "theoretical_loss": 4.2356271636156855, + "tokens_seen": 265355264 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 105380, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6966397762298584, + "objective/train/theoretical_loss": 4.235502828283035, + "objective/train/tokens_used": 265420800, + "theoretical_loss": 4.235502828283035, + "tokens_seen": 265420800 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046443329989969914, + "loss": 1.7159, + "theoretical_loss": 4.235502828283035, + "tokens_seen": 265420800 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046442326980942827, + "loss": 1.8085, + "theoretical_loss": 4.235378532240283, + "tokens_seen": 265486336 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004644132397191575, + "loss": 1.7347, + "theoretical_loss": 4.235254275465317, + "tokens_seen": 265551872 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004644032096288867, + "loss": 1.6261, + "theoretical_loss": 4.235130057936045, + "tokens_seen": 265617408 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046439317953861586, + "loss": 1.6006, + "theoretical_loss": 4.235005879630395, + "tokens_seen": 265682944 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046438314944834505, + "loss": 1.6933, + "theoretical_loss": 4.23488174052631, + "tokens_seen": 265748480 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004643731193580742, + "loss": 1.6639, + "theoretical_loss": 4.234757640601751, + "tokens_seen": 265814016 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004643630892678034, + "loss": 1.5759, + "theoretical_loss": 4.234633579834698, + "tokens_seen": 265879552 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046435305917753264, + "loss": 1.7019, + "theoretical_loss": 4.234509558203149, + "tokens_seen": 265945088 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046434302908726177, + "loss": 1.7063, + "theoretical_loss": 4.234385575685116, + "tokens_seen": 266010624 + }, + { + "epoch": 0.08, + "learning_rate": 0.000464332998996991, + "loss": 1.7586, + "theoretical_loss": 4.234261632258636, + "tokens_seen": 266076160 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046432296890672013, + "loss": 1.7552, + "theoretical_loss": 4.234137727901757, + "tokens_seen": 266141696 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046431293881644937, + "loss": 1.5973, + "theoretical_loss": 4.234013862592548, + "tokens_seen": 266207232 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046430290872617855, + "loss": 1.7945, + "theoretical_loss": 4.233890036309095, + "tokens_seen": 266272768 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046429287863590773, + "loss": 1.7553, + "theoretical_loss": 4.233766249029501, + "tokens_seen": 266338304 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004642828485456369, + "loss": 1.6776, + "theoretical_loss": 4.233642500731889, + "tokens_seen": 266403840 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046427281845536615, + "loss": 1.6588, + "theoretical_loss": 4.233518791394396, + "tokens_seen": 266469376 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004642627883650953, + "loss": 1.6196, + "theoretical_loss": 4.233395120995181, + "tokens_seen": 266534912 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004642527582748245, + "loss": 1.6783, + "theoretical_loss": 4.233271489512416, + "tokens_seen": 266600448 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046424272818455364, + "loss": 1.7195, + "theoretical_loss": 4.233147896924294, + "tokens_seen": 266665984 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046423269809428287, + "loss": 1.7604, + "theoretical_loss": 4.233024343209023, + "tokens_seen": 266731520 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046422266800401205, + "loss": 1.6127, + "theoretical_loss": 4.232900828344833, + "tokens_seen": 266797056 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046421263791374123, + "loss": 1.7739, + "theoretical_loss": 4.232777352309966, + "tokens_seen": 266862592 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004642026078234704, + "loss": 1.7512, + "theoretical_loss": 4.232653915082684, + "tokens_seen": 266928128 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004641925777331996, + "loss": 1.6751, + "theoretical_loss": 4.232530516641267, + "tokens_seen": 266993664 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004641825476429288, + "loss": 1.6752, + "theoretical_loss": 4.232407156964014, + "tokens_seen": 267059200 + }, + { + "epoch": 0.08, + "learning_rate": 0.000464172517552658, + "loss": 1.6141, + "theoretical_loss": 4.232283836029237, + "tokens_seen": 267124736 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046416248746238714, + "loss": 1.6299, + "theoretical_loss": 4.232160553815269, + "tokens_seen": 267190272 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004641524573721164, + "loss": 1.6926, + "theoretical_loss": 4.232037310300458, + "tokens_seen": 267255808 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004641424272818455, + "loss": 1.6511, + "theoretical_loss": 4.231914105463173, + "tokens_seen": 267321344 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046413239719157474, + "loss": 1.7545, + "theoretical_loss": 4.231790939281797, + "tokens_seen": 267386880 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004641223671013039, + "loss": 1.7372, + "theoretical_loss": 4.231667811734732, + "tokens_seen": 267452416 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004641123370110331, + "loss": 1.6178, + "theoretical_loss": 4.231544722800395, + "tokens_seen": 267517952 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004641023069207623, + "loss": 1.5862, + "theoretical_loss": 4.231421672457225, + "tokens_seen": 267583488 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004640922768304915, + "loss": 1.6637, + "theoretical_loss": 4.231298660683675, + "tokens_seen": 267649024 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046408224674022064, + "loss": 1.7941, + "theoretical_loss": 4.231175687458214, + "tokens_seen": 267714560 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004640722166499499, + "loss": 1.7227, + "theoretical_loss": 4.231052752759332, + "tokens_seen": 267780096 + }, + { + "epoch": 0.08, + "learning_rate": 0.000464062186559679, + "loss": 1.7096, + "theoretical_loss": 4.230929856565535, + "tokens_seen": 267845632 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046405215646940824, + "loss": 1.7177, + "theoretical_loss": 4.230806998855343, + "tokens_seen": 267911168 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004640421263791374, + "loss": 1.7424, + "theoretical_loss": 4.230684179607297, + "tokens_seen": 267976704 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004640320962888666, + "loss": 1.7293, + "theoretical_loss": 4.2305613987999555, + "tokens_seen": 268042240 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004640220661985958, + "loss": 1.7528, + "theoretical_loss": 4.230438656411891, + "tokens_seen": 268107776 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046401203610832496, + "loss": 1.7027, + "theoretical_loss": 4.230315952421697, + "tokens_seen": 268173312 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046400200601805414, + "loss": 1.7109, + "theoretical_loss": 4.230193286807981, + "tokens_seen": 268238848 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004639919759277834, + "loss": 1.6465, + "theoretical_loss": 4.230070659549368, + "tokens_seen": 268304384 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004639819458375125, + "loss": 1.7396, + "theoretical_loss": 4.229948070624502, + "tokens_seen": 268369920 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046397191574724174, + "loss": 1.6714, + "theoretical_loss": 4.229825520012042, + "tokens_seen": 268435456 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004639618856569709, + "loss": 1.65, + "theoretical_loss": 4.2297030076906665, + "tokens_seen": 268500992 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004639518555667001, + "loss": 1.6905, + "theoretical_loss": 4.229580533639068, + "tokens_seen": 268566528 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046394182547642934, + "loss": 1.6409, + "theoretical_loss": 4.229458097835959, + "tokens_seen": 268632064 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 106547, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7696818113327026, + "objective/train/theoretical_loss": 4.2293357002600676, + "objective/train/tokens_used": 268697600, + "theoretical_loss": 4.2293357002600676, + "tokens_seen": 268697600 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046393179538615847, + "loss": 1.7134, + "theoretical_loss": 4.2293357002600676, + "tokens_seen": 268697600 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004639217652958877, + "loss": 1.7845, + "theoretical_loss": 4.229213340890139, + "tokens_seen": 268763136 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004639117352056169, + "loss": 1.6845, + "theoretical_loss": 4.229091019704935, + "tokens_seen": 268828672 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046390170511534606, + "loss": 1.5504, + "theoretical_loss": 4.2289687366832345, + "tokens_seen": 268894208 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046389167502507525, + "loss": 1.6455, + "theoretical_loss": 4.228846491803835, + "tokens_seen": 268959744 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004638816449348044, + "loss": 1.6898, + "theoretical_loss": 4.228724285045549, + "tokens_seen": 269025280 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004638716148445336, + "loss": 1.564, + "theoretical_loss": 4.228602116387206, + "tokens_seen": 269090816 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046386158475426284, + "loss": 1.7127, + "theoretical_loss": 4.228479985807653, + "tokens_seen": 269156352 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046385155466399197, + "loss": 1.8527, + "theoretical_loss": 4.228357893285755, + "tokens_seen": 269221888 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004638415245737212, + "loss": 1.7551, + "theoretical_loss": 4.2282358388003916, + "tokens_seen": 269287424 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046383149448345033, + "loss": 1.6207, + "theoretical_loss": 4.228113822330461, + "tokens_seen": 269352960 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046382146439317957, + "loss": 1.7538, + "theoretical_loss": 4.227991843854878, + "tokens_seen": 269418496 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046381143430290875, + "loss": 1.7242, + "theoretical_loss": 4.227869903352572, + "tokens_seen": 269484032 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046380140421263793, + "loss": 1.6885, + "theoretical_loss": 4.227748000802493, + "tokens_seen": 269549568 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004637913741223671, + "loss": 1.6196, + "theoretical_loss": 4.227626136183606, + "tokens_seen": 269615104 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046378134403209635, + "loss": 1.5916, + "theoretical_loss": 4.227504309474892, + "tokens_seen": 269680640 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004637713139418255, + "loss": 1.7642, + "theoretical_loss": 4.227382520655348, + "tokens_seen": 269746176 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004637612838515547, + "loss": 1.7198, + "theoretical_loss": 4.227260769703991, + "tokens_seen": 269811712 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046375125376128384, + "loss": 1.7004, + "theoretical_loss": 4.227139056599851, + "tokens_seen": 269877248 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046374122367101307, + "loss": 1.6957, + "theoretical_loss": 4.2270173813219785, + "tokens_seen": 269942784 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046373119358074225, + "loss": 1.6521, + "theoretical_loss": 4.226895743849437, + "tokens_seen": 270008320 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046372116349047143, + "loss": 1.6903, + "theoretical_loss": 4.226774144161309, + "tokens_seen": 270073856 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004637111334002006, + "loss": 1.7107, + "theoretical_loss": 4.226652582236694, + "tokens_seen": 270139392 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004637011033099298, + "loss": 1.5766, + "theoretical_loss": 4.226531058054706, + "tokens_seen": 270204928 + }, + { + "epoch": 0.08, + "learning_rate": 0.000463691073219659, + "loss": 1.6141, + "theoretical_loss": 4.226409571594477, + "tokens_seen": 270270464 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004636810431293882, + "loss": 1.7098, + "theoretical_loss": 4.226288122835156, + "tokens_seen": 270336000 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046367101303911734, + "loss": 1.5607, + "theoretical_loss": 4.226166711755906, + "tokens_seen": 270401536 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004636609829488466, + "loss": 1.7744, + "theoretical_loss": 4.2260453383359105, + "tokens_seen": 270467072 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004636509528585757, + "loss": 1.7271, + "theoretical_loss": 4.2259240025543665, + "tokens_seen": 270532608 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046364092276830494, + "loss": 1.6856, + "theoretical_loss": 4.225802704390488, + "tokens_seen": 270598144 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004636308926780341, + "loss": 1.6275, + "theoretical_loss": 4.2256814438235075, + "tokens_seen": 270663680 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004636208625877633, + "loss": 1.6841, + "theoretical_loss": 4.225560220832672, + "tokens_seen": 270729216 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004636108324974925, + "loss": 1.701, + "theoretical_loss": 4.225439035397246, + "tokens_seen": 270794752 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004636008024072217, + "loss": 1.7277, + "theoretical_loss": 4.225317887496508, + "tokens_seen": 270860288 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046359077231695084, + "loss": 1.6489, + "theoretical_loss": 4.2251967771097565, + "tokens_seen": 270925824 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004635807422266801, + "loss": 1.6282, + "theoretical_loss": 4.225075704216305, + "tokens_seen": 270991360 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004635707121364092, + "loss": 1.7639, + "theoretical_loss": 4.224954668795483, + "tokens_seen": 271056896 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046356068204613844, + "loss": 1.6931, + "theoretical_loss": 4.224833670826636, + "tokens_seen": 271122432 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004635506519558676, + "loss": 1.6163, + "theoretical_loss": 4.224712710289127, + "tokens_seen": 271187968 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004635406218655968, + "loss": 1.6609, + "theoretical_loss": 4.2245917871623355, + "tokens_seen": 271253504 + }, + { + "epoch": 0.08, + "learning_rate": 0.000463530591775326, + "loss": 1.5118, + "theoretical_loss": 4.224470901425655, + "tokens_seen": 271319040 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046352056168505516, + "loss": 1.6932, + "theoretical_loss": 4.224350053058499, + "tokens_seen": 271384576 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046351053159478434, + "loss": 1.6178, + "theoretical_loss": 4.224229242040295, + "tokens_seen": 271450112 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004635005015045136, + "loss": 1.638, + "theoretical_loss": 4.224108468350485, + "tokens_seen": 271515648 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004634904714142427, + "loss": 1.7194, + "theoretical_loss": 4.223987731968532, + "tokens_seen": 271581184 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046348044132397194, + "loss": 1.6907, + "theoretical_loss": 4.223867032873912, + "tokens_seen": 271646720 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046347041123370107, + "loss": 1.8207, + "theoretical_loss": 4.223746371046116, + "tokens_seen": 271712256 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004634603811434303, + "loss": 1.5832, + "theoretical_loss": 4.223625746464656, + "tokens_seen": 271777792 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004634503510531595, + "loss": 1.6752, + "theoretical_loss": 4.223505159109054, + "tokens_seen": 271843328 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046344032096288867, + "loss": 1.7249, + "theoretical_loss": 4.223384608958855, + "tokens_seen": 271908864 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 107607, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7997349500656128, + "objective/train/theoretical_loss": 4.223264095993615, + "objective/train/tokens_used": 271974400, + "theoretical_loss": 4.223264095993615, + "tokens_seen": 271974400 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046343029087261785, + "loss": 1.7962, + "theoretical_loss": 4.223264095993615, + "tokens_seen": 271974400 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004634202607823471, + "loss": 1.673, + "theoretical_loss": 4.223143620192907, + "tokens_seen": 272039936 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004634102306920762, + "loss": 1.5485, + "theoretical_loss": 4.223023181536323, + "tokens_seen": 272105472 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046340020060180545, + "loss": 1.6354, + "theoretical_loss": 4.2229027800034675, + "tokens_seen": 272171008 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046339017051153457, + "loss": 1.613, + "theoretical_loss": 4.222782415573962, + "tokens_seen": 272236544 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004633801404212638, + "loss": 1.7214, + "theoretical_loss": 4.222662088227447, + "tokens_seen": 272302080 + }, + { + "epoch": 0.08, + "learning_rate": 0.000463370110330993, + "loss": 1.5507, + "theoretical_loss": 4.222541797943576, + "tokens_seen": 272367616 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046336008024072217, + "loss": 1.7722, + "theoretical_loss": 4.222421544702018, + "tokens_seen": 272433152 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046335005015045135, + "loss": 1.7998, + "theoretical_loss": 4.2223013284824615, + "tokens_seen": 272498688 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046334002006018053, + "loss": 1.7673, + "theoretical_loss": 4.222181149264608, + "tokens_seen": 272564224 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004633299899699097, + "loss": 1.6691, + "theoretical_loss": 4.222061007028175, + "tokens_seen": 272629760 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046331995987963895, + "loss": 1.6714, + "theoretical_loss": 4.221940901752899, + "tokens_seen": 272695296 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004633099297893681, + "loss": 1.7054, + "theoretical_loss": 4.221820833418529, + "tokens_seen": 272760832 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004632998996990973, + "loss": 1.6544, + "theoretical_loss": 4.221700802004832, + "tokens_seen": 272826368 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046328986960882644, + "loss": 1.7438, + "theoretical_loss": 4.2215808074915895, + "tokens_seen": 272891904 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004632798395185557, + "loss": 1.5827, + "theoretical_loss": 4.221460849858601, + "tokens_seen": 272957440 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046326980942828485, + "loss": 1.6381, + "theoretical_loss": 4.221340929085679, + "tokens_seen": 273022976 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046325977933801404, + "loss": 1.6013, + "theoretical_loss": 4.221221045152655, + "tokens_seen": 273088512 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004632497492477432, + "loss": 1.7288, + "theoretical_loss": 4.221101198039375, + "tokens_seen": 273154048 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046323971915747245, + "loss": 1.6504, + "theoretical_loss": 4.2209813877257, + "tokens_seen": 273219584 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004632296890672016, + "loss": 1.6798, + "theoretical_loss": 4.220861614191508, + "tokens_seen": 273285120 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004632196589769308, + "loss": 1.6889, + "theoretical_loss": 4.220741877416691, + "tokens_seen": 273350656 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046320962888666, + "loss": 1.758, + "theoretical_loss": 4.22062217738116, + "tokens_seen": 273416192 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004631995987963892, + "loss": 1.751, + "theoretical_loss": 4.22050251406484, + "tokens_seen": 273481728 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004631895687061184, + "loss": 1.6181, + "theoretical_loss": 4.220382887447671, + "tokens_seen": 273547264 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046317953861584754, + "loss": 1.6891, + "theoretical_loss": 4.22026329750961, + "tokens_seen": 273612800 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004631695085255768, + "loss": 1.6198, + "theoretical_loss": 4.220143744230629, + "tokens_seen": 273678336 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004631594784353059, + "loss": 1.725, + "theoretical_loss": 4.220024227590717, + "tokens_seen": 273743872 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046314944834503514, + "loss": 1.6381, + "theoretical_loss": 4.219904747569877, + "tokens_seen": 273809408 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004631394182547643, + "loss": 1.7394, + "theoretical_loss": 4.219785304148129, + "tokens_seen": 273874944 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004631293881644935, + "loss": 1.7242, + "theoretical_loss": 4.219665897305506, + "tokens_seen": 273940480 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004631193580742227, + "loss": 1.6699, + "theoretical_loss": 4.219546527022062, + "tokens_seen": 274006016 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004631093279839519, + "loss": 1.6277, + "theoretical_loss": 4.219427193277861, + "tokens_seen": 274071552 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046309929789368104, + "loss": 1.86, + "theoretical_loss": 4.219307896052987, + "tokens_seen": 274137088 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004630892678034103, + "loss": 1.8027, + "theoretical_loss": 4.219188635327535, + "tokens_seen": 274202624 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004630792377131394, + "loss": 1.6359, + "theoretical_loss": 4.219069411081622, + "tokens_seen": 274268160 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046306920762286864, + "loss": 1.7244, + "theoretical_loss": 4.218950223295374, + "tokens_seen": 274333696 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004630591775325978, + "loss": 1.6481, + "theoretical_loss": 4.218831071948936, + "tokens_seen": 274399232 + }, + { + "epoch": 0.08, + "learning_rate": 0.000463049147442327, + "loss": 1.6679, + "theoretical_loss": 4.218711957022469, + "tokens_seen": 274464768 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004630391173520562, + "loss": 1.7853, + "theoretical_loss": 4.218592878496148, + "tokens_seen": 274530304 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046302908726178536, + "loss": 1.6642, + "theoretical_loss": 4.218473836350164, + "tokens_seen": 274595840 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046301905717151455, + "loss": 1.711, + "theoretical_loss": 4.2183548305647225, + "tokens_seen": 274661376 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004630090270812438, + "loss": 1.7319, + "theoretical_loss": 4.218235861120047, + "tokens_seen": 274726912 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004629989969909729, + "loss": 1.6225, + "theoretical_loss": 4.218116927996375, + "tokens_seen": 274792448 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046298896690070214, + "loss": 1.6855, + "theoretical_loss": 4.217998031173959, + "tokens_seen": 274857984 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046297893681043127, + "loss": 1.5481, + "theoretical_loss": 4.217879170633067, + "tokens_seen": 274923520 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004629689067201605, + "loss": 1.7079, + "theoretical_loss": 4.217760346353984, + "tokens_seen": 274989056 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004629588766298897, + "loss": 1.6727, + "theoretical_loss": 4.217641558317008, + "tokens_seen": 275054592 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046294884653961887, + "loss": 1.6376, + "theoretical_loss": 4.217522806502455, + "tokens_seen": 275120128 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046293881644934805, + "loss": 1.5728, + "theoretical_loss": 4.217404090890653, + "tokens_seen": 275185664 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 108722, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7074024677276611, + "objective/train/theoretical_loss": 4.217285411461949, + "objective/train/tokens_used": 275251200, + "theoretical_loss": 4.217285411461949, + "tokens_seen": 275251200 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004629287863590773, + "loss": 1.7059, + "theoretical_loss": 4.217285411461949, + "tokens_seen": 275251200 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004629187562688064, + "loss": 1.6208, + "theoretical_loss": 4.217166768196704, + "tokens_seen": 275316736 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046290872617853565, + "loss": 1.7445, + "theoretical_loss": 4.217048161075294, + "tokens_seen": 275382272 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046289869608826477, + "loss": 1.6902, + "theoretical_loss": 4.2169295900781085, + "tokens_seen": 275447808 + }, + { + "epoch": 0.08, + "learning_rate": 0.000462888665997994, + "loss": 1.6393, + "theoretical_loss": 4.216811055185556, + "tokens_seen": 275513344 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004628786359077232, + "loss": 1.6397, + "theoretical_loss": 4.216692556378058, + "tokens_seen": 275578880 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046286860581745237, + "loss": 1.6112, + "theoretical_loss": 4.216574093636053, + "tokens_seen": 275644416 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046285857572718155, + "loss": 1.6083, + "theoretical_loss": 4.216455666939991, + "tokens_seen": 275709952 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046284854563691073, + "loss": 1.7433, + "theoretical_loss": 4.216337276270341, + "tokens_seen": 275775488 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004628385155466399, + "loss": 1.7487, + "theoretical_loss": 4.216218921607586, + "tokens_seen": 275841024 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046282848545636915, + "loss": 1.6601, + "theoretical_loss": 4.216100602932224, + "tokens_seen": 275906560 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004628184553660983, + "loss": 1.6673, + "theoretical_loss": 4.21598232022477, + "tokens_seen": 275972096 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004628084252758275, + "loss": 1.625, + "theoretical_loss": 4.215864073465749, + "tokens_seen": 276037632 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046279839518555664, + "loss": 1.7194, + "theoretical_loss": 4.215745862635708, + "tokens_seen": 276103168 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004627883650952859, + "loss": 1.6487, + "theoretical_loss": 4.2156276877152035, + "tokens_seen": 276168704 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046277833500501505, + "loss": 1.6485, + "theoretical_loss": 4.21550954868481, + "tokens_seen": 276234240 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046276830491474424, + "loss": 1.5641, + "theoretical_loss": 4.2153914455251185, + "tokens_seen": 276299776 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004627582748244734, + "loss": 1.6709, + "theoretical_loss": 4.215273378216731, + "tokens_seen": 276365312 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046274824473420265, + "loss": 1.7036, + "theoretical_loss": 4.215155346740268, + "tokens_seen": 276430848 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004627382146439318, + "loss": 1.6294, + "theoretical_loss": 4.215037351076363, + "tokens_seen": 276496384 + }, + { + "epoch": 0.08, + "learning_rate": 0.000462728184553661, + "loss": 1.6955, + "theoretical_loss": 4.214919391205665, + "tokens_seen": 276561920 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046271815446339014, + "loss": 1.7044, + "theoretical_loss": 4.21480146710884, + "tokens_seen": 276627456 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004627081243731194, + "loss": 1.672, + "theoretical_loss": 4.214683578766566, + "tokens_seen": 276692992 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046269809428284856, + "loss": 1.6074, + "theoretical_loss": 4.214565726159539, + "tokens_seen": 276758528 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046268806419257774, + "loss": 1.7116, + "theoretical_loss": 4.214447909268467, + "tokens_seen": 276824064 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004626780341023069, + "loss": 1.6776, + "theoretical_loss": 4.214330128074074, + "tokens_seen": 276889600 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004626680040120361, + "loss": 1.731, + "theoretical_loss": 4.214212382557101, + "tokens_seen": 276955136 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004626579739217653, + "loss": 1.6593, + "theoretical_loss": 4.214094672698302, + "tokens_seen": 277020672 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004626479438314945, + "loss": 1.6016, + "theoretical_loss": 4.213976998478445, + "tokens_seen": 277086208 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046263791374122364, + "loss": 1.6216, + "theoretical_loss": 4.213859359878316, + "tokens_seen": 277151744 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004626278836509529, + "loss": 1.6708, + "theoretical_loss": 4.213741756878713, + "tokens_seen": 277217280 + }, + { + "epoch": 0.08, + "learning_rate": 0.000462617853560682, + "loss": 1.5288, + "theoretical_loss": 4.213624189460449, + "tokens_seen": 277282816 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046260782347041124, + "loss": 1.7567, + "theoretical_loss": 4.213506657604356, + "tokens_seen": 277348352 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004625977933801404, + "loss": 1.6332, + "theoretical_loss": 4.213389161291273, + "tokens_seen": 277413888 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004625877632898696, + "loss": 1.6626, + "theoretical_loss": 4.213271700502062, + "tokens_seen": 277479424 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004625777331995988, + "loss": 1.6723, + "theoretical_loss": 4.213154275217597, + "tokens_seen": 277544960 + }, + { + "epoch": 0.08, + "learning_rate": 0.000462567703109328, + "loss": 1.6144, + "theoretical_loss": 4.213036885418763, + "tokens_seen": 277610496 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046255767301905715, + "loss": 1.622, + "theoretical_loss": 4.212919531086466, + "tokens_seen": 277676032 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004625476429287864, + "loss": 1.7117, + "theoretical_loss": 4.2128022122016215, + "tokens_seen": 277741568 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004625376128385155, + "loss": 1.7512, + "theoretical_loss": 4.212684928745163, + "tokens_seen": 277807104 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046252758274824475, + "loss": 1.7439, + "theoretical_loss": 4.212567680698038, + "tokens_seen": 277872640 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004625175526579739, + "loss": 1.5757, + "theoretical_loss": 4.212450468041208, + "tokens_seen": 277938176 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004625075225677031, + "loss": 1.6309, + "theoretical_loss": 4.21233329075565, + "tokens_seen": 278003712 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004624974924774323, + "loss": 1.6232, + "theoretical_loss": 4.212216148822355, + "tokens_seen": 278069248 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046248746238716147, + "loss": 1.6041, + "theoretical_loss": 4.212099042222331, + "tokens_seen": 278134784 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046247743229689065, + "loss": 1.548, + "theoretical_loss": 4.211981970936595, + "tokens_seen": 278200320 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004624674022066199, + "loss": 1.6003, + "theoretical_loss": 4.2118649349461865, + "tokens_seen": 278265856 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046245737211634907, + "loss": 1.5724, + "theoretical_loss": 4.211747934232153, + "tokens_seen": 278331392 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046244734202607825, + "loss": 1.6027, + "theoretical_loss": 4.211630968775561, + "tokens_seen": 278396928 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004624373119358075, + "loss": 1.7056, + "theoretical_loss": 4.211514038557488, + "tokens_seen": 278462464 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 110249, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.63548743724823, + "objective/train/theoretical_loss": 4.211397143559029, + "objective/train/tokens_used": 278528000, + "theoretical_loss": 4.211397143559029, + "tokens_seen": 278528000 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004624272818455366, + "loss": 1.6347, + "theoretical_loss": 4.211397143559029, + "tokens_seen": 278528000 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046241725175526585, + "loss": 1.5401, + "theoretical_loss": 4.211280283761294, + "tokens_seen": 278593536 + }, + { + "epoch": 0.08, + "learning_rate": 0.000462407221664995, + "loss": 1.6755, + "theoretical_loss": 4.211163459145403, + "tokens_seen": 278659072 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004623971915747242, + "loss": 1.709, + "theoretical_loss": 4.211046669692496, + "tokens_seen": 278724608 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004623871614844534, + "loss": 1.5563, + "theoretical_loss": 4.2109299153837245, + "tokens_seen": 278790144 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046237713139418257, + "loss": 1.6141, + "theoretical_loss": 4.210813196200255, + "tokens_seen": 278855680 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046236710130391175, + "loss": 1.7641, + "theoretical_loss": 4.21069651212327, + "tokens_seen": 278921216 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046235707121364093, + "loss": 1.6373, + "theoretical_loss": 4.2105798631339635, + "tokens_seen": 278986752 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004623470411233701, + "loss": 1.6674, + "theoretical_loss": 4.2104632492135465, + "tokens_seen": 279052288 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046233701103309935, + "loss": 1.6615, + "theoretical_loss": 4.210346670343244, + "tokens_seen": 279117824 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004623269809428285, + "loss": 1.721, + "theoretical_loss": 4.210230126504296, + "tokens_seen": 279183360 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004623169508525577, + "loss": 1.633, + "theoretical_loss": 4.210113617677955, + "tokens_seen": 279248896 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046230692076228684, + "loss": 1.683, + "theoretical_loss": 4.20999714384549, + "tokens_seen": 279314432 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004622968906720161, + "loss": 1.6522, + "theoretical_loss": 4.209880704988181, + "tokens_seen": 279379968 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046228686058174525, + "loss": 1.7099, + "theoretical_loss": 4.209764301087327, + "tokens_seen": 279445504 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046227683049147444, + "loss": 1.7306, + "theoretical_loss": 4.20964793212424, + "tokens_seen": 279511040 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004622668004012036, + "loss": 1.6972, + "theoretical_loss": 4.209531598080245, + "tokens_seen": 279576576 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046225677031093285, + "loss": 1.644, + "theoretical_loss": 4.20941529893668, + "tokens_seen": 279642112 + }, + { + "epoch": 0.08, + "learning_rate": 0.000462246740220662, + "loss": 1.5525, + "theoretical_loss": 4.209299034674902, + "tokens_seen": 279707648 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004622367101303912, + "loss": 1.606, + "theoretical_loss": 4.209182805276278, + "tokens_seen": 279773184 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046222668004012034, + "loss": 1.761, + "theoretical_loss": 4.209066610722192, + "tokens_seen": 279838720 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004622166499498496, + "loss": 1.5687, + "theoretical_loss": 4.208950450994041, + "tokens_seen": 279904256 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046220661985957876, + "loss": 1.6863, + "theoretical_loss": 4.2088343260732355, + "tokens_seen": 279969792 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046219658976930794, + "loss": 1.628, + "theoretical_loss": 4.208718235941204, + "tokens_seen": 280035328 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004621865596790371, + "loss": 1.7519, + "theoretical_loss": 4.2086021805793825, + "tokens_seen": 280100864 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004621765295887663, + "loss": 1.73, + "theoretical_loss": 4.208486159969229, + "tokens_seen": 280166400 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004621664994984955, + "loss": 1.7183, + "theoretical_loss": 4.20837017409221, + "tokens_seen": 280231936 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004621564694082247, + "loss": 1.6784, + "theoretical_loss": 4.20825422292981, + "tokens_seen": 280297472 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046214643931795384, + "loss": 1.5934, + "theoretical_loss": 4.208138306463524, + "tokens_seen": 280363008 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004621364092276831, + "loss": 1.6931, + "theoretical_loss": 4.208022424674864, + "tokens_seen": 280428544 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004621263791374122, + "loss": 1.7002, + "theoretical_loss": 4.207906577545356, + "tokens_seen": 280494080 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046211634904714144, + "loss": 1.6042, + "theoretical_loss": 4.207790765056538, + "tokens_seen": 280559616 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004621063189568706, + "loss": 1.7058, + "theoretical_loss": 4.207674987189965, + "tokens_seen": 280625152 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004620962888665998, + "loss": 1.8906, + "theoretical_loss": 4.2075592439272045, + "tokens_seen": 280690688 + }, + { + "epoch": 0.09, + "learning_rate": 0.000462086258776329, + "loss": 1.6257, + "theoretical_loss": 4.207443535249838, + "tokens_seen": 280756224 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004620762286860582, + "loss": 1.5809, + "theoretical_loss": 4.207327861139461, + "tokens_seen": 280821760 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046206619859578735, + "loss": 1.7312, + "theoretical_loss": 4.207212221577684, + "tokens_seen": 280887296 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004620561685055166, + "loss": 1.7429, + "theoretical_loss": 4.207096616546132, + "tokens_seen": 280952832 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004620461384152457, + "loss": 1.6491, + "theoretical_loss": 4.206981046026442, + "tokens_seen": 281018368 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046203610832497495, + "loss": 1.5835, + "theoretical_loss": 4.2068655100002665, + "tokens_seen": 281083904 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004620260782347041, + "loss": 1.598, + "theoretical_loss": 4.206750008449272, + "tokens_seen": 281149440 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004620160481444333, + "loss": 1.6652, + "theoretical_loss": 4.206634541355138, + "tokens_seen": 281214976 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004620060180541625, + "loss": 1.7501, + "theoretical_loss": 4.20651910869956, + "tokens_seen": 281280512 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046199598796389167, + "loss": 1.6591, + "theoretical_loss": 4.2064037104642455, + "tokens_seen": 281346048 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046198595787362085, + "loss": 1.6004, + "theoretical_loss": 4.206288346630917, + "tokens_seen": 281411584 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004619759277833501, + "loss": 1.652, + "theoretical_loss": 4.206173017181309, + "tokens_seen": 281477120 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004619658976930792, + "loss": 1.7249, + "theoretical_loss": 4.206057722097174, + "tokens_seen": 281542656 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046195586760280845, + "loss": 1.531, + "theoretical_loss": 4.205942461360276, + "tokens_seen": 281608192 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046194583751253763, + "loss": 1.5578, + "theoretical_loss": 4.205827234952391, + "tokens_seen": 281673728 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004619358074222668, + "loss": 1.6118, + "theoretical_loss": 4.205712042855312, + "tokens_seen": 281739264 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 111427, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5737414360046387, + "objective/train/theoretical_loss": 4.205596885050846, + "objective/train/tokens_used": 281804800, + "theoretical_loss": 4.205596885050846, + "tokens_seen": 281804800 + }, + { + "epoch": 0.09, + "learning_rate": 0.000461925777331996, + "loss": 1.5716, + "theoretical_loss": 4.205596885050846, + "tokens_seen": 281804800 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004619157472417252, + "loss": 1.6417, + "theoretical_loss": 4.205481761520811, + "tokens_seen": 281870336 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046190571715145435, + "loss": 1.7209, + "theoretical_loss": 4.20536667224704, + "tokens_seen": 281935872 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004618956870611836, + "loss": 1.6293, + "theoretical_loss": 4.205251617211382, + "tokens_seen": 282001408 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004618856569709127, + "loss": 1.7085, + "theoretical_loss": 4.205136596395698, + "tokens_seen": 282066944 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046187562688064195, + "loss": 1.6463, + "theoretical_loss": 4.205021609781861, + "tokens_seen": 282132480 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004618655967903711, + "loss": 1.5892, + "theoretical_loss": 4.204906657351763, + "tokens_seen": 282198016 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004618555667001003, + "loss": 1.5665, + "theoretical_loss": 4.2047917390873035, + "tokens_seen": 282263552 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004618455366098295, + "loss": 1.5897, + "theoretical_loss": 4.204676854970401, + "tokens_seen": 282329088 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004618355065195587, + "loss": 1.7423, + "theoretical_loss": 4.204562004982986, + "tokens_seen": 282394624 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046182547642928786, + "loss": 1.6529, + "theoretical_loss": 4.204447189107, + "tokens_seen": 282460160 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046181544633901704, + "loss": 1.6199, + "theoretical_loss": 4.204332407324403, + "tokens_seen": 282525696 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004618054162487462, + "loss": 1.6609, + "theoretical_loss": 4.204217659617164, + "tokens_seen": 282591232 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046179538615847545, + "loss": 1.6127, + "theoretical_loss": 4.2041029459672705, + "tokens_seen": 282656768 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004617853560682046, + "loss": 1.587, + "theoretical_loss": 4.2039882663567205, + "tokens_seen": 282722304 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004617753259779338, + "loss": 1.6788, + "theoretical_loss": 4.203873620767526, + "tokens_seen": 282787840 + }, + { + "epoch": 0.09, + "learning_rate": 0.000461765295887663, + "loss": 1.6184, + "theoretical_loss": 4.203759009181714, + "tokens_seen": 282853376 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004617552657973922, + "loss": 1.8315, + "theoretical_loss": 4.203644431581323, + "tokens_seen": 282918912 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046174523570712136, + "loss": 1.6235, + "theoretical_loss": 4.203529887948408, + "tokens_seen": 282984448 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046173520561685054, + "loss": 1.6767, + "theoretical_loss": 4.203415378265034, + "tokens_seen": 283049984 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004617251755265797, + "loss": 1.5395, + "theoretical_loss": 4.203300902513284, + "tokens_seen": 283115520 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046171514543630896, + "loss": 1.6687, + "theoretical_loss": 4.203186460675251, + "tokens_seen": 283181056 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046170511534603814, + "loss": 1.588, + "theoretical_loss": 4.203072052733043, + "tokens_seen": 283246592 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004616950852557673, + "loss": 1.7012, + "theoretical_loss": 4.202957678668782, + "tokens_seen": 283312128 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004616850551654965, + "loss": 1.6562, + "theoretical_loss": 4.202843338464602, + "tokens_seen": 283377664 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004616750250752257, + "loss": 1.7215, + "theoretical_loss": 4.202729032102653, + "tokens_seen": 283443200 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004616649949849549, + "loss": 1.6929, + "theoretical_loss": 4.202614759565096, + "tokens_seen": 283508736 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046165496489468404, + "loss": 1.6852, + "theoretical_loss": 4.202500520834106, + "tokens_seen": 283574272 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004616449348044133, + "loss": 1.6509, + "theoretical_loss": 4.202386315891874, + "tokens_seen": 283639808 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004616349047141424, + "loss": 1.6796, + "theoretical_loss": 4.202272144720602, + "tokens_seen": 283705344 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046162487462387164, + "loss": 1.7137, + "theoretical_loss": 4.202158007302506, + "tokens_seen": 283770880 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004616148445336008, + "loss": 1.6678, + "theoretical_loss": 4.202043903619814, + "tokens_seen": 283836416 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046160481444333, + "loss": 1.6511, + "theoretical_loss": 4.201929833654772, + "tokens_seen": 283901952 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004615947843530592, + "loss": 1.6635, + "theoretical_loss": 4.201815797389634, + "tokens_seen": 283967488 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004615847542627884, + "loss": 1.6314, + "theoretical_loss": 4.201701794806672, + "tokens_seen": 284033024 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046157472417251755, + "loss": 1.674, + "theoretical_loss": 4.201587825888167, + "tokens_seen": 284098560 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004615646940822468, + "loss": 1.651, + "theoretical_loss": 4.201473890616418, + "tokens_seen": 284164096 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004615546639919759, + "loss": 1.6095, + "theoretical_loss": 4.201359988973733, + "tokens_seen": 284229632 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046154463390170515, + "loss": 1.6924, + "theoretical_loss": 4.201246120942438, + "tokens_seen": 284295168 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004615346038114343, + "loss": 1.652, + "theoretical_loss": 4.201132286504867, + "tokens_seen": 284360704 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004615245737211635, + "loss": 1.697, + "theoretical_loss": 4.201018485643373, + "tokens_seen": 284426240 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004615145436308927, + "loss": 1.7817, + "theoretical_loss": 4.200904718340317, + "tokens_seen": 284491776 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046150451354062187, + "loss": 1.6408, + "theoretical_loss": 4.200790984578078, + "tokens_seen": 284557312 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046149448345035105, + "loss": 1.673, + "theoretical_loss": 4.200677284339045, + "tokens_seen": 284622848 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004614844533600803, + "loss": 1.6601, + "theoretical_loss": 4.2005636176056225, + "tokens_seen": 284688384 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004614744232698094, + "loss": 1.646, + "theoretical_loss": 4.200449984360226, + "tokens_seen": 284753920 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046146439317953865, + "loss": 1.6519, + "theoretical_loss": 4.200336384585286, + "tokens_seen": 284819456 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046145436308926783, + "loss": 1.6732, + "theoretical_loss": 4.200222818263246, + "tokens_seen": 284884992 + }, + { + "epoch": 0.09, + "learning_rate": 0.000461444332998997, + "loss": 1.6287, + "theoretical_loss": 4.200109285376563, + "tokens_seen": 284950528 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004614343029087262, + "loss": 1.5752, + "theoretical_loss": 4.1999957859077055, + "tokens_seen": 285016064 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 112523, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7507089376449585, + "objective/train/theoretical_loss": 4.1998823198391575, + "objective/train/tokens_used": 285081600, + "theoretical_loss": 4.1998823198391575, + "tokens_seen": 285081600 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004614242728184554, + "loss": 1.6997, + "theoretical_loss": 4.1998823198391575, + "tokens_seen": 285081600 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046141424272818455, + "loss": 1.5748, + "theoretical_loss": 4.199768887153415, + "tokens_seen": 285147136 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004614042126379138, + "loss": 1.785, + "theoretical_loss": 4.199655487832987, + "tokens_seen": 285212672 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004613941825476429, + "loss": 1.6275, + "theoretical_loss": 4.199542121860397, + "tokens_seen": 285278208 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046138415245737215, + "loss": 1.5781, + "theoretical_loss": 4.199428789218181, + "tokens_seen": 285343744 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004613741223671013, + "loss": 1.6165, + "theoretical_loss": 4.199315489888885, + "tokens_seen": 285409280 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004613640922768305, + "loss": 1.6722, + "theoretical_loss": 4.199202223855075, + "tokens_seen": 285474816 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004613540621865597, + "loss": 1.7112, + "theoretical_loss": 4.199088991099323, + "tokens_seen": 285540352 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004613440320962889, + "loss": 1.7305, + "theoretical_loss": 4.19897579160422, + "tokens_seen": 285605888 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046133400200601806, + "loss": 1.6023, + "theoretical_loss": 4.1988626253523655, + "tokens_seen": 285671424 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046132397191574724, + "loss": 1.5642, + "theoretical_loss": 4.198749492326374, + "tokens_seen": 285736960 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004613139418254764, + "loss": 1.5711, + "theoretical_loss": 4.198636392508874, + "tokens_seen": 285802496 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046130391173520566, + "loss": 1.6807, + "theoretical_loss": 4.198523325882505, + "tokens_seen": 285868032 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004612938816449348, + "loss": 1.6569, + "theoretical_loss": 4.198410292429922, + "tokens_seen": 285933568 + }, + { + "epoch": 0.09, + "learning_rate": 0.000461283851554664, + "loss": 1.6764, + "theoretical_loss": 4.19829729213379, + "tokens_seen": 285999104 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004612738214643932, + "loss": 1.6312, + "theoretical_loss": 4.198184324976791, + "tokens_seen": 286064640 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004612637913741224, + "loss": 1.6462, + "theoretical_loss": 4.198071390941616, + "tokens_seen": 286130176 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046125376128385156, + "loss": 1.6496, + "theoretical_loss": 4.197958490010971, + "tokens_seen": 286195712 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046124373119358074, + "loss": 1.6505, + "theoretical_loss": 4.197845622167575, + "tokens_seen": 286261248 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004612337011033099, + "loss": 1.6635, + "theoretical_loss": 4.197732787394161, + "tokens_seen": 286326784 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046122367101303916, + "loss": 1.7169, + "theoretical_loss": 4.197619985673471, + "tokens_seen": 286392320 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004612136409227683, + "loss": 1.6863, + "theoretical_loss": 4.197507216988265, + "tokens_seen": 286457856 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004612036108324975, + "loss": 1.6374, + "theoretical_loss": 4.197394481321312, + "tokens_seen": 286523392 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046119358074222665, + "loss": 1.625, + "theoretical_loss": 4.1972817786553955, + "tokens_seen": 286588928 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004611835506519559, + "loss": 1.7286, + "theoretical_loss": 4.197169108973313, + "tokens_seen": 286654464 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046117352056168506, + "loss": 1.6383, + "theoretical_loss": 4.197056472257874, + "tokens_seen": 286720000 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046116349047141425, + "loss": 1.6491, + "theoretical_loss": 4.196943868491899, + "tokens_seen": 286785536 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004611534603811434, + "loss": 1.6145, + "theoretical_loss": 4.196831297658225, + "tokens_seen": 286851072 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004611434302908726, + "loss": 1.6756, + "theoretical_loss": 4.1967187597396975, + "tokens_seen": 286916608 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004611334002006018, + "loss": 1.6028, + "theoretical_loss": 4.196606254719179, + "tokens_seen": 286982144 + }, + { + "epoch": 0.09, + "learning_rate": 0.000461123370110331, + "loss": 1.6497, + "theoretical_loss": 4.196493782579544, + "tokens_seen": 287047680 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046111334002006015, + "loss": 1.5304, + "theoretical_loss": 4.196381343303676, + "tokens_seen": 287113216 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004611033099297894, + "loss": 1.6773, + "theoretical_loss": 4.1962689368744766, + "tokens_seen": 287178752 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046109327983951857, + "loss": 1.7498, + "theoretical_loss": 4.196156563274858, + "tokens_seen": 287244288 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046108324974924775, + "loss": 1.7346, + "theoretical_loss": 4.1960442224877434, + "tokens_seen": 287309824 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046107321965897693, + "loss": 1.6377, + "theoretical_loss": 4.195931914496072, + "tokens_seen": 287375360 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004610631895687061, + "loss": 1.6347, + "theoretical_loss": 4.195819639282792, + "tokens_seen": 287440896 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004610531594784353, + "loss": 1.6579, + "theoretical_loss": 4.195707396830869, + "tokens_seen": 287506432 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004610431293881645, + "loss": 1.6895, + "theoretical_loss": 4.195595187123277, + "tokens_seen": 287571968 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046103309929789365, + "loss": 1.5214, + "theoretical_loss": 4.195483010143006, + "tokens_seen": 287637504 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004610230692076229, + "loss": 1.5498, + "theoretical_loss": 4.1953708658730555, + "tokens_seen": 287703040 + }, + { + "epoch": 0.09, + "learning_rate": 0.000461013039117352, + "loss": 1.5748, + "theoretical_loss": 4.195258754296441, + "tokens_seen": 287768576 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046100300902708125, + "loss": 1.5452, + "theoretical_loss": 4.195146675396189, + "tokens_seen": 287834112 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046099297893681043, + "loss": 1.536, + "theoretical_loss": 4.195034629155338, + "tokens_seen": 287899648 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004609829488465396, + "loss": 1.629, + "theoretical_loss": 4.194922615556941, + "tokens_seen": 287965184 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004609729187562688, + "loss": 1.6436, + "theoretical_loss": 4.194810634584061, + "tokens_seen": 288030720 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046096288866599803, + "loss": 1.6561, + "theoretical_loss": 4.194698686219778, + "tokens_seen": 288096256 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004609528585757272, + "loss": 1.624, + "theoretical_loss": 4.194586770447179, + "tokens_seen": 288161792 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004609428284854564, + "loss": 1.5704, + "theoretical_loss": 4.194474887249369, + "tokens_seen": 288227328 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004609327983951856, + "loss": 1.5855, + "theoretical_loss": 4.19436303660946, + "tokens_seen": 288292864 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 113697, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6115751266479492, + "objective/train/theoretical_loss": 4.194251218510583, + "objective/train/tokens_used": 288358400, + "theoretical_loss": 4.194251218510583, + "tokens_seen": 288358400 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046092276830491475, + "loss": 1.6454, + "theoretical_loss": 4.194251218510583, + "tokens_seen": 288358400 + }, + { + "epoch": 0.09, + "learning_rate": 0.000460912738214644, + "loss": 1.6719, + "theoretical_loss": 4.194139432935877, + "tokens_seen": 288423936 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004609027081243731, + "loss": 1.6238, + "theoretical_loss": 4.194027679868495, + "tokens_seen": 288489472 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046089267803410235, + "loss": 1.6613, + "theoretical_loss": 4.193915959291601, + "tokens_seen": 288555008 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004608826479438315, + "loss": 1.676, + "theoretical_loss": 4.193804271188376, + "tokens_seen": 288620544 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004608726178535607, + "loss": 1.6172, + "theoretical_loss": 4.193692615542007, + "tokens_seen": 288686080 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004608625877632899, + "loss": 1.7037, + "theoretical_loss": 4.193580992335699, + "tokens_seen": 288751616 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004608525576730191, + "loss": 1.6349, + "theoretical_loss": 4.193469401552667, + "tokens_seen": 288817152 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046084252758274826, + "loss": 1.6025, + "theoretical_loss": 4.193357843176139, + "tokens_seen": 288882688 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046083249749247744, + "loss": 1.6554, + "theoretical_loss": 4.193246317189356, + "tokens_seen": 288948224 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004608224674022066, + "loss": 1.5523, + "theoretical_loss": 4.19313482357557, + "tokens_seen": 289013760 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046081243731193586, + "loss": 1.5922, + "theoretical_loss": 4.193023362318047, + "tokens_seen": 289079296 + }, + { + "epoch": 0.09, + "learning_rate": 0.000460802407221665, + "loss": 1.7185, + "theoretical_loss": 4.1929119334000635, + "tokens_seen": 289144832 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004607923771313942, + "loss": 1.6527, + "theoretical_loss": 4.192800536804912, + "tokens_seen": 289210368 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004607823470411234, + "loss": 1.6893, + "theoretical_loss": 4.192689172515893, + "tokens_seen": 289275904 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004607723169508526, + "loss": 1.6617, + "theoretical_loss": 4.192577840516323, + "tokens_seen": 289341440 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046076228686058176, + "loss": 1.6854, + "theoretical_loss": 4.192466540789528, + "tokens_seen": 289406976 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046075225677031094, + "loss": 1.6255, + "theoretical_loss": 4.19235527331885, + "tokens_seen": 289472512 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004607422266800401, + "loss": 1.54, + "theoretical_loss": 4.19224403808764, + "tokens_seen": 289538048 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046073219658976936, + "loss": 1.7685, + "theoretical_loss": 4.192132835079262, + "tokens_seen": 289603584 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004607221664994985, + "loss": 1.6051, + "theoretical_loss": 4.192021664277094, + "tokens_seen": 289669120 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004607121364092277, + "loss": 1.6121, + "theoretical_loss": 4.191910525664524, + "tokens_seen": 289734656 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046070210631895685, + "loss": 1.7641, + "theoretical_loss": 4.191799419224955, + "tokens_seen": 289800192 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004606920762286861, + "loss": 1.5467, + "theoretical_loss": 4.1916883449418, + "tokens_seen": 289865728 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046068204613841526, + "loss": 1.5386, + "theoretical_loss": 4.1915773027984855, + "tokens_seen": 289931264 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046067201604814445, + "loss": 1.6652, + "theoretical_loss": 4.19146629277845, + "tokens_seen": 289996800 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004606619859578736, + "loss": 1.6831, + "theoretical_loss": 4.1913553148651435, + "tokens_seen": 290062336 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004606519558676028, + "loss": 1.6369, + "theoretical_loss": 4.191244369042029, + "tokens_seen": 290127872 + }, + { + "epoch": 0.09, + "learning_rate": 0.000460641925777332, + "loss": 1.6354, + "theoretical_loss": 4.191133455292583, + "tokens_seen": 290193408 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004606318956870612, + "loss": 1.6653, + "theoretical_loss": 4.191022573600293, + "tokens_seen": 290258944 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046062186559679035, + "loss": 1.5844, + "theoretical_loss": 4.190911723948657, + "tokens_seen": 290324480 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004606118355065196, + "loss": 1.5593, + "theoretical_loss": 4.190800906321188, + "tokens_seen": 290390016 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046060180541624877, + "loss": 1.7331, + "theoretical_loss": 4.190690120701412, + "tokens_seen": 290455552 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046059177532597795, + "loss": 1.5844, + "theoretical_loss": 4.190579367072861, + "tokens_seen": 290521088 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046058174523570713, + "loss": 1.6983, + "theoretical_loss": 4.190468645419086, + "tokens_seen": 290586624 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004605717151454363, + "loss": 1.6684, + "theoretical_loss": 4.190357955723649, + "tokens_seen": 290652160 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004605616850551655, + "loss": 1.718, + "theoretical_loss": 4.190247297970121, + "tokens_seen": 290717696 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004605516549648947, + "loss": 1.6358, + "theoretical_loss": 4.190136672142088, + "tokens_seen": 290783232 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046054162487462385, + "loss": 1.6526, + "theoretical_loss": 4.190026078223145, + "tokens_seen": 290848768 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004605315947843531, + "loss": 1.6943, + "theoretical_loss": 4.189915516196904, + "tokens_seen": 290914304 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004605215646940822, + "loss": 1.7163, + "theoretical_loss": 4.189804986046984, + "tokens_seen": 290979840 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046051153460381145, + "loss": 1.7076, + "theoretical_loss": 4.18969448775702, + "tokens_seen": 291045376 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046050150451354063, + "loss": 1.7543, + "theoretical_loss": 4.189584021310657, + "tokens_seen": 291110912 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004604914744232698, + "loss": 1.6126, + "theoretical_loss": 4.189473586691554, + "tokens_seen": 291176448 + }, + { + "epoch": 0.09, + "learning_rate": 0.000460481444332999, + "loss": 1.7164, + "theoretical_loss": 4.189363183883377, + "tokens_seen": 291241984 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046047141424272823, + "loss": 1.6146, + "theoretical_loss": 4.189252812869811, + "tokens_seen": 291307520 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046046138415245736, + "loss": 1.6668, + "theoretical_loss": 4.18914247363455, + "tokens_seen": 291373056 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004604513540621866, + "loss": 1.583, + "theoretical_loss": 4.189032166161297, + "tokens_seen": 291438592 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004604413239719157, + "loss": 1.6823, + "theoretical_loss": 4.188921890433772, + "tokens_seen": 291504128 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046043129388164495, + "loss": 1.6478, + "theoretical_loss": 4.1888116464357035, + "tokens_seen": 291569664 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 114979, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6656078100204468, + "objective/train/theoretical_loss": 4.188701434150834, + "objective/train/tokens_used": 291635200, + "theoretical_loss": 4.188701434150834, + "tokens_seen": 291635200 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046042126379137414, + "loss": 1.6678, + "theoretical_loss": 4.188701434150834, + "tokens_seen": 291635200 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004604112337011033, + "loss": 1.6758, + "theoretical_loss": 4.188591253562918, + "tokens_seen": 291700736 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004604012036108325, + "loss": 1.6705, + "theoretical_loss": 4.18848110465572, + "tokens_seen": 291766272 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004603911735205617, + "loss": 1.591, + "theoretical_loss": 4.188370987413018, + "tokens_seen": 291831808 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046038114343029086, + "loss": 1.5852, + "theoretical_loss": 4.188260901818602, + "tokens_seen": 291897344 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004603711133400201, + "loss": 1.5966, + "theoretical_loss": 4.1881508478562735, + "tokens_seen": 291962880 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004603610832497492, + "loss": 1.6114, + "theoretical_loss": 4.188040825509845, + "tokens_seen": 292028416 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046035105315947846, + "loss": 1.6468, + "theoretical_loss": 4.1879308347631445, + "tokens_seen": 292093952 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004603410230692076, + "loss": 1.609, + "theoretical_loss": 4.187820875600006, + "tokens_seen": 292159488 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004603309929789368, + "loss": 1.6591, + "theoretical_loss": 4.187710948004281, + "tokens_seen": 292225024 + }, + { + "epoch": 0.09, + "learning_rate": 0.000460320962888666, + "loss": 1.7203, + "theoretical_loss": 4.187601051959831, + "tokens_seen": 292290560 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004603109327983952, + "loss": 1.7449, + "theoretical_loss": 4.187491187450527, + "tokens_seen": 292356096 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046030090270812436, + "loss": 1.6597, + "theoretical_loss": 4.187381354460255, + "tokens_seen": 292421632 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004602908726178536, + "loss": 1.716, + "theoretical_loss": 4.187271552972911, + "tokens_seen": 292487168 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004602808425275827, + "loss": 1.6562, + "theoretical_loss": 4.187161782972405, + "tokens_seen": 292552704 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046027081243731196, + "loss": 1.739, + "theoretical_loss": 4.187052044442655, + "tokens_seen": 292618240 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004602607823470411, + "loss": 1.6258, + "theoretical_loss": 4.1869423373675945, + "tokens_seen": 292683776 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004602507522567703, + "loss": 1.698, + "theoretical_loss": 4.1868326617311675, + "tokens_seen": 292749312 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004602407221664995, + "loss": 1.6381, + "theoretical_loss": 4.18672301751733, + "tokens_seen": 292814848 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004602306920762287, + "loss": 1.6397, + "theoretical_loss": 4.186613404710048, + "tokens_seen": 292880384 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046022066198595787, + "loss": 1.6686, + "theoretical_loss": 4.186503823293303, + "tokens_seen": 292945920 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046021063189568705, + "loss": 1.7061, + "theoretical_loss": 4.186394273251084, + "tokens_seen": 293011456 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004602006018054163, + "loss": 1.7329, + "theoretical_loss": 4.186284754567395, + "tokens_seen": 293076992 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046019057171514546, + "loss": 1.6713, + "theoretical_loss": 4.1861752672262496, + "tokens_seen": 293142528 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046018054162487465, + "loss": 1.6553, + "theoretical_loss": 4.186065811211675, + "tokens_seen": 293208064 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004601705115346038, + "loss": 1.6219, + "theoretical_loss": 4.185956386507708, + "tokens_seen": 293273600 + }, + { + "epoch": 0.09, + "learning_rate": 0.000460160481444333, + "loss": 1.6075, + "theoretical_loss": 4.185846993098398, + "tokens_seen": 293339136 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004601504513540622, + "loss": 1.625, + "theoretical_loss": 4.185737630967808, + "tokens_seen": 293404672 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004601404212637914, + "loss": 1.6952, + "theoretical_loss": 4.18562830010001, + "tokens_seen": 293470208 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046013039117352055, + "loss": 1.7106, + "theoretical_loss": 4.185519000479089, + "tokens_seen": 293535744 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004601203610832498, + "loss": 1.5772, + "theoretical_loss": 4.18540973208914, + "tokens_seen": 293601280 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046011033099297897, + "loss": 1.7055, + "theoretical_loss": 4.185300494914273, + "tokens_seen": 293666816 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046010030090270815, + "loss": 1.5678, + "theoretical_loss": 4.185191288938606, + "tokens_seen": 293732352 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046009027081243733, + "loss": 1.6545, + "theoretical_loss": 4.185082114146271, + "tokens_seen": 293797888 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004600802407221665, + "loss": 1.7297, + "theoretical_loss": 4.184972970521409, + "tokens_seen": 293863424 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004600702106318957, + "loss": 1.5722, + "theoretical_loss": 4.184863858048177, + "tokens_seen": 293928960 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046006018054162493, + "loss": 1.6774, + "theoretical_loss": 4.18475477671074, + "tokens_seen": 293994496 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046005015045135405, + "loss": 1.6837, + "theoretical_loss": 4.184645726493275, + "tokens_seen": 294060032 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004600401203610833, + "loss": 1.5158, + "theoretical_loss": 4.184536707379972, + "tokens_seen": 294125568 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004600300902708124, + "loss": 1.6246, + "theoretical_loss": 4.184427719355031, + "tokens_seen": 294191104 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046002006018054165, + "loss": 1.6649, + "theoretical_loss": 4.184318762402664, + "tokens_seen": 294256640 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046001003009027083, + "loss": 1.5912, + "theoretical_loss": 4.184209836507096, + "tokens_seen": 294322176 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046, + "loss": 1.6861, + "theoretical_loss": 4.18410094165256, + "tokens_seen": 294387712 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004599899699097292, + "loss": 1.5414, + "theoretical_loss": 4.183992077823305, + "tokens_seen": 294453248 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045997993981945843, + "loss": 1.6137, + "theoretical_loss": 4.183883245003588, + "tokens_seen": 294518784 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045996990972918756, + "loss": 1.6096, + "theoretical_loss": 4.183774443177679, + "tokens_seen": 294584320 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004599598796389168, + "loss": 1.6523, + "theoretical_loss": 4.183665672329861, + "tokens_seen": 294649856 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004599498495486459, + "loss": 1.5347, + "theoretical_loss": 4.183556932444424, + "tokens_seen": 294715392 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045993981945837515, + "loss": 1.6186, + "theoretical_loss": 4.183448223505673, + "tokens_seen": 294780928 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045992978936810434, + "loss": 1.7213, + "theoretical_loss": 4.183339545497924, + "tokens_seen": 294846464 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 116257, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7284107208251953, + "objective/train/theoretical_loss": 4.1832308984055055, + "objective/train/tokens_used": 294912000, + "theoretical_loss": 4.1832308984055055, + "tokens_seen": 294912000 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004599197592778335, + "loss": 1.6852, + "theoretical_loss": 4.1832308984055055, + "tokens_seen": 294912000 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004599097291875627, + "loss": 1.5887, + "theoretical_loss": 4.183122282212754, + "tokens_seen": 294977536 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004598996990972919, + "loss": 1.6999, + "theoretical_loss": 4.183013696904018, + "tokens_seen": 295043072 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045988966900702106, + "loss": 1.6029, + "theoretical_loss": 4.182905142463662, + "tokens_seen": 295108608 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004598796389167503, + "loss": 1.4938, + "theoretical_loss": 4.182796618876057, + "tokens_seen": 295174144 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004598696088264794, + "loss": 1.7283, + "theoretical_loss": 4.182688126125587, + "tokens_seen": 295239680 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045985957873620866, + "loss": 1.6662, + "theoretical_loss": 4.182579664196648, + "tokens_seen": 295305216 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004598495486459378, + "loss": 1.657, + "theoretical_loss": 4.182471233073647, + "tokens_seen": 295370752 + }, + { + "epoch": 0.09, + "learning_rate": 0.000459839518555667, + "loss": 1.5484, + "theoretical_loss": 4.182362832741001, + "tokens_seen": 295436288 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004598294884653962, + "loss": 1.6432, + "theoretical_loss": 4.182254463183139, + "tokens_seen": 295501824 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004598194583751254, + "loss": 1.6776, + "theoretical_loss": 4.182146124384504, + "tokens_seen": 295567360 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045980942828485456, + "loss": 1.6899, + "theoretical_loss": 4.182037816329547, + "tokens_seen": 295632896 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004597993981945838, + "loss": 1.6236, + "theoretical_loss": 4.181929539002731, + "tokens_seen": 295698432 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004597893681043129, + "loss": 1.7164, + "theoretical_loss": 4.1818212923885305, + "tokens_seen": 295763968 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045977933801404216, + "loss": 1.6835, + "theoretical_loss": 4.181713076471432, + "tokens_seen": 295829504 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004597693079237713, + "loss": 1.5767, + "theoretical_loss": 4.181604891235933, + "tokens_seen": 295895040 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004597592778335005, + "loss": 1.5968, + "theoretical_loss": 4.181496736666542, + "tokens_seen": 295960576 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004597492477432297, + "loss": 1.7113, + "theoretical_loss": 4.181388612747778, + "tokens_seen": 296026112 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004597392176529589, + "loss": 1.6237, + "theoretical_loss": 4.181280519464172, + "tokens_seen": 296091648 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045972918756268807, + "loss": 1.5792, + "theoretical_loss": 4.181172456800267, + "tokens_seen": 296157184 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045971915747241725, + "loss": 1.6714, + "theoretical_loss": 4.181064424740617, + "tokens_seen": 296222720 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045970912738214643, + "loss": 1.5484, + "theoretical_loss": 4.180956423269784, + "tokens_seen": 296288256 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045969909729187566, + "loss": 1.581, + "theoretical_loss": 4.180848452372347, + "tokens_seen": 296353792 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004596890672016048, + "loss": 1.6687, + "theoretical_loss": 4.180740512032891, + "tokens_seen": 296419328 + }, + { + "epoch": 0.09, + "learning_rate": 0.000459679037111334, + "loss": 1.6364, + "theoretical_loss": 4.180632602236014, + "tokens_seen": 296484864 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045966900702106315, + "loss": 1.5334, + "theoretical_loss": 4.180524722966327, + "tokens_seen": 296550400 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004596589769307924, + "loss": 1.6823, + "theoretical_loss": 4.180416874208449, + "tokens_seen": 296615936 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045964894684052157, + "loss": 1.6341, + "theoretical_loss": 4.180309055947012, + "tokens_seen": 296681472 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045963891675025075, + "loss": 1.6797, + "theoretical_loss": 4.18020126816666, + "tokens_seen": 296747008 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045962888665997993, + "loss": 1.6833, + "theoretical_loss": 4.180093510852045, + "tokens_seen": 296812544 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045961885656970917, + "loss": 1.7263, + "theoretical_loss": 4.179985783987834, + "tokens_seen": 296878080 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004596088264794383, + "loss": 1.6393, + "theoretical_loss": 4.179878087558702, + "tokens_seen": 296943616 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045959879638916753, + "loss": 1.5704, + "theoretical_loss": 4.179770421549336, + "tokens_seen": 297009152 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045958876629889666, + "loss": 1.6592, + "theoretical_loss": 4.179662785944435, + "tokens_seen": 297074688 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004595787362086259, + "loss": 1.5107, + "theoretical_loss": 4.179555180728708, + "tokens_seen": 297140224 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004595687061183551, + "loss": 1.6687, + "theoretical_loss": 4.179447605886876, + "tokens_seen": 297205760 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045955867602808425, + "loss": 1.6509, + "theoretical_loss": 4.1793400614036695, + "tokens_seen": 297271296 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045954864593781344, + "loss": 1.6868, + "theoretical_loss": 4.179232547263831, + "tokens_seen": 297336832 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004595386158475426, + "loss": 1.6702, + "theoretical_loss": 4.1791250634521155, + "tokens_seen": 297402368 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004595285857572718, + "loss": 1.7591, + "theoretical_loss": 4.179017609953287, + "tokens_seen": 297467904 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045951855566700103, + "loss": 1.6884, + "theoretical_loss": 4.17891018675212, + "tokens_seen": 297533440 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045950852557673016, + "loss": 1.7116, + "theoretical_loss": 4.178802793833402, + "tokens_seen": 297598976 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004594984954864594, + "loss": 1.5065, + "theoretical_loss": 4.17869543118193, + "tokens_seen": 297664512 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004594884653961885, + "loss": 1.6582, + "theoretical_loss": 4.178588098782513, + "tokens_seen": 297730048 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045947843530591776, + "loss": 1.5576, + "theoretical_loss": 4.178480796619971, + "tokens_seen": 297795584 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045946840521564694, + "loss": 1.7219, + "theoretical_loss": 4.178373524679134, + "tokens_seen": 297861120 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004594583751253761, + "loss": 1.6977, + "theoretical_loss": 4.1782662829448425, + "tokens_seen": 297926656 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045944834503510536, + "loss": 1.6702, + "theoretical_loss": 4.17815907140195, + "tokens_seen": 297992192 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045943831494483454, + "loss": 1.5838, + "theoretical_loss": 4.178051890035319, + "tokens_seen": 298057728 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004594282848545637, + "loss": 1.6394, + "theoretical_loss": 4.177944738829825, + "tokens_seen": 298123264 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 117476, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.532116413116455, + "objective/train/theoretical_loss": 4.177837617770352, + "objective/train/tokens_used": 298188800, + "theoretical_loss": 4.177837617770352, + "tokens_seen": 298188800 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004594182547642929, + "loss": 1.5819, + "theoretical_loss": 4.177837617770352, + "tokens_seen": 298188800 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004594082246740221, + "loss": 1.7609, + "theoretical_loss": 4.1777305268417955, + "tokens_seen": 298254336 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045939819458375126, + "loss": 1.5927, + "theoretical_loss": 4.177623466029064, + "tokens_seen": 298319872 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004593881644934805, + "loss": 1.5925, + "theoretical_loss": 4.177516435317074, + "tokens_seen": 298385408 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004593781344032096, + "loss": 1.6878, + "theoretical_loss": 4.177409434690754, + "tokens_seen": 298450944 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045936810431293886, + "loss": 1.6294, + "theoretical_loss": 4.177302464135044, + "tokens_seen": 298516480 + }, + { + "epoch": 0.09, + "learning_rate": 0.000459358074222668, + "loss": 1.6035, + "theoretical_loss": 4.177195523634895, + "tokens_seen": 298582016 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004593480441323972, + "loss": 1.6935, + "theoretical_loss": 4.1770886131752665, + "tokens_seen": 298647552 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004593380140421264, + "loss": 1.6128, + "theoretical_loss": 4.176981732741131, + "tokens_seen": 298713088 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004593279839518556, + "loss": 1.7059, + "theoretical_loss": 4.176874882317472, + "tokens_seen": 298778624 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045931795386158476, + "loss": 1.5656, + "theoretical_loss": 4.176768061889283, + "tokens_seen": 298844160 + }, + { + "epoch": 0.09, + "learning_rate": 0.000459307923771314, + "loss": 1.6349, + "theoretical_loss": 4.176661271441568, + "tokens_seen": 298909696 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004592978936810431, + "loss": 1.6502, + "theoretical_loss": 4.176554510959342, + "tokens_seen": 298975232 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045928786359077236, + "loss": 1.7652, + "theoretical_loss": 4.1764477804276305, + "tokens_seen": 299040768 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004592778335005015, + "loss": 1.5318, + "theoretical_loss": 4.176341079831471, + "tokens_seen": 299106304 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004592678034102307, + "loss": 1.7183, + "theoretical_loss": 4.176234409155912, + "tokens_seen": 299171840 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004592577733199599, + "loss": 1.6861, + "theoretical_loss": 4.176127768386008, + "tokens_seen": 299237376 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004592477432296891, + "loss": 1.5711, + "theoretical_loss": 4.176021157506832, + "tokens_seen": 299302912 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045923771313941827, + "loss": 1.6442, + "theoretical_loss": 4.1759145765034615, + "tokens_seen": 299368448 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045922768304914745, + "loss": 1.673, + "theoretical_loss": 4.175808025360988, + "tokens_seen": 299433984 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045921765295887663, + "loss": 1.6553, + "theoretical_loss": 4.175701504064511, + "tokens_seen": 299499520 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045920762286860586, + "loss": 1.6059, + "theoretical_loss": 4.175595012599143, + "tokens_seen": 299565056 + }, + { + "epoch": 0.09, + "learning_rate": 0.000459197592778335, + "loss": 1.6531, + "theoretical_loss": 4.175488550950007, + "tokens_seen": 299630592 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004591875626880642, + "loss": 1.6394, + "theoretical_loss": 4.175382119102235, + "tokens_seen": 299696128 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045917753259779335, + "loss": 1.7436, + "theoretical_loss": 4.175275717040972, + "tokens_seen": 299761664 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004591675025075226, + "loss": 1.5948, + "theoretical_loss": 4.175169344751371, + "tokens_seen": 299827200 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045915747241725177, + "loss": 1.7066, + "theoretical_loss": 4.175063002218598, + "tokens_seen": 299892736 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045914744232698095, + "loss": 1.6168, + "theoretical_loss": 4.174956689427829, + "tokens_seen": 299958272 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045913741223671013, + "loss": 1.6259, + "theoretical_loss": 4.17485040636425, + "tokens_seen": 300023808 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045912738214643937, + "loss": 1.6589, + "theoretical_loss": 4.174744153013057, + "tokens_seen": 300089344 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004591173520561685, + "loss": 1.6104, + "theoretical_loss": 4.174637929359458, + "tokens_seen": 300154880 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045910732196589773, + "loss": 1.6799, + "theoretical_loss": 4.174531735388673, + "tokens_seen": 300220416 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045909729187562686, + "loss": 1.5264, + "theoretical_loss": 4.174425571085927, + "tokens_seen": 300285952 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004590872617853561, + "loss": 1.6579, + "theoretical_loss": 4.174319436436462, + "tokens_seen": 300351488 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004590772316950853, + "loss": 1.6112, + "theoretical_loss": 4.174213331425527, + "tokens_seen": 300417024 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045906720160481445, + "loss": 1.5842, + "theoretical_loss": 4.174107256038383, + "tokens_seen": 300482560 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045905717151454364, + "loss": 1.5477, + "theoretical_loss": 4.1740012102603, + "tokens_seen": 300548096 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004590471414242728, + "loss": 1.5508, + "theoretical_loss": 4.17389519407656, + "tokens_seen": 300613632 + }, + { + "epoch": 0.09, + "learning_rate": 0.000459037111334002, + "loss": 1.6588, + "theoretical_loss": 4.173789207472455, + "tokens_seen": 300679168 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045902708124373123, + "loss": 1.7207, + "theoretical_loss": 4.173683250433287, + "tokens_seen": 300744704 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045901705115346036, + "loss": 1.7324, + "theoretical_loss": 4.17357732294437, + "tokens_seen": 300810240 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004590070210631896, + "loss": 1.6289, + "theoretical_loss": 4.1734714249910265, + "tokens_seen": 300875776 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004589969909729187, + "loss": 1.4539, + "theoretical_loss": 4.173365556558591, + "tokens_seen": 300941312 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045898696088264796, + "loss": 1.6936, + "theoretical_loss": 4.1732597176324076, + "tokens_seen": 301006848 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045897693079237714, + "loss": 1.6817, + "theoretical_loss": 4.173153908197831, + "tokens_seen": 301072384 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004589669007021063, + "loss": 1.5554, + "theoretical_loss": 4.173048128240227, + "tokens_seen": 301137920 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004589568706118355, + "loss": 1.6182, + "theoretical_loss": 4.172942377744972, + "tokens_seen": 301203456 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045894684052156474, + "loss": 1.6058, + "theoretical_loss": 4.17283665669745, + "tokens_seen": 301268992 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045893681043129386, + "loss": 1.6311, + "theoretical_loss": 4.172730965083061, + "tokens_seen": 301334528 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004589267803410231, + "loss": 1.7419, + "theoretical_loss": 4.1726253028872105, + "tokens_seen": 301400064 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 118641, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5080492496490479, + "objective/train/theoretical_loss": 4.172519670095315, + "objective/train/tokens_used": 301465600, + "theoretical_loss": 4.172519670095315, + "tokens_seen": 301465600 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004589167502507522, + "loss": 1.5165, + "theoretical_loss": 4.172519670095315, + "tokens_seen": 301465600 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045890672016048146, + "loss": 1.7492, + "theoretical_loss": 4.172414066692804, + "tokens_seen": 301531136 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045889669007021064, + "loss": 1.6785, + "theoretical_loss": 4.172308492665114, + "tokens_seen": 301596672 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004588866599799398, + "loss": 1.691, + "theoretical_loss": 4.172202947997695, + "tokens_seen": 301662208 + }, + { + "epoch": 0.09, + "learning_rate": 0.000458876629889669, + "loss": 1.5723, + "theoretical_loss": 4.172097432676006, + "tokens_seen": 301727744 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004588665997993982, + "loss": 1.592, + "theoretical_loss": 4.171991946685516, + "tokens_seen": 301793280 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045885656970912737, + "loss": 1.6757, + "theoretical_loss": 4.171886490011705, + "tokens_seen": 301858816 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004588465396188566, + "loss": 1.6972, + "theoretical_loss": 4.171781062640063, + "tokens_seen": 301924352 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045883650952858573, + "loss": 1.7607, + "theoretical_loss": 4.17167566455609, + "tokens_seen": 301989888 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045882647943831496, + "loss": 1.6218, + "theoretical_loss": 4.171570295745297, + "tokens_seen": 302055424 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045881644934804415, + "loss": 1.6333, + "theoretical_loss": 4.171464956193205, + "tokens_seen": 302120960 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004588064192577733, + "loss": 1.7224, + "theoretical_loss": 4.171359645885346, + "tokens_seen": 302186496 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004587963891675025, + "loss": 1.6172, + "theoretical_loss": 4.17125436480726, + "tokens_seen": 302252032 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004587863590772317, + "loss": 1.5661, + "theoretical_loss": 4.171149112944501, + "tokens_seen": 302317568 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045877632898696087, + "loss": 1.672, + "theoretical_loss": 4.171043890282629, + "tokens_seen": 302383104 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004587662988966901, + "loss": 1.6253, + "theoretical_loss": 4.170938696807218, + "tokens_seen": 302448640 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045875626880641923, + "loss": 1.5491, + "theoretical_loss": 4.17083353250385, + "tokens_seen": 302514176 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045874623871614847, + "loss": 1.629, + "theoretical_loss": 4.170728397358118, + "tokens_seen": 302579712 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004587362086258776, + "loss": 1.7158, + "theoretical_loss": 4.170623291355626, + "tokens_seen": 302645248 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045872617853560683, + "loss": 1.6165, + "theoretical_loss": 4.1705182144819855, + "tokens_seen": 302710784 + }, + { + "epoch": 0.09, + "learning_rate": 0.000458716148445336, + "loss": 1.609, + "theoretical_loss": 4.170413166722822, + "tokens_seen": 302776320 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004587061183550652, + "loss": 1.7342, + "theoretical_loss": 4.170308148063769, + "tokens_seen": 302841856 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004586960882647944, + "loss": 1.7079, + "theoretical_loss": 4.1702031584904695, + "tokens_seen": 302907392 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045868605817452355, + "loss": 1.5792, + "theoretical_loss": 4.170098197988578, + "tokens_seen": 302972928 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004586760280842528, + "loss": 1.6964, + "theoretical_loss": 4.169993266543759, + "tokens_seen": 303038464 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045866599799398197, + "loss": 1.6971, + "theoretical_loss": 4.169888364141689, + "tokens_seen": 303104000 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045865596790371115, + "loss": 1.6229, + "theoretical_loss": 4.169783490768049, + "tokens_seen": 303169536 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045864593781344033, + "loss": 1.7269, + "theoretical_loss": 4.169678646408536, + "tokens_seen": 303235072 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045863590772316957, + "loss": 1.5735, + "theoretical_loss": 4.169573831048854, + "tokens_seen": 303300608 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004586258776328987, + "loss": 1.6174, + "theoretical_loss": 4.16946904467472, + "tokens_seen": 303366144 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045861584754262793, + "loss": 1.6615, + "theoretical_loss": 4.169364287271858, + "tokens_seen": 303431680 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045860581745235706, + "loss": 1.6648, + "theoretical_loss": 4.169259558826003, + "tokens_seen": 303497216 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004585957873620863, + "loss": 1.7775, + "theoretical_loss": 4.169154859322901, + "tokens_seen": 303562752 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004585857572718155, + "loss": 1.6608, + "theoretical_loss": 4.169050188748307, + "tokens_seen": 303628288 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045857572718154465, + "loss": 1.6266, + "theoretical_loss": 4.168945547087987, + "tokens_seen": 303693824 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045856569709127384, + "loss": 1.6521, + "theoretical_loss": 4.168840934327717, + "tokens_seen": 303759360 + }, + { + "epoch": 0.09, + "learning_rate": 0.000458555667001003, + "loss": 1.7355, + "theoretical_loss": 4.1687363504532815, + "tokens_seen": 303824896 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004585456369107322, + "loss": 1.5736, + "theoretical_loss": 4.168631795450478, + "tokens_seen": 303890432 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045853560682046143, + "loss": 1.7311, + "theoretical_loss": 4.168527269305112, + "tokens_seen": 303955968 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045852557673019056, + "loss": 1.6627, + "theoretical_loss": 4.168422772002998, + "tokens_seen": 304021504 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004585155466399198, + "loss": 1.6325, + "theoretical_loss": 4.168318303529963, + "tokens_seen": 304087040 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004585055165496489, + "loss": 1.5595, + "theoretical_loss": 4.168213863871843, + "tokens_seen": 304152576 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045849548645937816, + "loss": 1.6241, + "theoretical_loss": 4.168109453014484, + "tokens_seen": 304218112 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045848545636910734, + "loss": 1.6467, + "theoretical_loss": 4.1680050709437415, + "tokens_seen": 304283648 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004584754262788365, + "loss": 1.6596, + "theoretical_loss": 4.167900717645482, + "tokens_seen": 304349184 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004584653961885657, + "loss": 1.6297, + "theoretical_loss": 4.16779639310558, + "tokens_seen": 304414720 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045845536609829494, + "loss": 1.6991, + "theoretical_loss": 4.167692097309923, + "tokens_seen": 304480256 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045844533600802406, + "loss": 1.5692, + "theoretical_loss": 4.167587830244408, + "tokens_seen": 304545792 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004584353059177533, + "loss": 1.6499, + "theoretical_loss": 4.167483591894937, + "tokens_seen": 304611328 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004584252758274824, + "loss": 1.6293, + "theoretical_loss": 4.167379382247429, + "tokens_seen": 304676864 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 119817, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6829400062561035, + "objective/train/theoretical_loss": 4.167275201287808, + "objective/train/tokens_used": 304742400, + "theoretical_loss": 4.167275201287808, + "tokens_seen": 304742400 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045841524573721166, + "loss": 1.7078, + "theoretical_loss": 4.167275201287808, + "tokens_seen": 304742400 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045840521564694084, + "loss": 1.5721, + "theoretical_loss": 4.16717104900201, + "tokens_seen": 304807936 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045839518555667, + "loss": 1.5925, + "theoretical_loss": 4.167066925375981, + "tokens_seen": 304873472 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004583851554663992, + "loss": 1.5842, + "theoretical_loss": 4.1669628303956765, + "tokens_seen": 304939008 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004583751253761284, + "loss": 1.6002, + "theoretical_loss": 4.166858764047062, + "tokens_seen": 305004544 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045836509528585757, + "loss": 1.6123, + "theoretical_loss": 4.166754726316111, + "tokens_seen": 305070080 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004583550651955868, + "loss": 1.6112, + "theoretical_loss": 4.16665071718881, + "tokens_seen": 305135616 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045834503510531593, + "loss": 1.6073, + "theoretical_loss": 4.1665467366511555, + "tokens_seen": 305201152 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045833500501504516, + "loss": 1.6741, + "theoretical_loss": 4.166442784689149, + "tokens_seen": 305266688 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045832497492477435, + "loss": 1.6874, + "theoretical_loss": 4.166338861288808, + "tokens_seen": 305332224 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004583149448345035, + "loss": 1.6084, + "theoretical_loss": 4.166234966436155, + "tokens_seen": 305397760 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004583049147442327, + "loss": 1.6354, + "theoretical_loss": 4.166131100117226, + "tokens_seen": 305463296 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004582948846539619, + "loss": 1.705, + "theoretical_loss": 4.166027262318064, + "tokens_seen": 305528832 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045828485456369107, + "loss": 1.5271, + "theoretical_loss": 4.165923453024724, + "tokens_seen": 305594368 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004582748244734203, + "loss": 1.6939, + "theoretical_loss": 4.165819672223269, + "tokens_seen": 305659904 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045826479438314943, + "loss": 1.6673, + "theoretical_loss": 4.165715919899773, + "tokens_seen": 305725440 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045825476429287867, + "loss": 1.5994, + "theoretical_loss": 4.165612196040319, + "tokens_seen": 305790976 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004582447342026078, + "loss": 1.7809, + "theoretical_loss": 4.1655085006310015, + "tokens_seen": 305856512 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045823470411233703, + "loss": 1.5933, + "theoretical_loss": 4.165404833657923, + "tokens_seen": 305922048 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004582246740220662, + "loss": 1.7069, + "theoretical_loss": 4.165301195107195, + "tokens_seen": 305987584 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004582146439317954, + "loss": 1.5783, + "theoretical_loss": 4.1651975849649405, + "tokens_seen": 306053120 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045820461384152457, + "loss": 1.6189, + "theoretical_loss": 4.165094003217293, + "tokens_seen": 306118656 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045819458375125375, + "loss": 1.7249, + "theoretical_loss": 4.164990449850393, + "tokens_seen": 306184192 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045818455366098294, + "loss": 1.7461, + "theoretical_loss": 4.164886924850393, + "tokens_seen": 306249728 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045817452357071217, + "loss": 1.5778, + "theoretical_loss": 4.164783428203453, + "tokens_seen": 306315264 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004581644934804413, + "loss": 1.6641, + "theoretical_loss": 4.1646799598957465, + "tokens_seen": 306380800 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045815446339017053, + "loss": 1.6359, + "theoretical_loss": 4.164576519913453, + "tokens_seen": 306446336 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004581444332998997, + "loss": 1.6613, + "theoretical_loss": 4.164473108242763, + "tokens_seen": 306511872 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004581344032096289, + "loss": 1.7131, + "theoretical_loss": 4.164369724869878, + "tokens_seen": 306577408 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004581243731193581, + "loss": 1.6846, + "theoretical_loss": 4.164266369781005, + "tokens_seen": 306642944 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045811434302908726, + "loss": 1.6136, + "theoretical_loss": 4.164163042962366, + "tokens_seen": 306708480 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045810431293881644, + "loss": 1.6647, + "theoretical_loss": 4.1640597444001894, + "tokens_seen": 306774016 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004580942828485457, + "loss": 1.6518, + "theoretical_loss": 4.163956474080715, + "tokens_seen": 306839552 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004580842527582748, + "loss": 1.6812, + "theoretical_loss": 4.16385323199019, + "tokens_seen": 306905088 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045807422266800404, + "loss": 1.6442, + "theoretical_loss": 4.1637500181148726, + "tokens_seen": 306970624 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045806419257773316, + "loss": 1.6653, + "theoretical_loss": 4.163646832441031, + "tokens_seen": 307036160 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004580541624874624, + "loss": 1.6706, + "theoretical_loss": 4.163543674954942, + "tokens_seen": 307101696 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004580441323971916, + "loss": 1.6486, + "theoretical_loss": 4.163440545642894, + "tokens_seen": 307167232 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045803410230692076, + "loss": 1.6707, + "theoretical_loss": 4.163337444491183, + "tokens_seen": 307232768 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045802407221664994, + "loss": 1.6056, + "theoretical_loss": 4.163234371486114, + "tokens_seen": 307298304 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004580140421263791, + "loss": 1.5561, + "theoretical_loss": 4.163131326614003, + "tokens_seen": 307363840 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004580040120361083, + "loss": 1.6937, + "theoretical_loss": 4.163028309861176, + "tokens_seen": 307429376 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045799398194583754, + "loss": 1.6514, + "theoretical_loss": 4.162925321213968, + "tokens_seen": 307494912 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045798395185556667, + "loss": 1.6001, + "theoretical_loss": 4.162822360658723, + "tokens_seen": 307560448 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004579739217652959, + "loss": 1.6873, + "theoretical_loss": 4.162719428181794, + "tokens_seen": 307625984 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004579638916750251, + "loss": 1.6563, + "theoretical_loss": 4.162616523769547, + "tokens_seen": 307691520 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045795386158475426, + "loss": 1.5855, + "theoretical_loss": 4.1625136474083515, + "tokens_seen": 307757056 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004579438314944835, + "loss": 1.6838, + "theoretical_loss": 4.1624107990845935, + "tokens_seen": 307822592 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004579338014042126, + "loss": 1.6242, + "theoretical_loss": 4.162307978784662, + "tokens_seen": 307888128 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045792377131394186, + "loss": 1.626, + "theoretical_loss": 4.162205186494962, + "tokens_seen": 307953664 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 120992, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5657944679260254, + "objective/train/theoretical_loss": 4.162102422201901, + "objective/train/tokens_used": 308019200, + "theoretical_loss": 4.162102422201901, + "tokens_seen": 308019200 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045791374122367104, + "loss": 1.525, + "theoretical_loss": 4.162102422201901, + "tokens_seen": 308019200 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004579037111334002, + "loss": 1.7519, + "theoretical_loss": 4.161999685891901, + "tokens_seen": 308084736 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004578936810431294, + "loss": 1.5885, + "theoretical_loss": 4.161896977551393, + "tokens_seen": 308150272 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004578836509528586, + "loss": 1.6268, + "theoretical_loss": 4.161794297166815, + "tokens_seen": 308215808 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045787362086258777, + "loss": 1.6833, + "theoretical_loss": 4.161691644724616, + "tokens_seen": 308281344 + }, + { + "epoch": 0.09, + "learning_rate": 0.000457863590772317, + "loss": 1.4875, + "theoretical_loss": 4.161589020211256, + "tokens_seen": 308346880 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045785356068204613, + "loss": 1.6642, + "theoretical_loss": 4.1614864236132, + "tokens_seen": 308412416 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045784353059177536, + "loss": 1.6967, + "theoretical_loss": 4.161383854916926, + "tokens_seen": 308477952 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045783350050150455, + "loss": 1.6211, + "theoretical_loss": 4.161281314108923, + "tokens_seen": 308543488 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004578234704112337, + "loss": 1.5726, + "theoretical_loss": 4.1611788011756845, + "tokens_seen": 308609024 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004578134403209629, + "loss": 1.5628, + "theoretical_loss": 4.161076316103717, + "tokens_seen": 308674560 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004578034102306921, + "loss": 1.6727, + "theoretical_loss": 4.160973858879535, + "tokens_seen": 308740096 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045779338014042127, + "loss": 1.5475, + "theoretical_loss": 4.160871429489663, + "tokens_seen": 308805632 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004577833500501505, + "loss": 1.5975, + "theoretical_loss": 4.160769027920635, + "tokens_seen": 308871168 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045777331995987963, + "loss": 1.7357, + "theoretical_loss": 4.160666654158993, + "tokens_seen": 308936704 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045776328986960887, + "loss": 1.6982, + "theoretical_loss": 4.160564308191289, + "tokens_seen": 309002240 + }, + { + "epoch": 0.09, + "learning_rate": 0.000457753259779338, + "loss": 1.6087, + "theoretical_loss": 4.1604619900040865, + "tokens_seen": 309067776 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045774322968906723, + "loss": 1.6312, + "theoretical_loss": 4.160359699583955, + "tokens_seen": 309133312 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004577331995987964, + "loss": 1.5883, + "theoretical_loss": 4.160257436917476, + "tokens_seen": 309198848 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004577231695085256, + "loss": 1.589, + "theoretical_loss": 4.160155201991238, + "tokens_seen": 309264384 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004577131394182548, + "loss": 1.5793, + "theoretical_loss": 4.160052994791841, + "tokens_seen": 309329920 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045770310932798395, + "loss": 1.6709, + "theoretical_loss": 4.159950815305893, + "tokens_seen": 309395456 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045769307923771314, + "loss": 1.726, + "theoretical_loss": 4.1598486635200125, + "tokens_seen": 309460992 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045768304914744237, + "loss": 1.6151, + "theoretical_loss": 4.159746539420825, + "tokens_seen": 309526528 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004576730190571715, + "loss": 1.5442, + "theoretical_loss": 4.159644442994967, + "tokens_seen": 309592064 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045766298896690073, + "loss": 1.7418, + "theoretical_loss": 4.159542374229084, + "tokens_seen": 309657600 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004576529588766299, + "loss": 1.5619, + "theoretical_loss": 4.159440333109832, + "tokens_seen": 309723136 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004576429287863591, + "loss": 1.7123, + "theoretical_loss": 4.159338319623874, + "tokens_seen": 309788672 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004576328986960883, + "loss": 1.6477, + "theoretical_loss": 4.1592363337578835, + "tokens_seen": 309854208 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045762286860581746, + "loss": 1.6097, + "theoretical_loss": 4.159134375498542, + "tokens_seen": 309919744 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045761283851554664, + "loss": 1.5044, + "theoretical_loss": 4.159032444832543, + "tokens_seen": 309985280 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004576028084252759, + "loss": 1.633, + "theoretical_loss": 4.158930541746587, + "tokens_seen": 310050816 + }, + { + "epoch": 0.09, + "learning_rate": 0.000457592778335005, + "loss": 1.665, + "theoretical_loss": 4.158828666227383, + "tokens_seen": 310116352 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045758274824473424, + "loss": 1.6462, + "theoretical_loss": 4.158726818261652, + "tokens_seen": 310181888 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045757271815446336, + "loss": 1.6558, + "theoretical_loss": 4.158624997836123, + "tokens_seen": 310247424 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004575626880641926, + "loss": 1.7397, + "theoretical_loss": 4.158523204937531, + "tokens_seen": 310312960 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004575526579739218, + "loss": 1.6543, + "theoretical_loss": 4.158421439552626, + "tokens_seen": 310378496 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045754262788365096, + "loss": 1.6402, + "theoretical_loss": 4.158319701668162, + "tokens_seen": 310444032 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045753259779338014, + "loss": 1.636, + "theoretical_loss": 4.158217991270906, + "tokens_seen": 310509568 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004575225677031093, + "loss": 1.8042, + "theoretical_loss": 4.158116308347632, + "tokens_seen": 310575104 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004575125376128385, + "loss": 1.7418, + "theoretical_loss": 4.1580146528851225, + "tokens_seen": 310640640 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045750250752256774, + "loss": 1.5758, + "theoretical_loss": 4.157913024870172, + "tokens_seen": 310706176 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045749247743229687, + "loss": 1.6217, + "theoretical_loss": 4.157811424289582, + "tokens_seen": 310771712 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004574824473420261, + "loss": 1.6949, + "theoretical_loss": 4.1577098511301624, + "tokens_seen": 310837248 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004574724172517553, + "loss": 1.63, + "theoretical_loss": 4.157608305378735, + "tokens_seen": 310902784 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045746238716148446, + "loss": 1.7518, + "theoretical_loss": 4.157506787022128, + "tokens_seen": 310968320 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045745235707121364, + "loss": 1.6203, + "theoretical_loss": 4.15740529604718, + "tokens_seen": 311033856 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004574423269809428, + "loss": 1.5373, + "theoretical_loss": 4.15730383244074, + "tokens_seen": 311099392 + }, + { + "epoch": 0.09, + "learning_rate": 0.000457432296890672, + "loss": 1.6087, + "theoretical_loss": 4.157202396189662, + "tokens_seen": 311164928 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045742226680040124, + "loss": 1.6083, + "theoretical_loss": 4.157100987280814, + "tokens_seen": 311230464 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 122179, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7349475622177124, + "objective/train/theoretical_loss": 4.15699960570107, + "objective/train/tokens_used": 311296000, + "theoretical_loss": 4.15699960570107, + "tokens_seen": 311296000 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045741223671013037, + "loss": 1.6916, + "theoretical_loss": 4.15699960570107, + "tokens_seen": 311296000 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004574022066198596, + "loss": 1.7401, + "theoretical_loss": 4.156898251437312, + "tokens_seen": 311361536 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045739217652958873, + "loss": 1.6452, + "theoretical_loss": 4.156796924476435, + "tokens_seen": 311427072 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045738214643931797, + "loss": 1.7236, + "theoretical_loss": 4.156695624805341, + "tokens_seen": 311492608 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045737211634904715, + "loss": 1.6542, + "theoretical_loss": 4.15659435241094, + "tokens_seen": 311558144 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045736208625877633, + "loss": 1.5559, + "theoretical_loss": 4.15649310728015, + "tokens_seen": 311623680 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004573520561685055, + "loss": 1.7129, + "theoretical_loss": 4.156391889399904, + "tokens_seen": 311689216 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045734202607823475, + "loss": 1.5943, + "theoretical_loss": 4.156290698757139, + "tokens_seen": 311754752 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045733199598796387, + "loss": 1.5935, + "theoretical_loss": 4.156189535338799, + "tokens_seen": 311820288 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004573219658976931, + "loss": 1.5943, + "theoretical_loss": 4.156088399131844, + "tokens_seen": 311885824 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045731193580742223, + "loss": 1.4944, + "theoretical_loss": 4.155987290123236, + "tokens_seen": 311951360 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045730190571715147, + "loss": 1.5703, + "theoretical_loss": 4.15588620829995, + "tokens_seen": 312016896 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045729187562688065, + "loss": 1.6605, + "theoretical_loss": 4.15578515364897, + "tokens_seen": 312082432 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045728184553660983, + "loss": 1.6768, + "theoretical_loss": 4.155684126157286, + "tokens_seen": 312147968 + }, + { + "epoch": 0.09, + "learning_rate": 0.000457271815446339, + "loss": 1.642, + "theoretical_loss": 4.1555831258119005, + "tokens_seen": 312213504 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004572617853560682, + "loss": 1.6353, + "theoretical_loss": 4.155482152599823, + "tokens_seen": 312279040 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004572517552657974, + "loss": 1.653, + "theoretical_loss": 4.155381206508071, + "tokens_seen": 312344576 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004572417251755266, + "loss": 1.7573, + "theoretical_loss": 4.155280287523674, + "tokens_seen": 312410112 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045723169508525574, + "loss": 1.6446, + "theoretical_loss": 4.155179395633667, + "tokens_seen": 312475648 + }, + { + "epoch": 0.09, + "learning_rate": 0.000457221664994985, + "loss": 1.6834, + "theoretical_loss": 4.1550785308250955, + "tokens_seen": 312541184 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004572116349047141, + "loss": 1.6097, + "theoretical_loss": 4.154977693085016, + "tokens_seen": 312606720 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045720160481444334, + "loss": 1.6497, + "theoretical_loss": 4.154876882400489, + "tokens_seen": 312672256 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045719157472417257, + "loss": 1.6117, + "theoretical_loss": 4.154776098758589, + "tokens_seen": 312737792 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004571815446339017, + "loss": 1.6142, + "theoretical_loss": 4.154675342146397, + "tokens_seen": 312803328 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045717151454363093, + "loss": 1.7798, + "theoretical_loss": 4.154574612551001, + "tokens_seen": 312868864 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004571614844533601, + "loss": 1.6149, + "theoretical_loss": 4.154473909959503, + "tokens_seen": 312934400 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004571514543630893, + "loss": 1.7271, + "theoretical_loss": 4.154373234359007, + "tokens_seen": 312999936 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004571414242728185, + "loss": 1.7701, + "theoretical_loss": 4.154272585736632, + "tokens_seen": 313065472 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045713139418254766, + "loss": 1.6391, + "theoretical_loss": 4.154171964079504, + "tokens_seen": 313131008 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045712136409227684, + "loss": 1.6958, + "theoretical_loss": 4.154071369374755, + "tokens_seen": 313196544 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004571113340020061, + "loss": 1.5459, + "theoretical_loss": 4.15397080160953, + "tokens_seen": 313262080 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004571013039117352, + "loss": 1.5976, + "theoretical_loss": 4.153870260770979, + "tokens_seen": 313327616 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045709127382146444, + "loss": 1.6923, + "theoretical_loss": 4.153769746846265, + "tokens_seen": 313393152 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045708124373119356, + "loss": 1.8205, + "theoretical_loss": 4.153669259822556, + "tokens_seen": 313458688 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004570712136409228, + "loss": 1.7369, + "theoretical_loss": 4.153568799687031, + "tokens_seen": 313524224 + }, + { + "epoch": 0.1, + "learning_rate": 0.000457061183550652, + "loss": 1.7065, + "theoretical_loss": 4.153468366426877, + "tokens_seen": 313589760 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045705115346038116, + "loss": 1.6046, + "theoretical_loss": 4.153367960029288, + "tokens_seen": 313655296 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045704112337011034, + "loss": 1.6031, + "theoretical_loss": 4.153267580481472, + "tokens_seen": 313720832 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004570310932798395, + "loss": 1.5812, + "theoretical_loss": 4.15316722777064, + "tokens_seen": 313786368 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004570210631895687, + "loss": 1.6071, + "theoretical_loss": 4.153066901884016, + "tokens_seen": 313851904 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045701103309929794, + "loss": 1.7358, + "theoretical_loss": 4.15296660280883, + "tokens_seen": 313917440 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045700100300902707, + "loss": 1.638, + "theoretical_loss": 4.152866330532321, + "tokens_seen": 313982976 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004569909729187563, + "loss": 1.7157, + "theoretical_loss": 4.152766085041737, + "tokens_seen": 314048512 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004569809428284855, + "loss": 1.6652, + "theoretical_loss": 4.152665866324337, + "tokens_seen": 314114048 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045697091273821466, + "loss": 1.553, + "theoretical_loss": 4.152565674367387, + "tokens_seen": 314179584 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045696088264794384, + "loss": 1.6469, + "theoretical_loss": 4.15246550915816, + "tokens_seen": 314245120 + }, + { + "epoch": 0.1, + "learning_rate": 0.000456950852557673, + "loss": 1.6488, + "theoretical_loss": 4.15236537068394, + "tokens_seen": 314310656 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004569408224674022, + "loss": 1.6272, + "theoretical_loss": 4.152265258932019, + "tokens_seen": 314376192 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045693079237713144, + "loss": 1.5579, + "theoretical_loss": 4.152165173889697, + "tokens_seen": 314441728 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045692076228686057, + "loss": 1.5626, + "theoretical_loss": 4.152065115544284, + "tokens_seen": 314507264 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 123650, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5360088348388672, + "objective/train/theoretical_loss": 4.151965083883098, + "objective/train/tokens_used": 314572800, + "theoretical_loss": 4.151965083883098, + "tokens_seen": 314572800 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004569107321965898, + "loss": 1.5777, + "theoretical_loss": 4.151965083883098, + "tokens_seen": 314572800 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045690070210631893, + "loss": 1.5741, + "theoretical_loss": 4.151865078893466, + "tokens_seen": 314638336 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045689067201604817, + "loss": 1.629, + "theoretical_loss": 4.151765100562723, + "tokens_seen": 314703872 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045688064192577735, + "loss": 1.7092, + "theoretical_loss": 4.151665148878212, + "tokens_seen": 314769408 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045687061183550653, + "loss": 1.6197, + "theoretical_loss": 4.151565223827288, + "tokens_seen": 314834944 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004568605817452357, + "loss": 1.5697, + "theoretical_loss": 4.15146532539731, + "tokens_seen": 314900480 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045685055165496495, + "loss": 1.6387, + "theoretical_loss": 4.151365453575648, + "tokens_seen": 314966016 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045684052156469407, + "loss": 1.617, + "theoretical_loss": 4.151265608349682, + "tokens_seen": 315031552 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004568304914744233, + "loss": 1.5826, + "theoretical_loss": 4.151165789706798, + "tokens_seen": 315097088 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045682046138415243, + "loss": 1.6451, + "theoretical_loss": 4.151065997634393, + "tokens_seen": 315162624 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045681043129388167, + "loss": 1.5292, + "theoretical_loss": 4.1509662321198695, + "tokens_seen": 315228160 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045680040120361085, + "loss": 1.5924, + "theoretical_loss": 4.150866493150642, + "tokens_seen": 315293696 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045679037111334003, + "loss": 1.5916, + "theoretical_loss": 4.150766780714131, + "tokens_seen": 315359232 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004567803410230692, + "loss": 1.7312, + "theoretical_loss": 4.150667094797766, + "tokens_seen": 315424768 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004567703109327984, + "loss": 1.6409, + "theoretical_loss": 4.150567435388988, + "tokens_seen": 315490304 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004567602808425276, + "loss": 1.6565, + "theoretical_loss": 4.150467802475241, + "tokens_seen": 315555840 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004567502507522568, + "loss": 1.5528, + "theoretical_loss": 4.150368196043983, + "tokens_seen": 315621376 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045674022066198594, + "loss": 1.6034, + "theoretical_loss": 4.150268616082678, + "tokens_seen": 315686912 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004567301905717152, + "loss": 1.6453, + "theoretical_loss": 4.150169062578798, + "tokens_seen": 315752448 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004567201604814443, + "loss": 1.6486, + "theoretical_loss": 4.150069535519825, + "tokens_seen": 315817984 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045671013039117354, + "loss": 1.6611, + "theoretical_loss": 4.149970034893249, + "tokens_seen": 315883520 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004567001003009027, + "loss": 1.5998, + "theoretical_loss": 4.149870560686567, + "tokens_seen": 315949056 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004566900702106319, + "loss": 1.663, + "theoretical_loss": 4.149771112887288, + "tokens_seen": 316014592 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004566800401203611, + "loss": 1.5825, + "theoretical_loss": 4.149671691482926, + "tokens_seen": 316080128 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004566700100300903, + "loss": 1.7401, + "theoretical_loss": 4.149572296461005, + "tokens_seen": 316145664 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045665997993981944, + "loss": 1.6033, + "theoretical_loss": 4.149472927809057, + "tokens_seen": 316211200 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004566499498495487, + "loss": 1.687, + "theoretical_loss": 4.149373585514625, + "tokens_seen": 316276736 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004566399197592778, + "loss": 1.6815, + "theoretical_loss": 4.149274269565256, + "tokens_seen": 316342272 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045662988966900704, + "loss": 1.6856, + "theoretical_loss": 4.149174979948507, + "tokens_seen": 316407808 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004566198595787362, + "loss": 1.5956, + "theoretical_loss": 4.1490757166519465, + "tokens_seen": 316473344 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004566098294884654, + "loss": 1.5781, + "theoretical_loss": 4.148976479663148, + "tokens_seen": 316538880 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004565997993981946, + "loss": 1.5267, + "theoretical_loss": 4.148877268969694, + "tokens_seen": 316604416 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045658976930792376, + "loss": 1.6504, + "theoretical_loss": 4.1487780845591775, + "tokens_seen": 316669952 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045657973921765294, + "loss": 1.6525, + "theoretical_loss": 4.148678926419197, + "tokens_seen": 316735488 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004565697091273822, + "loss": 1.6226, + "theoretical_loss": 4.148579794537362, + "tokens_seen": 316801024 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004565596790371113, + "loss": 1.6011, + "theoretical_loss": 4.148480688901288, + "tokens_seen": 316866560 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045654964894684054, + "loss": 1.7076, + "theoretical_loss": 4.148381609498602, + "tokens_seen": 316932096 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045653961885656967, + "loss": 1.5711, + "theoretical_loss": 4.148282556316936, + "tokens_seen": 316997632 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004565295887662989, + "loss": 1.6311, + "theoretical_loss": 4.148183529343932, + "tokens_seen": 317063168 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004565195586760281, + "loss": 1.5747, + "theoretical_loss": 4.148084528567241, + "tokens_seen": 317128704 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045650952858575727, + "loss": 1.6642, + "theoretical_loss": 4.147985553974522, + "tokens_seen": 317194240 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045649949849548645, + "loss": 1.5669, + "theoretical_loss": 4.14788660555344, + "tokens_seen": 317259776 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004564894684052157, + "loss": 1.5502, + "theoretical_loss": 4.147787683291672, + "tokens_seen": 317325312 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004564794383149448, + "loss": 1.7189, + "theoretical_loss": 4.147688787176902, + "tokens_seen": 317390848 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045646940822467405, + "loss": 1.5937, + "theoretical_loss": 4.147589917196821, + "tokens_seen": 317456384 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045645937813440317, + "loss": 1.5763, + "theoretical_loss": 4.14749107333913, + "tokens_seen": 317521920 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004564493480441324, + "loss": 1.6355, + "theoretical_loss": 4.147392255591538, + "tokens_seen": 317587456 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045643931795386164, + "loss": 1.8, + "theoretical_loss": 4.147293463941762, + "tokens_seen": 317652992 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045642928786359077, + "loss": 1.4901, + "theoretical_loss": 4.147194698377526, + "tokens_seen": 317718528 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045641925777332, + "loss": 1.651, + "theoretical_loss": 4.147095958886566, + "tokens_seen": 317784064 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 124777, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.660879373550415, + "objective/train/theoretical_loss": 4.146997245456623, + "objective/train/tokens_used": 317849600, + "theoretical_loss": 4.146997245456623, + "tokens_seen": 317849600 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045640922768304913, + "loss": 1.6741, + "theoretical_loss": 4.146997245456623, + "tokens_seen": 317849600 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045639919759277837, + "loss": 1.6595, + "theoretical_loss": 4.146898558075446, + "tokens_seen": 317915136 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045638916750250755, + "loss": 1.6058, + "theoretical_loss": 4.146799896730794, + "tokens_seen": 317980672 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045637913741223673, + "loss": 1.5584, + "theoretical_loss": 4.146701261410435, + "tokens_seen": 318046208 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004563691073219659, + "loss": 1.5973, + "theoretical_loss": 4.1466026521021435, + "tokens_seen": 318111744 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045635907723169515, + "loss": 1.7282, + "theoretical_loss": 4.146504068793703, + "tokens_seen": 318177280 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045634904714142427, + "loss": 1.6262, + "theoretical_loss": 4.146405511472905, + "tokens_seen": 318242816 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004563390170511535, + "loss": 1.6586, + "theoretical_loss": 4.146306980127548, + "tokens_seen": 318308352 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045632898696088263, + "loss": 1.6878, + "theoretical_loss": 4.146208474745442, + "tokens_seen": 318373888 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045631895687061187, + "loss": 1.6648, + "theoretical_loss": 4.1461099953144025, + "tokens_seen": 318439424 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045630892678034105, + "loss": 1.55, + "theoretical_loss": 4.146011541822253, + "tokens_seen": 318504960 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045629889669007023, + "loss": 1.5503, + "theoretical_loss": 4.145913114256829, + "tokens_seen": 318570496 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004562888665997994, + "loss": 1.5879, + "theoretical_loss": 4.145814712605969, + "tokens_seen": 318636032 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004562788365095286, + "loss": 1.6157, + "theoretical_loss": 4.1457163368575225, + "tokens_seen": 318701568 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004562688064192578, + "loss": 1.5648, + "theoretical_loss": 4.145617986999348, + "tokens_seen": 318767104 + }, + { + "epoch": 0.1, + "learning_rate": 0.000456258776328987, + "loss": 1.643, + "theoretical_loss": 4.145519663019309, + "tokens_seen": 318832640 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045624874623871614, + "loss": 1.5466, + "theoretical_loss": 4.145421364905282, + "tokens_seen": 318898176 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004562387161484454, + "loss": 1.6176, + "theoretical_loss": 4.145323092645146, + "tokens_seen": 318963712 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004562286860581745, + "loss": 1.6646, + "theoretical_loss": 4.145224846226793, + "tokens_seen": 319029248 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045621865596790374, + "loss": 1.6995, + "theoretical_loss": 4.14512662563812, + "tokens_seen": 319094784 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004562086258776329, + "loss": 1.5878, + "theoretical_loss": 4.145028430867035, + "tokens_seen": 319160320 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004561985957873621, + "loss": 1.6336, + "theoretical_loss": 4.144930261901451, + "tokens_seen": 319225856 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004561885656970913, + "loss": 1.5615, + "theoretical_loss": 4.144832118729291, + "tokens_seen": 319291392 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004561785356068205, + "loss": 1.6166, + "theoretical_loss": 4.144734001338486, + "tokens_seen": 319356928 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045616850551654964, + "loss": 1.7197, + "theoretical_loss": 4.144635909716976, + "tokens_seen": 319422464 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004561584754262789, + "loss": 1.6238, + "theoretical_loss": 4.144537843852705, + "tokens_seen": 319488000 + }, + { + "epoch": 0.1, + "learning_rate": 0.000456148445336008, + "loss": 1.4932, + "theoretical_loss": 4.144439803733631, + "tokens_seen": 319553536 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045613841524573724, + "loss": 1.5746, + "theoretical_loss": 4.144341789347715, + "tokens_seen": 319619072 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004561283851554664, + "loss": 1.6571, + "theoretical_loss": 4.144243800682931, + "tokens_seen": 319684608 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004561183550651956, + "loss": 1.6134, + "theoretical_loss": 4.1441458377272555, + "tokens_seen": 319750144 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004561083249749248, + "loss": 1.6498, + "theoretical_loss": 4.144047900468678, + "tokens_seen": 319815680 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045609829488465396, + "loss": 1.6669, + "theoretical_loss": 4.143949988895193, + "tokens_seen": 319881216 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045608826479438314, + "loss": 1.6652, + "theoretical_loss": 4.1438521029948046, + "tokens_seen": 319946752 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004560782347041124, + "loss": 1.5744, + "theoretical_loss": 4.1437542427555245, + "tokens_seen": 320012288 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004560682046138415, + "loss": 1.6, + "theoretical_loss": 4.143656408165373, + "tokens_seen": 320077824 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045605817452357074, + "loss": 1.6559, + "theoretical_loss": 4.143558599212376, + "tokens_seen": 320143360 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045604814443329987, + "loss": 1.6289, + "theoretical_loss": 4.143460815884571, + "tokens_seen": 320208896 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004560381143430291, + "loss": 1.6519, + "theoretical_loss": 4.143363058170001, + "tokens_seen": 320274432 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004560280842527583, + "loss": 1.6537, + "theoretical_loss": 4.143265326056719, + "tokens_seen": 320339968 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045601805416248747, + "loss": 1.6994, + "theoretical_loss": 4.1431676195327825, + "tokens_seen": 320405504 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045600802407221665, + "loss": 1.7414, + "theoretical_loss": 4.143069938586263, + "tokens_seen": 320471040 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004559979939819459, + "loss": 1.7822, + "theoretical_loss": 4.142972283205233, + "tokens_seen": 320536576 + }, + { + "epoch": 0.1, + "learning_rate": 0.000455987963891675, + "loss": 1.5796, + "theoretical_loss": 4.142874653377778, + "tokens_seen": 320602112 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045597793380140425, + "loss": 1.6344, + "theoretical_loss": 4.14277704909199, + "tokens_seen": 320667648 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045596790371113337, + "loss": 1.5363, + "theoretical_loss": 4.142679470335969, + "tokens_seen": 320733184 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004559578736208626, + "loss": 1.5941, + "theoretical_loss": 4.142581917097821, + "tokens_seen": 320798720 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004559478435305918, + "loss": 1.6738, + "theoretical_loss": 4.142484389365664, + "tokens_seen": 320864256 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045593781344032097, + "loss": 1.6029, + "theoretical_loss": 4.14238688712762, + "tokens_seen": 320929792 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045592778335005015, + "loss": 1.6692, + "theoretical_loss": 4.142289410371823, + "tokens_seen": 320995328 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045591775325977933, + "loss": 1.68, + "theoretical_loss": 4.14219195908641, + "tokens_seen": 321060864 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 125851, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5645902156829834, + "objective/train/theoretical_loss": 4.14209453325953, + "objective/train/tokens_used": 321126400, + "theoretical_loss": 4.14209453325953, + "tokens_seen": 321126400 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004559077231695085, + "loss": 1.5728, + "theoretical_loss": 4.14209453325953, + "tokens_seen": 321126400 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045589769307923775, + "loss": 1.6167, + "theoretical_loss": 4.1419971328793395, + "tokens_seen": 321191936 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004558876629889669, + "loss": 1.6568, + "theoretical_loss": 4.141899757934, + "tokens_seen": 321257472 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004558776328986961, + "loss": 1.5345, + "theoretical_loss": 4.141802408411683, + "tokens_seen": 321323008 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045586760280842524, + "loss": 1.6141, + "theoretical_loss": 4.14170508430057, + "tokens_seen": 321388544 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004558575727181545, + "loss": 1.6999, + "theoretical_loss": 4.141607785588846, + "tokens_seen": 321454080 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045584754262788365, + "loss": 1.6624, + "theoretical_loss": 4.141510512264707, + "tokens_seen": 321519616 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045583751253761284, + "loss": 1.6686, + "theoretical_loss": 4.141413264316356, + "tokens_seen": 321585152 + }, + { + "epoch": 0.1, + "learning_rate": 0.000455827482447342, + "loss": 1.7413, + "theoretical_loss": 4.141316041732004, + "tokens_seen": 321650688 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045581745235707125, + "loss": 1.6717, + "theoretical_loss": 4.141218844499869, + "tokens_seen": 321716224 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004558074222668004, + "loss": 1.6153, + "theoretical_loss": 4.141121672608178, + "tokens_seen": 321781760 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004557973921765296, + "loss": 1.6181, + "theoretical_loss": 4.141024526045166, + "tokens_seen": 321847296 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045578736208625874, + "loss": 1.7368, + "theoretical_loss": 4.140927404799075, + "tokens_seen": 321912832 + }, + { + "epoch": 0.1, + "learning_rate": 0.000455777331995988, + "loss": 1.6581, + "theoretical_loss": 4.140830308858154, + "tokens_seen": 321978368 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045576730190571716, + "loss": 1.6001, + "theoretical_loss": 4.140733238210664, + "tokens_seen": 322043904 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045575727181544634, + "loss": 1.5362, + "theoretical_loss": 4.140636192844867, + "tokens_seen": 322109440 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004557472417251755, + "loss": 1.6921, + "theoretical_loss": 4.1405391727490395, + "tokens_seen": 322174976 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004557372116349047, + "loss": 1.6719, + "theoretical_loss": 4.140442177911462, + "tokens_seen": 322240512 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004557271815446339, + "loss": 1.5746, + "theoretical_loss": 4.140345208320425, + "tokens_seen": 322306048 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004557171514543631, + "loss": 1.6474, + "theoretical_loss": 4.1402482639642235, + "tokens_seen": 322371584 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045570712136409224, + "loss": 1.5638, + "theoretical_loss": 4.140151344831164, + "tokens_seen": 322437120 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004556970912738215, + "loss": 1.6393, + "theoretical_loss": 4.14005445090956, + "tokens_seen": 322502656 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004556870611835507, + "loss": 1.6387, + "theoretical_loss": 4.13995758218773, + "tokens_seen": 322568192 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045567703109327984, + "loss": 1.7056, + "theoretical_loss": 4.1398607386540025, + "tokens_seen": 322633728 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004556670010030091, + "loss": 1.6969, + "theoretical_loss": 4.139763920296716, + "tokens_seen": 322699264 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004556569709127382, + "loss": 1.6115, + "theoretical_loss": 4.139667127104211, + "tokens_seen": 322764800 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045564694082246744, + "loss": 1.5822, + "theoretical_loss": 4.139570359064842, + "tokens_seen": 322830336 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004556369107321966, + "loss": 1.557, + "theoretical_loss": 4.139473616166967, + "tokens_seen": 322895872 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004556268806419258, + "loss": 1.479, + "theoretical_loss": 4.139376898398953, + "tokens_seen": 322961408 + }, + { + "epoch": 0.1, + "learning_rate": 0.000455616850551655, + "loss": 1.6946, + "theoretical_loss": 4.139280205749175, + "tokens_seen": 323026944 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045560682046138416, + "loss": 1.4767, + "theoretical_loss": 4.139183538206017, + "tokens_seen": 323092480 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045559679037111334, + "loss": 1.7856, + "theoretical_loss": 4.139086895757868, + "tokens_seen": 323158016 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004555867602808426, + "loss": 1.6485, + "theoretical_loss": 4.138990278393125, + "tokens_seen": 323223552 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004555767301905717, + "loss": 1.5461, + "theoretical_loss": 4.138893686100196, + "tokens_seen": 323289088 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045556670010030094, + "loss": 1.5446, + "theoretical_loss": 4.138797118867493, + "tokens_seen": 323354624 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045555667001003007, + "loss": 1.6227, + "theoretical_loss": 4.138700576683437, + "tokens_seen": 323420160 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004555466399197593, + "loss": 1.5492, + "theoretical_loss": 4.138604059536458, + "tokens_seen": 323485696 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004555366098294885, + "loss": 1.5797, + "theoretical_loss": 4.138507567414993, + "tokens_seen": 323551232 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045552657973921767, + "loss": 1.5947, + "theoretical_loss": 4.138411100307485, + "tokens_seen": 323616768 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045551654964894685, + "loss": 1.682, + "theoretical_loss": 4.138314658202385, + "tokens_seen": 323682304 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004555065195586761, + "loss": 1.6614, + "theoretical_loss": 4.138218241088156, + "tokens_seen": 323747840 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004554964894684052, + "loss": 1.5583, + "theoretical_loss": 4.138121848953261, + "tokens_seen": 323813376 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045548645937813445, + "loss": 1.6599, + "theoretical_loss": 4.138025481786178, + "tokens_seen": 323878912 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045547642928786357, + "loss": 1.6228, + "theoretical_loss": 4.137929139575387, + "tokens_seen": 323944448 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004554663991975928, + "loss": 1.6947, + "theoretical_loss": 4.137832822309382, + "tokens_seen": 324009984 + }, + { + "epoch": 0.1, + "learning_rate": 0.000455456369107322, + "loss": 1.6845, + "theoretical_loss": 4.1377365299766575, + "tokens_seen": 324075520 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045544633901705117, + "loss": 1.6114, + "theoretical_loss": 4.13764026256572, + "tokens_seen": 324141056 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045543630892678035, + "loss": 1.7091, + "theoretical_loss": 4.137544020065082, + "tokens_seen": 324206592 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045542627883650953, + "loss": 1.6321, + "theoretical_loss": 4.1374478024632655, + "tokens_seen": 324272128 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004554162487462387, + "loss": 1.7791, + "theoretical_loss": 4.137351609748798, + "tokens_seen": 324337664 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 127012, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7011569738388062, + "objective/train/theoretical_loss": 4.137255441910215, + "objective/train/tokens_used": 324403200, + "theoretical_loss": 4.137255441910215, + "tokens_seen": 324403200 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045540621865596795, + "loss": 1.6946, + "theoretical_loss": 4.137255441910215, + "tokens_seen": 324403200 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004553961885656971, + "loss": 1.7009, + "theoretical_loss": 4.13715929893606, + "tokens_seen": 324468736 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004553861584754263, + "loss": 1.7005, + "theoretical_loss": 4.137063180814886, + "tokens_seen": 324534272 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045537612838515544, + "loss": 1.5577, + "theoretical_loss": 4.136967087535249, + "tokens_seen": 324599808 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004553660982948847, + "loss": 1.4735, + "theoretical_loss": 4.136871019085717, + "tokens_seen": 324665344 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045535606820461385, + "loss": 1.6462, + "theoretical_loss": 4.136774975454863, + "tokens_seen": 324730880 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045534603811434304, + "loss": 1.5749, + "theoretical_loss": 4.136678956631269, + "tokens_seen": 324796416 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004553360080240722, + "loss": 1.5852, + "theoretical_loss": 4.136582962603523, + "tokens_seen": 324861952 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045532597793380145, + "loss": 1.6568, + "theoretical_loss": 4.136486993360222, + "tokens_seen": 324927488 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004553159478435306, + "loss": 1.5925, + "theoretical_loss": 4.136391048889971, + "tokens_seen": 324993024 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004553059177532598, + "loss": 1.6282, + "theoretical_loss": 4.136295129181381, + "tokens_seen": 325058560 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045529588766298894, + "loss": 1.6446, + "theoretical_loss": 4.136199234223069, + "tokens_seen": 325124096 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004552858575727182, + "loss": 1.5948, + "theoretical_loss": 4.136103364003665, + "tokens_seen": 325189632 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045527582748244736, + "loss": 1.6307, + "theoretical_loss": 4.1360075185118, + "tokens_seen": 325255168 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045526579739217654, + "loss": 1.7372, + "theoretical_loss": 4.135911697736118, + "tokens_seen": 325320704 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004552557673019057, + "loss": 1.5461, + "theoretical_loss": 4.135815901665267, + "tokens_seen": 325386240 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004552457372116349, + "loss": 1.6747, + "theoretical_loss": 4.135720130287904, + "tokens_seen": 325451776 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004552357071213641, + "loss": 1.685, + "theoretical_loss": 4.135624383592694, + "tokens_seen": 325517312 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004552256770310933, + "loss": 1.5986, + "theoretical_loss": 4.135528661568307, + "tokens_seen": 325582848 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045521564694082244, + "loss": 1.6179, + "theoretical_loss": 4.135432964203422, + "tokens_seen": 325648384 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004552056168505517, + "loss": 1.7741, + "theoretical_loss": 4.135337291486727, + "tokens_seen": 325713920 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004551955867602808, + "loss": 1.668, + "theoretical_loss": 4.135241643406915, + "tokens_seen": 325779456 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045518555667001004, + "loss": 1.6599, + "theoretical_loss": 4.135146019952688, + "tokens_seen": 325844992 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004551755265797392, + "loss": 1.5603, + "theoretical_loss": 4.135050421112755, + "tokens_seen": 325910528 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004551654964894684, + "loss": 1.6695, + "theoretical_loss": 4.134954846875832, + "tokens_seen": 325976064 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004551554663991976, + "loss": 1.7053, + "theoretical_loss": 4.134859297230644, + "tokens_seen": 326041600 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004551454363089268, + "loss": 1.6398, + "theoretical_loss": 4.13476377216592, + "tokens_seen": 326107136 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045513540621865595, + "loss": 1.6777, + "theoretical_loss": 4.134668271670401, + "tokens_seen": 326172672 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004551253761283852, + "loss": 1.6265, + "theoretical_loss": 4.134572795732832, + "tokens_seen": 326238208 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004551153460381143, + "loss": 1.6586, + "theoretical_loss": 4.134477344341967, + "tokens_seen": 326303744 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045510531594784354, + "loss": 1.7001, + "theoretical_loss": 4.134381917486565, + "tokens_seen": 326369280 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004550952858575727, + "loss": 1.6104, + "theoretical_loss": 4.134286515155398, + "tokens_seen": 326434816 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004550852557673019, + "loss": 1.6129, + "theoretical_loss": 4.134191137337238, + "tokens_seen": 326500352 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004550752256770311, + "loss": 1.6391, + "theoretical_loss": 4.134095784020871, + "tokens_seen": 326565888 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045506519558676027, + "loss": 1.5627, + "theoretical_loss": 4.1340004551950855, + "tokens_seen": 326631424 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045505516549648945, + "loss": 1.5658, + "theoretical_loss": 4.133905150848681, + "tokens_seen": 326696960 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004550451354062187, + "loss": 1.6147, + "theoretical_loss": 4.133809870970461, + "tokens_seen": 326762496 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004550351053159478, + "loss": 1.6437, + "theoretical_loss": 4.1337146155492395, + "tokens_seen": 326828032 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045502507522567705, + "loss": 1.5042, + "theoretical_loss": 4.133619384573836, + "tokens_seen": 326893568 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045501504513540623, + "loss": 1.5817, + "theoretical_loss": 4.1335241780330785, + "tokens_seen": 326959104 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004550050150451354, + "loss": 1.5707, + "theoretical_loss": 4.1334289959158, + "tokens_seen": 327024640 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004549949849548646, + "loss": 1.5857, + "theoretical_loss": 4.133333838210844, + "tokens_seen": 327090176 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045498495486459377, + "loss": 1.7915, + "theoretical_loss": 4.133238704907059, + "tokens_seen": 327155712 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045497492477432295, + "loss": 1.6668, + "theoretical_loss": 4.133143595993303, + "tokens_seen": 327221248 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004549648946840522, + "loss": 1.6884, + "theoretical_loss": 4.133048511458438, + "tokens_seen": 327286784 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045495486459378137, + "loss": 1.7294, + "theoretical_loss": 4.132953451291337, + "tokens_seen": 327352320 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045494483450351055, + "loss": 1.5907, + "theoretical_loss": 4.132858415480877, + "tokens_seen": 327417856 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045493480441323973, + "loss": 1.6257, + "theoretical_loss": 4.132763404015946, + "tokens_seen": 327483392 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004549247743229689, + "loss": 1.6113, + "theoretical_loss": 4.132668416885435, + "tokens_seen": 327548928 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045491474423269815, + "loss": 1.6004, + "theoretical_loss": 4.132573454078246, + "tokens_seen": 327614464 + }, + { + "debugging/length": 449.25, + "debugging/num_segments": 8, + "debugging/score": 0.0038660800931859442, + "debugging/score_std": 0.002461787614057621, + "epoch": 0.1, + "objective/train/docs_used": 128204, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5223948955535889, + "objective/train/theoretical_loss": 4.132478515583286, + "objective/train/tokens_used": 327680000, + "theoretical_loss": 4.132478515583286, + "tokens_seen": 327680000 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004549047141424273, + "loss": 1.5925, + "theoretical_loss": 4.132478515583286, + "tokens_seen": 327680000 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004548946840521565, + "loss": 1.6808, + "theoretical_loss": 4.132383601389471, + "tokens_seen": 327745536 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045488465396188564, + "loss": 1.733, + "theoretical_loss": 4.132288711485721, + "tokens_seen": 327811072 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004548746238716149, + "loss": 1.7361, + "theoretical_loss": 4.132193845860969, + "tokens_seen": 327876608 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045486459378134405, + "loss": 1.5053, + "theoretical_loss": 4.132099004504149, + "tokens_seen": 327942144 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045485456369107324, + "loss": 1.6884, + "theoretical_loss": 4.132004187404206, + "tokens_seen": 328007680 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004548445336008024, + "loss": 1.6149, + "theoretical_loss": 4.131909394550092, + "tokens_seen": 328073216 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045483450351053165, + "loss": 1.6735, + "theoretical_loss": 4.131814625930764, + "tokens_seen": 328138752 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004548244734202608, + "loss": 1.5961, + "theoretical_loss": 4.13171988153519, + "tokens_seen": 328204288 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045481444332999, + "loss": 1.6449, + "theoretical_loss": 4.131625161352342, + "tokens_seen": 328269824 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045480441323971914, + "loss": 1.5737, + "theoretical_loss": 4.131530465371199, + "tokens_seen": 328335360 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004547943831494484, + "loss": 1.612, + "theoretical_loss": 4.131435793580749, + "tokens_seen": 328400896 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045478435305917756, + "loss": 1.5476, + "theoretical_loss": 4.131341145969989, + "tokens_seen": 328466432 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045477432296890674, + "loss": 1.6517, + "theoretical_loss": 4.131246522527918, + "tokens_seen": 328531968 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004547642928786359, + "loss": 1.7022, + "theoretical_loss": 4.1311519232435465, + "tokens_seen": 328597504 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004547542627883651, + "loss": 1.6194, + "theoretical_loss": 4.131057348105889, + "tokens_seen": 328663040 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004547442326980943, + "loss": 1.5875, + "theoretical_loss": 4.1309627971039715, + "tokens_seen": 328728576 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004547342026078235, + "loss": 1.5283, + "theoretical_loss": 4.130868270226824, + "tokens_seen": 328794112 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045472417251755264, + "loss": 1.6344, + "theoretical_loss": 4.130773767463483, + "tokens_seen": 328859648 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004547141424272819, + "loss": 1.5778, + "theoretical_loss": 4.130679288802993, + "tokens_seen": 328925184 + }, + { + "epoch": 0.1, + "learning_rate": 0.000454704112337011, + "loss": 1.6235, + "theoretical_loss": 4.1305848342344085, + "tokens_seen": 328990720 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045469408224674024, + "loss": 1.6224, + "theoretical_loss": 4.130490403746787, + "tokens_seen": 329056256 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004546840521564694, + "loss": 1.6694, + "theoretical_loss": 4.130395997329196, + "tokens_seen": 329121792 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004546740220661986, + "loss": 1.6765, + "theoretical_loss": 4.1303016149707075, + "tokens_seen": 329187328 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004546639919759278, + "loss": 1.7046, + "theoretical_loss": 4.130207256660404, + "tokens_seen": 329252864 + }, + { + "epoch": 0.1, + "learning_rate": 0.000454653961885657, + "loss": 1.5927, + "theoretical_loss": 4.130112922387372, + "tokens_seen": 329318400 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045464393179538615, + "loss": 1.5638, + "theoretical_loss": 4.130018612140707, + "tokens_seen": 329383936 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004546339017051154, + "loss": 1.6259, + "theoretical_loss": 4.12992432590951, + "tokens_seen": 329449472 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004546238716148445, + "loss": 1.5598, + "theoretical_loss": 4.129830063682892, + "tokens_seen": 329515008 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045461384152457374, + "loss": 1.6204, + "theoretical_loss": 4.129735825449968, + "tokens_seen": 329580544 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004546038114343029, + "loss": 1.5708, + "theoretical_loss": 4.129641611199862, + "tokens_seen": 329646080 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004545937813440321, + "loss": 1.5211, + "theoretical_loss": 4.1295474209217025, + "tokens_seen": 329711616 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004545837512537613, + "loss": 1.6534, + "theoretical_loss": 4.12945325460463, + "tokens_seen": 329777152 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045457372116349047, + "loss": 1.7038, + "theoretical_loss": 4.129359112237787, + "tokens_seen": 329842688 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045456369107321965, + "loss": 1.6102, + "theoretical_loss": 4.1292649938103265, + "tokens_seen": 329908224 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004545536609829489, + "loss": 1.6224, + "theoretical_loss": 4.129170899311406, + "tokens_seen": 329973760 + }, + { + "epoch": 0.1, + "learning_rate": 0.000454543630892678, + "loss": 1.6972, + "theoretical_loss": 4.129076828730192, + "tokens_seen": 330039296 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045453360080240725, + "loss": 1.5797, + "theoretical_loss": 4.128982782055858, + "tokens_seen": 330104832 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045452357071213643, + "loss": 1.6205, + "theoretical_loss": 4.128888759277583, + "tokens_seen": 330170368 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004545135406218656, + "loss": 1.6121, + "theoretical_loss": 4.1287947603845545, + "tokens_seen": 330235904 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004545035105315948, + "loss": 1.5993, + "theoretical_loss": 4.128700785365965, + "tokens_seen": 330301440 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045449348044132397, + "loss": 1.5275, + "theoretical_loss": 4.128606834211018, + "tokens_seen": 330366976 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045448345035105315, + "loss": 1.6216, + "theoretical_loss": 4.128512906908919, + "tokens_seen": 330432512 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004544734202607824, + "loss": 1.6955, + "theoretical_loss": 4.128419003448886, + "tokens_seen": 330498048 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004544633901705115, + "loss": 1.5958, + "theoretical_loss": 4.128325123820138, + "tokens_seen": 330563584 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045445336008024075, + "loss": 1.6646, + "theoretical_loss": 4.1282312680119055, + "tokens_seen": 330629120 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004544433299899699, + "loss": 1.6694, + "theoretical_loss": 4.128137436013425, + "tokens_seen": 330694656 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004544332998996991, + "loss": 1.5965, + "theoretical_loss": 4.128043627813939, + "tokens_seen": 330760192 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004544232698094283, + "loss": 1.6022, + "theoretical_loss": 4.127949843402699, + "tokens_seen": 330825728 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004544132397191575, + "loss": 1.6707, + "theoretical_loss": 4.12785608276896, + "tokens_seen": 330891264 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 129428, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6354544162750244, + "objective/train/theoretical_loss": 4.127762345901987, + "objective/train/tokens_used": 330956800, + "theoretical_loss": 4.127762345901987, + "tokens_seen": 330956800 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045440320962888666, + "loss": 1.604, + "theoretical_loss": 4.127762345901987, + "tokens_seen": 330956800 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045439317953861584, + "loss": 1.6097, + "theoretical_loss": 4.1276686327910515, + "tokens_seen": 331022336 + }, + { + "epoch": 0.1, + "learning_rate": 0.000454383149448345, + "loss": 1.6558, + "theoretical_loss": 4.127574943425429, + "tokens_seen": 331087872 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045437311935807425, + "loss": 1.7237, + "theoretical_loss": 4.127481277794408, + "tokens_seen": 331153408 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004543630892678034, + "loss": 1.6348, + "theoretical_loss": 4.1273876358872785, + "tokens_seen": 331218944 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004543530591775326, + "loss": 1.5056, + "theoretical_loss": 4.127294017693339, + "tokens_seen": 331284480 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004543430290872618, + "loss": 1.6164, + "theoretical_loss": 4.127200423201897, + "tokens_seen": 331350016 + }, + { + "epoch": 0.1, + "learning_rate": 0.000454332998996991, + "loss": 1.6046, + "theoretical_loss": 4.127106852402264, + "tokens_seen": 331415552 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045432296890672016, + "loss": 1.6002, + "theoretical_loss": 4.127013305283759, + "tokens_seen": 331481088 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045431293881644934, + "loss": 1.6866, + "theoretical_loss": 4.12691978183571, + "tokens_seen": 331546624 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004543029087261785, + "loss": 1.6054, + "theoretical_loss": 4.126826282047449, + "tokens_seen": 331612160 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045429287863590776, + "loss": 1.6101, + "theoretical_loss": 4.1267328059083175, + "tokens_seen": 331677696 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004542828485456369, + "loss": 1.7318, + "theoretical_loss": 4.126639353407663, + "tokens_seen": 331743232 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004542728184553661, + "loss": 1.602, + "theoretical_loss": 4.12654592453484, + "tokens_seen": 331808768 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045426278836509525, + "loss": 1.6267, + "theoretical_loss": 4.126452519279208, + "tokens_seen": 331874304 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004542527582748245, + "loss": 1.6002, + "theoretical_loss": 4.126359137630136, + "tokens_seen": 331939840 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045424272818455366, + "loss": 1.657, + "theoretical_loss": 4.126265779576999, + "tokens_seen": 332005376 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045423269809428284, + "loss": 1.6979, + "theoretical_loss": 4.126172445109178, + "tokens_seen": 332070912 + }, + { + "epoch": 0.1, + "learning_rate": 0.000454222668004012, + "loss": 1.5851, + "theoretical_loss": 4.126079134216063, + "tokens_seen": 332136448 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004542126379137412, + "loss": 1.6593, + "theoretical_loss": 4.125985846887049, + "tokens_seen": 332201984 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045420260782347044, + "loss": 1.5788, + "theoretical_loss": 4.125892583111538, + "tokens_seen": 332267520 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004541925777331996, + "loss": 1.687, + "theoretical_loss": 4.125799342878939, + "tokens_seen": 332333056 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004541825476429288, + "loss": 1.6403, + "theoretical_loss": 4.1257061261786685, + "tokens_seen": 332398592 + }, + { + "epoch": 0.1, + "learning_rate": 0.000454172517552658, + "loss": 1.7607, + "theoretical_loss": 4.12561293300015, + "tokens_seen": 332464128 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004541624874623872, + "loss": 1.4906, + "theoretical_loss": 4.125519763332813, + "tokens_seen": 332529664 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045415245737211635, + "loss": 1.6415, + "theoretical_loss": 4.1254266171660925, + "tokens_seen": 332595200 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004541424272818456, + "loss": 1.5989, + "theoretical_loss": 4.125333494489434, + "tokens_seen": 332660736 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004541323971915747, + "loss": 1.5144, + "theoretical_loss": 4.125240395292286, + "tokens_seen": 332726272 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045412236710130395, + "loss": 1.6418, + "theoretical_loss": 4.125147319564107, + "tokens_seen": 332791808 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004541123370110331, + "loss": 1.7451, + "theoretical_loss": 4.12505426729436, + "tokens_seen": 332857344 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004541023069207623, + "loss": 1.6762, + "theoretical_loss": 4.124961238472517, + "tokens_seen": 332922880 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004540922768304915, + "loss": 1.6138, + "theoretical_loss": 4.124868233088054, + "tokens_seen": 332988416 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045408224674022067, + "loss": 1.5364, + "theoretical_loss": 4.124775251130455, + "tokens_seen": 333053952 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045407221664994985, + "loss": 1.5957, + "theoretical_loss": 4.124682292589212, + "tokens_seen": 333119488 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004540621865596791, + "loss": 1.6398, + "theoretical_loss": 4.124589357453823, + "tokens_seen": 333185024 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004540521564694082, + "loss": 1.6487, + "theoretical_loss": 4.124496445713792, + "tokens_seen": 333250560 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045404212637913745, + "loss": 1.6201, + "theoretical_loss": 4.12440355735863, + "tokens_seen": 333316096 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045403209628886663, + "loss": 1.7443, + "theoretical_loss": 4.124310692377857, + "tokens_seen": 333381632 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004540220661985958, + "loss": 1.6028, + "theoretical_loss": 4.124217850760996, + "tokens_seen": 333447168 + }, + { + "epoch": 0.1, + "learning_rate": 0.000454012036108325, + "loss": 1.6917, + "theoretical_loss": 4.124125032497578, + "tokens_seen": 333512704 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045400200601805417, + "loss": 1.7121, + "theoretical_loss": 4.124032237577143, + "tokens_seen": 333578240 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045399197592778335, + "loss": 1.473, + "theoretical_loss": 4.1239394659892366, + "tokens_seen": 333643776 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004539819458375126, + "loss": 1.6693, + "theoretical_loss": 4.12384671772341, + "tokens_seen": 333709312 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004539719157472417, + "loss": 1.7498, + "theoretical_loss": 4.12375399276922, + "tokens_seen": 333774848 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045396188565697095, + "loss": 1.6212, + "theoretical_loss": 4.1236612911162345, + "tokens_seen": 333840384 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004539518555667001, + "loss": 1.6911, + "theoretical_loss": 4.123568612754024, + "tokens_seen": 333905920 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004539418254764293, + "loss": 1.6958, + "theoretical_loss": 4.123475957672168, + "tokens_seen": 333971456 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004539317953861585, + "loss": 1.5458, + "theoretical_loss": 4.123383325860252, + "tokens_seen": 334036992 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004539217652958877, + "loss": 1.6729, + "theoretical_loss": 4.123290717307867, + "tokens_seen": 334102528 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045391173520561686, + "loss": 1.7045, + "theoretical_loss": 4.123198132004614, + "tokens_seen": 334168064 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 130593, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5800645351409912, + "objective/train/theoretical_loss": 4.123105569940097, + "objective/train/tokens_used": 334233600, + "theoretical_loss": 4.123105569940097, + "tokens_seen": 334233600 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045390170511534604, + "loss": 1.7278, + "theoretical_loss": 4.123105569940097, + "tokens_seen": 334233600 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004538916750250752, + "loss": 1.6394, + "theoretical_loss": 4.123013031103927, + "tokens_seen": 334299136 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045388164493480445, + "loss": 1.5647, + "theoretical_loss": 4.122920515485726, + "tokens_seen": 334364672 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004538716148445336, + "loss": 1.7103, + "theoretical_loss": 4.122828023075117, + "tokens_seen": 334430208 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004538615847542628, + "loss": 1.5626, + "theoretical_loss": 4.122735553861733, + "tokens_seen": 334495744 + }, + { + "epoch": 0.1, + "learning_rate": 0.000453851554663992, + "loss": 1.5786, + "theoretical_loss": 4.122643107835213, + "tokens_seen": 334561280 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004538415245737212, + "loss": 1.7251, + "theoretical_loss": 4.122550684985203, + "tokens_seen": 334626816 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045383149448345036, + "loss": 1.6231, + "theoretical_loss": 4.122458285301355, + "tokens_seen": 334692352 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045382146439317954, + "loss": 1.5597, + "theoretical_loss": 4.122365908773327, + "tokens_seen": 334757888 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004538114343029087, + "loss": 1.6911, + "theoretical_loss": 4.1222735553907865, + "tokens_seen": 334823424 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045380140421263796, + "loss": 1.5911, + "theoretical_loss": 4.122181225143404, + "tokens_seen": 334888960 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004537913741223671, + "loss": 1.638, + "theoretical_loss": 4.122088918020857, + "tokens_seen": 334954496 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004537813440320963, + "loss": 1.6211, + "theoretical_loss": 4.121996634012834, + "tokens_seen": 335020032 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045377131394182545, + "loss": 1.6687, + "theoretical_loss": 4.121904373109024, + "tokens_seen": 335085568 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004537612838515547, + "loss": 1.5649, + "theoretical_loss": 4.121812135299129, + "tokens_seen": 335151104 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045375125376128386, + "loss": 1.6696, + "theoretical_loss": 4.1217199205728505, + "tokens_seen": 335216640 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045374122367101304, + "loss": 1.5809, + "theoretical_loss": 4.121627728919903, + "tokens_seen": 335282176 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004537311935807422, + "loss": 1.7101, + "theoretical_loss": 4.121535560330004, + "tokens_seen": 335347712 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004537211634904714, + "loss": 1.6414, + "theoretical_loss": 4.1214434147928785, + "tokens_seen": 335413248 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004537111334002006, + "loss": 1.6611, + "theoretical_loss": 4.121351292298257, + "tokens_seen": 335478784 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004537011033099298, + "loss": 1.6184, + "theoretical_loss": 4.12125919283588, + "tokens_seen": 335544320 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045369107321965895, + "loss": 1.7824, + "theoretical_loss": 4.12116711639549, + "tokens_seen": 335609856 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004536810431293882, + "loss": 1.576, + "theoretical_loss": 4.12107506296684, + "tokens_seen": 335675392 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045367101303911737, + "loss": 1.6582, + "theoretical_loss": 4.120983032539687, + "tokens_seen": 335740928 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045366098294884655, + "loss": 1.5228, + "theoretical_loss": 4.120891025103796, + "tokens_seen": 335806464 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045365095285857573, + "loss": 1.6485, + "theoretical_loss": 4.1207990406489365, + "tokens_seen": 335872000 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004536409227683049, + "loss": 1.6674, + "theoretical_loss": 4.120707079164887, + "tokens_seen": 335937536 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004536308926780341, + "loss": 1.6651, + "theoretical_loss": 4.120615140641433, + "tokens_seen": 336003072 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004536208625877633, + "loss": 1.6594, + "theoretical_loss": 4.1205232250683625, + "tokens_seen": 336068608 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045361083249749245, + "loss": 1.7326, + "theoretical_loss": 4.120431332435475, + "tokens_seen": 336134144 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004536008024072217, + "loss": 1.5878, + "theoretical_loss": 4.120339462732572, + "tokens_seen": 336199680 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004535907723169508, + "loss": 1.6663, + "theoretical_loss": 4.120247615949465, + "tokens_seen": 336265216 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045358074222668005, + "loss": 1.7422, + "theoretical_loss": 4.12015579207597, + "tokens_seen": 336330752 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045357071213640923, + "loss": 1.614, + "theoretical_loss": 4.12006399110191, + "tokens_seen": 336396288 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004535606820461384, + "loss": 1.7251, + "theoretical_loss": 4.119972213017116, + "tokens_seen": 336461824 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004535506519558676, + "loss": 1.5741, + "theoretical_loss": 4.119880457811424, + "tokens_seen": 336527360 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045354062186559683, + "loss": 1.6825, + "theoretical_loss": 4.119788725474675, + "tokens_seen": 336592896 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045353059177532596, + "loss": 1.6511, + "theoretical_loss": 4.119697015996719, + "tokens_seen": 336658432 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004535205616850552, + "loss": 1.6057, + "theoretical_loss": 4.119605329367413, + "tokens_seen": 336723968 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004535105315947843, + "loss": 1.6297, + "theoretical_loss": 4.119513665576617, + "tokens_seen": 336789504 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045350050150451355, + "loss": 1.6999, + "theoretical_loss": 4.119422024614201, + "tokens_seen": 336855040 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045349047141424274, + "loss": 1.6161, + "theoretical_loss": 4.11933040647004, + "tokens_seen": 336920576 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004534804413239719, + "loss": 1.6153, + "theoretical_loss": 4.119238811134014, + "tokens_seen": 336986112 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004534704112337011, + "loss": 1.6689, + "theoretical_loss": 4.119147238596014, + "tokens_seen": 337051648 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004534603811434303, + "loss": 1.6814, + "theoretical_loss": 4.119055688845931, + "tokens_seen": 337117184 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004534503510531595, + "loss": 1.6358, + "theoretical_loss": 4.118964161873667, + "tokens_seen": 337182720 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004534403209628887, + "loss": 1.6984, + "theoretical_loss": 4.118872657669131, + "tokens_seen": 337248256 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004534302908726179, + "loss": 1.5555, + "theoretical_loss": 4.118781176222235, + "tokens_seen": 337313792 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045342026078234706, + "loss": 1.6186, + "theoretical_loss": 4.1186897175229, + "tokens_seen": 337379328 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045341023069207624, + "loss": 1.6347, + "theoretical_loss": 4.1185982815610505, + "tokens_seen": 337444864 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 131700, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7377758026123047, + "objective/train/theoretical_loss": 4.118506868326623, + "objective/train/tokens_used": 337510400, + "theoretical_loss": 4.118506868326623, + "tokens_seen": 337510400 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004534002006018054, + "loss": 1.5977, + "theoretical_loss": 4.118506868326623, + "tokens_seen": 337510400 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045339017051153465, + "loss": 1.6485, + "theoretical_loss": 4.118415477809553, + "tokens_seen": 337575936 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004533801404212638, + "loss": 1.639, + "theoretical_loss": 4.11832410999979, + "tokens_seen": 337641472 + }, + { + "epoch": 0.1, + "learning_rate": 0.000453370110330993, + "loss": 1.6671, + "theoretical_loss": 4.118232764887285, + "tokens_seen": 337707008 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004533600802407222, + "loss": 1.6034, + "theoretical_loss": 4.118141442461995, + "tokens_seen": 337772544 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004533500501504514, + "loss": 1.5916, + "theoretical_loss": 4.118050142713886, + "tokens_seen": 337838080 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045334002006018056, + "loss": 1.6537, + "theoretical_loss": 4.1179588656329305, + "tokens_seen": 337903616 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045332998996990974, + "loss": 1.6468, + "theoretical_loss": 4.117867611209104, + "tokens_seen": 337969152 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004533199598796389, + "loss": 1.5847, + "theoretical_loss": 4.117776379432392, + "tokens_seen": 338034688 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045330992978936816, + "loss": 1.6847, + "theoretical_loss": 4.117685170292786, + "tokens_seen": 338100224 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004532998996990973, + "loss": 1.6608, + "theoretical_loss": 4.1175939837802815, + "tokens_seen": 338165760 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004532898696088265, + "loss": 1.5852, + "theoretical_loss": 4.117502819884882, + "tokens_seen": 338231296 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045327983951855565, + "loss": 1.6036, + "theoretical_loss": 4.117411678596596, + "tokens_seen": 338296832 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004532698094282849, + "loss": 1.5663, + "theoretical_loss": 4.117320559905441, + "tokens_seen": 338362368 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045325977933801406, + "loss": 1.5775, + "theoretical_loss": 4.117229463801439, + "tokens_seen": 338427904 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045324974924774324, + "loss": 1.6569, + "theoretical_loss": 4.117138390274619, + "tokens_seen": 338493440 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004532397191574724, + "loss": 1.7173, + "theoretical_loss": 4.117047339315014, + "tokens_seen": 338558976 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004532296890672016, + "loss": 1.6035, + "theoretical_loss": 4.116956310912666, + "tokens_seen": 338624512 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004532196589769308, + "loss": 1.6322, + "theoretical_loss": 4.116865305057625, + "tokens_seen": 338690048 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045320962888666, + "loss": 1.7003, + "theoretical_loss": 4.116774321739942, + "tokens_seen": 338755584 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045319959879638915, + "loss": 1.6805, + "theoretical_loss": 4.116683360949678, + "tokens_seen": 338821120 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004531895687061184, + "loss": 1.676, + "theoretical_loss": 4.1165924226769, + "tokens_seen": 338886656 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045317953861584757, + "loss": 1.6916, + "theoretical_loss": 4.11650150691168, + "tokens_seen": 338952192 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045316950852557675, + "loss": 1.6231, + "theoretical_loss": 4.116410613644098, + "tokens_seen": 339017728 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045315947843530593, + "loss": 1.6264, + "theoretical_loss": 4.116319742864239, + "tokens_seen": 339083264 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004531494483450351, + "loss": 1.6038, + "theoretical_loss": 4.1162288945621945, + "tokens_seen": 339148800 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004531394182547643, + "loss": 1.5817, + "theoretical_loss": 4.116138068728063, + "tokens_seen": 339214336 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004531293881644935, + "loss": 1.4855, + "theoretical_loss": 4.1160472653519475, + "tokens_seen": 339279872 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045311935807422265, + "loss": 1.6239, + "theoretical_loss": 4.11595648442396, + "tokens_seen": 339345408 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004531093279839519, + "loss": 1.6541, + "theoretical_loss": 4.115865725934217, + "tokens_seen": 339410944 + }, + { + "epoch": 0.1, + "learning_rate": 0.000453099297893681, + "loss": 1.6159, + "theoretical_loss": 4.11577498987284, + "tokens_seen": 339476480 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045308926780341025, + "loss": 1.592, + "theoretical_loss": 4.11568427622996, + "tokens_seen": 339542016 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045307923771313943, + "loss": 1.7189, + "theoretical_loss": 4.115593584995712, + "tokens_seen": 339607552 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004530692076228686, + "loss": 1.6534, + "theoretical_loss": 4.115502916160237, + "tokens_seen": 339673088 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004530591775325978, + "loss": 1.6255, + "theoretical_loss": 4.115412269713684, + "tokens_seen": 339738624 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045304914744232703, + "loss": 1.631, + "theoretical_loss": 4.115321645646207, + "tokens_seen": 339804160 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045303911735205616, + "loss": 1.591, + "theoretical_loss": 4.115231043947968, + "tokens_seen": 339869696 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004530290872617854, + "loss": 1.6594, + "theoretical_loss": 4.11514046460913, + "tokens_seen": 339935232 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004530190571715145, + "loss": 1.5468, + "theoretical_loss": 4.115049907619868, + "tokens_seen": 340000768 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045300902708124375, + "loss": 1.6143, + "theoretical_loss": 4.114959372970362, + "tokens_seen": 340066304 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045299899699097294, + "loss": 1.7306, + "theoretical_loss": 4.114868860650796, + "tokens_seen": 340131840 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004529889669007021, + "loss": 1.597, + "theoretical_loss": 4.114778370651361, + "tokens_seen": 340197376 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004529789368104313, + "loss": 1.6019, + "theoretical_loss": 4.114687902962258, + "tokens_seen": 340262912 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004529689067201605, + "loss": 1.6451, + "theoretical_loss": 4.114597457573687, + "tokens_seen": 340328448 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045295887662988966, + "loss": 1.6665, + "theoretical_loss": 4.11450703447586, + "tokens_seen": 340393984 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004529488465396189, + "loss": 1.6291, + "theoretical_loss": 4.114416633658992, + "tokens_seen": 340459520 + }, + { + "epoch": 0.1, + "learning_rate": 0.000452938816449348, + "loss": 1.5703, + "theoretical_loss": 4.114326255113308, + "tokens_seen": 340525056 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045292878635907726, + "loss": 1.7362, + "theoretical_loss": 4.114235898829034, + "tokens_seen": 340590592 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004529187562688064, + "loss": 1.5455, + "theoretical_loss": 4.114145564796406, + "tokens_seen": 340656128 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004529087261785356, + "loss": 1.6518, + "theoretical_loss": 4.114055253005665, + "tokens_seen": 340721664 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 132897, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5353009700775146, + "objective/train/theoretical_loss": 4.113964963447058, + "objective/train/tokens_used": 340787200, + "theoretical_loss": 4.113964963447058, + "tokens_seen": 340787200 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004528986960882648, + "loss": 1.5543, + "theoretical_loss": 4.113964963447058, + "tokens_seen": 340787200 + }, + { + "epoch": 0.1, + "learning_rate": 0.000452888665997994, + "loss": 1.6896, + "theoretical_loss": 4.113874696110838, + "tokens_seen": 340852736 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045287863590772316, + "loss": 1.704, + "theoretical_loss": 4.113784450987264, + "tokens_seen": 340918272 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004528686058174524, + "loss": 1.5522, + "theoretical_loss": 4.113694228066604, + "tokens_seen": 340983808 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004528585757271815, + "loss": 1.7009, + "theoretical_loss": 4.113604027339126, + "tokens_seen": 341049344 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045284854563691076, + "loss": 1.6975, + "theoretical_loss": 4.11351384879511, + "tokens_seen": 341114880 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004528385155466399, + "loss": 1.5545, + "theoretical_loss": 4.11342369242484, + "tokens_seen": 341180416 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004528284854563691, + "loss": 1.6505, + "theoretical_loss": 4.1133335582186055, + "tokens_seen": 341245952 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004528184553660983, + "loss": 1.5308, + "theoretical_loss": 4.1132434461667025, + "tokens_seen": 341311488 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004528084252758275, + "loss": 1.5249, + "theoretical_loss": 4.113153356259433, + "tokens_seen": 341377024 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045279839518555667, + "loss": 1.5866, + "theoretical_loss": 4.1130632884871074, + "tokens_seen": 341442560 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045278836509528585, + "loss": 1.5997, + "theoretical_loss": 4.112973242840038, + "tokens_seen": 341508096 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045277833500501503, + "loss": 1.6168, + "theoretical_loss": 4.1128832193085465, + "tokens_seen": 341573632 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045276830491474426, + "loss": 1.7204, + "theoretical_loss": 4.11279321788296, + "tokens_seen": 341639168 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004527582748244734, + "loss": 1.6904, + "theoretical_loss": 4.112703238553609, + "tokens_seen": 341704704 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004527482447342026, + "loss": 1.672, + "theoretical_loss": 4.112613281310835, + "tokens_seen": 341770240 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045273821464393175, + "loss": 1.6556, + "theoretical_loss": 4.1125233461449815, + "tokens_seen": 341835776 + }, + { + "epoch": 0.1, + "learning_rate": 0.000452728184553661, + "loss": 1.6173, + "theoretical_loss": 4.1124334330464, + "tokens_seen": 341901312 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045271815446339017, + "loss": 1.7135, + "theoretical_loss": 4.112343542005447, + "tokens_seen": 341966848 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045270812437311935, + "loss": 1.6631, + "theoretical_loss": 4.112253673012486, + "tokens_seen": 342032384 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004526980942828486, + "loss": 1.6837, + "theoretical_loss": 4.112163826057886, + "tokens_seen": 342097920 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045268806419257777, + "loss": 1.6086, + "theoretical_loss": 4.112074001132023, + "tokens_seen": 342163456 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045267803410230695, + "loss": 1.6718, + "theoretical_loss": 4.111984198225278, + "tokens_seen": 342228992 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045266800401203613, + "loss": 1.6062, + "theoretical_loss": 4.1118944173280365, + "tokens_seen": 342294528 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004526579739217653, + "loss": 1.5623, + "theoretical_loss": 4.111804658430694, + "tokens_seen": 342360064 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004526479438314945, + "loss": 1.6959, + "theoretical_loss": 4.11171492152365, + "tokens_seen": 342425600 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004526379137412237, + "loss": 1.6594, + "theoretical_loss": 4.111625206597309, + "tokens_seen": 342491136 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045262788365095285, + "loss": 1.6422, + "theoretical_loss": 4.111535513642082, + "tokens_seen": 342556672 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004526178535606821, + "loss": 1.6507, + "theoretical_loss": 4.111445842648387, + "tokens_seen": 342622208 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004526078234704112, + "loss": 1.5304, + "theoretical_loss": 4.111356193606648, + "tokens_seen": 342687744 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045259779338014045, + "loss": 1.5299, + "theoretical_loss": 4.111266566507294, + "tokens_seen": 342753280 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045258776328986963, + "loss": 1.5828, + "theoretical_loss": 4.11117696134076, + "tokens_seen": 342818816 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004525777331995988, + "loss": 1.6547, + "theoretical_loss": 4.111087378097488, + "tokens_seen": 342884352 + }, + { + "epoch": 0.1, + "learning_rate": 0.000452567703109328, + "loss": 1.5142, + "theoretical_loss": 4.1109978167679255, + "tokens_seen": 342949888 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045255767301905723, + "loss": 1.6678, + "theoretical_loss": 4.110908277342526, + "tokens_seen": 343015424 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045254764292878636, + "loss": 1.7182, + "theoretical_loss": 4.110818759811748, + "tokens_seen": 343080960 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004525376128385156, + "loss": 1.5995, + "theoretical_loss": 4.110729264166058, + "tokens_seen": 343146496 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004525275827482447, + "loss": 1.6564, + "theoretical_loss": 4.110639790395926, + "tokens_seen": 343212032 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045251755265797395, + "loss": 1.5997, + "theoretical_loss": 4.110550338491832, + "tokens_seen": 343277568 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045250752256770314, + "loss": 1.5582, + "theoretical_loss": 4.1104609084442565, + "tokens_seen": 343343104 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004524974924774323, + "loss": 1.6593, + "theoretical_loss": 4.11037150024369, + "tokens_seen": 343408640 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004524874623871615, + "loss": 1.6774, + "theoretical_loss": 4.110282113880627, + "tokens_seen": 343474176 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004524774322968907, + "loss": 1.6147, + "theoretical_loss": 4.11019274934557, + "tokens_seen": 343539712 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045246740220661986, + "loss": 1.5375, + "theoretical_loss": 4.110103406629026, + "tokens_seen": 343605248 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004524573721163491, + "loss": 1.6497, + "theoretical_loss": 4.110014085721506, + "tokens_seen": 343670784 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004524473420260782, + "loss": 1.6157, + "theoretical_loss": 4.109924786613531, + "tokens_seen": 343736320 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045243731193580746, + "loss": 1.5162, + "theoretical_loss": 4.109835509295625, + "tokens_seen": 343801856 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004524272818455366, + "loss": 1.5906, + "theoretical_loss": 4.10974625375832, + "tokens_seen": 343867392 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004524172517552658, + "loss": 1.7054, + "theoretical_loss": 4.1096570199921505, + "tokens_seen": 343932928 + }, + { + "epoch": 0.1, + "learning_rate": 0.000452407221664995, + "loss": 1.603, + "theoretical_loss": 4.109567807987661, + "tokens_seen": 343998464 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 134155, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6215081214904785, + "objective/train/theoretical_loss": 4.109478617735399, + "objective/train/tokens_used": 344064000, + "theoretical_loss": 4.109478617735399, + "tokens_seen": 344064000 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004523971915747242, + "loss": 1.6398, + "theoretical_loss": 4.109478617735399, + "tokens_seen": 344064000 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045238716148445336, + "loss": 1.5878, + "theoretical_loss": 4.10938944922592, + "tokens_seen": 344129536 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004523771313941826, + "loss": 1.616, + "theoretical_loss": 4.109300302449784, + "tokens_seen": 344195072 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004523671013039117, + "loss": 1.684, + "theoretical_loss": 4.109211177397556, + "tokens_seen": 344260608 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045235707121364096, + "loss": 1.5869, + "theoretical_loss": 4.109122074059811, + "tokens_seen": 344326144 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004523470411233701, + "loss": 1.5992, + "theoretical_loss": 4.1090329924271245, + "tokens_seen": 344391680 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004523370110330993, + "loss": 1.5337, + "theoretical_loss": 4.108943932490082, + "tokens_seen": 344457216 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004523269809428285, + "loss": 1.6068, + "theoretical_loss": 4.108854894239271, + "tokens_seen": 344522752 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004523169508525577, + "loss": 1.6181, + "theoretical_loss": 4.10876587766529, + "tokens_seen": 344588288 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045230692076228687, + "loss": 1.6638, + "theoretical_loss": 4.108676882758738, + "tokens_seen": 344653824 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045229689067201605, + "loss": 1.5901, + "theoretical_loss": 4.1085879095102245, + "tokens_seen": 344719360 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045228686058174523, + "loss": 1.6013, + "theoretical_loss": 4.108498957910362, + "tokens_seen": 344784896 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045227683049147446, + "loss": 1.4628, + "theoretical_loss": 4.108410027949769, + "tokens_seen": 344850432 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004522668004012036, + "loss": 1.5065, + "theoretical_loss": 4.10832111961907, + "tokens_seen": 344915968 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004522567703109328, + "loss": 1.5791, + "theoretical_loss": 4.1082322329088985, + "tokens_seen": 344981504 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045224674022066195, + "loss": 1.6004, + "theoretical_loss": 4.108143367809888, + "tokens_seen": 345047040 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004522367101303912, + "loss": 1.6338, + "theoretical_loss": 4.108054524312683, + "tokens_seen": 345112576 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045222668004012037, + "loss": 1.5444, + "theoretical_loss": 4.107965702407931, + "tokens_seen": 345178112 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045221664994984955, + "loss": 1.6712, + "theoretical_loss": 4.107876902086286, + "tokens_seen": 345243648 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045220661985957873, + "loss": 1.6503, + "theoretical_loss": 4.107788123338408, + "tokens_seen": 345309184 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045219658976930797, + "loss": 1.7155, + "theoretical_loss": 4.1076993661549634, + "tokens_seen": 345374720 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004521865596790371, + "loss": 1.5733, + "theoretical_loss": 4.107610630526622, + "tokens_seen": 345440256 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045217652958876633, + "loss": 1.5609, + "theoretical_loss": 4.107521916444064, + "tokens_seen": 345505792 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045216649949849546, + "loss": 1.6962, + "theoretical_loss": 4.10743322389797, + "tokens_seen": 345571328 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004521564694082247, + "loss": 1.5156, + "theoretical_loss": 4.107344552879029, + "tokens_seen": 345636864 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045214643931795387, + "loss": 1.5808, + "theoretical_loss": 4.107255903377938, + "tokens_seen": 345702400 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045213640922768305, + "loss": 1.6249, + "theoretical_loss": 4.107167275385395, + "tokens_seen": 345767936 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045212637913741223, + "loss": 1.4924, + "theoretical_loss": 4.107078668892108, + "tokens_seen": 345833472 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004521163490471414, + "loss": 1.5855, + "theoretical_loss": 4.106990083888789, + "tokens_seen": 345899008 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004521063189568706, + "loss": 1.6063, + "theoretical_loss": 4.106901520366154, + "tokens_seen": 345964544 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045209628886659983, + "loss": 1.5503, + "theoretical_loss": 4.106812978314929, + "tokens_seen": 346030080 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045208625877632896, + "loss": 1.6287, + "theoretical_loss": 4.106724457725842, + "tokens_seen": 346095616 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004520762286860582, + "loss": 1.5584, + "theoretical_loss": 4.106635958589628, + "tokens_seen": 346161152 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004520661985957873, + "loss": 1.624, + "theoretical_loss": 4.106547480897029, + "tokens_seen": 346226688 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045205616850551656, + "loss": 1.6159, + "theoretical_loss": 4.106459024638791, + "tokens_seen": 346292224 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045204613841524574, + "loss": 1.5464, + "theoretical_loss": 4.106370589805666, + "tokens_seen": 346357760 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004520361083249749, + "loss": 1.5321, + "theoretical_loss": 4.1062821763884125, + "tokens_seen": 346423296 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004520260782347041, + "loss": 1.6541, + "theoretical_loss": 4.106193784377794, + "tokens_seen": 346488832 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045201604814443334, + "loss": 1.6049, + "theoretical_loss": 4.106105413764581, + "tokens_seen": 346554368 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045200601805416246, + "loss": 1.6552, + "theoretical_loss": 4.106017064539548, + "tokens_seen": 346619904 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004519959879638917, + "loss": 1.6041, + "theoretical_loss": 4.1059287366934765, + "tokens_seen": 346685440 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004519859578736208, + "loss": 1.5638, + "theoretical_loss": 4.105840430217153, + "tokens_seen": 346750976 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045197592778335006, + "loss": 1.5953, + "theoretical_loss": 4.10575214510137, + "tokens_seen": 346816512 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045196589769307924, + "loss": 1.5797, + "theoretical_loss": 4.105663881336925, + "tokens_seen": 346882048 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004519558676028084, + "loss": 1.6553, + "theoretical_loss": 4.105575638914623, + "tokens_seen": 346947584 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045194583751253766, + "loss": 1.5948, + "theoretical_loss": 4.105487417825273, + "tokens_seen": 347013120 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004519358074222668, + "loss": 1.4931, + "theoretical_loss": 4.10539921805969, + "tokens_seen": 347078656 + }, + { + "epoch": 0.11, + "learning_rate": 0.000451925777331996, + "loss": 1.5911, + "theoretical_loss": 4.105311039608695, + "tokens_seen": 347144192 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004519157472417252, + "loss": 1.6819, + "theoretical_loss": 4.105222882463115, + "tokens_seen": 347209728 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004519057171514544, + "loss": 1.6212, + "theoretical_loss": 4.105134746613783, + "tokens_seen": 347275264 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 135262, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7094647884368896, + "objective/train/theoretical_loss": 4.105046632051535, + "objective/train/tokens_used": 347340800, + "theoretical_loss": 4.105046632051535, + "tokens_seen": 347340800 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045189568706118356, + "loss": 1.6674, + "theoretical_loss": 4.105046632051535, + "tokens_seen": 347340800 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004518856569709128, + "loss": 1.5784, + "theoretical_loss": 4.1049585387672165, + "tokens_seen": 347406336 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004518756268806419, + "loss": 1.5813, + "theoretical_loss": 4.1048704667516756, + "tokens_seen": 347471872 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045186559679037116, + "loss": 1.6348, + "theoretical_loss": 4.104782415995768, + "tokens_seen": 347537408 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004518555667001003, + "loss": 1.4536, + "theoretical_loss": 4.104694386490354, + "tokens_seen": 347602944 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004518455366098295, + "loss": 1.5432, + "theoretical_loss": 4.1046063782263, + "tokens_seen": 347668480 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004518355065195587, + "loss": 1.4624, + "theoretical_loss": 4.104518391194477, + "tokens_seen": 347734016 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004518254764292879, + "loss": 1.688, + "theoretical_loss": 4.104430425385765, + "tokens_seen": 347799552 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045181544633901707, + "loss": 1.619, + "theoretical_loss": 4.104342480791044, + "tokens_seen": 347865088 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045180541624874625, + "loss": 1.6477, + "theoretical_loss": 4.104254557401205, + "tokens_seen": 347930624 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045179538615847543, + "loss": 1.6315, + "theoretical_loss": 4.104166655207142, + "tokens_seen": 347996160 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045178535606820466, + "loss": 1.6673, + "theoretical_loss": 4.104078774199754, + "tokens_seen": 348061696 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004517753259779338, + "loss": 1.5553, + "theoretical_loss": 4.10399091436995, + "tokens_seen": 348127232 + }, + { + "epoch": 0.11, + "learning_rate": 0.000451765295887663, + "loss": 1.6397, + "theoretical_loss": 4.103903075708637, + "tokens_seen": 348192768 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045175526579739215, + "loss": 1.6289, + "theoretical_loss": 4.103815258206735, + "tokens_seen": 348258304 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004517452357071214, + "loss": 1.5936, + "theoretical_loss": 4.103727461855165, + "tokens_seen": 348323840 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045173520561685057, + "loss": 1.5047, + "theoretical_loss": 4.103639686644856, + "tokens_seen": 348389376 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045172517552657975, + "loss": 1.7077, + "theoretical_loss": 4.103551932566741, + "tokens_seen": 348454912 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045171514543630893, + "loss": 1.6717, + "theoretical_loss": 4.103464199611759, + "tokens_seen": 348520448 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045170511534603817, + "loss": 1.5603, + "theoretical_loss": 4.1033764877708565, + "tokens_seen": 348585984 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004516950852557673, + "loss": 1.4777, + "theoretical_loss": 4.103288797034983, + "tokens_seen": 348651520 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045168505516549653, + "loss": 1.6448, + "theoretical_loss": 4.103201127395095, + "tokens_seen": 348717056 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045167502507522566, + "loss": 1.6443, + "theoretical_loss": 4.103113478842153, + "tokens_seen": 348782592 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004516649949849549, + "loss": 1.6048, + "theoretical_loss": 4.103025851367125, + "tokens_seen": 348848128 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045165496489468407, + "loss": 1.5937, + "theoretical_loss": 4.102938244960985, + "tokens_seen": 348913664 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045164493480441325, + "loss": 1.6799, + "theoretical_loss": 4.102850659614709, + "tokens_seen": 348979200 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045163490471414244, + "loss": 1.8311, + "theoretical_loss": 4.102763095319282, + "tokens_seen": 349044736 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004516248746238716, + "loss": 1.6292, + "theoretical_loss": 4.102675552065694, + "tokens_seen": 349110272 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004516148445336008, + "loss": 1.5782, + "theoretical_loss": 4.10258802984494, + "tokens_seen": 349175808 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045160481444333003, + "loss": 1.6289, + "theoretical_loss": 4.102500528648019, + "tokens_seen": 349241344 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045159478435305916, + "loss": 1.5734, + "theoretical_loss": 4.102413048465938, + "tokens_seen": 349306880 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004515847542627884, + "loss": 1.5566, + "theoretical_loss": 4.10232558928971, + "tokens_seen": 349372416 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004515747241725175, + "loss": 1.6235, + "theoretical_loss": 4.10223815111035, + "tokens_seen": 349437952 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045156469408224676, + "loss": 1.6353, + "theoretical_loss": 4.102150733918882, + "tokens_seen": 349503488 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045155466399197594, + "loss": 1.6734, + "theoretical_loss": 4.102063337706333, + "tokens_seen": 349569024 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004515446339017051, + "loss": 1.6886, + "theoretical_loss": 4.101975962463738, + "tokens_seen": 349634560 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004515346038114343, + "loss": 1.577, + "theoretical_loss": 4.101888608182136, + "tokens_seen": 349700096 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045152457372116354, + "loss": 1.6256, + "theoretical_loss": 4.101801274852571, + "tokens_seen": 349765632 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045151454363089266, + "loss": 1.5243, + "theoretical_loss": 4.1017139624660945, + "tokens_seen": 349831168 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004515045135406219, + "loss": 1.65, + "theoretical_loss": 4.1016266710137606, + "tokens_seen": 349896704 + }, + { + "epoch": 0.11, + "learning_rate": 0.000451494483450351, + "loss": 1.7435, + "theoretical_loss": 4.101539400486631, + "tokens_seen": 349962240 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045148445336008026, + "loss": 1.5551, + "theoretical_loss": 4.101452150875773, + "tokens_seen": 350027776 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045147442326980944, + "loss": 1.6061, + "theoretical_loss": 4.101364922172259, + "tokens_seen": 350093312 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004514643931795386, + "loss": 1.6238, + "theoretical_loss": 4.101277714367166, + "tokens_seen": 350158848 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004514543630892678, + "loss": 1.6454, + "theoretical_loss": 4.1011905274515765, + "tokens_seen": 350224384 + }, + { + "epoch": 0.11, + "learning_rate": 0.000451444332998997, + "loss": 1.6467, + "theoretical_loss": 4.101103361416581, + "tokens_seen": 350289920 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045143430290872617, + "loss": 1.5624, + "theoretical_loss": 4.101016216253273, + "tokens_seen": 350355456 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004514242728184554, + "loss": 1.7226, + "theoretical_loss": 4.10092909195275, + "tokens_seen": 350420992 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045141424272818453, + "loss": 1.5295, + "theoretical_loss": 4.10084198850612, + "tokens_seen": 350486528 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045140421263791376, + "loss": 1.6043, + "theoretical_loss": 4.100754905904491, + "tokens_seen": 350552064 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 136425, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7181216478347778, + "objective/train/theoretical_loss": 4.100667844138981, + "objective/train/tokens_used": 350617600, + "theoretical_loss": 4.100667844138981, + "tokens_seen": 350617600 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045139418254764294, + "loss": 1.606, + "theoretical_loss": 4.100667844138981, + "tokens_seen": 350617600 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004513841524573721, + "loss": 1.5922, + "theoretical_loss": 4.10058080320071, + "tokens_seen": 350683136 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004513741223671013, + "loss": 1.7682, + "theoretical_loss": 4.100493783080806, + "tokens_seen": 350748672 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004513640922768305, + "loss": 1.6071, + "theoretical_loss": 4.1004067837704, + "tokens_seen": 350814208 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045135406218655967, + "loss": 1.6375, + "theoretical_loss": 4.1003198052606304, + "tokens_seen": 350879744 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004513440320962889, + "loss": 1.4804, + "theoretical_loss": 4.10023284754264, + "tokens_seen": 350945280 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045133400200601803, + "loss": 1.685, + "theoretical_loss": 4.100145910607576, + "tokens_seen": 351010816 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045132397191574727, + "loss": 1.5736, + "theoretical_loss": 4.100058994446596, + "tokens_seen": 351076352 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004513139418254764, + "loss": 1.588, + "theoretical_loss": 4.099972099050856, + "tokens_seen": 351141888 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045130391173520563, + "loss": 1.5371, + "theoretical_loss": 4.099885224411522, + "tokens_seen": 351207424 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004512938816449348, + "loss": 1.5948, + "theoretical_loss": 4.099798370519764, + "tokens_seen": 351272960 + }, + { + "epoch": 0.11, + "learning_rate": 0.000451283851554664, + "loss": 1.6046, + "theoretical_loss": 4.099711537366758, + "tokens_seen": 351338496 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045127382146439317, + "loss": 1.611, + "theoretical_loss": 4.099624724943685, + "tokens_seen": 351404032 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045126379137412235, + "loss": 1.5681, + "theoretical_loss": 4.099537933241731, + "tokens_seen": 351469568 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045125376128385153, + "loss": 1.6065, + "theoretical_loss": 4.0994511622520875, + "tokens_seen": 351535104 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045124373119358077, + "loss": 1.6477, + "theoretical_loss": 4.099364411965952, + "tokens_seen": 351600640 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004512337011033099, + "loss": 1.6884, + "theoretical_loss": 4.0992776823745265, + "tokens_seen": 351666176 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045122367101303913, + "loss": 1.6003, + "theoretical_loss": 4.09919097346902, + "tokens_seen": 351731712 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004512136409227683, + "loss": 1.6057, + "theoretical_loss": 4.099104285240646, + "tokens_seen": 351797248 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004512036108324975, + "loss": 1.657, + "theoretical_loss": 4.099017617680621, + "tokens_seen": 351862784 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045119358074222673, + "loss": 1.6184, + "theoretical_loss": 4.098930970780172, + "tokens_seen": 351928320 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045118355065195586, + "loss": 1.6378, + "theoretical_loss": 4.098844344530526, + "tokens_seen": 351993856 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004511735205616851, + "loss": 1.4792, + "theoretical_loss": 4.098757738922918, + "tokens_seen": 352059392 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004511634904714143, + "loss": 1.5763, + "theoretical_loss": 4.09867115394859, + "tokens_seen": 352124928 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045115346038114345, + "loss": 1.5227, + "theoretical_loss": 4.098584589598786, + "tokens_seen": 352190464 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045114343029087264, + "loss": 1.7476, + "theoretical_loss": 4.098498045864756, + "tokens_seen": 352256000 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004511334002006018, + "loss": 1.6475, + "theoretical_loss": 4.0984115227377576, + "tokens_seen": 352321536 + }, + { + "epoch": 0.11, + "learning_rate": 0.000451123370110331, + "loss": 1.586, + "theoretical_loss": 4.098325020209051, + "tokens_seen": 352387072 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045111334002006023, + "loss": 1.6471, + "theoretical_loss": 4.098238538269904, + "tokens_seen": 352452608 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045110330992978936, + "loss": 1.5824, + "theoretical_loss": 4.098152076911588, + "tokens_seen": 352518144 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004510932798395186, + "loss": 1.5294, + "theoretical_loss": 4.098065636125381, + "tokens_seen": 352583680 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004510832497492477, + "loss": 1.6206, + "theoretical_loss": 4.097979215902566, + "tokens_seen": 352649216 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045107321965897696, + "loss": 1.653, + "theoretical_loss": 4.097892816234429, + "tokens_seen": 352714752 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045106318956870614, + "loss": 1.6304, + "theoretical_loss": 4.097806437112267, + "tokens_seen": 352780288 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004510531594784353, + "loss": 1.6695, + "theoretical_loss": 4.0977200785273755, + "tokens_seen": 352845824 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004510431293881645, + "loss": 1.5541, + "theoretical_loss": 4.09763374047106, + "tokens_seen": 352911360 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045103309929789374, + "loss": 1.6352, + "theoretical_loss": 4.097547422934629, + "tokens_seen": 352976896 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045102306920762286, + "loss": 1.5854, + "theoretical_loss": 4.097461125909398, + "tokens_seen": 353042432 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004510130391173521, + "loss": 1.5577, + "theoretical_loss": 4.097374849386685, + "tokens_seen": 353107968 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004510030090270812, + "loss": 1.6965, + "theoretical_loss": 4.097288593357817, + "tokens_seen": 353173504 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045099297893681046, + "loss": 1.6496, + "theoretical_loss": 4.097202357814124, + "tokens_seen": 353239040 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045098294884653964, + "loss": 1.6417, + "theoretical_loss": 4.097116142746941, + "tokens_seen": 353304576 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004509729187562688, + "loss": 1.548, + "theoretical_loss": 4.09702994814761, + "tokens_seen": 353370112 + }, + { + "epoch": 0.11, + "learning_rate": 0.000450962888665998, + "loss": 1.6034, + "theoretical_loss": 4.096943774007476, + "tokens_seen": 353435648 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004509528585757272, + "loss": 1.6217, + "theoretical_loss": 4.096857620317892, + "tokens_seen": 353501184 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045094282848545637, + "loss": 1.6567, + "theoretical_loss": 4.096771487070214, + "tokens_seen": 353566720 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004509327983951856, + "loss": 1.6481, + "theoretical_loss": 4.096685374255803, + "tokens_seen": 353632256 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045092276830491473, + "loss": 1.6026, + "theoretical_loss": 4.096599281866028, + "tokens_seen": 353697792 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045091273821464396, + "loss": 1.6196, + "theoretical_loss": 4.09651320989226, + "tokens_seen": 353763328 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045090270812437314, + "loss": 1.5346, + "theoretical_loss": 4.096427158325877, + "tokens_seen": 353828864 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 138080, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5399386882781982, + "objective/train/theoretical_loss": 4.096341127158263, + "objective/train/tokens_used": 353894400, + "theoretical_loss": 4.096341127158263, + "tokens_seen": 353894400 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004508926780341023, + "loss": 1.5374, + "theoretical_loss": 4.096341127158263, + "tokens_seen": 353894400 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004508826479438315, + "loss": 1.6001, + "theoretical_loss": 4.096255116380806, + "tokens_seen": 353959936 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004508726178535607, + "loss": 1.5105, + "theoretical_loss": 4.096169125984899, + "tokens_seen": 354025472 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045086258776328987, + "loss": 1.6584, + "theoretical_loss": 4.0960831559619395, + "tokens_seen": 354091008 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004508525576730191, + "loss": 1.6416, + "theoretical_loss": 4.095997206303333, + "tokens_seen": 354156544 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045084252758274823, + "loss": 1.6537, + "theoretical_loss": 4.095911277000488, + "tokens_seen": 354222080 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045083249749247747, + "loss": 1.674, + "theoretical_loss": 4.095825368044819, + "tokens_seen": 354287616 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004508224674022066, + "loss": 1.5785, + "theoretical_loss": 4.0957394794277455, + "tokens_seen": 354353152 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045081243731193583, + "loss": 1.5984, + "theoretical_loss": 4.095653611140692, + "tokens_seen": 354418688 + }, + { + "epoch": 0.11, + "learning_rate": 0.000450802407221665, + "loss": 1.4898, + "theoretical_loss": 4.09556776317509, + "tokens_seen": 354484224 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004507923771313942, + "loss": 1.6572, + "theoretical_loss": 4.095481935522372, + "tokens_seen": 354549760 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045078234704112337, + "loss": 1.5135, + "theoretical_loss": 4.09539612817398, + "tokens_seen": 354615296 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045077231695085255, + "loss": 1.5775, + "theoretical_loss": 4.095310341121358, + "tokens_seen": 354680832 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045076228686058173, + "loss": 1.4659, + "theoretical_loss": 4.095224574355958, + "tokens_seen": 354746368 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045075225677031097, + "loss": 1.5973, + "theoretical_loss": 4.095138827869237, + "tokens_seen": 354811904 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004507422266800401, + "loss": 1.5902, + "theoretical_loss": 4.0950531016526535, + "tokens_seen": 354877440 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045073219658976933, + "loss": 1.5541, + "theoretical_loss": 4.094967395697675, + "tokens_seen": 354942976 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004507221664994985, + "loss": 1.6618, + "theoretical_loss": 4.094881709995773, + "tokens_seen": 355008512 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004507121364092277, + "loss": 1.5188, + "theoretical_loss": 4.094796044538423, + "tokens_seen": 355074048 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004507021063189569, + "loss": 1.6095, + "theoretical_loss": 4.094710399317109, + "tokens_seen": 355139584 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045069207622868606, + "loss": 1.6509, + "theoretical_loss": 4.094624774323315, + "tokens_seen": 355205120 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045068204613841524, + "loss": 1.6521, + "theoretical_loss": 4.094539169548535, + "tokens_seen": 355270656 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004506720160481445, + "loss": 1.5547, + "theoretical_loss": 4.094453584984265, + "tokens_seen": 355336192 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004506619859578736, + "loss": 1.5987, + "theoretical_loss": 4.094368020622008, + "tokens_seen": 355401728 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045065195586760284, + "loss": 1.6845, + "theoretical_loss": 4.094282476453272, + "tokens_seen": 355467264 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045064192577733196, + "loss": 1.5689, + "theoretical_loss": 4.094196952469568, + "tokens_seen": 355532800 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004506318956870612, + "loss": 1.6492, + "theoretical_loss": 4.0941114486624155, + "tokens_seen": 355598336 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004506218655967904, + "loss": 1.6662, + "theoretical_loss": 4.094025965023335, + "tokens_seen": 355663872 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045061183550651956, + "loss": 1.6538, + "theoretical_loss": 4.093940501543857, + "tokens_seen": 355729408 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045060180541624874, + "loss": 1.6486, + "theoretical_loss": 4.093855058215513, + "tokens_seen": 355794944 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004505917753259779, + "loss": 1.5441, + "theoretical_loss": 4.093769635029841, + "tokens_seen": 355860480 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004505817452357071, + "loss": 1.5696, + "theoretical_loss": 4.0936842319783855, + "tokens_seen": 355926016 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045057171514543634, + "loss": 1.5766, + "theoretical_loss": 4.093598849052694, + "tokens_seen": 355991552 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045056168505516547, + "loss": 1.4921, + "theoretical_loss": 4.093513486244321, + "tokens_seen": 356057088 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004505516549648947, + "loss": 1.5437, + "theoretical_loss": 4.093428143544823, + "tokens_seen": 356122624 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004505416248746239, + "loss": 1.639, + "theoretical_loss": 4.093342820945766, + "tokens_seen": 356188160 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045053159478435306, + "loss": 1.5368, + "theoretical_loss": 4.093257518438718, + "tokens_seen": 356253696 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045052156469408224, + "loss": 1.5652, + "theoretical_loss": 4.093172236015253, + "tokens_seen": 356319232 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004505115346038114, + "loss": 1.552, + "theoretical_loss": 4.093086973666949, + "tokens_seen": 356384768 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004505015045135406, + "loss": 1.5035, + "theoretical_loss": 4.0930017313853915, + "tokens_seen": 356450304 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045049147442326984, + "loss": 1.554, + "theoretical_loss": 4.092916509162169, + "tokens_seen": 356515840 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045048144433299897, + "loss": 1.525, + "theoretical_loss": 4.092831306988875, + "tokens_seen": 356581376 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004504714142427282, + "loss": 1.6785, + "theoretical_loss": 4.0927461248571095, + "tokens_seen": 356646912 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045046138415245733, + "loss": 1.5919, + "theoretical_loss": 4.092660962758477, + "tokens_seen": 356712448 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045045135406218657, + "loss": 1.5882, + "theoretical_loss": 4.092575820684586, + "tokens_seen": 356777984 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004504413239719158, + "loss": 1.5967, + "theoretical_loss": 4.092490698627052, + "tokens_seen": 356843520 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045043129388164493, + "loss": 1.584, + "theoretical_loss": 4.0924055965774935, + "tokens_seen": 356909056 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045042126379137416, + "loss": 1.514, + "theoretical_loss": 4.092320514527535, + "tokens_seen": 356974592 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045041123370110334, + "loss": 1.5394, + "theoretical_loss": 4.092235452468808, + "tokens_seen": 357040128 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004504012036108325, + "loss": 1.5556, + "theoretical_loss": 4.092150410392945, + "tokens_seen": 357105664 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 139226, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4913017749786377, + "objective/train/theoretical_loss": 4.092065388291585, + "objective/train/tokens_used": 357171200, + "theoretical_loss": 4.092065388291585, + "tokens_seen": 357171200 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004503911735205617, + "loss": 1.5457, + "theoretical_loss": 4.092065388291585, + "tokens_seen": 357171200 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004503811434302909, + "loss": 1.501, + "theoretical_loss": 4.091980386156376, + "tokens_seen": 357236736 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045037111334002007, + "loss": 1.6416, + "theoretical_loss": 4.091895403978965, + "tokens_seen": 357302272 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004503610832497493, + "loss": 1.5475, + "theoretical_loss": 4.091810441751007, + "tokens_seen": 357367808 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045035105315947843, + "loss": 1.5356, + "theoretical_loss": 4.091725499464163, + "tokens_seen": 357433344 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045034102306920767, + "loss": 1.6212, + "theoretical_loss": 4.091640577110096, + "tokens_seen": 357498880 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004503309929789368, + "loss": 1.5061, + "theoretical_loss": 4.091555674680476, + "tokens_seen": 357564416 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045032096288866603, + "loss": 1.6768, + "theoretical_loss": 4.09147079216698, + "tokens_seen": 357629952 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004503109327983952, + "loss": 1.5306, + "theoretical_loss": 4.0913859295612856, + "tokens_seen": 357695488 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004503009027081244, + "loss": 1.5765, + "theoretical_loss": 4.091301086855078, + "tokens_seen": 357761024 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045029087261785357, + "loss": 1.7298, + "theoretical_loss": 4.091216264040048, + "tokens_seen": 357826560 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045028084252758275, + "loss": 1.5925, + "theoretical_loss": 4.091131461107889, + "tokens_seen": 357892096 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045027081243731193, + "loss": 1.4583, + "theoretical_loss": 4.091046678050301, + "tokens_seen": 357957632 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045026078234704117, + "loss": 1.6479, + "theoretical_loss": 4.090961914858989, + "tokens_seen": 358023168 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004502507522567703, + "loss": 1.6242, + "theoretical_loss": 4.090877171525663, + "tokens_seen": 358088704 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045024072216649953, + "loss": 1.6207, + "theoretical_loss": 4.090792448042038, + "tokens_seen": 358154240 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004502306920762287, + "loss": 1.6107, + "theoretical_loss": 4.0907077443998325, + "tokens_seen": 358219776 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004502206619859579, + "loss": 1.4861, + "theoretical_loss": 4.090623060590772, + "tokens_seen": 358285312 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004502106318956871, + "loss": 1.584, + "theoretical_loss": 4.090538396606585, + "tokens_seen": 358350848 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045020060180541626, + "loss": 1.5368, + "theoretical_loss": 4.090453752439008, + "tokens_seen": 358416384 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045019057171514544, + "loss": 1.5517, + "theoretical_loss": 4.09036912807978, + "tokens_seen": 358481920 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004501805416248747, + "loss": 1.6078, + "theoretical_loss": 4.090284523520644, + "tokens_seen": 358547456 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004501705115346038, + "loss": 1.6363, + "theoretical_loss": 4.09019993875335, + "tokens_seen": 358612992 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045016048144433304, + "loss": 1.6208, + "theoretical_loss": 4.090115373769653, + "tokens_seen": 358678528 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045015045135406216, + "loss": 1.6677, + "theoretical_loss": 4.0900308285613125, + "tokens_seen": 358744064 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004501404212637914, + "loss": 1.6484, + "theoretical_loss": 4.089946303120092, + "tokens_seen": 358809600 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004501303911735206, + "loss": 1.5458, + "theoretical_loss": 4.089861797437761, + "tokens_seen": 358875136 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045012036108324976, + "loss": 1.548, + "theoretical_loss": 4.089777311506094, + "tokens_seen": 358940672 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045011033099297894, + "loss": 1.5395, + "theoretical_loss": 4.089692845316868, + "tokens_seen": 359006208 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004501003009027081, + "loss": 1.6593, + "theoretical_loss": 4.08960839886187, + "tokens_seen": 359071744 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004500902708124373, + "loss": 1.5254, + "theoretical_loss": 4.089523972132886, + "tokens_seen": 359137280 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045008024072216654, + "loss": 1.5728, + "theoretical_loss": 4.089439565121712, + "tokens_seen": 359202816 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045007021063189567, + "loss": 1.5987, + "theoretical_loss": 4.089355177820146, + "tokens_seen": 359268352 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004500601805416249, + "loss": 1.6776, + "theoretical_loss": 4.0892708102199915, + "tokens_seen": 359333888 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004500501504513541, + "loss": 1.5344, + "theoretical_loss": 4.089186462313056, + "tokens_seen": 359399424 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045004012036108326, + "loss": 1.4645, + "theoretical_loss": 4.089102134091155, + "tokens_seen": 359464960 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045003009027081244, + "loss": 1.6339, + "theoretical_loss": 4.0890178255461045, + "tokens_seen": 359530496 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004500200601805416, + "loss": 1.5919, + "theoretical_loss": 4.08893353666973, + "tokens_seen": 359596032 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004500100300902708, + "loss": 1.7397, + "theoretical_loss": 4.088849267453858, + "tokens_seen": 359661568 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045000000000000004, + "loss": 1.6093, + "theoretical_loss": 4.088765017890322, + "tokens_seen": 359727104 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044998996990972917, + "loss": 1.6332, + "theoretical_loss": 4.08868078797096, + "tokens_seen": 359792640 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004499799398194584, + "loss": 1.5605, + "theoretical_loss": 4.088596577687614, + "tokens_seen": 359858176 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044996990972918753, + "loss": 1.5881, + "theoretical_loss": 4.088512387032132, + "tokens_seen": 359923712 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044995987963891677, + "loss": 1.5395, + "theoretical_loss": 4.088428215996366, + "tokens_seen": 359989248 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044994984954864595, + "loss": 1.6044, + "theoretical_loss": 4.0883440645721745, + "tokens_seen": 360054784 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044993981945837513, + "loss": 1.6386, + "theoretical_loss": 4.088259932751419, + "tokens_seen": 360120320 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004499297893681043, + "loss": 1.5714, + "theoretical_loss": 4.088175820525966, + "tokens_seen": 360185856 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044991975927783355, + "loss": 1.6227, + "theoretical_loss": 4.088091727887688, + "tokens_seen": 360251392 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044990972918756267, + "loss": 1.6552, + "theoretical_loss": 4.088007654828463, + "tokens_seen": 360316928 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004498996990972919, + "loss": 1.6357, + "theoretical_loss": 4.087923601340169, + "tokens_seen": 360382464 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 140424, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5372493267059326, + "objective/train/theoretical_loss": 4.0878395674146955, + "objective/train/tokens_used": 360448000, + "theoretical_loss": 4.0878395674146955, + "tokens_seen": 360448000 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044988966900702103, + "loss": 1.5165, + "theoretical_loss": 4.0878395674146955, + "tokens_seen": 360448000 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044987963891675027, + "loss": 1.559, + "theoretical_loss": 4.087755553043934, + "tokens_seen": 360513536 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044986960882647945, + "loss": 1.67, + "theoretical_loss": 4.087671558219778, + "tokens_seen": 360579072 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044985957873620863, + "loss": 1.5665, + "theoretical_loss": 4.087587582934131, + "tokens_seen": 360644608 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004498495486459378, + "loss": 1.6447, + "theoretical_loss": 4.087503627178897, + "tokens_seen": 360710144 + }, + { + "epoch": 0.11, + "learning_rate": 0.000449839518555667, + "loss": 1.6435, + "theoretical_loss": 4.087419690945988, + "tokens_seen": 360775680 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004498294884653962, + "loss": 1.6547, + "theoretical_loss": 4.087335774227318, + "tokens_seen": 360841216 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004498194583751254, + "loss": 1.6213, + "theoretical_loss": 4.0872518770148085, + "tokens_seen": 360906752 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044980942828485454, + "loss": 1.5505, + "theoretical_loss": 4.087167999300384, + "tokens_seen": 360972288 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044979939819458377, + "loss": 1.6932, + "theoretical_loss": 4.087084141075974, + "tokens_seen": 361037824 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004497893681043129, + "loss": 1.6082, + "theoretical_loss": 4.087000302333513, + "tokens_seen": 361103360 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044977933801404213, + "loss": 1.6003, + "theoretical_loss": 4.086916483064941, + "tokens_seen": 361168896 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004497693079237713, + "loss": 1.5356, + "theoretical_loss": 4.086832683262203, + "tokens_seen": 361234432 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004497592778335005, + "loss": 1.5432, + "theoretical_loss": 4.086748902917246, + "tokens_seen": 361299968 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004497492477432297, + "loss": 1.5538, + "theoretical_loss": 4.086665142022025, + "tokens_seen": 361365504 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004497392176529589, + "loss": 1.5316, + "theoretical_loss": 4.086581400568499, + "tokens_seen": 361431040 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044972918756268804, + "loss": 1.559, + "theoretical_loss": 4.086497678548631, + "tokens_seen": 361496576 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004497191574724173, + "loss": 1.5492, + "theoretical_loss": 4.086413975954389, + "tokens_seen": 361562112 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004497091273821464, + "loss": 1.5338, + "theoretical_loss": 4.086330292777745, + "tokens_seen": 361627648 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044969909729187564, + "loss": 1.5915, + "theoretical_loss": 4.086246629010679, + "tokens_seen": 361693184 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004496890672016049, + "loss": 1.6034, + "theoretical_loss": 4.086162984645172, + "tokens_seen": 361758720 + }, + { + "epoch": 0.11, + "learning_rate": 0.000449679037111334, + "loss": 1.5357, + "theoretical_loss": 4.086079359673212, + "tokens_seen": 361824256 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044966900702106324, + "loss": 1.6249, + "theoretical_loss": 4.085995754086789, + "tokens_seen": 361889792 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044965897693079236, + "loss": 1.6182, + "theoretical_loss": 4.085912167877902, + "tokens_seen": 361955328 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004496489468405216, + "loss": 1.5946, + "theoretical_loss": 4.085828601038552, + "tokens_seen": 362020864 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004496389167502508, + "loss": 1.528, + "theoretical_loss": 4.0857450535607445, + "tokens_seen": 362086400 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044962888665997996, + "loss": 1.6314, + "theoretical_loss": 4.085661525436492, + "tokens_seen": 362151936 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044961885656970914, + "loss": 1.5553, + "theoretical_loss": 4.0855780166578075, + "tokens_seen": 362217472 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004496088264794383, + "loss": 1.5835, + "theoretical_loss": 4.0854945272167145, + "tokens_seen": 362283008 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004495987963891675, + "loss": 1.6818, + "theoretical_loss": 4.085411057105237, + "tokens_seen": 362348544 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044958876629889674, + "loss": 1.6405, + "theoretical_loss": 4.0853276063154045, + "tokens_seen": 362414080 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044957873620862587, + "loss": 1.6051, + "theoretical_loss": 4.085244174839252, + "tokens_seen": 362479616 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004495687061183551, + "loss": 1.5709, + "theoretical_loss": 4.085160762668819, + "tokens_seen": 362545152 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004495586760280843, + "loss": 1.7403, + "theoretical_loss": 4.085077369796149, + "tokens_seen": 362610688 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044954864593781346, + "loss": 1.6405, + "theoretical_loss": 4.084993996213292, + "tokens_seen": 362676224 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044953861584754264, + "loss": 1.5936, + "theoretical_loss": 4.084910641912301, + "tokens_seen": 362741760 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004495285857572718, + "loss": 1.6818, + "theoretical_loss": 4.084827306885234, + "tokens_seen": 362807296 + }, + { + "epoch": 0.11, + "learning_rate": 0.000449518555667001, + "loss": 1.6384, + "theoretical_loss": 4.084743991124155, + "tokens_seen": 362872832 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044950852557673024, + "loss": 1.5308, + "theoretical_loss": 4.08466069462113, + "tokens_seen": 362938368 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044949849548645937, + "loss": 1.6932, + "theoretical_loss": 4.084577417368232, + "tokens_seen": 363003904 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004494884653961886, + "loss": 1.6928, + "theoretical_loss": 4.084494159357538, + "tokens_seen": 363069440 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044947843530591773, + "loss": 1.5897, + "theoretical_loss": 4.0844109205811305, + "tokens_seen": 363134976 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044946840521564697, + "loss": 1.5581, + "theoretical_loss": 4.084327701031095, + "tokens_seen": 363200512 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044945837512537615, + "loss": 1.5592, + "theoretical_loss": 4.084244500699524, + "tokens_seen": 363266048 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044944834503510533, + "loss": 1.5478, + "theoretical_loss": 4.084161319578511, + "tokens_seen": 363331584 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004494383149448345, + "loss": 1.5299, + "theoretical_loss": 4.084078157660159, + "tokens_seen": 363397120 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044942828485456375, + "loss": 1.5667, + "theoretical_loss": 4.083995014936571, + "tokens_seen": 363462656 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044941825476429287, + "loss": 1.6238, + "theoretical_loss": 4.0839118913998576, + "tokens_seen": 363528192 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004494082246740221, + "loss": 1.6197, + "theoretical_loss": 4.083828787042134, + "tokens_seen": 363593728 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044939819458375123, + "loss": 1.5494, + "theoretical_loss": 4.083745701855518, + "tokens_seen": 363659264 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 141566, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.553741216659546, + "objective/train/theoretical_loss": 4.083662635832135, + "objective/train/tokens_used": 363724800, + "theoretical_loss": 4.083662635832135, + "tokens_seen": 363724800 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044938816449348047, + "loss": 1.5295, + "theoretical_loss": 4.083662635832135, + "tokens_seen": 363724800 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044937813440320965, + "loss": 1.67, + "theoretical_loss": 4.0835795889641116, + "tokens_seen": 363790336 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044936810431293883, + "loss": 1.5025, + "theoretical_loss": 4.0834965612435825, + "tokens_seen": 363855872 + }, + { + "epoch": 0.11, + "learning_rate": 0.000449358074222668, + "loss": 1.6155, + "theoretical_loss": 4.083413552662685, + "tokens_seen": 363921408 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004493480441323972, + "loss": 1.5901, + "theoretical_loss": 4.08333056321356, + "tokens_seen": 363986944 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004493380140421264, + "loss": 1.6095, + "theoretical_loss": 4.083247592888357, + "tokens_seen": 364052480 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004493279839518556, + "loss": 1.6446, + "theoretical_loss": 4.083164641679225, + "tokens_seen": 364118016 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044931795386158474, + "loss": 1.552, + "theoretical_loss": 4.083081709578323, + "tokens_seen": 364183552 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044930792377131397, + "loss": 1.5892, + "theoretical_loss": 4.082998796577811, + "tokens_seen": 364249088 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004492978936810431, + "loss": 1.5199, + "theoretical_loss": 4.082915902669853, + "tokens_seen": 364314624 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044928786359077234, + "loss": 1.6055, + "theoretical_loss": 4.082833027846622, + "tokens_seen": 364380160 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004492778335005015, + "loss": 1.6525, + "theoretical_loss": 4.08275017210029, + "tokens_seen": 364445696 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004492678034102307, + "loss": 1.6668, + "theoretical_loss": 4.082667335423038, + "tokens_seen": 364511232 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004492577733199599, + "loss": 1.6608, + "theoretical_loss": 4.0825845178070495, + "tokens_seen": 364576768 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004492477432296891, + "loss": 1.658, + "theoretical_loss": 4.0825017192445126, + "tokens_seen": 364642304 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044923771313941824, + "loss": 1.6233, + "theoretical_loss": 4.082418939727622, + "tokens_seen": 364707840 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004492276830491475, + "loss": 1.539, + "theoretical_loss": 4.082336179248574, + "tokens_seen": 364773376 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004492176529588766, + "loss": 1.5664, + "theoretical_loss": 4.082253437799572, + "tokens_seen": 364838912 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044920762286860584, + "loss": 1.6154, + "theoretical_loss": 4.082170715372823, + "tokens_seen": 364904448 + }, + { + "epoch": 0.11, + "learning_rate": 0.000449197592778335, + "loss": 1.7171, + "theoretical_loss": 4.082088011960538, + "tokens_seen": 364969984 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004491875626880642, + "loss": 1.6639, + "theoretical_loss": 4.082005327554934, + "tokens_seen": 365035520 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004491775325977934, + "loss": 1.7553, + "theoretical_loss": 4.081922662148231, + "tokens_seen": 365101056 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044916750250752256, + "loss": 1.5944, + "theoretical_loss": 4.081840015732654, + "tokens_seen": 365166592 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044915747241725174, + "loss": 1.6236, + "theoretical_loss": 4.081757388300435, + "tokens_seen": 365232128 + }, + { + "epoch": 0.11, + "learning_rate": 0.000449147442326981, + "loss": 1.5986, + "theoretical_loss": 4.081674779843806, + "tokens_seen": 365297664 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004491374122367101, + "loss": 1.7044, + "theoretical_loss": 4.081592190355008, + "tokens_seen": 365363200 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044912738214643934, + "loss": 1.5543, + "theoretical_loss": 4.081509619826283, + "tokens_seen": 365428736 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044911735205616847, + "loss": 1.5675, + "theoretical_loss": 4.081427068249881, + "tokens_seen": 365494272 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004491073219658977, + "loss": 1.7351, + "theoretical_loss": 4.081344535618053, + "tokens_seen": 365559808 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004490972918756269, + "loss": 1.581, + "theoretical_loss": 4.081262021923058, + "tokens_seen": 365625344 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044908726178535607, + "loss": 1.6704, + "theoretical_loss": 4.081179527157156, + "tokens_seen": 365690880 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044907723169508525, + "loss": 1.6073, + "theoretical_loss": 4.081097051312614, + "tokens_seen": 365756416 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004490672016048145, + "loss": 1.6013, + "theoretical_loss": 4.081014594381705, + "tokens_seen": 365821952 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004490571715145436, + "loss": 1.6857, + "theoretical_loss": 4.0809321563567025, + "tokens_seen": 365887488 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044904714142427284, + "loss": 1.6359, + "theoretical_loss": 4.080849737229887, + "tokens_seen": 365953024 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044903711133400197, + "loss": 1.5874, + "theoretical_loss": 4.080767336993542, + "tokens_seen": 366018560 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004490270812437312, + "loss": 1.5796, + "theoretical_loss": 4.080684955639958, + "tokens_seen": 366084096 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004490170511534604, + "loss": 1.587, + "theoretical_loss": 4.080602593161428, + "tokens_seen": 366149632 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044900702106318957, + "loss": 1.601, + "theoretical_loss": 4.0805202495502515, + "tokens_seen": 366215168 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044899699097291875, + "loss": 1.6574, + "theoretical_loss": 4.080437924798729, + "tokens_seen": 366280704 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044898696088264793, + "loss": 1.6079, + "theoretical_loss": 4.080355618899169, + "tokens_seen": 366346240 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004489769307923771, + "loss": 1.5725, + "theoretical_loss": 4.080273331843884, + "tokens_seen": 366411776 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044896690070210635, + "loss": 1.5608, + "theoretical_loss": 4.080191063625188, + "tokens_seen": 366477312 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004489568706118355, + "loss": 1.6016, + "theoretical_loss": 4.080108814235404, + "tokens_seen": 366542848 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004489468405215647, + "loss": 1.6069, + "theoretical_loss": 4.080026583666855, + "tokens_seen": 366608384 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044893681043129395, + "loss": 1.6201, + "theoretical_loss": 4.079944371911871, + "tokens_seen": 366673920 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044892678034102307, + "loss": 1.5644, + "theoretical_loss": 4.07986217896279, + "tokens_seen": 366739456 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004489167502507523, + "loss": 1.613, + "theoretical_loss": 4.079780004811946, + "tokens_seen": 366804992 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044890672016048143, + "loss": 1.5405, + "theoretical_loss": 4.079697849451684, + "tokens_seen": 366870528 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044889669007021067, + "loss": 1.3653, + "theoretical_loss": 4.079615712874352, + "tokens_seen": 366936064 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 142675, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6674308776855469, + "objective/train/theoretical_loss": 4.079533595072301, + "objective/train/tokens_used": 367001600, + "theoretical_loss": 4.079533595072301, + "tokens_seen": 367001600 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044888665997993985, + "loss": 1.6822, + "theoretical_loss": 4.079533595072301, + "tokens_seen": 367001600 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044887662988966903, + "loss": 1.54, + "theoretical_loss": 4.0794514960378905, + "tokens_seen": 367067136 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004488665997993982, + "loss": 1.6375, + "theoretical_loss": 4.079369415763479, + "tokens_seen": 367132672 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004488565697091274, + "loss": 1.6096, + "theoretical_loss": 4.079287354241432, + "tokens_seen": 367198208 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004488465396188566, + "loss": 1.766, + "theoretical_loss": 4.07920531146412, + "tokens_seen": 367263744 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004488365095285858, + "loss": 1.6479, + "theoretical_loss": 4.0791232874239185, + "tokens_seen": 367329280 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044882647943831494, + "loss": 1.6894, + "theoretical_loss": 4.079041282113206, + "tokens_seen": 367394816 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004488164493480442, + "loss": 1.6858, + "theoretical_loss": 4.078959295524365, + "tokens_seen": 367460352 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004488064192577733, + "loss": 1.5625, + "theoretical_loss": 4.078877327649785, + "tokens_seen": 367525888 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044879638916750254, + "loss": 1.5157, + "theoretical_loss": 4.078795378481858, + "tokens_seen": 367591424 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004487863590772317, + "loss": 1.6092, + "theoretical_loss": 4.07871344801298, + "tokens_seen": 367656960 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004487763289869609, + "loss": 1.5639, + "theoretical_loss": 4.0786315362355525, + "tokens_seen": 367722496 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004487662988966901, + "loss": 1.485, + "theoretical_loss": 4.078549643141981, + "tokens_seen": 367788032 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004487562688064193, + "loss": 1.6242, + "theoretical_loss": 4.078467768724677, + "tokens_seen": 367853568 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044874623871614844, + "loss": 1.5067, + "theoretical_loss": 4.078385912976053, + "tokens_seen": 367919104 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004487362086258777, + "loss": 1.6079, + "theoretical_loss": 4.078304075888529, + "tokens_seen": 367984640 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004487261785356068, + "loss": 1.6734, + "theoretical_loss": 4.078222257454529, + "tokens_seen": 368050176 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044871614844533604, + "loss": 1.5804, + "theoretical_loss": 4.07814045766648, + "tokens_seen": 368115712 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004487061183550652, + "loss": 1.6381, + "theoretical_loss": 4.0780586765168145, + "tokens_seen": 368181248 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004486960882647944, + "loss": 1.5511, + "theoretical_loss": 4.07797691399797, + "tokens_seen": 368246784 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004486860581745236, + "loss": 1.6974, + "theoretical_loss": 4.077895170102386, + "tokens_seen": 368312320 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044867602808425276, + "loss": 1.5138, + "theoretical_loss": 4.077813444822509, + "tokens_seen": 368377856 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044866599799398194, + "loss": 1.7139, + "theoretical_loss": 4.077731738150789, + "tokens_seen": 368443392 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004486559679037112, + "loss": 1.62, + "theoretical_loss": 4.077650050079681, + "tokens_seen": 368508928 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004486459378134403, + "loss": 1.7, + "theoretical_loss": 4.077568380601642, + "tokens_seen": 368574464 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044863590772316954, + "loss": 1.5455, + "theoretical_loss": 4.077486729709136, + "tokens_seen": 368640000 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044862587763289867, + "loss": 1.5636, + "theoretical_loss": 4.077405097394631, + "tokens_seen": 368705536 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004486158475426279, + "loss": 1.5961, + "theoretical_loss": 4.077323483650598, + "tokens_seen": 368771072 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004486058174523571, + "loss": 1.5595, + "theoretical_loss": 4.077241888469514, + "tokens_seen": 368836608 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044859578736208627, + "loss": 1.6865, + "theoretical_loss": 4.077160311843858, + "tokens_seen": 368902144 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044858575727181545, + "loss": 1.5256, + "theoretical_loss": 4.077078753766118, + "tokens_seen": 368967680 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004485757271815447, + "loss": 1.6149, + "theoretical_loss": 4.076997214228782, + "tokens_seen": 369033216 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004485656970912738, + "loss": 1.7822, + "theoretical_loss": 4.076915693224343, + "tokens_seen": 369098752 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044855566700100304, + "loss": 1.5789, + "theoretical_loss": 4.0768341907453, + "tokens_seen": 369164288 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044854563691073217, + "loss": 1.4386, + "theoretical_loss": 4.0767527067841565, + "tokens_seen": 369229824 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004485356068204614, + "loss": 1.6026, + "theoretical_loss": 4.076671241333417, + "tokens_seen": 369295360 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004485255767301906, + "loss": 1.5338, + "theoretical_loss": 4.076589794385596, + "tokens_seen": 369360896 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044851554663991977, + "loss": 1.6727, + "theoretical_loss": 4.076508365933206, + "tokens_seen": 369426432 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044850551654964895, + "loss": 1.6704, + "theoretical_loss": 4.076426955968769, + "tokens_seen": 369491968 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044849548645937813, + "loss": 1.5974, + "theoretical_loss": 4.076345564484807, + "tokens_seen": 369557504 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004484854563691073, + "loss": 1.5516, + "theoretical_loss": 4.076264191473852, + "tokens_seen": 369623040 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044847542627883655, + "loss": 1.6187, + "theoretical_loss": 4.076182836928435, + "tokens_seen": 369688576 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004484653961885657, + "loss": 1.6161, + "theoretical_loss": 4.0761015008410935, + "tokens_seen": 369754112 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004484553660982949, + "loss": 1.6973, + "theoretical_loss": 4.07602018320437, + "tokens_seen": 369819648 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044844533600802404, + "loss": 1.6715, + "theoretical_loss": 4.075938884010809, + "tokens_seen": 369885184 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044843530591775327, + "loss": 1.6053, + "theoretical_loss": 4.075857603252963, + "tokens_seen": 369950720 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044842527582748245, + "loss": 1.5537, + "theoretical_loss": 4.075776340923385, + "tokens_seen": 370016256 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044841524573721163, + "loss": 1.6516, + "theoretical_loss": 4.075695097014635, + "tokens_seen": 370081792 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004484052156469408, + "loss": 1.6181, + "theoretical_loss": 4.075613871519276, + "tokens_seen": 370147328 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044839518555667005, + "loss": 1.6015, + "theoretical_loss": 4.0755326644298755, + "tokens_seen": 370212864 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 143749, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4950112104415894, + "objective/train/theoretical_loss": 4.075451475739007, + "objective/train/tokens_used": 370278400, + "theoretical_loss": 4.075451475739007, + "tokens_seen": 370278400 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004483851554663992, + "loss": 1.5254, + "theoretical_loss": 4.075451475739007, + "tokens_seen": 370278400 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004483751253761284, + "loss": 1.6164, + "theoretical_loss": 4.075370305439243, + "tokens_seen": 370343936 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044836509528585754, + "loss": 1.6486, + "theoretical_loss": 4.075289153523168, + "tokens_seen": 370409472 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004483550651955868, + "loss": 1.6318, + "theoretical_loss": 4.075208019983366, + "tokens_seen": 370475008 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044834503510531596, + "loss": 1.5971, + "theoretical_loss": 4.075126904812425, + "tokens_seen": 370540544 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044833500501504514, + "loss": 1.6103, + "theoretical_loss": 4.0750458080029395, + "tokens_seen": 370606080 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004483249749247743, + "loss": 1.5622, + "theoretical_loss": 4.074964729547506, + "tokens_seen": 370671616 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004483149448345035, + "loss": 1.6382, + "theoretical_loss": 4.074883669438729, + "tokens_seen": 370737152 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004483049147442327, + "loss": 1.6665, + "theoretical_loss": 4.074802627669213, + "tokens_seen": 370802688 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004482948846539619, + "loss": 1.5937, + "theoretical_loss": 4.07472160423157, + "tokens_seen": 370868224 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044828485456369104, + "loss": 1.614, + "theoretical_loss": 4.074640599118413, + "tokens_seen": 370933760 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004482748244734203, + "loss": 1.6208, + "theoretical_loss": 4.074559612322363, + "tokens_seen": 370999296 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004482647943831494, + "loss": 1.6117, + "theoretical_loss": 4.074478643836043, + "tokens_seen": 371064832 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044825476429287864, + "loss": 1.5939, + "theoretical_loss": 4.074397693652081, + "tokens_seen": 371130368 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004482447342026078, + "loss": 1.4925, + "theoretical_loss": 4.074316761763107, + "tokens_seen": 371195904 + }, + { + "epoch": 0.11, + "learning_rate": 0.000448234704112337, + "loss": 1.7206, + "theoretical_loss": 4.074235848161761, + "tokens_seen": 371261440 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004482246740220662, + "loss": 1.5784, + "theoretical_loss": 4.074154952840681, + "tokens_seen": 371326976 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004482146439317954, + "loss": 1.5577, + "theoretical_loss": 4.074074075792513, + "tokens_seen": 371392512 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044820461384152455, + "loss": 1.5762, + "theoretical_loss": 4.073993217009905, + "tokens_seen": 371458048 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004481945837512538, + "loss": 1.6166, + "theoretical_loss": 4.073912376485511, + "tokens_seen": 371523584 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044818455366098296, + "loss": 1.59, + "theoretical_loss": 4.073831554211989, + "tokens_seen": 371589120 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044817452357071214, + "loss": 1.5665, + "theoretical_loss": 4.0737507501820005, + "tokens_seen": 371654656 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004481644934804414, + "loss": 1.5518, + "theoretical_loss": 4.073669964388212, + "tokens_seen": 371720192 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004481544633901705, + "loss": 1.6013, + "theoretical_loss": 4.073589196823293, + "tokens_seen": 371785728 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044814443329989974, + "loss": 1.5353, + "theoretical_loss": 4.073508447479918, + "tokens_seen": 371851264 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044813440320962887, + "loss": 1.7013, + "theoretical_loss": 4.073427716350767, + "tokens_seen": 371916800 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004481243731193581, + "loss": 1.5459, + "theoretical_loss": 4.0733470034285215, + "tokens_seen": 371982336 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004481143430290873, + "loss": 1.5573, + "theoretical_loss": 4.07326630870587, + "tokens_seen": 372047872 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044810431293881647, + "loss": 1.6166, + "theoretical_loss": 4.073185632175504, + "tokens_seen": 372113408 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044809428284854565, + "loss": 1.5977, + "theoretical_loss": 4.073104973830118, + "tokens_seen": 372178944 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004480842527582749, + "loss": 1.5994, + "theoretical_loss": 4.073024333662413, + "tokens_seen": 372244480 + }, + { + "epoch": 0.11, + "learning_rate": 0.000448074222668004, + "loss": 1.6167, + "theoretical_loss": 4.072943711665093, + "tokens_seen": 372310016 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044806419257773324, + "loss": 1.6151, + "theoretical_loss": 4.072863107830866, + "tokens_seen": 372375552 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044805416248746237, + "loss": 1.6409, + "theoretical_loss": 4.072782522152444, + "tokens_seen": 372441088 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004480441323971916, + "loss": 1.4824, + "theoretical_loss": 4.072701954622546, + "tokens_seen": 372506624 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004480341023069208, + "loss": 1.6083, + "theoretical_loss": 4.072621405233891, + "tokens_seen": 372572160 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044802407221664997, + "loss": 1.6239, + "theoretical_loss": 4.072540873979205, + "tokens_seen": 372637696 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044801404212637915, + "loss": 1.6525, + "theoretical_loss": 4.072460360851217, + "tokens_seen": 372703232 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044800401203610833, + "loss": 1.5684, + "theoretical_loss": 4.07237986584266, + "tokens_seen": 372768768 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004479939819458375, + "loss": 1.6867, + "theoretical_loss": 4.0722993889462735, + "tokens_seen": 372834304 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044798395185556675, + "loss": 1.5445, + "theoretical_loss": 4.072218930154797, + "tokens_seen": 372899840 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004479739217652959, + "loss": 1.6251, + "theoretical_loss": 4.072138489460979, + "tokens_seen": 372965376 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004479638916750251, + "loss": 1.614, + "theoretical_loss": 4.0720580668575685, + "tokens_seen": 373030912 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044795386158475424, + "loss": 1.6284, + "theoretical_loss": 4.0719776623373205, + "tokens_seen": 373096448 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044794383149448347, + "loss": 1.6345, + "theoretical_loss": 4.071897275892994, + "tokens_seen": 373161984 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044793380140421265, + "loss": 1.6108, + "theoretical_loss": 4.071816907517349, + "tokens_seen": 373227520 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044792377131394183, + "loss": 1.5843, + "theoretical_loss": 4.071736557203157, + "tokens_seen": 373293056 + }, + { + "epoch": 0.11, + "learning_rate": 0.000447913741223671, + "loss": 1.57, + "theoretical_loss": 4.071656224943186, + "tokens_seen": 373358592 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044790371113340025, + "loss": 1.6044, + "theoretical_loss": 4.071575910730212, + "tokens_seen": 373424128 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004478936810431294, + "loss": 1.6212, + "theoretical_loss": 4.071495614557016, + "tokens_seen": 373489664 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 144870, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6372146606445312, + "objective/train/theoretical_loss": 4.071415336416378, + "objective/train/tokens_used": 373555200, + "theoretical_loss": 4.071415336416378, + "tokens_seen": 373555200 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004478836509528586, + "loss": 1.6484, + "theoretical_loss": 4.071415336416378, + "tokens_seen": 373555200 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044787362086258774, + "loss": 1.5808, + "theoretical_loss": 4.071335076301089, + "tokens_seen": 373620736 + }, + { + "epoch": 0.11, + "learning_rate": 0.000447863590772317, + "loss": 1.5521, + "theoretical_loss": 4.071254834203941, + "tokens_seen": 373686272 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044785356068204616, + "loss": 1.6024, + "theoretical_loss": 4.071174610117729, + "tokens_seen": 373751808 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044784353059177534, + "loss": 1.5648, + "theoretical_loss": 4.071094404035253, + "tokens_seen": 373817344 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004478335005015045, + "loss": 1.5557, + "theoretical_loss": 4.071014215949317, + "tokens_seen": 373882880 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004478234704112337, + "loss": 1.626, + "theoretical_loss": 4.070934045852731, + "tokens_seen": 373948416 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004478134403209629, + "loss": 1.6523, + "theoretical_loss": 4.070853893738307, + "tokens_seen": 374013952 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004478034102306921, + "loss": 1.4717, + "theoretical_loss": 4.070773759598861, + "tokens_seen": 374079488 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044779338014042124, + "loss": 1.607, + "theoretical_loss": 4.070693643427215, + "tokens_seen": 374145024 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004477833500501505, + "loss": 1.6276, + "theoretical_loss": 4.070613545216193, + "tokens_seen": 374210560 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004477733199598796, + "loss": 1.578, + "theoretical_loss": 4.070533464958626, + "tokens_seen": 374276096 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044776328986960884, + "loss": 1.5748, + "theoretical_loss": 4.070453402647344, + "tokens_seen": 374341632 + }, + { + "epoch": 0.11, + "learning_rate": 0.000447753259779338, + "loss": 1.5853, + "theoretical_loss": 4.070373358275187, + "tokens_seen": 374407168 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004477432296890672, + "loss": 1.6318, + "theoretical_loss": 4.070293331834996, + "tokens_seen": 374472704 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004477331995987964, + "loss": 1.4533, + "theoretical_loss": 4.070213323319616, + "tokens_seen": 374538240 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004477231695085256, + "loss": 1.6133, + "theoretical_loss": 4.070133332721897, + "tokens_seen": 374603776 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044771313941825475, + "loss": 1.5593, + "theoretical_loss": 4.070053360034693, + "tokens_seen": 374669312 + }, + { + "epoch": 0.11, + "learning_rate": 0.000447703109327984, + "loss": 1.5328, + "theoretical_loss": 4.069973405250862, + "tokens_seen": 374734848 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004476930792377131, + "loss": 1.579, + "theoretical_loss": 4.069893468363265, + "tokens_seen": 374800384 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044768304914744234, + "loss": 1.5884, + "theoretical_loss": 4.06981354936477, + "tokens_seen": 374865920 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004476730190571715, + "loss": 1.6637, + "theoretical_loss": 4.069733648248246, + "tokens_seen": 374931456 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004476629889669007, + "loss": 1.5164, + "theoretical_loss": 4.069653765006565, + "tokens_seen": 374996992 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004476529588766299, + "loss": 1.5651, + "theoretical_loss": 4.06957389963261, + "tokens_seen": 375062528 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044764292878635907, + "loss": 1.5697, + "theoretical_loss": 4.069494052119261, + "tokens_seen": 375128064 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044763289869608825, + "loss": 1.4781, + "theoretical_loss": 4.069414222459403, + "tokens_seen": 375193600 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004476228686058175, + "loss": 1.5872, + "theoretical_loss": 4.069334410645928, + "tokens_seen": 375259136 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004476128385155466, + "loss": 1.6029, + "theoretical_loss": 4.069254616671732, + "tokens_seen": 375324672 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044760280842527585, + "loss": 1.7451, + "theoretical_loss": 4.069174840529712, + "tokens_seen": 375390208 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044759277833500503, + "loss": 1.6387, + "theoretical_loss": 4.069095082212772, + "tokens_seen": 375455744 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004475827482447342, + "loss": 1.666, + "theoretical_loss": 4.0690153417138175, + "tokens_seen": 375521280 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004475727181544634, + "loss": 1.6416, + "theoretical_loss": 4.06893561902576, + "tokens_seen": 375586816 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044756268806419257, + "loss": 1.5299, + "theoretical_loss": 4.068855914141515, + "tokens_seen": 375652352 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044755265797392175, + "loss": 1.5868, + "theoretical_loss": 4.068776227054, + "tokens_seen": 375717888 + }, + { + "epoch": 0.11, + "learning_rate": 0.000447542627883651, + "loss": 1.6332, + "theoretical_loss": 4.06869655775614, + "tokens_seen": 375783424 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004475325977933801, + "loss": 1.5923, + "theoretical_loss": 4.068616906240861, + "tokens_seen": 375848960 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044752256770310935, + "loss": 1.6367, + "theoretical_loss": 4.068537272501095, + "tokens_seen": 375914496 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004475125376128385, + "loss": 1.5534, + "theoretical_loss": 4.068457656529775, + "tokens_seen": 375980032 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004475025075225677, + "loss": 1.6601, + "theoretical_loss": 4.068378058319843, + "tokens_seen": 376045568 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004474924774322969, + "loss": 1.5495, + "theoretical_loss": 4.068298477864241, + "tokens_seen": 376111104 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004474824473420261, + "loss": 1.7125, + "theoretical_loss": 4.068218915155917, + "tokens_seen": 376176640 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044747241725175526, + "loss": 1.7024, + "theoretical_loss": 4.068139370187821, + "tokens_seen": 376242176 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044746238716148444, + "loss": 1.5838, + "theoretical_loss": 4.06805984295291, + "tokens_seen": 376307712 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004474523570712136, + "loss": 1.5802, + "theoretical_loss": 4.067980333444142, + "tokens_seen": 376373248 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044744232698094285, + "loss": 1.6067, + "theoretical_loss": 4.067900841654481, + "tokens_seen": 376438784 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044743229689067204, + "loss": 1.642, + "theoretical_loss": 4.067821367576894, + "tokens_seen": 376504320 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004474222668004012, + "loss": 1.6398, + "theoretical_loss": 4.067741911204354, + "tokens_seen": 376569856 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044741223671013045, + "loss": 1.511, + "theoretical_loss": 4.067662472529834, + "tokens_seen": 376635392 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004474022066198596, + "loss": 1.5843, + "theoretical_loss": 4.067583051546316, + "tokens_seen": 376700928 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004473921765295888, + "loss": 1.5867, + "theoretical_loss": 4.067503648246781, + "tokens_seen": 376766464 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 145961, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5557124614715576, + "objective/train/theoretical_loss": 4.067424262624218, + "objective/train/tokens_used": 376832000, + "theoretical_loss": 4.067424262624218, + "tokens_seen": 376832000 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044738214643931794, + "loss": 1.5467, + "theoretical_loss": 4.067424262624218, + "tokens_seen": 376832000 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004473721163490472, + "loss": 1.5723, + "theoretical_loss": 4.06734489467162, + "tokens_seen": 376897536 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044736208625877636, + "loss": 1.6343, + "theoretical_loss": 4.0672655443819785, + "tokens_seen": 376963072 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044735205616850554, + "loss": 1.6042, + "theoretical_loss": 4.067186211748297, + "tokens_seen": 377028608 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004473420260782347, + "loss": 1.6607, + "theoretical_loss": 4.067106896763575, + "tokens_seen": 377094144 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004473319959879639, + "loss": 1.6569, + "theoretical_loss": 4.067027599420824, + "tokens_seen": 377159680 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004473219658976931, + "loss": 1.5561, + "theoretical_loss": 4.0669483197130525, + "tokens_seen": 377225216 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004473119358074223, + "loss": 1.5476, + "theoretical_loss": 4.066869057633277, + "tokens_seen": 377290752 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044730190571715144, + "loss": 1.6472, + "theoretical_loss": 4.066789813174517, + "tokens_seen": 377356288 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004472918756268807, + "loss": 1.6266, + "theoretical_loss": 4.066710586329796, + "tokens_seen": 377421824 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004472818455366098, + "loss": 1.6523, + "theoretical_loss": 4.066631377092141, + "tokens_seen": 377487360 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044727181544633904, + "loss": 1.6226, + "theoretical_loss": 4.066552185454582, + "tokens_seen": 377552896 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004472617853560682, + "loss": 1.5627, + "theoretical_loss": 4.066473011410157, + "tokens_seen": 377618432 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004472517552657974, + "loss": 1.6544, + "theoretical_loss": 4.066393854951903, + "tokens_seen": 377683968 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004472417251755266, + "loss": 1.5318, + "theoretical_loss": 4.066314716072864, + "tokens_seen": 377749504 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004472316950852558, + "loss": 1.5947, + "theoretical_loss": 4.066235594766088, + "tokens_seen": 377815040 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044722166499498495, + "loss": 1.5385, + "theoretical_loss": 4.066156491024625, + "tokens_seen": 377880576 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004472116349047142, + "loss": 1.7391, + "theoretical_loss": 4.06607740484153, + "tokens_seen": 377946112 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004472016048144433, + "loss": 1.6656, + "theoretical_loss": 4.065998336209863, + "tokens_seen": 378011648 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044719157472417254, + "loss": 1.6673, + "theoretical_loss": 4.065919285122686, + "tokens_seen": 378077184 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004471815446339017, + "loss": 1.588, + "theoretical_loss": 4.065840251573066, + "tokens_seen": 378142720 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004471715145436309, + "loss": 1.641, + "theoretical_loss": 4.065761235554075, + "tokens_seen": 378208256 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004471614844533601, + "loss": 1.4877, + "theoretical_loss": 4.065682237058787, + "tokens_seen": 378273792 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044715145436308927, + "loss": 1.6583, + "theoretical_loss": 4.06560325608028, + "tokens_seen": 378339328 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044714142427281845, + "loss": 1.5781, + "theoretical_loss": 4.065524292611637, + "tokens_seen": 378404864 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004471313941825477, + "loss": 1.6101, + "theoretical_loss": 4.065445346645945, + "tokens_seen": 378470400 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004471213640922768, + "loss": 1.5426, + "theoretical_loss": 4.0653664181762945, + "tokens_seen": 378535936 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044711133400200605, + "loss": 1.5907, + "theoretical_loss": 4.06528750719578, + "tokens_seen": 378601472 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044710130391173523, + "loss": 1.5562, + "theoretical_loss": 4.0652086136975, + "tokens_seen": 378667008 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004470912738214644, + "loss": 1.5798, + "theoretical_loss": 4.065129737674556, + "tokens_seen": 378732544 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004470812437311936, + "loss": 1.5283, + "theoretical_loss": 4.065050879120054, + "tokens_seen": 378798080 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044707121364092277, + "loss": 1.5579, + "theoretical_loss": 4.064972038027106, + "tokens_seen": 378863616 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044706118355065195, + "loss": 1.6934, + "theoretical_loss": 4.064893214388825, + "tokens_seen": 378929152 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004470511534603812, + "loss": 1.5969, + "theoretical_loss": 4.064814408198327, + "tokens_seen": 378994688 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004470411233701103, + "loss": 1.6435, + "theoretical_loss": 4.0647356194487365, + "tokens_seen": 379060224 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044703109327983955, + "loss": 1.6781, + "theoretical_loss": 4.064656848133178, + "tokens_seen": 379125760 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004470210631895687, + "loss": 1.6406, + "theoretical_loss": 4.064578094244782, + "tokens_seen": 379191296 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004470110330992979, + "loss": 1.5982, + "theoretical_loss": 4.06449935777668, + "tokens_seen": 379256832 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004470010030090271, + "loss": 1.493, + "theoretical_loss": 4.0644206387220105, + "tokens_seen": 379322368 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004469909729187563, + "loss": 1.719, + "theoretical_loss": 4.064341937073916, + "tokens_seen": 379387904 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044698094282848546, + "loss": 1.5433, + "theoretical_loss": 4.0642632528255405, + "tokens_seen": 379453440 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044697091273821464, + "loss": 1.5634, + "theoretical_loss": 4.064184585970033, + "tokens_seen": 379518976 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004469608826479438, + "loss": 1.604, + "theoretical_loss": 4.064105936500547, + "tokens_seen": 379584512 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044695085255767305, + "loss": 1.5771, + "theoretical_loss": 4.064027304410239, + "tokens_seen": 379650048 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004469408224674022, + "loss": 1.4709, + "theoretical_loss": 4.063948689692269, + "tokens_seen": 379715584 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004469307923771314, + "loss": 1.5597, + "theoretical_loss": 4.063870092339803, + "tokens_seen": 379781120 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004469207622868606, + "loss": 1.5643, + "theoretical_loss": 4.0637915123460076, + "tokens_seen": 379846656 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004469107321965898, + "loss": 1.5566, + "theoretical_loss": 4.0637129497040565, + "tokens_seen": 379912192 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044690070210631896, + "loss": 1.6714, + "theoretical_loss": 4.063634404407126, + "tokens_seen": 379977728 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044689067201604814, + "loss": 1.5788, + "theoretical_loss": 4.063555876448395, + "tokens_seen": 380043264 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 147136, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6121628284454346, + "objective/train/theoretical_loss": 4.063477365821048, + "objective/train/tokens_used": 380108800, + "theoretical_loss": 4.063477365821048, + "tokens_seen": 380108800 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004468806419257773, + "loss": 1.6223, + "theoretical_loss": 4.063477365821048, + "tokens_seen": 380108800 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044687061183550656, + "loss": 1.5011, + "theoretical_loss": 4.063398872518273, + "tokens_seen": 380174336 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004468605817452357, + "loss": 1.6382, + "theoretical_loss": 4.0633203965332605, + "tokens_seen": 380239872 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004468505516549649, + "loss": 1.7055, + "theoretical_loss": 4.063241937859207, + "tokens_seen": 380305408 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044684052156469405, + "loss": 1.6894, + "theoretical_loss": 4.063163496489312, + "tokens_seen": 380370944 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004468304914744233, + "loss": 1.6345, + "theoretical_loss": 4.0630850724167775, + "tokens_seen": 380436480 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044682046138415246, + "loss": 1.5616, + "theoretical_loss": 4.063006665634811, + "tokens_seen": 380502016 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044681043129388164, + "loss": 1.5733, + "theoretical_loss": 4.062928276136624, + "tokens_seen": 380567552 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004468004012036108, + "loss": 1.5107, + "theoretical_loss": 4.0628499039154296, + "tokens_seen": 380633088 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044679037111334, + "loss": 1.6411, + "theoretical_loss": 4.062771548964447, + "tokens_seen": 380698624 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004467803410230692, + "loss": 1.687, + "theoretical_loss": 4.062693211276899, + "tokens_seen": 380764160 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004467703109327984, + "loss": 1.697, + "theoretical_loss": 4.062614890846012, + "tokens_seen": 380829696 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044676028084252755, + "loss": 1.5426, + "theoretical_loss": 4.062536587665014, + "tokens_seen": 380895232 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004467502507522568, + "loss": 1.5672, + "theoretical_loss": 4.0624583017271405, + "tokens_seen": 380960768 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044674022066198597, + "loss": 1.6477, + "theoretical_loss": 4.062380033025629, + "tokens_seen": 381026304 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044673019057171515, + "loss": 1.6217, + "theoretical_loss": 4.062301781553719, + "tokens_seen": 381091840 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044672016048144433, + "loss": 1.5891, + "theoretical_loss": 4.062223547304658, + "tokens_seen": 381157376 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004467101303911735, + "loss": 1.6276, + "theoretical_loss": 4.0621453302716946, + "tokens_seen": 381222912 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004467001003009027, + "loss": 1.5085, + "theoretical_loss": 4.062067130448081, + "tokens_seen": 381288448 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004466900702106319, + "loss": 1.5838, + "theoretical_loss": 4.061988947827073, + "tokens_seen": 381353984 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004466800401203611, + "loss": 1.6911, + "theoretical_loss": 4.061910782401933, + "tokens_seen": 381419520 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004466700100300903, + "loss": 1.5325, + "theoretical_loss": 4.061832634165923, + "tokens_seen": 381485056 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044665997993981947, + "loss": 1.6159, + "theoretical_loss": 4.061754503112313, + "tokens_seen": 381550592 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044664994984954865, + "loss": 1.6204, + "theoretical_loss": 4.061676389234375, + "tokens_seen": 381616128 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004466399197592779, + "loss": 1.5786, + "theoretical_loss": 4.061598292525382, + "tokens_seen": 381681664 + }, + { + "epoch": 0.12, + "learning_rate": 0.000446629889669007, + "loss": 1.5809, + "theoretical_loss": 4.061520212978616, + "tokens_seen": 381747200 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044661985957873625, + "loss": 1.5883, + "theoretical_loss": 4.061442150587358, + "tokens_seen": 381812736 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044660982948846543, + "loss": 1.6765, + "theoretical_loss": 4.0613641053448974, + "tokens_seen": 381878272 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004465997993981946, + "loss": 1.5999, + "theoretical_loss": 4.061286077244523, + "tokens_seen": 381943808 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004465897693079238, + "loss": 1.5541, + "theoretical_loss": 4.06120806627953, + "tokens_seen": 382009344 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044657973921765297, + "loss": 1.5166, + "theoretical_loss": 4.061130072443216, + "tokens_seen": 382074880 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044656970912738215, + "loss": 1.6052, + "theoretical_loss": 4.0610520957288845, + "tokens_seen": 382140416 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004465596790371114, + "loss": 1.58, + "theoretical_loss": 4.06097413612984, + "tokens_seen": 382205952 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004465496489468405, + "loss": 1.6113, + "theoretical_loss": 4.060896193639392, + "tokens_seen": 382271488 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044653961885656975, + "loss": 1.6082, + "theoretical_loss": 4.060818268250855, + "tokens_seen": 382337024 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004465295887662989, + "loss": 1.6576, + "theoretical_loss": 4.060740359957546, + "tokens_seen": 382402560 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004465195586760281, + "loss": 1.554, + "theoretical_loss": 4.060662468752784, + "tokens_seen": 382468096 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004465095285857573, + "loss": 1.6249, + "theoretical_loss": 4.060584594629895, + "tokens_seen": 382533632 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004464994984954865, + "loss": 1.663, + "theoretical_loss": 4.060506737582208, + "tokens_seen": 382599168 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044648946840521566, + "loss": 1.5775, + "theoretical_loss": 4.060428897603054, + "tokens_seen": 382664704 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044647943831494484, + "loss": 1.7107, + "theoretical_loss": 4.060351074685769, + "tokens_seen": 382730240 + }, + { + "epoch": 0.12, + "learning_rate": 0.000446469408224674, + "loss": 1.5318, + "theoretical_loss": 4.060273268823693, + "tokens_seen": 382795776 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044645937813440325, + "loss": 1.6782, + "theoretical_loss": 4.060195480010169, + "tokens_seen": 382861312 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004464493480441324, + "loss": 1.624, + "theoretical_loss": 4.060117708238544, + "tokens_seen": 382926848 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004464393179538616, + "loss": 1.5451, + "theoretical_loss": 4.060039953502169, + "tokens_seen": 382992384 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004464292878635908, + "loss": 1.6264, + "theoretical_loss": 4.059962215794399, + "tokens_seen": 383057920 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044641925777332, + "loss": 1.6132, + "theoretical_loss": 4.059884495108591, + "tokens_seen": 383123456 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044640922768304916, + "loss": 1.6938, + "theoretical_loss": 4.059806791438108, + "tokens_seen": 383188992 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044639919759277834, + "loss": 1.5708, + "theoretical_loss": 4.059729104776316, + "tokens_seen": 383254528 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004463891675025075, + "loss": 1.6384, + "theoretical_loss": 4.059651435116583, + "tokens_seen": 383320064 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 148269, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6788184642791748, + "objective/train/theoretical_loss": 4.059573782452284, + "objective/train/tokens_used": 383385600, + "theoretical_loss": 4.059573782452284, + "tokens_seen": 383385600 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044637913741223676, + "loss": 1.6216, + "theoretical_loss": 4.059573782452284, + "tokens_seen": 383385600 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004463691073219659, + "loss": 1.6931, + "theoretical_loss": 4.0594961467767945, + "tokens_seen": 383451136 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004463590772316951, + "loss": 1.7128, + "theoretical_loss": 4.059418528083495, + "tokens_seen": 383516672 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044634904714142425, + "loss": 1.5162, + "theoretical_loss": 4.0593409263657705, + "tokens_seen": 383582208 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004463390170511535, + "loss": 1.6165, + "theoretical_loss": 4.059263341617009, + "tokens_seen": 383647744 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044632898696088266, + "loss": 1.6503, + "theoretical_loss": 4.059185773830602, + "tokens_seen": 383713280 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044631895687061184, + "loss": 1.5914, + "theoretical_loss": 4.059108222999944, + "tokens_seen": 383778816 + }, + { + "epoch": 0.12, + "learning_rate": 0.000446308926780341, + "loss": 1.5773, + "theoretical_loss": 4.059030689118436, + "tokens_seen": 383844352 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004462988966900702, + "loss": 1.5181, + "theoretical_loss": 4.058953172179479, + "tokens_seen": 383909888 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004462888665997994, + "loss": 1.6957, + "theoretical_loss": 4.05887567217648, + "tokens_seen": 383975424 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004462788365095286, + "loss": 1.6036, + "theoretical_loss": 4.05879818910285, + "tokens_seen": 384040960 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044626880641925775, + "loss": 1.6467, + "theoretical_loss": 4.058720722952002, + "tokens_seen": 384106496 + }, + { + "epoch": 0.12, + "learning_rate": 0.000446258776328987, + "loss": 1.4849, + "theoretical_loss": 4.058643273717355, + "tokens_seen": 384172032 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044624874623871617, + "loss": 1.5467, + "theoretical_loss": 4.058565841392328, + "tokens_seen": 384237568 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044623871614844535, + "loss": 1.6036, + "theoretical_loss": 4.058488425970347, + "tokens_seen": 384303104 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044622868605817453, + "loss": 1.4782, + "theoretical_loss": 4.058411027444841, + "tokens_seen": 384368640 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004462186559679037, + "loss": 1.6036, + "theoretical_loss": 4.05833364580924, + "tokens_seen": 384434176 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004462086258776329, + "loss": 1.5803, + "theoretical_loss": 4.058256281056984, + "tokens_seen": 384499712 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004461985957873621, + "loss": 1.4187, + "theoretical_loss": 4.058178933181509, + "tokens_seen": 384565248 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044618856569709125, + "loss": 1.5035, + "theoretical_loss": 4.05810160217626, + "tokens_seen": 384630784 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004461785356068205, + "loss": 1.692, + "theoretical_loss": 4.058024288034684, + "tokens_seen": 384696320 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004461685055165496, + "loss": 1.621, + "theoretical_loss": 4.057946990750231, + "tokens_seen": 384761856 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044615847542627885, + "loss": 1.5984, + "theoretical_loss": 4.057869710316355, + "tokens_seen": 384827392 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044614844533600803, + "loss": 1.5286, + "theoretical_loss": 4.0577924467265145, + "tokens_seen": 384892928 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004461384152457372, + "loss": 1.7038, + "theoretical_loss": 4.057715199974171, + "tokens_seen": 384958464 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004461283851554664, + "loss": 1.4001, + "theoretical_loss": 4.05763797005279, + "tokens_seen": 385024000 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044611835506519563, + "loss": 1.6016, + "theoretical_loss": 4.05756075695584, + "tokens_seen": 385089536 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044610832497492476, + "loss": 1.6193, + "theoretical_loss": 4.0574835606767925, + "tokens_seen": 385155072 + }, + { + "epoch": 0.12, + "learning_rate": 0.000446098294884654, + "loss": 1.54, + "theoretical_loss": 4.057406381209126, + "tokens_seen": 385220608 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004460882647943831, + "loss": 1.6847, + "theoretical_loss": 4.057329218546318, + "tokens_seen": 385286144 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044607823470411235, + "loss": 1.5337, + "theoretical_loss": 4.0572520726818535, + "tokens_seen": 385351680 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044606820461384153, + "loss": 1.556, + "theoretical_loss": 4.057174943609218, + "tokens_seen": 385417216 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004460581745235707, + "loss": 1.6561, + "theoretical_loss": 4.057097831321904, + "tokens_seen": 385482752 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004460481444332999, + "loss": 1.5698, + "theoretical_loss": 4.057020735813405, + "tokens_seen": 385548288 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004460381143430291, + "loss": 1.4442, + "theoretical_loss": 4.056943657077219, + "tokens_seen": 385613824 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044602808425275826, + "loss": 1.8255, + "theoretical_loss": 4.056866595106847, + "tokens_seen": 385679360 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004460180541624875, + "loss": 1.5506, + "theoretical_loss": 4.056789549895795, + "tokens_seen": 385744896 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004460080240722166, + "loss": 1.6073, + "theoretical_loss": 4.056712521437571, + "tokens_seen": 385810432 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044599799398194586, + "loss": 1.5807, + "theoretical_loss": 4.056635509725689, + "tokens_seen": 385875968 + }, + { + "epoch": 0.12, + "learning_rate": 0.000445987963891675, + "loss": 1.648, + "theoretical_loss": 4.056558514753664, + "tokens_seen": 385941504 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004459779338014042, + "loss": 1.6357, + "theoretical_loss": 4.056481536515014, + "tokens_seen": 386007040 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004459679037111334, + "loss": 1.6107, + "theoretical_loss": 4.056404575003266, + "tokens_seen": 386072576 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004459578736208626, + "loss": 1.6333, + "theoretical_loss": 4.0563276302119435, + "tokens_seen": 386138112 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044594784353059176, + "loss": 1.6179, + "theoretical_loss": 4.056250702134578, + "tokens_seen": 386203648 + }, + { + "epoch": 0.12, + "learning_rate": 0.000445937813440321, + "loss": 1.782, + "theoretical_loss": 4.0561737907647055, + "tokens_seen": 386269184 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004459277833500502, + "loss": 1.5863, + "theoretical_loss": 4.056096896095861, + "tokens_seen": 386334720 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044591775325977936, + "loss": 1.5541, + "theoretical_loss": 4.056020018121587, + "tokens_seen": 386400256 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044590772316950854, + "loss": 1.6078, + "theoretical_loss": 4.055943156835428, + "tokens_seen": 386465792 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004458976930792377, + "loss": 1.5904, + "theoretical_loss": 4.055866312230933, + "tokens_seen": 386531328 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044588766298896696, + "loss": 1.6412, + "theoretical_loss": 4.055789484301653, + "tokens_seen": 386596864 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 149397, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7365185022354126, + "objective/train/theoretical_loss": 4.055712673041144, + "objective/train/tokens_used": 386662400, + "theoretical_loss": 4.055712673041144, + "tokens_seen": 386662400 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004458776328986961, + "loss": 1.6587, + "theoretical_loss": 4.055712673041144, + "tokens_seen": 386662400 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004458676028084253, + "loss": 1.6664, + "theoretical_loss": 4.055635878442967, + "tokens_seen": 386727936 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044585757271815445, + "loss": 1.5783, + "theoretical_loss": 4.055559100500682, + "tokens_seen": 386793472 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004458475426278837, + "loss": 1.6645, + "theoretical_loss": 4.055482339207858, + "tokens_seen": 386859008 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044583751253761286, + "loss": 1.6141, + "theoretical_loss": 4.055405594558062, + "tokens_seen": 386924544 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044582748244734204, + "loss": 1.5221, + "theoretical_loss": 4.05532886654487, + "tokens_seen": 386990080 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004458174523570712, + "loss": 1.6275, + "theoretical_loss": 4.055252155161858, + "tokens_seen": 387055616 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004458074222668004, + "loss": 1.5572, + "theoretical_loss": 4.055175460402607, + "tokens_seen": 387121152 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004457973921765296, + "loss": 1.6077, + "theoretical_loss": 4.055098782260701, + "tokens_seen": 387186688 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004457873620862588, + "loss": 1.5348, + "theoretical_loss": 4.055022120729728, + "tokens_seen": 387252224 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044577733199598795, + "loss": 1.6882, + "theoretical_loss": 4.054945475803278, + "tokens_seen": 387317760 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004457673019057172, + "loss": 1.5017, + "theoretical_loss": 4.054868847474948, + "tokens_seen": 387383296 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044575727181544637, + "loss": 1.6517, + "theoretical_loss": 4.054792235738335, + "tokens_seen": 387448832 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044574724172517555, + "loss": 1.5774, + "theoretical_loss": 4.054715640587042, + "tokens_seen": 387514368 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044573721163490473, + "loss": 1.6065, + "theoretical_loss": 4.054639062014673, + "tokens_seen": 387579904 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004457271815446339, + "loss": 1.6844, + "theoretical_loss": 4.054562500014839, + "tokens_seen": 387645440 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004457171514543631, + "loss": 1.5536, + "theoretical_loss": 4.054485954581152, + "tokens_seen": 387710976 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004457071213640923, + "loss": 1.5691, + "theoretical_loss": 4.054409425707227, + "tokens_seen": 387776512 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044569709127382145, + "loss": 1.5902, + "theoretical_loss": 4.054332913386686, + "tokens_seen": 387842048 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004456870611835507, + "loss": 1.567, + "theoretical_loss": 4.054256417613149, + "tokens_seen": 387907584 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004456770310932798, + "loss": 1.7172, + "theoretical_loss": 4.054179938380246, + "tokens_seen": 387973120 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044566700100300905, + "loss": 1.5912, + "theoretical_loss": 4.054103475681605, + "tokens_seen": 388038656 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044565697091273823, + "loss": 1.5867, + "theoretical_loss": 4.054027029510861, + "tokens_seen": 388104192 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004456469408224674, + "loss": 1.502, + "theoretical_loss": 4.053950599861651, + "tokens_seen": 388169728 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004456369107321966, + "loss": 1.5958, + "theoretical_loss": 4.053874186727617, + "tokens_seen": 388235264 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044562688064192583, + "loss": 1.4592, + "theoretical_loss": 4.053797790102402, + "tokens_seen": 388300800 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044561685055165496, + "loss": 1.4705, + "theoretical_loss": 4.053721409979654, + "tokens_seen": 388366336 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004456068204613842, + "loss": 1.5584, + "theoretical_loss": 4.053645046353025, + "tokens_seen": 388431872 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004455967903711133, + "loss": 1.5707, + "theoretical_loss": 4.05356869921617, + "tokens_seen": 388497408 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044558676028084255, + "loss": 1.6798, + "theoretical_loss": 4.053492368562747, + "tokens_seen": 388562944 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044557673019057173, + "loss": 1.6631, + "theoretical_loss": 4.0534160543864175, + "tokens_seen": 388628480 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004455667001003009, + "loss": 1.5914, + "theoretical_loss": 4.053339756680849, + "tokens_seen": 388694016 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004455566700100301, + "loss": 1.5916, + "theoretical_loss": 4.053263475439707, + "tokens_seen": 388759552 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004455466399197593, + "loss": 1.5088, + "theoretical_loss": 4.053187210656668, + "tokens_seen": 388825088 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044553660982948846, + "loss": 1.4766, + "theoretical_loss": 4.053110962325406, + "tokens_seen": 388890624 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004455265797392177, + "loss": 1.5854, + "theoretical_loss": 4.053034730439601, + "tokens_seen": 388956160 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004455165496489468, + "loss": 1.6212, + "theoretical_loss": 4.052958514992934, + "tokens_seen": 389021696 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044550651955867606, + "loss": 1.4902, + "theoretical_loss": 4.052882315979094, + "tokens_seen": 389087232 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004454964894684052, + "loss": 1.5163, + "theoretical_loss": 4.052806133391769, + "tokens_seen": 389152768 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004454864593781344, + "loss": 1.4746, + "theoretical_loss": 4.052729967224654, + "tokens_seen": 389218304 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004454764292878636, + "loss": 1.6108, + "theoretical_loss": 4.0526538174714455, + "tokens_seen": 389283840 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004454663991975928, + "loss": 1.5747, + "theoretical_loss": 4.052577684125843, + "tokens_seen": 389349376 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044545636910732196, + "loss": 1.6147, + "theoretical_loss": 4.052501567181552, + "tokens_seen": 389414912 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004454463390170512, + "loss": 1.5909, + "theoretical_loss": 4.052425466632278, + "tokens_seen": 389480448 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004454363089267803, + "loss": 1.5618, + "theoretical_loss": 4.052349382471734, + "tokens_seen": 389545984 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044542627883650956, + "loss": 1.5451, + "theoretical_loss": 4.052273314693632, + "tokens_seen": 389611520 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004454162487462387, + "loss": 1.5454, + "theoretical_loss": 4.05219726329169, + "tokens_seen": 389677056 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004454062186559679, + "loss": 1.5703, + "theoretical_loss": 4.052121228259631, + "tokens_seen": 389742592 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004453961885656971, + "loss": 1.6176, + "theoretical_loss": 4.052045209591179, + "tokens_seen": 389808128 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004453861584754263, + "loss": 1.5859, + "theoretical_loss": 4.05196920728006, + "tokens_seen": 389873664 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 150881, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.624035120010376, + "objective/train/theoretical_loss": 4.051893221320009, + "objective/train/tokens_used": 389939200, + "theoretical_loss": 4.051893221320009, + "tokens_seen": 389939200 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044537612838515547, + "loss": 1.6505, + "theoretical_loss": 4.051893221320009, + "tokens_seen": 389939200 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044536609829488465, + "loss": 1.7677, + "theoretical_loss": 4.05181725170476, + "tokens_seen": 390004736 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044535606820461383, + "loss": 1.524, + "theoretical_loss": 4.051741298428049, + "tokens_seen": 390070272 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044534603811434306, + "loss": 1.5968, + "theoretical_loss": 4.0516653614836216, + "tokens_seen": 390135808 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004453360080240722, + "loss": 1.6589, + "theoretical_loss": 4.05158944086522, + "tokens_seen": 390201344 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004453259779338014, + "loss": 1.6302, + "theoretical_loss": 4.0515135365665955, + "tokens_seen": 390266880 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044531594784353055, + "loss": 1.7504, + "theoretical_loss": 4.0514376485815, + "tokens_seen": 390332416 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004453059177532598, + "loss": 1.6371, + "theoretical_loss": 4.051361776903688, + "tokens_seen": 390397952 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044529588766298897, + "loss": 1.5773, + "theoretical_loss": 4.051285921526919, + "tokens_seen": 390463488 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044528585757271815, + "loss": 1.5795, + "theoretical_loss": 4.051210082444956, + "tokens_seen": 390529024 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044527582748244733, + "loss": 1.5989, + "theoretical_loss": 4.051134259651565, + "tokens_seen": 390594560 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044526579739217657, + "loss": 1.5361, + "theoretical_loss": 4.0510584531405165, + "tokens_seen": 390660096 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004452557673019057, + "loss": 1.571, + "theoretical_loss": 4.050982662905581, + "tokens_seen": 390725632 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044524573721163493, + "loss": 1.4943, + "theoretical_loss": 4.0509068889405375, + "tokens_seen": 390791168 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044523570712136406, + "loss": 1.6169, + "theoretical_loss": 4.050831131239164, + "tokens_seen": 390856704 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004452256770310933, + "loss": 1.5693, + "theoretical_loss": 4.050755389795244, + "tokens_seen": 390922240 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044521564694082247, + "loss": 1.6046, + "theoretical_loss": 4.050679664602564, + "tokens_seen": 390987776 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044520561685055165, + "loss": 1.5444, + "theoretical_loss": 4.050603955654914, + "tokens_seen": 391053312 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044519558676028083, + "loss": 1.6529, + "theoretical_loss": 4.050528262946088, + "tokens_seen": 391118848 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044518555667001, + "loss": 1.5743, + "theoretical_loss": 4.050452586469882, + "tokens_seen": 391184384 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044517552657973925, + "loss": 1.6018, + "theoretical_loss": 4.050376926220098, + "tokens_seen": 391249920 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044516549648946843, + "loss": 1.55, + "theoretical_loss": 4.050301282190537, + "tokens_seen": 391315456 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004451554663991976, + "loss": 1.61, + "theoretical_loss": 4.050225654375008, + "tokens_seen": 391380992 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004451454363089268, + "loss": 1.5826, + "theoretical_loss": 4.050150042767321, + "tokens_seen": 391446528 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044513540621865603, + "loss": 1.5376, + "theoretical_loss": 4.050074447361289, + "tokens_seen": 391512064 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044512537612838516, + "loss": 1.6918, + "theoretical_loss": 4.04999886815073, + "tokens_seen": 391577600 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004451153460381144, + "loss": 1.6764, + "theoretical_loss": 4.0499233051294645, + "tokens_seen": 391643136 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004451053159478435, + "loss": 1.581, + "theoretical_loss": 4.049847758291317, + "tokens_seen": 391708672 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044509528585757275, + "loss": 1.6737, + "theoretical_loss": 4.049772227630114, + "tokens_seen": 391774208 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044508525576730194, + "loss": 1.6105, + "theoretical_loss": 4.049696713139687, + "tokens_seen": 391839744 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004450752256770311, + "loss": 1.6753, + "theoretical_loss": 4.049621214813869, + "tokens_seen": 391905280 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004450651955867603, + "loss": 1.5276, + "theoretical_loss": 4.0495457326465, + "tokens_seen": 391970816 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004450551654964895, + "loss": 1.5696, + "theoretical_loss": 4.049470266631419, + "tokens_seen": 392036352 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044504513540621866, + "loss": 1.6787, + "theoretical_loss": 4.04939481676247, + "tokens_seen": 392101888 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004450351053159479, + "loss": 1.6004, + "theoretical_loss": 4.0493193830335015, + "tokens_seen": 392167424 + }, + { + "epoch": 0.12, + "learning_rate": 0.000445025075225677, + "loss": 1.5265, + "theoretical_loss": 4.0492439654383645, + "tokens_seen": 392232960 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044501504513540626, + "loss": 1.5564, + "theoretical_loss": 4.049168563970914, + "tokens_seen": 392298496 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004450050150451354, + "loss": 1.7721, + "theoretical_loss": 4.049093178625007, + "tokens_seen": 392364032 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004449949849548646, + "loss": 1.5781, + "theoretical_loss": 4.049017809394506, + "tokens_seen": 392429568 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004449849548645938, + "loss": 1.5805, + "theoretical_loss": 4.048942456273274, + "tokens_seen": 392495104 + }, + { + "epoch": 0.12, + "learning_rate": 0.000444974924774323, + "loss": 1.5882, + "theoretical_loss": 4.048867119255178, + "tokens_seen": 392560640 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044496489468405216, + "loss": 1.4988, + "theoretical_loss": 4.048791798334092, + "tokens_seen": 392626176 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004449548645937814, + "loss": 1.6314, + "theoretical_loss": 4.048716493503888, + "tokens_seen": 392691712 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004449448345035105, + "loss": 1.6615, + "theoretical_loss": 4.048641204758446, + "tokens_seen": 392757248 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044493480441323976, + "loss": 1.5564, + "theoretical_loss": 4.048565932091646, + "tokens_seen": 392822784 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004449247743229689, + "loss": 1.6106, + "theoretical_loss": 4.048490675497373, + "tokens_seen": 392888320 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004449147442326981, + "loss": 1.5339, + "theoretical_loss": 4.048415434969516, + "tokens_seen": 392953856 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004449047141424273, + "loss": 1.6407, + "theoretical_loss": 4.048340210501965, + "tokens_seen": 393019392 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004448946840521565, + "loss": 1.6579, + "theoretical_loss": 4.048265002088614, + "tokens_seen": 393084928 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044488465396188567, + "loss": 1.7206, + "theoretical_loss": 4.048189809723364, + "tokens_seen": 393150464 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 152158, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5732717514038086, + "objective/train/theoretical_loss": 4.048114633400113, + "objective/train/tokens_used": 393216000, + "theoretical_loss": 4.048114633400113, + "tokens_seen": 393216000 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044487462387161485, + "loss": 1.5976, + "theoretical_loss": 4.048114633400113, + "tokens_seen": 393216000 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044486459378134403, + "loss": 1.6524, + "theoretical_loss": 4.048039473112768, + "tokens_seen": 393281536 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044485456369107326, + "loss": 1.6409, + "theoretical_loss": 4.047964328855237, + "tokens_seen": 393347072 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004448445336008024, + "loss": 1.7036, + "theoretical_loss": 4.047889200621428, + "tokens_seen": 393412608 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004448345035105316, + "loss": 1.6099, + "theoretical_loss": 4.047814088405261, + "tokens_seen": 393478144 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044482447342026075, + "loss": 1.5359, + "theoretical_loss": 4.0477389922006495, + "tokens_seen": 393543680 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044481444332999, + "loss": 1.5648, + "theoretical_loss": 4.047663912001518, + "tokens_seen": 393609216 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044480441323971917, + "loss": 1.6106, + "theoretical_loss": 4.047588847801788, + "tokens_seen": 393674752 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044479438314944835, + "loss": 1.5497, + "theoretical_loss": 4.04751379959539, + "tokens_seen": 393740288 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044478435305917753, + "loss": 1.4762, + "theoretical_loss": 4.047438767376255, + "tokens_seen": 393805824 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044477432296890677, + "loss": 1.6591, + "theoretical_loss": 4.047363751138316, + "tokens_seen": 393871360 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004447642928786359, + "loss": 1.5945, + "theoretical_loss": 4.047288750875512, + "tokens_seen": 393936896 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044475426278836513, + "loss": 1.5623, + "theoretical_loss": 4.047213766581785, + "tokens_seen": 394002432 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044474423269809426, + "loss": 1.5722, + "theoretical_loss": 4.047138798251078, + "tokens_seen": 394067968 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004447342026078235, + "loss": 1.5894, + "theoretical_loss": 4.047063845877339, + "tokens_seen": 394133504 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044472417251755267, + "loss": 1.5608, + "theoretical_loss": 4.04698890945452, + "tokens_seen": 394199040 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044471414242728185, + "loss": 1.5482, + "theoretical_loss": 4.046913988976575, + "tokens_seen": 394264576 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044470411233701103, + "loss": 1.5369, + "theoretical_loss": 4.046839084437463, + "tokens_seen": 394330112 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004446940822467402, + "loss": 1.676, + "theoretical_loss": 4.046764195831142, + "tokens_seen": 394395648 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004446840521564694, + "loss": 1.5567, + "theoretical_loss": 4.046689323151578, + "tokens_seen": 394461184 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044467402206619863, + "loss": 1.6183, + "theoretical_loss": 4.0466144663927395, + "tokens_seen": 394526720 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044466399197592776, + "loss": 1.5788, + "theoretical_loss": 4.046539625548595, + "tokens_seen": 394592256 + }, + { + "epoch": 0.12, + "learning_rate": 0.000444653961885657, + "loss": 1.5817, + "theoretical_loss": 4.046464800613121, + "tokens_seen": 394657792 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004446439317953861, + "loss": 1.5619, + "theoretical_loss": 4.046389991580294, + "tokens_seen": 394723328 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044463390170511536, + "loss": 1.6653, + "theoretical_loss": 4.046315198444094, + "tokens_seen": 394788864 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044462387161484454, + "loss": 1.6109, + "theoretical_loss": 4.046240421198506, + "tokens_seen": 394854400 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004446138415245737, + "loss": 1.4833, + "theoretical_loss": 4.046165659837517, + "tokens_seen": 394919936 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004446038114343029, + "loss": 1.6001, + "theoretical_loss": 4.046090914355117, + "tokens_seen": 394985472 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044459378134403214, + "loss": 1.4923, + "theoretical_loss": 4.046016184745302, + "tokens_seen": 395051008 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044458375125376126, + "loss": 1.655, + "theoretical_loss": 4.045941471002066, + "tokens_seen": 395116544 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004445737211634905, + "loss": 1.6352, + "theoretical_loss": 4.0458667731194105, + "tokens_seen": 395182080 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004445636910732196, + "loss": 1.7141, + "theoretical_loss": 4.045792091091339, + "tokens_seen": 395247616 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044455366098294886, + "loss": 1.5, + "theoretical_loss": 4.04571742491186, + "tokens_seen": 395313152 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044454363089267804, + "loss": 1.6109, + "theoretical_loss": 4.045642774574981, + "tokens_seen": 395378688 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004445336008024072, + "loss": 1.5574, + "theoretical_loss": 4.045568140074717, + "tokens_seen": 395444224 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004445235707121364, + "loss": 1.5923, + "theoretical_loss": 4.045493521405085, + "tokens_seen": 395509760 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004445135406218656, + "loss": 1.6824, + "theoretical_loss": 4.045418918560104, + "tokens_seen": 395575296 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044450351053159477, + "loss": 1.695, + "theoretical_loss": 4.045344331533796, + "tokens_seen": 395640832 + }, + { + "epoch": 0.12, + "learning_rate": 0.000444493480441324, + "loss": 1.6179, + "theoretical_loss": 4.04526976032019, + "tokens_seen": 395706368 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044448345035105313, + "loss": 1.7344, + "theoretical_loss": 4.0451952049133135, + "tokens_seen": 395771904 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044447342026078236, + "loss": 1.6689, + "theoretical_loss": 4.0451206653072, + "tokens_seen": 395837440 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004444633901705115, + "loss": 1.6017, + "theoretical_loss": 4.045046141495887, + "tokens_seen": 395902976 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004444533600802407, + "loss": 1.5874, + "theoretical_loss": 4.044971633473412, + "tokens_seen": 395968512 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044444332998996996, + "loss": 1.6016, + "theoretical_loss": 4.044897141233818, + "tokens_seen": 396034048 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004444332998996991, + "loss": 1.4903, + "theoretical_loss": 4.044822664771151, + "tokens_seen": 396099584 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004444232698094283, + "loss": 1.5146, + "theoretical_loss": 4.044748204079459, + "tokens_seen": 396165120 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004444132397191575, + "loss": 1.5532, + "theoretical_loss": 4.044673759152797, + "tokens_seen": 396230656 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004444032096288867, + "loss": 1.4785, + "theoretical_loss": 4.0445993299852185, + "tokens_seen": 396296192 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044439317953861587, + "loss": 1.5233, + "theoretical_loss": 4.044524916570782, + "tokens_seen": 396361728 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044438314944834505, + "loss": 1.5798, + "theoretical_loss": 4.04445051890355, + "tokens_seen": 396427264 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 153338, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6462069749832153, + "objective/train/theoretical_loss": 4.044376136977588, + "objective/train/tokens_used": 396492800, + "theoretical_loss": 4.044376136977588, + "tokens_seen": 396492800 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044437311935807423, + "loss": 1.6717, + "theoretical_loss": 4.044376136977588, + "tokens_seen": 396492800 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044436308926780346, + "loss": 1.6687, + "theoretical_loss": 4.044301770786964, + "tokens_seen": 396558336 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004443530591775326, + "loss": 1.5274, + "theoretical_loss": 4.044227420325749, + "tokens_seen": 396623872 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004443430290872618, + "loss": 1.5455, + "theoretical_loss": 4.044153085588018, + "tokens_seen": 396689408 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044433299899699095, + "loss": 1.5833, + "theoretical_loss": 4.044078766567851, + "tokens_seen": 396754944 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004443229689067202, + "loss": 1.6088, + "theoretical_loss": 4.044004463259327, + "tokens_seen": 396820480 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044431293881644937, + "loss": 1.5251, + "theoretical_loss": 4.043930175656531, + "tokens_seen": 396886016 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044430290872617855, + "loss": 1.5608, + "theoretical_loss": 4.043855903753551, + "tokens_seen": 396951552 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044429287863590773, + "loss": 1.5045, + "theoretical_loss": 4.043781647544477, + "tokens_seen": 397017088 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044428284854563697, + "loss": 1.4917, + "theoretical_loss": 4.043707407023404, + "tokens_seen": 397082624 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004442728184553661, + "loss": 1.5342, + "theoretical_loss": 4.043633182184428, + "tokens_seen": 397148160 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044426278836509533, + "loss": 1.5672, + "theoretical_loss": 4.043558973021652, + "tokens_seen": 397213696 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044425275827482446, + "loss": 1.5455, + "theoretical_loss": 4.043484779529176, + "tokens_seen": 397279232 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004442427281845537, + "loss": 1.5362, + "theoretical_loss": 4.04341060170111, + "tokens_seen": 397344768 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044423269809428287, + "loss": 1.5876, + "theoretical_loss": 4.043336439531561, + "tokens_seen": 397410304 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044422266800401205, + "loss": 1.6904, + "theoretical_loss": 4.043262293014646, + "tokens_seen": 397475840 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044421263791374123, + "loss": 1.519, + "theoretical_loss": 4.043188162144478, + "tokens_seen": 397541376 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004442026078234704, + "loss": 1.5665, + "theoretical_loss": 4.0431140469151785, + "tokens_seen": 397606912 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004441925777331996, + "loss": 1.5995, + "theoretical_loss": 4.043039947320869, + "tokens_seen": 397672448 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044418254764292883, + "loss": 1.5819, + "theoretical_loss": 4.0429658633556755, + "tokens_seen": 397737984 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044417251755265796, + "loss": 1.7007, + "theoretical_loss": 4.042891795013729, + "tokens_seen": 397803520 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004441624874623872, + "loss": 1.6159, + "theoretical_loss": 4.042817742289159, + "tokens_seen": 397869056 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004441524573721163, + "loss": 1.5626, + "theoretical_loss": 4.042743705176102, + "tokens_seen": 397934592 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044414242728184556, + "loss": 1.5428, + "theoretical_loss": 4.042669683668698, + "tokens_seen": 398000128 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044413239719157474, + "loss": 1.5626, + "theoretical_loss": 4.042595677761087, + "tokens_seen": 398065664 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004441223671013039, + "loss": 1.4794, + "theoretical_loss": 4.0425216874474135, + "tokens_seen": 398131200 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004441123370110331, + "loss": 1.5325, + "theoretical_loss": 4.0424477127218275, + "tokens_seen": 398196736 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044410230692076234, + "loss": 1.6776, + "theoretical_loss": 4.042373753578479, + "tokens_seen": 398262272 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044409227683049146, + "loss": 1.6119, + "theoretical_loss": 4.042299810011523, + "tokens_seen": 398327808 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004440822467402207, + "loss": 1.5326, + "theoretical_loss": 4.042225882015115, + "tokens_seen": 398393344 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004440722166499498, + "loss": 1.6729, + "theoretical_loss": 4.0421519695834185, + "tokens_seen": 398458880 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044406218655967906, + "loss": 1.5231, + "theoretical_loss": 4.042078072710597, + "tokens_seen": 398524416 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044405215646940824, + "loss": 1.6879, + "theoretical_loss": 4.042004191390816, + "tokens_seen": 398589952 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004440421263791374, + "loss": 1.6079, + "theoretical_loss": 4.041930325618246, + "tokens_seen": 398655488 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004440320962888666, + "loss": 1.5125, + "theoretical_loss": 4.041856475387061, + "tokens_seen": 398721024 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004440220661985958, + "loss": 1.6023, + "theoretical_loss": 4.041782640691437, + "tokens_seen": 398786560 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044401203610832497, + "loss": 1.4814, + "theoretical_loss": 4.041708821525554, + "tokens_seen": 398852096 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004440020060180542, + "loss": 1.7191, + "theoretical_loss": 4.0416350178835945, + "tokens_seen": 398917632 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044399197592778333, + "loss": 1.721, + "theoretical_loss": 4.041561229759743, + "tokens_seen": 398983168 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044398194583751256, + "loss": 1.4997, + "theoretical_loss": 4.0414874571481905, + "tokens_seen": 399048704 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004439719157472417, + "loss": 1.6326, + "theoretical_loss": 4.041413700043128, + "tokens_seen": 399114240 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004439618856569709, + "loss": 1.6282, + "theoretical_loss": 4.041339958438751, + "tokens_seen": 399179776 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004439518555667001, + "loss": 1.6642, + "theoretical_loss": 4.041266232329257, + "tokens_seen": 399245312 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004439418254764293, + "loss": 1.5393, + "theoretical_loss": 4.041192521708849, + "tokens_seen": 399310848 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044393179538615847, + "loss": 1.5958, + "theoretical_loss": 4.041118826571731, + "tokens_seen": 399376384 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004439217652958877, + "loss": 1.657, + "theoretical_loss": 4.041045146912111, + "tokens_seen": 399441920 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044391173520561683, + "loss": 1.6468, + "theoretical_loss": 4.040971482724199, + "tokens_seen": 399507456 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044390170511534607, + "loss": 1.5559, + "theoretical_loss": 4.040897834002209, + "tokens_seen": 399572992 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004438916750250752, + "loss": 1.5963, + "theoretical_loss": 4.040824200740358, + "tokens_seen": 399638528 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044388164493480443, + "loss": 1.5875, + "theoretical_loss": 4.040750582932868, + "tokens_seen": 399704064 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 154461, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5596903562545776, + "objective/train/theoretical_loss": 4.04067698057396, + "objective/train/tokens_used": 399769600, + "theoretical_loss": 4.04067698057396, + "tokens_seen": 399769600 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004438716148445336, + "loss": 1.5392, + "theoretical_loss": 4.04067698057396, + "tokens_seen": 399769600 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004438615847542628, + "loss": 1.498, + "theoretical_loss": 4.040603393657862, + "tokens_seen": 399835136 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044385155466399197, + "loss": 1.6916, + "theoretical_loss": 4.040529822178802, + "tokens_seen": 399900672 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044384152457372115, + "loss": 1.5755, + "theoretical_loss": 4.040456266131013, + "tokens_seen": 399966208 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044383149448345033, + "loss": 1.5705, + "theoretical_loss": 4.0403827255087315, + "tokens_seen": 400031744 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044382146439317957, + "loss": 1.572, + "theoretical_loss": 4.040309200306195, + "tokens_seen": 400097280 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004438114343029087, + "loss": 1.6325, + "theoretical_loss": 4.0402356905176475, + "tokens_seen": 400162816 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044380140421263793, + "loss": 1.7122, + "theoretical_loss": 4.040162196137332, + "tokens_seen": 400228352 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004437913741223671, + "loss": 1.4918, + "theoretical_loss": 4.040088717159497, + "tokens_seen": 400293888 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004437813440320963, + "loss": 1.6879, + "theoretical_loss": 4.040015253578394, + "tokens_seen": 400359424 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004437713139418255, + "loss": 1.5604, + "theoretical_loss": 4.039941805388277, + "tokens_seen": 400424960 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044376128385155466, + "loss": 1.5746, + "theoretical_loss": 4.0398683725834035, + "tokens_seen": 400490496 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044375125376128384, + "loss": 1.4717, + "theoretical_loss": 4.039794955158033, + "tokens_seen": 400556032 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044374122367101307, + "loss": 1.5711, + "theoretical_loss": 4.039721553106431, + "tokens_seen": 400621568 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004437311935807422, + "loss": 1.6256, + "theoretical_loss": 4.039648166422863, + "tokens_seen": 400687104 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044372116349047143, + "loss": 1.522, + "theoretical_loss": 4.039574795101597, + "tokens_seen": 400752640 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044371113340020056, + "loss": 1.5607, + "theoretical_loss": 4.039501439136908, + "tokens_seen": 400818176 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004437011033099298, + "loss": 1.6338, + "theoretical_loss": 4.0394280985230715, + "tokens_seen": 400883712 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044369107321965903, + "loss": 1.5923, + "theoretical_loss": 4.039354773254365, + "tokens_seen": 400949248 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044368104312938816, + "loss": 1.5989, + "theoretical_loss": 4.0392814633250715, + "tokens_seen": 401014784 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004436710130391174, + "loss": 1.6112, + "theoretical_loss": 4.039208168729475, + "tokens_seen": 401080320 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004436609829488465, + "loss": 1.6042, + "theoretical_loss": 4.039134889461866, + "tokens_seen": 401145856 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044365095285857576, + "loss": 1.5745, + "theoretical_loss": 4.039061625516532, + "tokens_seen": 401211392 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044364092276830494, + "loss": 1.6576, + "theoretical_loss": 4.038988376887771, + "tokens_seen": 401276928 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004436308926780341, + "loss": 1.6301, + "theoretical_loss": 4.038915143569876, + "tokens_seen": 401342464 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004436208625877633, + "loss": 1.6166, + "theoretical_loss": 4.038841925557151, + "tokens_seen": 401408000 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044361083249749254, + "loss": 1.6363, + "theoretical_loss": 4.038768722843898, + "tokens_seen": 401473536 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044360080240722166, + "loss": 1.6491, + "theoretical_loss": 4.038695535424422, + "tokens_seen": 401539072 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004435907723169509, + "loss": 1.6207, + "theoretical_loss": 4.038622363293034, + "tokens_seen": 401604608 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044358074222668, + "loss": 1.5705, + "theoretical_loss": 4.0385492064440465, + "tokens_seen": 401670144 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044357071213640926, + "loss": 1.6656, + "theoretical_loss": 4.038476064871775, + "tokens_seen": 401735680 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044356068204613844, + "loss": 1.4736, + "theoretical_loss": 4.0384029385705364, + "tokens_seen": 401801216 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004435506519558676, + "loss": 1.6665, + "theoretical_loss": 4.038329827534654, + "tokens_seen": 401866752 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004435406218655968, + "loss": 1.4909, + "theoretical_loss": 4.038256731758453, + "tokens_seen": 401932288 + }, + { + "epoch": 0.12, + "learning_rate": 0.000443530591775326, + "loss": 1.5834, + "theoretical_loss": 4.038183651236259, + "tokens_seen": 401997824 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044352056168505517, + "loss": 1.5833, + "theoretical_loss": 4.038110585962404, + "tokens_seen": 402063360 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004435105315947844, + "loss": 1.5603, + "theoretical_loss": 4.038037535931221, + "tokens_seen": 402128896 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044350050150451353, + "loss": 1.6142, + "theoretical_loss": 4.037964501137048, + "tokens_seen": 402194432 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044349047141424276, + "loss": 1.5615, + "theoretical_loss": 4.037891481574224, + "tokens_seen": 402259968 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004434804413239719, + "loss": 1.6453, + "theoretical_loss": 4.037818477237091, + "tokens_seen": 402325504 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004434704112337011, + "loss": 1.6285, + "theoretical_loss": 4.037745488119996, + "tokens_seen": 402391040 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004434603811434303, + "loss": 1.5885, + "theoretical_loss": 4.037672514217288, + "tokens_seen": 402456576 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004434503510531595, + "loss": 1.5615, + "theoretical_loss": 4.037599555523318, + "tokens_seen": 402522112 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044344032096288867, + "loss": 1.5966, + "theoretical_loss": 4.037526612032441, + "tokens_seen": 402587648 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004434302908726179, + "loss": 1.5584, + "theoretical_loss": 4.037453683739015, + "tokens_seen": 402653184 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044342026078234703, + "loss": 1.6806, + "theoretical_loss": 4.037380770637402, + "tokens_seen": 402718720 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044341023069207627, + "loss": 1.543, + "theoretical_loss": 4.037307872721964, + "tokens_seen": 402784256 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004434002006018054, + "loss": 1.5374, + "theoretical_loss": 4.037234989987069, + "tokens_seen": 402849792 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044339017051153463, + "loss": 1.5172, + "theoretical_loss": 4.0371621224270875, + "tokens_seen": 402915328 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004433801404212638, + "loss": 1.5663, + "theoretical_loss": 4.037089270036391, + "tokens_seen": 402980864 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 155674, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5914859771728516, + "objective/train/theoretical_loss": 4.037016432809357, + "objective/train/tokens_used": 403046400, + "theoretical_loss": 4.037016432809357, + "tokens_seen": 403046400 + }, + { + "epoch": 0.12, + "learning_rate": 0.000443370110330993, + "loss": 1.5974, + "theoretical_loss": 4.037016432809357, + "tokens_seen": 403046400 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044336008024072217, + "loss": 1.4496, + "theoretical_loss": 4.036943610740362, + "tokens_seen": 403111936 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044335005015045135, + "loss": 1.5284, + "theoretical_loss": 4.036870803823791, + "tokens_seen": 403177472 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044334002006018053, + "loss": 1.6356, + "theoretical_loss": 4.036798012054027, + "tokens_seen": 403243008 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044332998996990977, + "loss": 1.4868, + "theoretical_loss": 4.036725235425458, + "tokens_seen": 403308544 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004433199598796389, + "loss": 1.5914, + "theoretical_loss": 4.0366524739324765, + "tokens_seen": 403374080 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044330992978936813, + "loss": 1.5773, + "theoretical_loss": 4.036579727569475, + "tokens_seen": 403439616 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004432998996990973, + "loss": 1.6568, + "theoretical_loss": 4.036506996330849, + "tokens_seen": 403505152 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004432898696088265, + "loss": 1.6951, + "theoretical_loss": 4.036434280211001, + "tokens_seen": 403570688 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004432798395185557, + "loss": 1.6416, + "theoretical_loss": 4.036361579204334, + "tokens_seen": 403636224 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044326980942828486, + "loss": 1.5815, + "theoretical_loss": 4.036288893305251, + "tokens_seen": 403701760 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044325977933801404, + "loss": 1.5694, + "theoretical_loss": 4.036216222508163, + "tokens_seen": 403767296 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044324974924774327, + "loss": 1.4911, + "theoretical_loss": 4.036143566807481, + "tokens_seen": 403832832 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004432397191574724, + "loss": 1.5694, + "theoretical_loss": 4.03607092619762, + "tokens_seen": 403898368 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044322968906720163, + "loss": 1.5425, + "theoretical_loss": 4.035998300672999, + "tokens_seen": 403963904 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044321965897693076, + "loss": 1.591, + "theoretical_loss": 4.035925690228035, + "tokens_seen": 404029440 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044320962888666, + "loss": 1.5948, + "theoretical_loss": 4.035853094857156, + "tokens_seen": 404094976 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004431995987963892, + "loss": 1.591, + "theoretical_loss": 4.035780514554787, + "tokens_seen": 404160512 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044318956870611836, + "loss": 1.5639, + "theoretical_loss": 4.0357079493153565, + "tokens_seen": 404226048 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044317953861584754, + "loss": 1.6002, + "theoretical_loss": 4.035635399133298, + "tokens_seen": 404291584 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004431695085255767, + "loss": 1.5231, + "theoretical_loss": 4.035562864003047, + "tokens_seen": 404357120 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004431594784353059, + "loss": 1.5879, + "theoretical_loss": 4.035490343919042, + "tokens_seen": 404422656 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044314944834503514, + "loss": 1.6344, + "theoretical_loss": 4.035417838875725, + "tokens_seen": 404488192 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044313941825476427, + "loss": 1.5394, + "theoretical_loss": 4.035345348867539, + "tokens_seen": 404553728 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004431293881644935, + "loss": 1.5883, + "theoretical_loss": 4.035272873888932, + "tokens_seen": 404619264 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004431193580742227, + "loss": 1.6583, + "theoretical_loss": 4.035200413934355, + "tokens_seen": 404684800 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044310932798395186, + "loss": 1.6014, + "theoretical_loss": 4.03512796899826, + "tokens_seen": 404750336 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044309929789368104, + "loss": 1.6625, + "theoretical_loss": 4.035055539075104, + "tokens_seen": 404815872 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004430892678034102, + "loss": 1.686, + "theoretical_loss": 4.034983124159346, + "tokens_seen": 404881408 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004430792377131394, + "loss": 1.6035, + "theoretical_loss": 4.0349107242454485, + "tokens_seen": 404946944 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044306920762286864, + "loss": 1.5189, + "theoretical_loss": 4.034838339327876, + "tokens_seen": 405012480 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044305917753259777, + "loss": 1.6219, + "theoretical_loss": 4.034765969401096, + "tokens_seen": 405078016 + }, + { + "epoch": 0.12, + "learning_rate": 0.000443049147442327, + "loss": 1.5554, + "theoretical_loss": 4.03469361445958, + "tokens_seen": 405143552 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044303911735205613, + "loss": 1.5979, + "theoretical_loss": 4.034621274497802, + "tokens_seen": 405209088 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044302908726178537, + "loss": 1.6011, + "theoretical_loss": 4.0345489495102385, + "tokens_seen": 405274624 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044301905717151455, + "loss": 1.5614, + "theoretical_loss": 4.034476639491368, + "tokens_seen": 405340160 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044300902708124373, + "loss": 1.581, + "theoretical_loss": 4.034404344435676, + "tokens_seen": 405405696 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004429989969909729, + "loss": 1.5712, + "theoretical_loss": 4.0343320643376455, + "tokens_seen": 405471232 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004429889669007021, + "loss": 1.4967, + "theoretical_loss": 4.034259799191765, + "tokens_seen": 405536768 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044297893681043127, + "loss": 1.5769, + "theoretical_loss": 4.034187548992527, + "tokens_seen": 405602304 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004429689067201605, + "loss": 1.6181, + "theoretical_loss": 4.034115313734426, + "tokens_seen": 405667840 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044295887662988963, + "loss": 1.5151, + "theoretical_loss": 4.034043093411958, + "tokens_seen": 405733376 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044294884653961887, + "loss": 1.5589, + "theoretical_loss": 4.033970888019624, + "tokens_seen": 405798912 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004429388164493481, + "loss": 1.6184, + "theoretical_loss": 4.033898697551926, + "tokens_seen": 405864448 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044292878635907723, + "loss": 1.6428, + "theoretical_loss": 4.0338265220033716, + "tokens_seen": 405929984 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044291875626880647, + "loss": 1.5708, + "theoretical_loss": 4.0337543613684685, + "tokens_seen": 405995520 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004429087261785356, + "loss": 1.5455, + "theoretical_loss": 4.033682215641729, + "tokens_seen": 406061056 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044289869608826483, + "loss": 1.6317, + "theoretical_loss": 4.033610084817667, + "tokens_seen": 406126592 + }, + { + "epoch": 0.12, + "learning_rate": 0.000442888665997994, + "loss": 1.4989, + "theoretical_loss": 4.0335379688908, + "tokens_seen": 406192128 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004428786359077232, + "loss": 1.55, + "theoretical_loss": 4.03346586785565, + "tokens_seen": 406257664 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 156877, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4918584823608398, + "objective/train/theoretical_loss": 4.03339378170674, + "objective/train/tokens_used": 406323200, + "theoretical_loss": 4.03339378170674, + "tokens_seen": 406323200 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044286860581745237, + "loss": 1.625, + "theoretical_loss": 4.03339378170674, + "tokens_seen": 406323200 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044285857572718155, + "loss": 1.569, + "theoretical_loss": 4.033321710438594, + "tokens_seen": 406388736 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044284854563691073, + "loss": 1.5869, + "theoretical_loss": 4.033249654045743, + "tokens_seen": 406454272 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044283851554663997, + "loss": 1.4823, + "theoretical_loss": 4.033177612522721, + "tokens_seen": 406519808 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004428284854563691, + "loss": 1.6755, + "theoretical_loss": 4.0331055858640585, + "tokens_seen": 406585344 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044281845536609833, + "loss": 1.6777, + "theoretical_loss": 4.033033574064297, + "tokens_seen": 406650880 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004428084252758275, + "loss": 1.5694, + "theoretical_loss": 4.032961577117976, + "tokens_seen": 406716416 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004427983951855567, + "loss": 1.5614, + "theoretical_loss": 4.032889595019638, + "tokens_seen": 406781952 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004427883650952859, + "loss": 1.6066, + "theoretical_loss": 4.032817627763832, + "tokens_seen": 406847488 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044277833500501506, + "loss": 1.6083, + "theoretical_loss": 4.0327456753451045, + "tokens_seen": 406913024 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044276830491474424, + "loss": 1.5497, + "theoretical_loss": 4.03267373775801, + "tokens_seen": 406978560 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044275827482447347, + "loss": 1.5334, + "theoretical_loss": 4.032601814997102, + "tokens_seen": 407044096 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004427482447342026, + "loss": 1.5703, + "theoretical_loss": 4.03252990705694, + "tokens_seen": 407109632 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044273821464393184, + "loss": 1.6674, + "theoretical_loss": 4.032458013932084, + "tokens_seen": 407175168 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044272818455366096, + "loss": 1.5719, + "theoretical_loss": 4.032386135617098, + "tokens_seen": 407240704 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004427181544633902, + "loss": 1.5175, + "theoretical_loss": 4.032314272106548, + "tokens_seen": 407306240 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004427081243731194, + "loss": 1.5531, + "theoretical_loss": 4.0322424233950045, + "tokens_seen": 407371776 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044269809428284856, + "loss": 1.646, + "theoretical_loss": 4.032170589477039, + "tokens_seen": 407437312 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044268806419257774, + "loss": 1.6682, + "theoretical_loss": 4.032098770347228, + "tokens_seen": 407502848 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004426780341023069, + "loss": 1.5499, + "theoretical_loss": 4.0320269660001475, + "tokens_seen": 407568384 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004426680040120361, + "loss": 1.5089, + "theoretical_loss": 4.0319551764303805, + "tokens_seen": 407633920 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044265797392176534, + "loss": 1.6451, + "theoretical_loss": 4.03188340163251, + "tokens_seen": 407699456 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044264794383149447, + "loss": 1.6031, + "theoretical_loss": 4.031811641601122, + "tokens_seen": 407764992 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004426379137412237, + "loss": 1.5219, + "theoretical_loss": 4.031739896330807, + "tokens_seen": 407830528 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004426278836509529, + "loss": 1.5535, + "theoretical_loss": 4.031668165816156, + "tokens_seen": 407896064 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044261785356068206, + "loss": 1.6014, + "theoretical_loss": 4.031596450051766, + "tokens_seen": 407961600 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044260782347041124, + "loss": 1.6276, + "theoretical_loss": 4.031524749032234, + "tokens_seen": 408027136 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004425977933801404, + "loss": 1.686, + "theoretical_loss": 4.031453062752161, + "tokens_seen": 408092672 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004425877632898696, + "loss": 1.5442, + "theoretical_loss": 4.031381391206151, + "tokens_seen": 408158208 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044257773319959884, + "loss": 1.6052, + "theoretical_loss": 4.03130973438881, + "tokens_seen": 408223744 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044256770310932797, + "loss": 1.6263, + "theoretical_loss": 4.031238092294748, + "tokens_seen": 408289280 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004425576730190572, + "loss": 1.5916, + "theoretical_loss": 4.031166464918576, + "tokens_seen": 408354816 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044254764292878633, + "loss": 1.5904, + "theoretical_loss": 4.0310948522549115, + "tokens_seen": 408420352 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044253761283851557, + "loss": 1.5723, + "theoretical_loss": 4.031023254298371, + "tokens_seen": 408485888 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044252758274824475, + "loss": 1.6016, + "theoretical_loss": 4.030951671043574, + "tokens_seen": 408551424 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044251755265797393, + "loss": 1.5401, + "theoretical_loss": 4.030880102485147, + "tokens_seen": 408616960 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004425075225677031, + "loss": 1.5355, + "theoretical_loss": 4.030808548617714, + "tokens_seen": 408682496 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004424974924774323, + "loss": 1.6041, + "theoretical_loss": 4.030737009435906, + "tokens_seen": 408748032 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044248746238716147, + "loss": 1.567, + "theoretical_loss": 4.030665484934353, + "tokens_seen": 408813568 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004424774322968907, + "loss": 1.5022, + "theoretical_loss": 4.030593975107692, + "tokens_seen": 408879104 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044246740220661983, + "loss": 1.5263, + "theoretical_loss": 4.03052247995056, + "tokens_seen": 408944640 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044245737211634907, + "loss": 1.6137, + "theoretical_loss": 4.030450999457596, + "tokens_seen": 409010176 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044244734202607825, + "loss": 1.5153, + "theoretical_loss": 4.030379533623445, + "tokens_seen": 409075712 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044243731193580743, + "loss": 1.6396, + "theoretical_loss": 4.030308082442753, + "tokens_seen": 409141248 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004424272818455366, + "loss": 1.5654, + "theoretical_loss": 4.030236645910169, + "tokens_seen": 409206784 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004424172517552658, + "loss": 1.6037, + "theoretical_loss": 4.030165224020345, + "tokens_seen": 409272320 + }, + { + "epoch": 0.12, + "learning_rate": 0.000442407221664995, + "loss": 1.6483, + "theoretical_loss": 4.030093816767934, + "tokens_seen": 409337856 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004423971915747242, + "loss": 1.6522, + "theoretical_loss": 4.030022424147595, + "tokens_seen": 409403392 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044238716148445334, + "loss": 1.5115, + "theoretical_loss": 4.029951046153988, + "tokens_seen": 409468928 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044237713139418257, + "loss": 1.6527, + "theoretical_loss": 4.029879682781776, + "tokens_seen": 409534464 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 158085, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5842972993850708, + "objective/train/theoretical_loss": 4.029808334025624, + "objective/train/tokens_used": 409600000, + "theoretical_loss": 4.029808334025624, + "tokens_seen": 409600000 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004423671013039117, + "loss": 1.6558, + "theoretical_loss": 4.029808334025624, + "tokens_seen": 409600000 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044235707121364093, + "loss": 1.5592, + "theoretical_loss": 4.0297369998802015, + "tokens_seen": 409665536 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004423470411233701, + "loss": 1.4817, + "theoretical_loss": 4.0296656803401785, + "tokens_seen": 409731072 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004423370110330993, + "loss": 1.5172, + "theoretical_loss": 4.029594375400231, + "tokens_seen": 409796608 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004423269809428285, + "loss": 1.6072, + "theoretical_loss": 4.029523085055036, + "tokens_seen": 409862144 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004423169508525577, + "loss": 1.501, + "theoretical_loss": 4.029451809299273, + "tokens_seen": 409927680 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044230692076228684, + "loss": 1.4771, + "theoretical_loss": 4.029380548127622, + "tokens_seen": 409993216 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004422968906720161, + "loss": 1.5675, + "theoretical_loss": 4.029309301534772, + "tokens_seen": 410058752 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004422868605817452, + "loss": 1.5196, + "theoretical_loss": 4.02923806951541, + "tokens_seen": 410124288 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044227683049147444, + "loss": 1.5449, + "theoretical_loss": 4.029166852064225, + "tokens_seen": 410189824 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004422668004012036, + "loss": 1.5316, + "theoretical_loss": 4.029095649175913, + "tokens_seen": 410255360 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004422567703109328, + "loss": 1.5462, + "theoretical_loss": 4.02902446084517, + "tokens_seen": 410320896 + }, + { + "epoch": 0.12, + "learning_rate": 0.000442246740220662, + "loss": 1.6166, + "theoretical_loss": 4.028953287066695, + "tokens_seen": 410386432 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044223671013039116, + "loss": 1.5366, + "theoretical_loss": 4.02888212783519, + "tokens_seen": 410451968 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044222668004012034, + "loss": 1.6212, + "theoretical_loss": 4.028810983145361, + "tokens_seen": 410517504 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004422166499498496, + "loss": 1.4884, + "theoretical_loss": 4.0287398529919125, + "tokens_seen": 410583040 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004422066198595787, + "loss": 1.4748, + "theoretical_loss": 4.028668737369558, + "tokens_seen": 410648576 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044219658976930794, + "loss": 1.5334, + "theoretical_loss": 4.028597636273009, + "tokens_seen": 410714112 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004421865596790371, + "loss": 1.5665, + "theoretical_loss": 4.028526549696982, + "tokens_seen": 410779648 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004421765295887663, + "loss": 1.5466, + "theoretical_loss": 4.0284554776361965, + "tokens_seen": 410845184 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044216649949849554, + "loss": 1.554, + "theoretical_loss": 4.028384420085372, + "tokens_seen": 410910720 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044215646940822467, + "loss": 1.4902, + "theoretical_loss": 4.028313377039234, + "tokens_seen": 410976256 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004421464393179539, + "loss": 1.5452, + "theoretical_loss": 4.028242348492509, + "tokens_seen": 411041792 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004421364092276831, + "loss": 1.5022, + "theoretical_loss": 4.028171334439927, + "tokens_seen": 411107328 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044212637913741226, + "loss": 1.6517, + "theoretical_loss": 4.02810033487622, + "tokens_seen": 411172864 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044211634904714144, + "loss": 1.4666, + "theoretical_loss": 4.028029349796125, + "tokens_seen": 411238400 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004421063189568706, + "loss": 1.6111, + "theoretical_loss": 4.027958379194376, + "tokens_seen": 411303936 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004420962888665998, + "loss": 1.5652, + "theoretical_loss": 4.027887423065717, + "tokens_seen": 411369472 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044208625877632904, + "loss": 1.6735, + "theoretical_loss": 4.027816481404892, + "tokens_seen": 411435008 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044207622868605817, + "loss": 1.5165, + "theoretical_loss": 4.0277455542066445, + "tokens_seen": 411500544 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004420661985957874, + "loss": 1.5177, + "theoretical_loss": 4.027674641465725, + "tokens_seen": 411566080 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044205616850551653, + "loss": 1.5844, + "theoretical_loss": 4.027603743176885, + "tokens_seen": 411631616 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044204613841524577, + "loss": 1.5185, + "theoretical_loss": 4.02753285933488, + "tokens_seen": 411697152 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044203610832497495, + "loss": 1.5788, + "theoretical_loss": 4.027461989934466, + "tokens_seen": 411762688 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044202607823470413, + "loss": 1.4866, + "theoretical_loss": 4.027391134970402, + "tokens_seen": 411828224 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004420160481444333, + "loss": 1.4839, + "theoretical_loss": 4.027320294437452, + "tokens_seen": 411893760 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004420060180541625, + "loss": 1.5915, + "theoretical_loss": 4.027249468330381, + "tokens_seen": 411959296 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044199598796389167, + "loss": 1.4566, + "theoretical_loss": 4.027178656643957, + "tokens_seen": 412024832 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004419859578736209, + "loss": 1.5913, + "theoretical_loss": 4.0271078593729515, + "tokens_seen": 412090368 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044197592778335003, + "loss": 1.5022, + "theoretical_loss": 4.027037076512137, + "tokens_seen": 412155904 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044196589769307927, + "loss": 1.6921, + "theoretical_loss": 4.026966308056291, + "tokens_seen": 412221440 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044195586760280845, + "loss": 1.5794, + "theoretical_loss": 4.026895554000192, + "tokens_seen": 412286976 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044194583751253763, + "loss": 1.6427, + "theoretical_loss": 4.026824814338621, + "tokens_seen": 412352512 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004419358074222668, + "loss": 1.6472, + "theoretical_loss": 4.026754089066364, + "tokens_seen": 412418048 + }, + { + "epoch": 0.12, + "learning_rate": 0.000441925777331996, + "loss": 1.5518, + "theoretical_loss": 4.026683378178206, + "tokens_seen": 412483584 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004419157472417252, + "loss": 1.67, + "theoretical_loss": 4.026612681668939, + "tokens_seen": 412549120 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004419057171514544, + "loss": 1.662, + "theoretical_loss": 4.026541999533355, + "tokens_seen": 412614656 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044189568706118354, + "loss": 1.555, + "theoretical_loss": 4.026471331766249, + "tokens_seen": 412680192 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044188565697091277, + "loss": 1.5986, + "theoretical_loss": 4.026400678362418, + "tokens_seen": 412745728 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004418756268806419, + "loss": 1.4865, + "theoretical_loss": 4.026330039316665, + "tokens_seen": 412811264 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 159288, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5327322483062744, + "objective/train/theoretical_loss": 4.026259414623792, + "objective/train/tokens_used": 412876800, + "theoretical_loss": 4.026259414623792, + "tokens_seen": 412876800 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044186559679037113, + "loss": 1.4466, + "theoretical_loss": 4.026259414623792, + "tokens_seen": 412876800 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004418555667001003, + "loss": 1.6004, + "theoretical_loss": 4.026188804278605, + "tokens_seen": 412942336 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004418455366098295, + "loss": 1.524, + "theoretical_loss": 4.026118208275914, + "tokens_seen": 413007872 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004418355065195587, + "loss": 1.5302, + "theoretical_loss": 4.02604762661053, + "tokens_seen": 413073408 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004418254764292879, + "loss": 1.479, + "theoretical_loss": 4.025977059277267, + "tokens_seen": 413138944 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044181544633901704, + "loss": 1.566, + "theoretical_loss": 4.025906506270941, + "tokens_seen": 413204480 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004418054162487463, + "loss": 1.5841, + "theoretical_loss": 4.0258359675863735, + "tokens_seen": 413270016 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004417953861584754, + "loss": 1.6021, + "theoretical_loss": 4.025765443218386, + "tokens_seen": 413335552 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044178535606820464, + "loss": 1.5623, + "theoretical_loss": 4.025694933161804, + "tokens_seen": 413401088 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004417753259779338, + "loss": 1.6008, + "theoretical_loss": 4.025624437411455, + "tokens_seen": 413466624 + }, + { + "epoch": 0.13, + "learning_rate": 0.000441765295887663, + "loss": 1.5926, + "theoretical_loss": 4.0255539559621685, + "tokens_seen": 413532160 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004417552657973922, + "loss": 1.5915, + "theoretical_loss": 4.025483488808779, + "tokens_seen": 413597696 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044174523570712136, + "loss": 1.5687, + "theoretical_loss": 4.025413035946121, + "tokens_seen": 413663232 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044173520561685054, + "loss": 1.5105, + "theoretical_loss": 4.025342597369034, + "tokens_seen": 413728768 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004417251755265798, + "loss": 1.5094, + "theoretical_loss": 4.025272173072359, + "tokens_seen": 413794304 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004417151454363089, + "loss": 1.632, + "theoretical_loss": 4.025201763050939, + "tokens_seen": 413859840 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044170511534603814, + "loss": 1.4594, + "theoretical_loss": 4.025131367299621, + "tokens_seen": 413925376 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044169508525576727, + "loss": 1.6094, + "theoretical_loss": 4.025060985813255, + "tokens_seen": 413990912 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004416850551654965, + "loss": 1.5552, + "theoretical_loss": 4.024990618586692, + "tokens_seen": 414056448 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004416750250752257, + "loss": 1.5283, + "theoretical_loss": 4.024920265614787, + "tokens_seen": 414121984 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044166499498495487, + "loss": 1.6545, + "theoretical_loss": 4.024849926892396, + "tokens_seen": 414187520 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044165496489468405, + "loss": 1.6542, + "theoretical_loss": 4.02477960241438, + "tokens_seen": 414253056 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004416449348044133, + "loss": 1.57, + "theoretical_loss": 4.024709292175602, + "tokens_seen": 414318592 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004416349047141424, + "loss": 1.5299, + "theoretical_loss": 4.024638996170927, + "tokens_seen": 414384128 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044162487462387164, + "loss": 1.6873, + "theoretical_loss": 4.024568714395221, + "tokens_seen": 414449664 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044161484453360077, + "loss": 1.5304, + "theoretical_loss": 4.0244984468433564, + "tokens_seen": 414515200 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044160481444333, + "loss": 1.5816, + "theoretical_loss": 4.0244281935102055, + "tokens_seen": 414580736 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004415947843530592, + "loss": 1.5743, + "theoretical_loss": 4.024357954390645, + "tokens_seen": 414646272 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044158475426278837, + "loss": 1.5566, + "theoretical_loss": 4.0242877294795525, + "tokens_seen": 414711808 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044157472417251755, + "loss": 1.6023, + "theoretical_loss": 4.0242175187718106, + "tokens_seen": 414777344 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044156469408224673, + "loss": 1.5374, + "theoretical_loss": 4.024147322262301, + "tokens_seen": 414842880 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004415546639919759, + "loss": 1.5734, + "theoretical_loss": 4.024077139945912, + "tokens_seen": 414908416 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044154463390170515, + "loss": 1.6211, + "theoretical_loss": 4.024006971817532, + "tokens_seen": 414973952 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004415346038114343, + "loss": 1.5504, + "theoretical_loss": 4.023936817872053, + "tokens_seen": 415039488 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004415245737211635, + "loss": 1.6544, + "theoretical_loss": 4.023866678104368, + "tokens_seen": 415105024 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044151454363089264, + "loss": 1.59, + "theoretical_loss": 4.023796552509376, + "tokens_seen": 415170560 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044150451354062187, + "loss": 1.566, + "theoretical_loss": 4.023726441081976, + "tokens_seen": 415236096 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044149448345035105, + "loss": 1.6711, + "theoretical_loss": 4.02365634381707, + "tokens_seen": 415301632 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044148445336008023, + "loss": 1.4352, + "theoretical_loss": 4.023586260709564, + "tokens_seen": 415367168 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004414744232698094, + "loss": 1.5901, + "theoretical_loss": 4.023516191754363, + "tokens_seen": 415432704 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044146439317953865, + "loss": 1.6904, + "theoretical_loss": 4.02344613694638, + "tokens_seen": 415498240 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004414543630892678, + "loss": 1.6018, + "theoretical_loss": 4.023376096280527, + "tokens_seen": 415563776 + }, + { + "epoch": 0.13, + "learning_rate": 0.000441444332998997, + "loss": 1.6066, + "theoretical_loss": 4.02330606975172, + "tokens_seen": 415629312 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004414343029087262, + "loss": 1.5874, + "theoretical_loss": 4.023236057354875, + "tokens_seen": 415694848 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004414242728184554, + "loss": 1.5523, + "theoretical_loss": 4.023166059084916, + "tokens_seen": 415760384 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004414142427281846, + "loss": 1.5297, + "theoretical_loss": 4.023096074936764, + "tokens_seen": 415825920 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044140421263791374, + "loss": 1.5549, + "theoretical_loss": 4.0230261049053455, + "tokens_seen": 415891456 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044139418254764297, + "loss": 1.5383, + "theoretical_loss": 4.022956148985589, + "tokens_seen": 415956992 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004413841524573721, + "loss": 1.4823, + "theoretical_loss": 4.0228862071724265, + "tokens_seen": 416022528 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044137412236710133, + "loss": 1.6685, + "theoretical_loss": 4.022816279460791, + "tokens_seen": 416088064 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 160506, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4579585790634155, + "objective/train/theoretical_loss": 4.02274636584562, + "objective/train/tokens_used": 416153600, + "theoretical_loss": 4.02274636584562, + "tokens_seen": 416153600 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004413640922768305, + "loss": 1.5199, + "theoretical_loss": 4.02274636584562, + "tokens_seen": 416153600 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004413540621865597, + "loss": 1.5968, + "theoretical_loss": 4.022676466321853, + "tokens_seen": 416219136 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004413440320962889, + "loss": 1.6423, + "theoretical_loss": 4.022606580884429, + "tokens_seen": 416284672 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004413340020060181, + "loss": 1.5581, + "theoretical_loss": 4.0225367095282945, + "tokens_seen": 416350208 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044132397191574724, + "loss": 1.5552, + "theoretical_loss": 4.022466852248396, + "tokens_seen": 416415744 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004413139418254765, + "loss": 1.6184, + "theoretical_loss": 4.0223970090396834, + "tokens_seen": 416481280 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004413039117352056, + "loss": 1.5939, + "theoretical_loss": 4.0223271798971085, + "tokens_seen": 416546816 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044129388164493484, + "loss": 1.5309, + "theoretical_loss": 4.0222573648156255, + "tokens_seen": 416612352 + }, + { + "epoch": 0.13, + "learning_rate": 0.000441283851554664, + "loss": 1.5422, + "theoretical_loss": 4.022187563790193, + "tokens_seen": 416677888 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004412738214643932, + "loss": 1.5679, + "theoretical_loss": 4.02211777681577, + "tokens_seen": 416743424 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004412637913741224, + "loss": 1.5383, + "theoretical_loss": 4.022048003887319, + "tokens_seen": 416808960 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044125376128385156, + "loss": 1.4451, + "theoretical_loss": 4.021978244999806, + "tokens_seen": 416874496 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044124373119358074, + "loss": 1.5019, + "theoretical_loss": 4.021908500148198, + "tokens_seen": 416940032 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044123370110331, + "loss": 1.5519, + "theoretical_loss": 4.0218387693274655, + "tokens_seen": 417005568 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004412236710130391, + "loss": 1.6778, + "theoretical_loss": 4.021769052532582, + "tokens_seen": 417071104 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044121364092276834, + "loss": 1.4431, + "theoretical_loss": 4.0216993497585225, + "tokens_seen": 417136640 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044120361083249747, + "loss": 1.5734, + "theoretical_loss": 4.021629661000265, + "tokens_seen": 417202176 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004411935807422267, + "loss": 1.5407, + "theoretical_loss": 4.02155998625279, + "tokens_seen": 417267712 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004411835506519559, + "loss": 1.6788, + "theoretical_loss": 4.021490325511082, + "tokens_seen": 417333248 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044117352056168507, + "loss": 1.5191, + "theoretical_loss": 4.021420678770125, + "tokens_seen": 417398784 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044116349047141425, + "loss": 1.568, + "theoretical_loss": 4.02135104602491, + "tokens_seen": 417464320 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004411534603811435, + "loss": 1.6131, + "theoretical_loss": 4.021281427270425, + "tokens_seen": 417529856 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004411434302908726, + "loss": 1.559, + "theoretical_loss": 4.021211822501666, + "tokens_seen": 417595392 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044113340020060184, + "loss": 1.5462, + "theoretical_loss": 4.021142231713629, + "tokens_seen": 417660928 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044112337011033097, + "loss": 1.5035, + "theoretical_loss": 4.021072654901312, + "tokens_seen": 417726464 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004411133400200602, + "loss": 1.6668, + "theoretical_loss": 4.021003092059716, + "tokens_seen": 417792000 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004411033099297894, + "loss": 1.4788, + "theoretical_loss": 4.020933543183845, + "tokens_seen": 417857536 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044109327983951857, + "loss": 1.5421, + "theoretical_loss": 4.020864008268706, + "tokens_seen": 417923072 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044108324974924775, + "loss": 1.5295, + "theoretical_loss": 4.020794487309308, + "tokens_seen": 417988608 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044107321965897693, + "loss": 1.5422, + "theoretical_loss": 4.020724980300663, + "tokens_seen": 418054144 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004410631895687061, + "loss": 1.5617, + "theoretical_loss": 4.020655487237785, + "tokens_seen": 418119680 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044105315947843535, + "loss": 1.6127, + "theoretical_loss": 4.02058600811569, + "tokens_seen": 418185216 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004410431293881645, + "loss": 1.6723, + "theoretical_loss": 4.020516542929398, + "tokens_seen": 418250752 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004410330992978937, + "loss": 1.5939, + "theoretical_loss": 4.0204470916739306, + "tokens_seen": 418316288 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044102306920762284, + "loss": 1.5386, + "theoretical_loss": 4.020377654344312, + "tokens_seen": 418381824 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044101303911735207, + "loss": 1.5501, + "theoretical_loss": 4.020308230935569, + "tokens_seen": 418447360 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044100300902708125, + "loss": 1.6412, + "theoretical_loss": 4.020238821442733, + "tokens_seen": 418512896 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044099297893681043, + "loss": 1.5769, + "theoretical_loss": 4.020169425860834, + "tokens_seen": 418578432 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004409829488465396, + "loss": 1.6127, + "theoretical_loss": 4.0201000441849075, + "tokens_seen": 418643968 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044097291875626885, + "loss": 1.6888, + "theoretical_loss": 4.0200306764099905, + "tokens_seen": 418709504 + }, + { + "epoch": 0.13, + "learning_rate": 0.000440962888665998, + "loss": 1.558, + "theoretical_loss": 4.019961322531122, + "tokens_seen": 418775040 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004409528585757272, + "loss": 1.6325, + "theoretical_loss": 4.019891982543346, + "tokens_seen": 418840576 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044094282848545634, + "loss": 1.4702, + "theoretical_loss": 4.019822656441706, + "tokens_seen": 418906112 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004409327983951856, + "loss": 1.5748, + "theoretical_loss": 4.01975334422125, + "tokens_seen": 418971648 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044092276830491476, + "loss": 1.6056, + "theoretical_loss": 4.019684045877027, + "tokens_seen": 419037184 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044091273821464394, + "loss": 1.5605, + "theoretical_loss": 4.019614761404091, + "tokens_seen": 419102720 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004409027081243731, + "loss": 1.5763, + "theoretical_loss": 4.019545490797495, + "tokens_seen": 419168256 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004408926780341023, + "loss": 1.5606, + "theoretical_loss": 4.0194762340522985, + "tokens_seen": 419233792 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004408826479438315, + "loss": 1.5819, + "theoretical_loss": 4.01940699116356, + "tokens_seen": 419299328 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004408726178535607, + "loss": 1.6557, + "theoretical_loss": 4.019337762126343, + "tokens_seen": 419364864 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 162117, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5481873750686646, + "objective/train/theoretical_loss": 4.019268546935713, + "objective/train/tokens_used": 419430400, + "theoretical_loss": 4.019268546935713, + "tokens_seen": 419430400 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044086258776328984, + "loss": 1.6545, + "theoretical_loss": 4.019268546935713, + "tokens_seen": 419430400 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004408525576730191, + "loss": 1.5244, + "theoretical_loss": 4.0191993455867365, + "tokens_seen": 419495936 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004408425275827482, + "loss": 1.5229, + "theoretical_loss": 4.019130158074484, + "tokens_seen": 419561472 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044083249749247744, + "loss": 1.612, + "theoretical_loss": 4.019060984394028, + "tokens_seen": 419627008 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004408224674022066, + "loss": 1.6334, + "theoretical_loss": 4.018991824540445, + "tokens_seen": 419692544 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004408124373119358, + "loss": 1.5218, + "theoretical_loss": 4.018922678508812, + "tokens_seen": 419758080 + }, + { + "epoch": 0.13, + "learning_rate": 0.000440802407221665, + "loss": 1.6553, + "theoretical_loss": 4.018853546294208, + "tokens_seen": 419823616 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004407923771313942, + "loss": 1.5668, + "theoretical_loss": 4.018784427891718, + "tokens_seen": 419889152 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044078234704112335, + "loss": 1.6837, + "theoretical_loss": 4.018715323296425, + "tokens_seen": 419954688 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004407723169508526, + "loss": 1.5945, + "theoretical_loss": 4.018646232503419, + "tokens_seen": 420020224 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004407622868605817, + "loss": 1.6805, + "theoretical_loss": 4.0185771555077885, + "tokens_seen": 420085760 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044075225677031094, + "loss": 1.6055, + "theoretical_loss": 4.018508092304628, + "tokens_seen": 420151296 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004407422266800401, + "loss": 1.6362, + "theoretical_loss": 4.018439042889032, + "tokens_seen": 420216832 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004407321965897693, + "loss": 1.5412, + "theoretical_loss": 4.018370007256098, + "tokens_seen": 420282368 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004407221664994985, + "loss": 1.5265, + "theoretical_loss": 4.018300985400927, + "tokens_seen": 420347904 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044071213640922767, + "loss": 1.548, + "theoretical_loss": 4.018231977318622, + "tokens_seen": 420413440 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044070210631895685, + "loss": 1.5832, + "theoretical_loss": 4.018162983004287, + "tokens_seen": 420478976 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004406920762286861, + "loss": 1.616, + "theoretical_loss": 4.018094002453031, + "tokens_seen": 420544512 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044068204613841527, + "loss": 1.5157, + "theoretical_loss": 4.018025035659965, + "tokens_seen": 420610048 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044067201604814445, + "loss": 1.6948, + "theoretical_loss": 4.017956082620201, + "tokens_seen": 420675584 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004406619859578737, + "loss": 1.5132, + "theoretical_loss": 4.0178871433288545, + "tokens_seen": 420741120 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004406519558676028, + "loss": 1.6125, + "theoretical_loss": 4.017818217781044, + "tokens_seen": 420806656 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044064192577733204, + "loss": 1.4806, + "theoretical_loss": 4.017749305971889, + "tokens_seen": 420872192 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044063189568706117, + "loss": 1.6229, + "theoretical_loss": 4.0176804078965125, + "tokens_seen": 420937728 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004406218655967904, + "loss": 1.5825, + "theoretical_loss": 4.017611523550039, + "tokens_seen": 421003264 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004406118355065196, + "loss": 1.4869, + "theoretical_loss": 4.017542652927599, + "tokens_seen": 421068800 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044060180541624877, + "loss": 1.6081, + "theoretical_loss": 4.01747379602432, + "tokens_seen": 421134336 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044059177532597795, + "loss": 1.6298, + "theoretical_loss": 4.017404952835336, + "tokens_seen": 421199872 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044058174523570713, + "loss": 1.5973, + "theoretical_loss": 4.017336123355783, + "tokens_seen": 421265408 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004405717151454363, + "loss": 1.5597, + "theoretical_loss": 4.017267307580798, + "tokens_seen": 421330944 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044056168505516555, + "loss": 1.6274, + "theoretical_loss": 4.017198505505521, + "tokens_seen": 421396480 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004405516549648947, + "loss": 1.6327, + "theoretical_loss": 4.017129717125094, + "tokens_seen": 421462016 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004405416248746239, + "loss": 1.5537, + "theoretical_loss": 4.017060942434664, + "tokens_seen": 421527552 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044053159478435304, + "loss": 1.4704, + "theoretical_loss": 4.016992181429377, + "tokens_seen": 421593088 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044052156469408227, + "loss": 1.6261, + "theoretical_loss": 4.016923434104386, + "tokens_seen": 421658624 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044051153460381145, + "loss": 1.6091, + "theoretical_loss": 4.01685470045484, + "tokens_seen": 421724160 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044050150451354063, + "loss": 1.5867, + "theoretical_loss": 4.0167859804758965, + "tokens_seen": 421789696 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004404914744232698, + "loss": 1.6685, + "theoretical_loss": 4.016717274162712, + "tokens_seen": 421855232 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044048144433299905, + "loss": 1.5374, + "theoretical_loss": 4.016648581510448, + "tokens_seen": 421920768 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004404714142427282, + "loss": 1.571, + "theoretical_loss": 4.016579902514264, + "tokens_seen": 421986304 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004404613841524574, + "loss": 1.5921, + "theoretical_loss": 4.016511237169329, + "tokens_seen": 422051840 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044045135406218654, + "loss": 1.5133, + "theoretical_loss": 4.016442585470807, + "tokens_seen": 422117376 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004404413239719158, + "loss": 1.6197, + "theoretical_loss": 4.01637394741387, + "tokens_seen": 422182912 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044043129388164496, + "loss": 1.5531, + "theoretical_loss": 4.016305322993689, + "tokens_seen": 422248448 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044042126379137414, + "loss": 1.5866, + "theoretical_loss": 4.016236712205439, + "tokens_seen": 422313984 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004404112337011033, + "loss": 1.5395, + "theoretical_loss": 4.016168115044298, + "tokens_seen": 422379520 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004404012036108325, + "loss": 1.5544, + "theoretical_loss": 4.016099531505445, + "tokens_seen": 422445056 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004403911735205617, + "loss": 1.5451, + "theoretical_loss": 4.016030961584063, + "tokens_seen": 422510592 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004403811434302909, + "loss": 1.4652, + "theoretical_loss": 4.015962405275336, + "tokens_seen": 422576128 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044037111334002004, + "loss": 1.4984, + "theoretical_loss": 4.015893862574451, + "tokens_seen": 422641664 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 163279, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6046137809753418, + "objective/train/theoretical_loss": 4.015825333476598, + "objective/train/tokens_used": 422707200, + "theoretical_loss": 4.015825333476598, + "tokens_seen": 422707200 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004403610832497493, + "loss": 1.5531, + "theoretical_loss": 4.015825333476598, + "tokens_seen": 422707200 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004403510531594784, + "loss": 1.5766, + "theoretical_loss": 4.0157568179769685, + "tokens_seen": 422772736 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044034102306920764, + "loss": 1.5834, + "theoretical_loss": 4.015688316070757, + "tokens_seen": 422838272 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004403309929789368, + "loss": 1.5706, + "theoretical_loss": 4.01561982775316, + "tokens_seen": 422903808 + }, + { + "epoch": 0.13, + "learning_rate": 0.000440320962888666, + "loss": 1.5114, + "theoretical_loss": 4.015551353019378, + "tokens_seen": 422969344 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004403109327983952, + "loss": 1.6439, + "theoretical_loss": 4.015482891864612, + "tokens_seen": 423034880 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004403009027081244, + "loss": 1.5311, + "theoretical_loss": 4.015414444284065, + "tokens_seen": 423100416 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044029087261785355, + "loss": 1.6116, + "theoretical_loss": 4.015346010272946, + "tokens_seen": 423165952 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004402808425275828, + "loss": 1.5208, + "theoretical_loss": 4.0152775898264625, + "tokens_seen": 423231488 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004402708124373119, + "loss": 1.5381, + "theoretical_loss": 4.0152091829398255, + "tokens_seen": 423297024 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044026078234704114, + "loss": 1.4343, + "theoretical_loss": 4.0151407896082505, + "tokens_seen": 423362560 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004402507522567703, + "loss": 1.5922, + "theoretical_loss": 4.015072409826953, + "tokens_seen": 423428096 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004402407221664995, + "loss": 1.5949, + "theoretical_loss": 4.015004043591151, + "tokens_seen": 423493632 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004402306920762287, + "loss": 1.5359, + "theoretical_loss": 4.014935690896067, + "tokens_seen": 423559168 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044022066198595787, + "loss": 1.5375, + "theoretical_loss": 4.014867351736924, + "tokens_seen": 423624704 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044021063189568705, + "loss": 1.6387, + "theoretical_loss": 4.014799026108948, + "tokens_seen": 423690240 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004402006018054163, + "loss": 1.5695, + "theoretical_loss": 4.014730714007367, + "tokens_seen": 423755776 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004401905717151454, + "loss": 1.6253, + "theoretical_loss": 4.014662415427413, + "tokens_seen": 423821312 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044018054162487465, + "loss": 1.5712, + "theoretical_loss": 4.014594130364319, + "tokens_seen": 423886848 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044017051153460383, + "loss": 1.5221, + "theoretical_loss": 4.01452585881332, + "tokens_seen": 423952384 + }, + { + "epoch": 0.13, + "learning_rate": 0.000440160481444333, + "loss": 1.6063, + "theoretical_loss": 4.014457600769655, + "tokens_seen": 424017920 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004401504513540622, + "loss": 1.4378, + "theoretical_loss": 4.0143893562285635, + "tokens_seen": 424083456 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044014042126379137, + "loss": 1.6358, + "theoretical_loss": 4.014321125185289, + "tokens_seen": 424148992 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044013039117352055, + "loss": 1.5027, + "theoretical_loss": 4.014252907635077, + "tokens_seen": 424214528 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004401203610832498, + "loss": 1.507, + "theoretical_loss": 4.014184703573176, + "tokens_seen": 424280064 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004401103309929789, + "loss": 1.4294, + "theoretical_loss": 4.014116512994835, + "tokens_seen": 424345600 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044010030090270815, + "loss": 1.598, + "theoretical_loss": 4.014048335895307, + "tokens_seen": 424411136 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004400902708124373, + "loss": 1.6551, + "theoretical_loss": 4.0139801722698465, + "tokens_seen": 424476672 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004400802407221665, + "loss": 1.6331, + "theoretical_loss": 4.013912022113711, + "tokens_seen": 424542208 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004400702106318957, + "loss": 1.4907, + "theoretical_loss": 4.013843885422162, + "tokens_seen": 424607744 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004400601805416249, + "loss": 1.5825, + "theoretical_loss": 4.01377576219046, + "tokens_seen": 424673280 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044005015045135406, + "loss": 1.6636, + "theoretical_loss": 4.01370765241387, + "tokens_seen": 424738816 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044004012036108324, + "loss": 1.5824, + "theoretical_loss": 4.013639556087659, + "tokens_seen": 424804352 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004400300902708124, + "loss": 1.5841, + "theoretical_loss": 4.0135714732070955, + "tokens_seen": 424869888 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044002006018054165, + "loss": 1.4964, + "theoretical_loss": 4.013503403767453, + "tokens_seen": 424935424 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004400100300902708, + "loss": 1.566, + "theoretical_loss": 4.013435347764005, + "tokens_seen": 425000960 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044, + "loss": 1.4724, + "theoretical_loss": 4.013367305192027, + "tokens_seen": 425066496 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004399899699097292, + "loss": 1.6129, + "theoretical_loss": 4.0132992760468, + "tokens_seen": 425132032 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004399799398194584, + "loss": 1.5025, + "theoretical_loss": 4.013231260323603, + "tokens_seen": 425197568 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043996990972918756, + "loss": 1.5166, + "theoretical_loss": 4.0131632580177214, + "tokens_seen": 425263104 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043995987963891674, + "loss": 1.5536, + "theoretical_loss": 4.01309526912444, + "tokens_seen": 425328640 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004399498495486459, + "loss": 1.4979, + "theoretical_loss": 4.013027293639049, + "tokens_seen": 425394176 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043993981945837516, + "loss": 1.5205, + "theoretical_loss": 4.0129593315568375, + "tokens_seen": 425459712 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043992978936810434, + "loss": 1.6292, + "theoretical_loss": 4.0128913828731, + "tokens_seen": 425525248 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004399197592778335, + "loss": 1.5375, + "theoretical_loss": 4.012823447583132, + "tokens_seen": 425590784 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004399097291875627, + "loss": 1.4429, + "theoretical_loss": 4.012755525682231, + "tokens_seen": 425656320 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004398996990972919, + "loss": 1.5594, + "theoretical_loss": 4.012687617165698, + "tokens_seen": 425721856 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004398896690070211, + "loss": 1.5205, + "theoretical_loss": 4.012619722028835, + "tokens_seen": 425787392 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043987963891675024, + "loss": 1.4937, + "theoretical_loss": 4.012551840266948, + "tokens_seen": 425852928 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004398696088264795, + "loss": 1.5226, + "theoretical_loss": 4.0124839718753424, + "tokens_seen": 425918464 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 164536, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5919736623764038, + "objective/train/theoretical_loss": 4.01241611684933, + "objective/train/tokens_used": 425984000, + "theoretical_loss": 4.01241611684933, + "tokens_seen": 425984000 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004398595787362086, + "loss": 1.5869, + "theoretical_loss": 4.01241611684933, + "tokens_seen": 425984000 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043984954864593784, + "loss": 1.6172, + "theoretical_loss": 4.012348275184223, + "tokens_seen": 426049536 + }, + { + "epoch": 0.13, + "learning_rate": 0.000439839518555667, + "loss": 1.5803, + "theoretical_loss": 4.012280446875336, + "tokens_seen": 426115072 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004398294884653962, + "loss": 1.5593, + "theoretical_loss": 4.012212631917986, + "tokens_seen": 426180608 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004398194583751254, + "loss": 1.6246, + "theoretical_loss": 4.012144830307491, + "tokens_seen": 426246144 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004398094282848546, + "loss": 1.5994, + "theoretical_loss": 4.012077042039174, + "tokens_seen": 426311680 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043979939819458375, + "loss": 1.5706, + "theoretical_loss": 4.012009267108358, + "tokens_seen": 426377216 + }, + { + "epoch": 0.13, + "learning_rate": 0.000439789368104313, + "loss": 1.6174, + "theoretical_loss": 4.01194150551037, + "tokens_seen": 426442752 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004397793380140421, + "loss": 1.6779, + "theoretical_loss": 4.011873757240539, + "tokens_seen": 426508288 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043976930792377134, + "loss": 1.4773, + "theoretical_loss": 4.011806022294196, + "tokens_seen": 426573824 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004397592778335005, + "loss": 1.449, + "theoretical_loss": 4.011738300666674, + "tokens_seen": 426639360 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004397492477432297, + "loss": 1.54, + "theoretical_loss": 4.011670592353309, + "tokens_seen": 426704896 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004397392176529589, + "loss": 1.4989, + "theoretical_loss": 4.011602897349439, + "tokens_seen": 426770432 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043972918756268807, + "loss": 1.4805, + "theoretical_loss": 4.011535215650405, + "tokens_seen": 426835968 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043971915747241725, + "loss": 1.5619, + "theoretical_loss": 4.0114675472515495, + "tokens_seen": 426901504 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004397091273821465, + "loss": 1.5867, + "theoretical_loss": 4.011399892148218, + "tokens_seen": 426967040 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004396990972918756, + "loss": 1.6101, + "theoretical_loss": 4.011332250335758, + "tokens_seen": 427032576 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043968906720160485, + "loss": 1.5066, + "theoretical_loss": 4.011264621809518, + "tokens_seen": 427098112 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043967903711133403, + "loss": 1.5092, + "theoretical_loss": 4.011197006564853, + "tokens_seen": 427163648 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004396690070210632, + "loss": 1.55, + "theoretical_loss": 4.011129404597114, + "tokens_seen": 427229184 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004396589769307924, + "loss": 1.5313, + "theoretical_loss": 4.011061815901661, + "tokens_seen": 427294720 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043964894684052157, + "loss": 1.4866, + "theoretical_loss": 4.010994240473853, + "tokens_seen": 427360256 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043963891675025075, + "loss": 1.5292, + "theoretical_loss": 4.0109266783090485, + "tokens_seen": 427425792 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043962888665998, + "loss": 1.5177, + "theoretical_loss": 4.010859129402615, + "tokens_seen": 427491328 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004396188565697091, + "loss": 1.6003, + "theoretical_loss": 4.010791593749916, + "tokens_seen": 427556864 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043960882647943835, + "loss": 1.5411, + "theoretical_loss": 4.010724071346322, + "tokens_seen": 427622400 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004395987963891675, + "loss": 1.5647, + "theoretical_loss": 4.0106565621872035, + "tokens_seen": 427687936 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004395887662988967, + "loss": 1.5514, + "theoretical_loss": 4.010589066267933, + "tokens_seen": 427753472 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004395787362086259, + "loss": 1.5597, + "theoretical_loss": 4.010521583583886, + "tokens_seen": 427819008 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004395687061183551, + "loss": 1.5374, + "theoretical_loss": 4.010454114130441, + "tokens_seen": 427884544 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043955867602808426, + "loss": 1.5145, + "theoretical_loss": 4.010386657902979, + "tokens_seen": 427950080 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043954864593781344, + "loss": 1.6433, + "theoretical_loss": 4.01031921489688, + "tokens_seen": 428015616 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004395386158475426, + "loss": 1.7175, + "theoretical_loss": 4.01025178510753, + "tokens_seen": 428081152 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043952858575727185, + "loss": 1.6106, + "theoretical_loss": 4.010184368530318, + "tokens_seen": 428146688 + }, + { + "epoch": 0.13, + "learning_rate": 0.000439518555667001, + "loss": 1.5704, + "theoretical_loss": 4.0101169651606305, + "tokens_seen": 428212224 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004395085255767302, + "loss": 1.5442, + "theoretical_loss": 4.010049574993861, + "tokens_seen": 428277760 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004394984954864594, + "loss": 1.5767, + "theoretical_loss": 4.009982198025404, + "tokens_seen": 428343296 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004394884653961886, + "loss": 1.5884, + "theoretical_loss": 4.0099148342506545, + "tokens_seen": 428408832 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043947843530591776, + "loss": 1.4817, + "theoretical_loss": 4.009847483665011, + "tokens_seen": 428474368 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043946840521564694, + "loss": 1.668, + "theoretical_loss": 4.009780146263877, + "tokens_seen": 428539904 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004394583751253761, + "loss": 1.6804, + "theoretical_loss": 4.009712822042652, + "tokens_seen": 428605440 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043944834503510536, + "loss": 1.6288, + "theoretical_loss": 4.009645510996744, + "tokens_seen": 428670976 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004394383149448345, + "loss": 1.4796, + "theoretical_loss": 4.0095782131215625, + "tokens_seen": 428736512 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004394282848545637, + "loss": 1.5612, + "theoretical_loss": 4.0095109284125146, + "tokens_seen": 428802048 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043941825476429285, + "loss": 1.5612, + "theoretical_loss": 4.009443656865013, + "tokens_seen": 428867584 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004394082246740221, + "loss": 1.5599, + "theoretical_loss": 4.009376398474475, + "tokens_seen": 428933120 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043939819458375126, + "loss": 1.436, + "theoretical_loss": 4.009309153236316, + "tokens_seen": 428998656 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043938816449348044, + "loss": 1.5108, + "theoretical_loss": 4.009241921145956, + "tokens_seen": 429064192 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004393781344032096, + "loss": 1.6849, + "theoretical_loss": 4.009174702198816, + "tokens_seen": 429129728 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004393681043129388, + "loss": 1.5307, + "theoretical_loss": 4.009107496390321, + "tokens_seen": 429195264 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 165667, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6264562606811523, + "objective/train/theoretical_loss": 4.009040303715897, + "objective/train/tokens_used": 429260800, + "theoretical_loss": 4.009040303715897, + "tokens_seen": 429260800 + }, + { + "epoch": 0.13, + "learning_rate": 0.000439358074222668, + "loss": 1.5752, + "theoretical_loss": 4.009040303715897, + "tokens_seen": 429260800 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004393480441323972, + "loss": 1.6282, + "theoretical_loss": 4.008973124170971, + "tokens_seen": 429326336 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043933801404212635, + "loss": 1.596, + "theoretical_loss": 4.008905957750976, + "tokens_seen": 429391872 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004393279839518556, + "loss": 1.6366, + "theoretical_loss": 4.0088388044513446, + "tokens_seen": 429457408 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043931795386158477, + "loss": 1.5681, + "theoretical_loss": 4.0087716642675115, + "tokens_seen": 429522944 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043930792377131395, + "loss": 1.4614, + "theoretical_loss": 4.008704537194916, + "tokens_seen": 429588480 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043929789368104313, + "loss": 1.6205, + "theoretical_loss": 4.008637423228996, + "tokens_seen": 429654016 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004392878635907723, + "loss": 1.6163, + "theoretical_loss": 4.008570322365195, + "tokens_seen": 429719552 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004392778335005015, + "loss": 1.4983, + "theoretical_loss": 4.008503234598958, + "tokens_seen": 429785088 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004392678034102307, + "loss": 1.6114, + "theoretical_loss": 4.008436159925731, + "tokens_seen": 429850624 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043925777331995985, + "loss": 1.5407, + "theoretical_loss": 4.008369098340963, + "tokens_seen": 429916160 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004392477432296891, + "loss": 1.6199, + "theoretical_loss": 4.008302049840106, + "tokens_seen": 429981696 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004392377131394182, + "loss": 1.5144, + "theoretical_loss": 4.008235014418614, + "tokens_seen": 430047232 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043922768304914745, + "loss": 1.5704, + "theoretical_loss": 4.008167992071941, + "tokens_seen": 430112768 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043921765295887663, + "loss": 1.5957, + "theoretical_loss": 4.008100982795549, + "tokens_seen": 430178304 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004392076228686058, + "loss": 1.5451, + "theoretical_loss": 4.008033986584895, + "tokens_seen": 430243840 + }, + { + "epoch": 0.13, + "learning_rate": 0.000439197592778335, + "loss": 1.5222, + "theoretical_loss": 4.007967003435443, + "tokens_seen": 430309376 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043918756268806423, + "loss": 1.4852, + "theoretical_loss": 4.007900033342658, + "tokens_seen": 430374912 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004391775325977934, + "loss": 1.4516, + "theoretical_loss": 4.007833076302006, + "tokens_seen": 430440448 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004391675025075226, + "loss": 1.5747, + "theoretical_loss": 4.00776613230896, + "tokens_seen": 430505984 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043915747241725177, + "loss": 1.5407, + "theoretical_loss": 4.007699201358989, + "tokens_seen": 430571520 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043914744232698095, + "loss": 1.615, + "theoretical_loss": 4.007632283447567, + "tokens_seen": 430637056 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004391374122367102, + "loss": 1.5685, + "theoretical_loss": 4.007565378570171, + "tokens_seen": 430702592 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004391273821464393, + "loss": 1.4843, + "theoretical_loss": 4.00749848672228, + "tokens_seen": 430768128 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043911735205616855, + "loss": 1.492, + "theoretical_loss": 4.007431607899375, + "tokens_seen": 430833664 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004391073219658977, + "loss": 1.5299, + "theoretical_loss": 4.007364742096938, + "tokens_seen": 430899200 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004390972918756269, + "loss": 1.624, + "theoretical_loss": 4.007297889310455, + "tokens_seen": 430964736 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004390872617853561, + "loss": 1.5395, + "theoretical_loss": 4.007231049535413, + "tokens_seen": 431030272 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004390772316950853, + "loss": 1.5927, + "theoretical_loss": 4.007164222767304, + "tokens_seen": 431095808 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043906720160481446, + "loss": 1.4842, + "theoretical_loss": 4.0070974090016165, + "tokens_seen": 431161344 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043905717151454364, + "loss": 1.5905, + "theoretical_loss": 4.007030608233848, + "tokens_seen": 431226880 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004390471414242728, + "loss": 1.4664, + "theoretical_loss": 4.006963820459494, + "tokens_seen": 431292416 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043903711133400205, + "loss": 1.4566, + "theoretical_loss": 4.006897045674053, + "tokens_seen": 431357952 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004390270812437312, + "loss": 1.5366, + "theoretical_loss": 4.006830283873027, + "tokens_seen": 431423488 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004390170511534604, + "loss": 1.446, + "theoretical_loss": 4.006763535051918, + "tokens_seen": 431489024 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004390070210631896, + "loss": 1.5198, + "theoretical_loss": 4.006696799206233, + "tokens_seen": 431554560 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004389969909729188, + "loss": 1.5922, + "theoretical_loss": 4.006630076331478, + "tokens_seen": 431620096 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043898696088264796, + "loss": 1.6065, + "theoretical_loss": 4.0065633664231655, + "tokens_seen": 431685632 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043897693079237714, + "loss": 1.5409, + "theoretical_loss": 4.006496669476805, + "tokens_seen": 431751168 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004389669007021063, + "loss": 1.7124, + "theoretical_loss": 4.006429985487913, + "tokens_seen": 431816704 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043895687061183556, + "loss": 1.4823, + "theoretical_loss": 4.006363314452007, + "tokens_seen": 431882240 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004389468405215647, + "loss": 1.5016, + "theoretical_loss": 4.006296656364603, + "tokens_seen": 431947776 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004389368104312939, + "loss": 1.5357, + "theoretical_loss": 4.006230011221225, + "tokens_seen": 432013312 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043892678034102305, + "loss": 1.5577, + "theoretical_loss": 4.006163379017394, + "tokens_seen": 432078848 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004389167502507523, + "loss": 1.5036, + "theoretical_loss": 4.0060967597486385, + "tokens_seen": 432144384 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043890672016048146, + "loss": 1.6047, + "theoretical_loss": 4.006030153410483, + "tokens_seen": 432209920 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043889669007021064, + "loss": 1.6109, + "theoretical_loss": 4.005963559998461, + "tokens_seen": 432275456 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004388866599799398, + "loss": 1.5974, + "theoretical_loss": 4.005896979508103, + "tokens_seen": 432340992 + }, + { + "epoch": 0.13, + "learning_rate": 0.000438876629889669, + "loss": 1.6126, + "theoretical_loss": 4.005830411934944, + "tokens_seen": 432406528 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004388665997993982, + "loss": 1.5858, + "theoretical_loss": 4.005763857274521, + "tokens_seen": 432472064 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 166832, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5915557146072388, + "objective/train/theoretical_loss": 4.005697315522371, + "objective/train/tokens_used": 432537600, + "theoretical_loss": 4.005697315522371, + "tokens_seen": 432537600 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004388565697091274, + "loss": 1.5765, + "theoretical_loss": 4.005697315522371, + "tokens_seen": 432537600 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043884653961885655, + "loss": 1.5719, + "theoretical_loss": 4.005630786674039, + "tokens_seen": 432603136 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004388365095285858, + "loss": 1.5999, + "theoretical_loss": 4.005564270725065, + "tokens_seen": 432668672 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043882647943831497, + "loss": 1.6003, + "theoretical_loss": 4.005497767670996, + "tokens_seen": 432734208 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043881644934804415, + "loss": 1.5001, + "theoretical_loss": 4.0054312775073795, + "tokens_seen": 432799744 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043880641925777333, + "loss": 1.5553, + "theoretical_loss": 4.005364800229766, + "tokens_seen": 432865280 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004387963891675025, + "loss": 1.5562, + "theoretical_loss": 4.005298335833708, + "tokens_seen": 432930816 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004387863590772317, + "loss": 1.6357, + "theoretical_loss": 4.005231884314759, + "tokens_seen": 432996352 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004387763289869609, + "loss": 1.6566, + "theoretical_loss": 4.005165445668477, + "tokens_seen": 433061888 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043876629889669005, + "loss": 1.4728, + "theoretical_loss": 4.0050990198904195, + "tokens_seen": 433127424 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004387562688064193, + "loss": 1.57, + "theoretical_loss": 4.005032606976149, + "tokens_seen": 433192960 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004387462387161484, + "loss": 1.6447, + "theoretical_loss": 4.004966206921227, + "tokens_seen": 433258496 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043873620862587765, + "loss": 1.6026, + "theoretical_loss": 4.004899819721221, + "tokens_seen": 433324032 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043872617853560683, + "loss": 1.5014, + "theoretical_loss": 4.004833445371697, + "tokens_seen": 433389568 + }, + { + "epoch": 0.13, + "learning_rate": 0.000438716148445336, + "loss": 1.6402, + "theoretical_loss": 4.004767083868226, + "tokens_seen": 433455104 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004387061183550652, + "loss": 1.6184, + "theoretical_loss": 4.004700735206381, + "tokens_seen": 433520640 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043869608826479443, + "loss": 1.6575, + "theoretical_loss": 4.004634399381733, + "tokens_seen": 433586176 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043868605817452356, + "loss": 1.596, + "theoretical_loss": 4.0045680763898615, + "tokens_seen": 433651712 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004386760280842528, + "loss": 1.5461, + "theoretical_loss": 4.004501766226344, + "tokens_seen": 433717248 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004386659979939819, + "loss": 1.5541, + "theoretical_loss": 4.004435468886761, + "tokens_seen": 433782784 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043865596790371115, + "loss": 1.4359, + "theoretical_loss": 4.0043691843666975, + "tokens_seen": 433848320 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043864593781344033, + "loss": 1.5921, + "theoretical_loss": 4.004302912661736, + "tokens_seen": 433913856 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004386359077231695, + "loss": 1.5826, + "theoretical_loss": 4.0042366537674665, + "tokens_seen": 433979392 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004386258776328987, + "loss": 1.6277, + "theoretical_loss": 4.004170407679476, + "tokens_seen": 434044928 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004386158475426279, + "loss": 1.4842, + "theoretical_loss": 4.004104174393359, + "tokens_seen": 434110464 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043860581745235706, + "loss": 1.4982, + "theoretical_loss": 4.004037953904708, + "tokens_seen": 434176000 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004385957873620863, + "loss": 1.5666, + "theoretical_loss": 4.003971746209119, + "tokens_seen": 434241536 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004385857572718154, + "loss": 1.4669, + "theoretical_loss": 4.003905551302191, + "tokens_seen": 434307072 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043857572718154466, + "loss": 1.5765, + "theoretical_loss": 4.0038393691795235, + "tokens_seen": 434372608 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004385656970912738, + "loss": 1.5262, + "theoretical_loss": 4.003773199836721, + "tokens_seen": 434438144 + }, + { + "epoch": 0.13, + "learning_rate": 0.000438555667001003, + "loss": 1.5616, + "theoretical_loss": 4.003707043269387, + "tokens_seen": 434503680 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004385456369107322, + "loss": 1.6307, + "theoretical_loss": 4.003640899473128, + "tokens_seen": 434569216 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004385356068204614, + "loss": 1.5174, + "theoretical_loss": 4.0035747684435545, + "tokens_seen": 434634752 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043852557673019056, + "loss": 1.5048, + "theoretical_loss": 4.003508650176278, + "tokens_seen": 434700288 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004385155466399198, + "loss": 1.6624, + "theoretical_loss": 4.003442544666911, + "tokens_seen": 434765824 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004385055165496489, + "loss": 1.6147, + "theoretical_loss": 4.003376451911071, + "tokens_seen": 434831360 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043849548645937816, + "loss": 1.6458, + "theoretical_loss": 4.003310371904373, + "tokens_seen": 434896896 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004384854563691073, + "loss": 1.6342, + "theoretical_loss": 4.00324430464244, + "tokens_seen": 434962432 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004384754262788365, + "loss": 1.4318, + "theoretical_loss": 4.003178250120892, + "tokens_seen": 435027968 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004384653961885657, + "loss": 1.4586, + "theoretical_loss": 4.003112208335355, + "tokens_seen": 435093504 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004384553660982949, + "loss": 1.4562, + "theoretical_loss": 4.003046179281455, + "tokens_seen": 435159040 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043844533600802407, + "loss": 1.6707, + "theoretical_loss": 4.00298016295482, + "tokens_seen": 435224576 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043843530591775325, + "loss": 1.5934, + "theoretical_loss": 4.002914159351082, + "tokens_seen": 435290112 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004384252758274825, + "loss": 1.5877, + "theoretical_loss": 4.0028481684658725, + "tokens_seen": 435355648 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043841524573721166, + "loss": 1.4928, + "theoretical_loss": 4.002782190294829, + "tokens_seen": 435421184 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043840521564694084, + "loss": 1.6188, + "theoretical_loss": 4.002716224833588, + "tokens_seen": 435486720 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043839518555667, + "loss": 1.566, + "theoretical_loss": 4.002650272077787, + "tokens_seen": 435552256 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004383851554663992, + "loss": 1.5512, + "theoretical_loss": 4.00258433202307, + "tokens_seen": 435617792 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004383751253761284, + "loss": 1.6206, + "theoretical_loss": 4.0025184046650795, + "tokens_seen": 435683328 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004383650952858576, + "loss": 1.5542, + "theoretical_loss": 4.002452489999463, + "tokens_seen": 435748864 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 167794, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.584423303604126, + "objective/train/theoretical_loss": 4.002386588021867, + "objective/train/tokens_used": 435814400, + "theoretical_loss": 4.002386588021867, + "tokens_seen": 435814400 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043835506519558675, + "loss": 1.6155, + "theoretical_loss": 4.002386588021867, + "tokens_seen": 435814400 + }, + { + "epoch": 0.13, + "learning_rate": 0.000438345035105316, + "loss": 1.4836, + "theoretical_loss": 4.002320698727942, + "tokens_seen": 435879936 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043833500501504517, + "loss": 1.5297, + "theoretical_loss": 4.002254822113341, + "tokens_seen": 435945472 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043832497492477435, + "loss": 1.5178, + "theoretical_loss": 4.0021889581737184, + "tokens_seen": 436011008 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043831494483450353, + "loss": 1.5809, + "theoretical_loss": 4.00212310690473, + "tokens_seen": 436076544 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004383049147442327, + "loss": 1.5047, + "theoretical_loss": 4.0020572683020355, + "tokens_seen": 436142080 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004382948846539619, + "loss": 1.5687, + "theoretical_loss": 4.0019914423612954, + "tokens_seen": 436207616 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004382848545636911, + "loss": 1.4473, + "theoretical_loss": 4.0019256290781735, + "tokens_seen": 436273152 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043827482447342025, + "loss": 1.6222, + "theoretical_loss": 4.001859828448334, + "tokens_seen": 436338688 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004382647943831495, + "loss": 1.4997, + "theoretical_loss": 4.001794040467446, + "tokens_seen": 436404224 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004382547642928786, + "loss": 1.6695, + "theoretical_loss": 4.0017282651311765, + "tokens_seen": 436469760 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043824473420260785, + "loss": 1.4816, + "theoretical_loss": 4.001662502435199, + "tokens_seen": 436535296 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043823470411233703, + "loss": 1.4823, + "theoretical_loss": 4.001596752375186, + "tokens_seen": 436600832 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004382246740220662, + "loss": 1.578, + "theoretical_loss": 4.001531014946815, + "tokens_seen": 436666368 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004382146439317954, + "loss": 1.6041, + "theoretical_loss": 4.001465290145763, + "tokens_seen": 436731904 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043820461384152463, + "loss": 1.4926, + "theoretical_loss": 4.00139957796771, + "tokens_seen": 436797440 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043819458375125376, + "loss": 1.4945, + "theoretical_loss": 4.001333878408339, + "tokens_seen": 436862976 + }, + { + "epoch": 0.13, + "learning_rate": 0.000438184553660983, + "loss": 1.5304, + "theoretical_loss": 4.001268191463334, + "tokens_seen": 436928512 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004381745235707121, + "loss": 1.5295, + "theoretical_loss": 4.00120251712838, + "tokens_seen": 436994048 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043816449348044135, + "loss": 1.5514, + "theoretical_loss": 4.001136855399168, + "tokens_seen": 437059584 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043815446339017053, + "loss": 1.6009, + "theoretical_loss": 4.001071206271388, + "tokens_seen": 437125120 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004381444332998997, + "loss": 1.5768, + "theoretical_loss": 4.001005569740732, + "tokens_seen": 437190656 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004381344032096289, + "loss": 1.6409, + "theoretical_loss": 4.000939945802896, + "tokens_seen": 437256192 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004381243731193581, + "loss": 1.5089, + "theoretical_loss": 4.000874334453577, + "tokens_seen": 437321728 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043811434302908726, + "loss": 1.4858, + "theoretical_loss": 4.000808735688474, + "tokens_seen": 437387264 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004381043129388165, + "loss": 1.497, + "theoretical_loss": 4.000743149503288, + "tokens_seen": 437452800 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004380942828485456, + "loss": 1.5085, + "theoretical_loss": 4.000677575893723, + "tokens_seen": 437518336 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043808425275827486, + "loss": 1.5386, + "theoretical_loss": 4.000612014855484, + "tokens_seen": 437583872 + }, + { + "epoch": 0.13, + "learning_rate": 0.000438074222668004, + "loss": 1.5464, + "theoretical_loss": 4.00054646638428, + "tokens_seen": 437649408 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004380641925777332, + "loss": 1.5465, + "theoretical_loss": 4.000480930475819, + "tokens_seen": 437714944 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004380541624874624, + "loss": 1.4773, + "theoretical_loss": 4.000415407125814, + "tokens_seen": 437780480 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004380441323971916, + "loss": 1.5537, + "theoretical_loss": 4.000349896329979, + "tokens_seen": 437846016 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043803410230692076, + "loss": 1.5254, + "theoretical_loss": 4.0002843980840295, + "tokens_seen": 437911552 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043802407221665, + "loss": 1.5703, + "theoretical_loss": 4.0002189123836835, + "tokens_seen": 437977088 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004380140421263791, + "loss": 1.4621, + "theoretical_loss": 4.000153439224662, + "tokens_seen": 438042624 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043800401203610836, + "loss": 1.4991, + "theoretical_loss": 4.0000879786026875, + "tokens_seen": 438108160 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004379939819458375, + "loss": 1.5752, + "theoretical_loss": 4.000022530513484, + "tokens_seen": 438173696 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004379839518555667, + "loss": 1.518, + "theoretical_loss": 3.999957094952779, + "tokens_seen": 438239232 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004379739217652959, + "loss": 1.4712, + "theoretical_loss": 3.9998916719163002, + "tokens_seen": 438304768 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004379638916750251, + "loss": 1.5442, + "theoretical_loss": 3.999826261399778, + "tokens_seen": 438370304 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043795386158475427, + "loss": 1.4954, + "theoretical_loss": 3.9997608633989463, + "tokens_seen": 438435840 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043794383149448345, + "loss": 1.5394, + "theoretical_loss": 3.9996954779095395, + "tokens_seen": 438501376 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043793380140421263, + "loss": 1.5452, + "theoretical_loss": 3.999630104927295, + "tokens_seen": 438566912 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043792377131394186, + "loss": 1.4269, + "theoretical_loss": 3.999564744447952, + "tokens_seen": 438632448 + }, + { + "epoch": 0.13, + "learning_rate": 0.000437913741223671, + "loss": 1.4628, + "theoretical_loss": 3.9994993964672516, + "tokens_seen": 438697984 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004379037111334002, + "loss": 1.5229, + "theoretical_loss": 3.9994340609809367, + "tokens_seen": 438763520 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043789368104312935, + "loss": 1.5135, + "theoretical_loss": 3.999368737984754, + "tokens_seen": 438829056 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004378836509528586, + "loss": 1.5254, + "theoretical_loss": 3.999303427474449, + "tokens_seen": 438894592 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043787362086258777, + "loss": 1.4771, + "theoretical_loss": 3.9992381294457733, + "tokens_seen": 438960128 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043786359077231695, + "loss": 1.591, + "theoretical_loss": 3.9991728438944767, + "tokens_seen": 439025664 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 168939, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6142783164978027, + "objective/train/theoretical_loss": 3.9991075708163146, + "objective/train/tokens_used": 439091200, + "theoretical_loss": 3.9991075708163146, + "tokens_seen": 439091200 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043785356068204613, + "loss": 1.602, + "theoretical_loss": 3.9991075708163146, + "tokens_seen": 439091200 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043784353059177537, + "loss": 1.5232, + "theoretical_loss": 3.9990423102070416, + "tokens_seen": 439156736 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004378335005015045, + "loss": 1.6132, + "theoretical_loss": 3.9989770620624165, + "tokens_seen": 439222272 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043782347041123373, + "loss": 1.5808, + "theoretical_loss": 3.9989118263781984, + "tokens_seen": 439287808 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043781344032096286, + "loss": 1.5032, + "theoretical_loss": 3.99884660315015, + "tokens_seen": 439353344 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004378034102306921, + "loss": 1.5632, + "theoretical_loss": 3.9987813923740356, + "tokens_seen": 439418880 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043779338014042127, + "loss": 1.5082, + "theoretical_loss": 3.9987161940456204, + "tokens_seen": 439484416 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043778335005015045, + "loss": 1.5853, + "theoretical_loss": 3.9986510081606736, + "tokens_seen": 439549952 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043777331995987963, + "loss": 1.4512, + "theoretical_loss": 3.998585834714965, + "tokens_seen": 439615488 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004377632898696088, + "loss": 1.4903, + "theoretical_loss": 3.9985206737042676, + "tokens_seen": 439681024 + }, + { + "epoch": 0.13, + "learning_rate": 0.000437753259779338, + "loss": 1.4373, + "theoretical_loss": 3.998455525124355, + "tokens_seen": 439746560 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043774322968906723, + "loss": 1.4967, + "theoretical_loss": 3.9983903889710044, + "tokens_seen": 439812096 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043773319959879636, + "loss": 1.4643, + "theoretical_loss": 3.998325265239994, + "tokens_seen": 439877632 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004377231695085256, + "loss": 1.5337, + "theoretical_loss": 3.998260153927105, + "tokens_seen": 439943168 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004377131394182547, + "loss": 1.4718, + "theoretical_loss": 3.998195055028119, + "tokens_seen": 440008704 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043770310932798396, + "loss": 1.5006, + "theoretical_loss": 3.998129968538822, + "tokens_seen": 440074240 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043769307923771314, + "loss": 1.447, + "theoretical_loss": 3.998064894455, + "tokens_seen": 440139776 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004376830491474423, + "loss": 1.5327, + "theoretical_loss": 3.997999832772443, + "tokens_seen": 440205312 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043767301905717155, + "loss": 1.608, + "theoretical_loss": 3.997934783486941, + "tokens_seen": 440270848 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043766298896690073, + "loss": 1.4861, + "theoretical_loss": 3.997869746594287, + "tokens_seen": 440336384 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004376529588766299, + "loss": 1.5147, + "theoretical_loss": 3.9978047220902764, + "tokens_seen": 440401920 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004376429287863591, + "loss": 1.533, + "theoretical_loss": 3.997739709970706, + "tokens_seen": 440467456 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004376328986960883, + "loss": 1.5573, + "theoretical_loss": 3.997674710231375, + "tokens_seen": 440532992 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043762286860581746, + "loss": 1.5724, + "theoretical_loss": 3.997609722868085, + "tokens_seen": 440598528 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004376128385155467, + "loss": 1.5542, + "theoretical_loss": 3.9975447478766393, + "tokens_seen": 440664064 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004376028084252758, + "loss": 1.7049, + "theoretical_loss": 3.997479785252842, + "tokens_seen": 440729600 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043759277833500506, + "loss": 1.4798, + "theoretical_loss": 3.997414834992502, + "tokens_seen": 440795136 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004375827482447342, + "loss": 1.5767, + "theoretical_loss": 3.9973498970914285, + "tokens_seen": 440860672 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004375727181544634, + "loss": 1.4825, + "theoretical_loss": 3.9972849715454317, + "tokens_seen": 440926208 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004375626880641926, + "loss": 1.5303, + "theoretical_loss": 3.9972200583503263, + "tokens_seen": 440991744 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004375526579739218, + "loss": 1.4483, + "theoretical_loss": 3.9971551575019273, + "tokens_seen": 441057280 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043754262788365096, + "loss": 1.5159, + "theoretical_loss": 3.997090268996052, + "tokens_seen": 441122816 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004375325977933802, + "loss": 1.4815, + "theoretical_loss": 3.9970253928285215, + "tokens_seen": 441188352 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004375225677031093, + "loss": 1.4571, + "theoretical_loss": 3.996960528995155, + "tokens_seen": 441253888 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043751253761283856, + "loss": 1.5067, + "theoretical_loss": 3.996895677491778, + "tokens_seen": 441319424 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004375025075225677, + "loss": 1.678, + "theoretical_loss": 3.996830838314216, + "tokens_seen": 441384960 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004374924774322969, + "loss": 1.4804, + "theoretical_loss": 3.996766011458296, + "tokens_seen": 441450496 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004374824473420261, + "loss": 1.5083, + "theoretical_loss": 3.9967011969198483, + "tokens_seen": 441516032 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004374724172517553, + "loss": 1.5546, + "theoretical_loss": 3.996636394694705, + "tokens_seen": 441581568 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043746238716148447, + "loss": 1.4766, + "theoretical_loss": 3.9965716047786985, + "tokens_seen": 441647104 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043745235707121365, + "loss": 1.6796, + "theoretical_loss": 3.996506827167667, + "tokens_seen": 441712640 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043744232698094283, + "loss": 1.4508, + "theoretical_loss": 3.9964420618574454, + "tokens_seen": 441778176 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043743229689067206, + "loss": 1.5177, + "theoretical_loss": 3.9963773088438765, + "tokens_seen": 441843712 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004374222668004012, + "loss": 1.5927, + "theoretical_loss": 3.996312568122801, + "tokens_seen": 441909248 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004374122367101304, + "loss": 1.4314, + "theoretical_loss": 3.9962478396900627, + "tokens_seen": 441974784 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043740220661985955, + "loss": 1.4877, + "theoretical_loss": 3.996183123541507, + "tokens_seen": 442040320 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004373921765295888, + "loss": 1.6468, + "theoretical_loss": 3.9961184196729835, + "tokens_seen": 442105856 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043738214643931797, + "loss": 1.5043, + "theoretical_loss": 3.9960537280803408, + "tokens_seen": 442171392 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043737211634904715, + "loss": 1.566, + "theoretical_loss": 3.995989048759432, + "tokens_seen": 442236928 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043736208625877633, + "loss": 1.614, + "theoretical_loss": 3.99592438170611, + "tokens_seen": 442302464 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 170173, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4622962474822998, + "objective/train/theoretical_loss": 3.9958597269162315, + "objective/train/tokens_used": 442368000, + "theoretical_loss": 3.9958597269162315, + "tokens_seen": 442368000 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043735205616850557, + "loss": 1.4305, + "theoretical_loss": 3.9958597269162315, + "tokens_seen": 442368000 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004373420260782347, + "loss": 1.5894, + "theoretical_loss": 3.9957950843856542, + "tokens_seen": 442433536 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043733199598796393, + "loss": 1.4558, + "theoretical_loss": 3.9957304541102396, + "tokens_seen": 442499072 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043732196589769306, + "loss": 1.4384, + "theoretical_loss": 3.995665836085848, + "tokens_seen": 442564608 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004373119358074223, + "loss": 1.5168, + "theoretical_loss": 3.9956012303083446, + "tokens_seen": 442630144 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043730190571715147, + "loss": 1.4664, + "theoretical_loss": 3.9955366367735947, + "tokens_seen": 442695680 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043729187562688065, + "loss": 1.4873, + "theoretical_loss": 3.995472055477467, + "tokens_seen": 442761216 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043728184553660983, + "loss": 1.5631, + "theoretical_loss": 3.9954074864158313, + "tokens_seen": 442826752 + }, + { + "epoch": 0.13, + "learning_rate": 0.000437271815446339, + "loss": 1.4863, + "theoretical_loss": 3.9953429295845604, + "tokens_seen": 442892288 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004372617853560682, + "loss": 1.5604, + "theoretical_loss": 3.9952783849795277, + "tokens_seen": 442957824 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043725175526579743, + "loss": 1.5181, + "theoretical_loss": 3.99521385259661, + "tokens_seen": 443023360 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043724172517552656, + "loss": 1.4915, + "theoretical_loss": 3.9951493324316845, + "tokens_seen": 443088896 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004372316950852558, + "loss": 1.5225, + "theoretical_loss": 3.9950848244806316, + "tokens_seen": 443154432 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004372216649949849, + "loss": 1.5074, + "theoretical_loss": 3.9950203287393347, + "tokens_seen": 443219968 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043721163490471416, + "loss": 1.5054, + "theoretical_loss": 3.994955845203676, + "tokens_seen": 443285504 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043720160481444334, + "loss": 1.4833, + "theoretical_loss": 3.994891373869543, + "tokens_seen": 443351040 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004371915747241725, + "loss": 1.5328, + "theoretical_loss": 3.9948269147328235, + "tokens_seen": 443416576 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004371815446339017, + "loss": 1.4666, + "theoretical_loss": 3.994762467789407, + "tokens_seen": 443482112 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043717151454363093, + "loss": 1.603, + "theoretical_loss": 3.9946980330351862, + "tokens_seen": 443547648 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043716148445336006, + "loss": 1.6262, + "theoretical_loss": 3.994633610466056, + "tokens_seen": 443613184 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004371514543630893, + "loss": 1.5235, + "theoretical_loss": 3.994569200077911, + "tokens_seen": 443678720 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004371414242728184, + "loss": 1.6226, + "theoretical_loss": 3.9945048018666496, + "tokens_seen": 443744256 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043713139418254766, + "loss": 1.5562, + "theoretical_loss": 3.9944404158281723, + "tokens_seen": 443809792 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043712136409227684, + "loss": 1.5157, + "theoretical_loss": 3.994376041958381, + "tokens_seen": 443875328 + }, + { + "epoch": 0.13, + "learning_rate": 0.000437111334002006, + "loss": 1.4606, + "theoretical_loss": 3.9943116802531797, + "tokens_seen": 443940864 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004371013039117352, + "loss": 1.448, + "theoretical_loss": 3.9942473307084745, + "tokens_seen": 444006400 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004370912738214644, + "loss": 1.5632, + "theoretical_loss": 3.9941829933201736, + "tokens_seen": 444071936 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043708124373119356, + "loss": 1.5335, + "theoretical_loss": 3.9941186680841865, + "tokens_seen": 444137472 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004370712136409228, + "loss": 1.5888, + "theoretical_loss": 3.9940543549964254, + "tokens_seen": 444203008 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004370611835506519, + "loss": 1.5417, + "theoretical_loss": 3.993990054052804, + "tokens_seen": 444268544 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043705115346038116, + "loss": 1.4698, + "theoretical_loss": 3.993925765249239, + "tokens_seen": 444334080 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004370411233701103, + "loss": 1.5104, + "theoretical_loss": 3.993861488581647, + "tokens_seen": 444399616 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004370310932798395, + "loss": 1.5163, + "theoretical_loss": 3.9937972240459487, + "tokens_seen": 444465152 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004370210631895687, + "loss": 1.5066, + "theoretical_loss": 3.9937329716380665, + "tokens_seen": 444530688 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004370110330992979, + "loss": 1.5038, + "theoretical_loss": 3.9936687313539228, + "tokens_seen": 444596224 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043700100300902707, + "loss": 1.5083, + "theoretical_loss": 3.9936045031894443, + "tokens_seen": 444661760 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004369909729187563, + "loss": 1.4999, + "theoretical_loss": 3.9935402871405588, + "tokens_seen": 444727296 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043698094282848543, + "loss": 1.5058, + "theoretical_loss": 3.9934760832031957, + "tokens_seen": 444792832 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043697091273821467, + "loss": 1.5778, + "theoretical_loss": 3.9934118913732863, + "tokens_seen": 444858368 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004369608826479438, + "loss": 1.5554, + "theoretical_loss": 3.9933477116467646, + "tokens_seen": 444923904 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043695085255767303, + "loss": 1.4906, + "theoretical_loss": 3.993283544019567, + "tokens_seen": 444989440 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004369408224674022, + "loss": 1.6085, + "theoretical_loss": 3.9932193884876304, + "tokens_seen": 445054976 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004369307923771314, + "loss": 1.4713, + "theoretical_loss": 3.9931552450468937, + "tokens_seen": 445120512 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004369207622868606, + "loss": 1.5638, + "theoretical_loss": 3.9930911136933, + "tokens_seen": 445186048 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043691073219658975, + "loss": 1.5507, + "theoretical_loss": 3.9930269944227907, + "tokens_seen": 445251584 + }, + { + "epoch": 0.13, + "learning_rate": 0.000436900702106319, + "loss": 1.5938, + "theoretical_loss": 3.9929628872313128, + "tokens_seen": 445317120 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043689067201604817, + "loss": 1.5322, + "theoretical_loss": 3.9928987921148127, + "tokens_seen": 445382656 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043688064192577735, + "loss": 1.6046, + "theoretical_loss": 3.992834709069241, + "tokens_seen": 445448192 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043687061183550653, + "loss": 1.5722, + "theoretical_loss": 3.9927706380905477, + "tokens_seen": 445513728 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043686058174523577, + "loss": 1.5426, + "theoretical_loss": 3.992706579174686, + "tokens_seen": 445579264 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 171375, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6765375137329102, + "objective/train/theoretical_loss": 3.9926425323176122, + "objective/train/tokens_used": 445644800, + "theoretical_loss": 3.9926425323176122, + "tokens_seen": 445644800 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004368505516549649, + "loss": 1.6168, + "theoretical_loss": 3.9926425323176122, + "tokens_seen": 445644800 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043684052156469413, + "loss": 1.4679, + "theoretical_loss": 3.992578497515283, + "tokens_seen": 445710336 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043683049147442326, + "loss": 1.3797, + "theoretical_loss": 3.9925144747636576, + "tokens_seen": 445775872 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004368204613841525, + "loss": 1.5488, + "theoretical_loss": 3.992450464058696, + "tokens_seen": 445841408 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043681043129388167, + "loss": 1.5574, + "theoretical_loss": 3.9923864653963625, + "tokens_seen": 445906944 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043680040120361085, + "loss": 1.4636, + "theoretical_loss": 3.9923224787726213, + "tokens_seen": 445972480 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043679037111334003, + "loss": 1.5433, + "theoretical_loss": 3.9922585041834395, + "tokens_seen": 446038016 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004367803410230692, + "loss": 1.4976, + "theoretical_loss": 3.9921945416247855, + "tokens_seen": 446103552 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004367703109327984, + "loss": 1.6106, + "theoretical_loss": 3.9921305910926312, + "tokens_seen": 446169088 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043676028084252763, + "loss": 1.5436, + "theoretical_loss": 3.992066652582948, + "tokens_seen": 446234624 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043675025075225676, + "loss": 1.52, + "theoretical_loss": 3.9920027260917115, + "tokens_seen": 446300160 + }, + { + "epoch": 0.14, + "learning_rate": 0.000436740220661986, + "loss": 1.4629, + "theoretical_loss": 3.9919388116148977, + "tokens_seen": 446365696 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004367301905717151, + "loss": 1.6331, + "theoretical_loss": 3.991874909148485, + "tokens_seen": 446431232 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043672016048144436, + "loss": 1.4695, + "theoretical_loss": 3.9918110186884546, + "tokens_seen": 446496768 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043671013039117354, + "loss": 1.7002, + "theoretical_loss": 3.991747140230788, + "tokens_seen": 446562304 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004367001003009027, + "loss": 1.6383, + "theoretical_loss": 3.9916832737714705, + "tokens_seen": 446627840 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004366900702106319, + "loss": 1.5104, + "theoretical_loss": 3.9916194193064873, + "tokens_seen": 446693376 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043668004012036113, + "loss": 1.57, + "theoretical_loss": 3.9915555768318276, + "tokens_seen": 446758912 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043667001003009026, + "loss": 1.5965, + "theoretical_loss": 3.9914917463434807, + "tokens_seen": 446824448 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004366599799398195, + "loss": 1.5636, + "theoretical_loss": 3.9914279278374387, + "tokens_seen": 446889984 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004366499498495486, + "loss": 1.4783, + "theoretical_loss": 3.9913641213096964, + "tokens_seen": 446955520 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043663991975927786, + "loss": 1.5578, + "theoretical_loss": 3.9913003267562486, + "tokens_seen": 447021056 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043662988966900704, + "loss": 1.5148, + "theoretical_loss": 3.9912365441730944, + "tokens_seen": 447086592 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004366198595787362, + "loss": 1.6083, + "theoretical_loss": 3.9911727735562326, + "tokens_seen": 447152128 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004366098294884654, + "loss": 1.5788, + "theoretical_loss": 3.9911090149016646, + "tokens_seen": 447217664 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004365997993981946, + "loss": 1.6343, + "theoretical_loss": 3.9910452682053954, + "tokens_seen": 447283200 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043658976930792376, + "loss": 1.5453, + "theoretical_loss": 3.9909815334634287, + "tokens_seen": 447348736 + }, + { + "epoch": 0.14, + "learning_rate": 0.000436579739217653, + "loss": 1.5203, + "theoretical_loss": 3.990917810671773, + "tokens_seen": 447414272 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043656970912738213, + "loss": 1.5214, + "theoretical_loss": 3.9908540998264375, + "tokens_seen": 447479808 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043655967903711136, + "loss": 1.5296, + "theoretical_loss": 3.9907904009234336, + "tokens_seen": 447545344 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004365496489468405, + "loss": 1.5334, + "theoretical_loss": 3.9907267139587743, + "tokens_seen": 447610880 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004365396188565697, + "loss": 1.6926, + "theoretical_loss": 3.9906630389284747, + "tokens_seen": 447676416 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004365295887662989, + "loss": 1.4845, + "theoretical_loss": 3.990599375828552, + "tokens_seen": 447741952 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004365195586760281, + "loss": 1.4999, + "theoretical_loss": 3.9905357246550253, + "tokens_seen": 447807488 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043650952858575727, + "loss": 1.5069, + "theoretical_loss": 3.9904720854039137, + "tokens_seen": 447873024 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004364994984954865, + "loss": 1.4281, + "theoretical_loss": 3.9904084580712427, + "tokens_seen": 447938560 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043648946840521563, + "loss": 1.5569, + "theoretical_loss": 3.9903448426530352, + "tokens_seen": 448004096 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043647943831494487, + "loss": 1.4059, + "theoretical_loss": 3.990281239145318, + "tokens_seen": 448069632 + }, + { + "epoch": 0.14, + "learning_rate": 0.000436469408224674, + "loss": 1.5426, + "theoretical_loss": 3.99021764754412, + "tokens_seen": 448135168 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043645937813440323, + "loss": 1.508, + "theoretical_loss": 3.990154067845472, + "tokens_seen": 448200704 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004364493480441324, + "loss": 1.4906, + "theoretical_loss": 3.990090500045404, + "tokens_seen": 448266240 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004364393179538616, + "loss": 1.4853, + "theoretical_loss": 3.990026944139953, + "tokens_seen": 448331776 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043642928786359077, + "loss": 1.4656, + "theoretical_loss": 3.989963400125153, + "tokens_seen": 448397312 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043641925777331995, + "loss": 1.4932, + "theoretical_loss": 3.9898998679970434, + "tokens_seen": 448462848 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043640922768304913, + "loss": 1.6376, + "theoretical_loss": 3.9898363477516634, + "tokens_seen": 448528384 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043639919759277837, + "loss": 1.4732, + "theoretical_loss": 3.9897728393850547, + "tokens_seen": 448593920 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004363891675025075, + "loss": 1.5454, + "theoretical_loss": 3.9897093428932613, + "tokens_seen": 448659456 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043637913741223673, + "loss": 1.5007, + "theoretical_loss": 3.989645858272328, + "tokens_seen": 448724992 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004363691073219659, + "loss": 1.483, + "theoretical_loss": 3.989582385518304, + "tokens_seen": 448790528 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004363590772316951, + "loss": 1.5906, + "theoretical_loss": 3.989518924627237, + "tokens_seen": 448856064 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 172451, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6724610328674316, + "objective/train/theoretical_loss": 3.9894554755951788, + "objective/train/tokens_used": 448921600, + "theoretical_loss": 3.9894554755951788, + "tokens_seen": 448921600 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004363490471414243, + "loss": 1.5445, + "theoretical_loss": 3.9894554755951788, + "tokens_seen": 448921600 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043633901705115346, + "loss": 1.5046, + "theoretical_loss": 3.989392038418182, + "tokens_seen": 448987136 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043632898696088264, + "loss": 1.4957, + "theoretical_loss": 3.9893286130923027, + "tokens_seen": 449052672 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043631895687061187, + "loss": 1.5645, + "theoretical_loss": 3.989265199613597, + "tokens_seen": 449118208 + }, + { + "epoch": 0.14, + "learning_rate": 0.000436308926780341, + "loss": 1.5276, + "theoretical_loss": 3.989201797978124, + "tokens_seen": 449183744 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043629889669007023, + "loss": 1.5545, + "theoretical_loss": 3.989138408181944, + "tokens_seen": 449249280 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043628886659979936, + "loss": 1.434, + "theoretical_loss": 3.9890750302211204, + "tokens_seen": 449314816 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004362788365095286, + "loss": 1.443, + "theoretical_loss": 3.9890116640917173, + "tokens_seen": 449380352 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004362688064192578, + "loss": 1.5339, + "theoretical_loss": 3.9889483097898006, + "tokens_seen": 449445888 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043625877632898696, + "loss": 1.419, + "theoretical_loss": 3.9888849673114386, + "tokens_seen": 449511424 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043624874623871614, + "loss": 1.5253, + "theoretical_loss": 3.988821636652702, + "tokens_seen": 449576960 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004362387161484453, + "loss": 1.5269, + "theoretical_loss": 3.988758317809662, + "tokens_seen": 449642496 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004362286860581745, + "loss": 1.5103, + "theoretical_loss": 3.988695010778393, + "tokens_seen": 449708032 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043621865596790374, + "loss": 1.4415, + "theoretical_loss": 3.9886317155549706, + "tokens_seen": 449773568 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043620862587763286, + "loss": 1.5833, + "theoretical_loss": 3.988568432135472, + "tokens_seen": 449839104 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004361985957873621, + "loss": 1.537, + "theoretical_loss": 3.988505160515978, + "tokens_seen": 449904640 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004361885656970913, + "loss": 1.5886, + "theoretical_loss": 3.9884419006925684, + "tokens_seen": 449970176 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043617853560682046, + "loss": 1.4724, + "theoretical_loss": 3.9883786526613267, + "tokens_seen": 450035712 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004361685055165497, + "loss": 1.6447, + "theoretical_loss": 3.988315416418339, + "tokens_seen": 450101248 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004361584754262788, + "loss": 1.512, + "theoretical_loss": 3.988252191959692, + "tokens_seen": 450166784 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043614844533600806, + "loss": 1.6044, + "theoretical_loss": 3.9881889792814733, + "tokens_seen": 450232320 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043613841524573724, + "loss": 1.5789, + "theoretical_loss": 3.9881257783797746, + "tokens_seen": 450297856 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004361283851554664, + "loss": 1.6341, + "theoretical_loss": 3.9880625892506885, + "tokens_seen": 450363392 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004361183550651956, + "loss": 1.5148, + "theoretical_loss": 3.9879994118903093, + "tokens_seen": 450428928 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004361083249749248, + "loss": 1.4788, + "theoretical_loss": 3.9879362462947334, + "tokens_seen": 450494464 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043609829488465397, + "loss": 1.4587, + "theoretical_loss": 3.9878730924600583, + "tokens_seen": 450560000 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004360882647943832, + "loss": 1.589, + "theoretical_loss": 3.9878099503823847, + "tokens_seen": 450625536 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043607823470411233, + "loss": 1.5458, + "theoretical_loss": 3.9877468200578146, + "tokens_seen": 450691072 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043606820461384156, + "loss": 1.5499, + "theoretical_loss": 3.9876837014824513, + "tokens_seen": 450756608 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004360581745235707, + "loss": 1.5431, + "theoretical_loss": 3.9876205946524, + "tokens_seen": 450822144 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004360481444332999, + "loss": 1.6488, + "theoretical_loss": 3.9875574995637697, + "tokens_seen": 450887680 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004360381143430291, + "loss": 1.504, + "theoretical_loss": 3.987494416212668, + "tokens_seen": 450953216 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004360280842527583, + "loss": 1.5018, + "theoretical_loss": 3.987431344595207, + "tokens_seen": 451018752 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043601805416248747, + "loss": 1.5562, + "theoretical_loss": 3.9873682847074994, + "tokens_seen": 451084288 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004360080240722167, + "loss": 1.6052, + "theoretical_loss": 3.9873052365456605, + "tokens_seen": 451149824 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043599799398194583, + "loss": 1.4546, + "theoretical_loss": 3.9872422001058063, + "tokens_seen": 451215360 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043598796389167507, + "loss": 1.5451, + "theoretical_loss": 3.987179175384056, + "tokens_seen": 451280896 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004359779338014042, + "loss": 1.5117, + "theoretical_loss": 3.9871161623765294, + "tokens_seen": 451346432 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043596790371113343, + "loss": 1.4878, + "theoretical_loss": 3.9870531610793494, + "tokens_seen": 451411968 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004359578736208626, + "loss": 1.5025, + "theoretical_loss": 3.9869901714886398, + "tokens_seen": 451477504 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004359478435305918, + "loss": 1.5418, + "theoretical_loss": 3.9869271936005264, + "tokens_seen": 451543040 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043593781344032097, + "loss": 1.4749, + "theoretical_loss": 3.9868642274111377, + "tokens_seen": 451608576 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043592778335005015, + "loss": 1.5087, + "theoretical_loss": 3.986801272916603, + "tokens_seen": 451674112 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043591775325977933, + "loss": 1.5218, + "theoretical_loss": 3.9867383301130532, + "tokens_seen": 451739648 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043590772316950857, + "loss": 1.5649, + "theoretical_loss": 3.9866753989966224, + "tokens_seen": 451805184 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004358976930792377, + "loss": 1.5204, + "theoretical_loss": 3.986612479563445, + "tokens_seen": 451870720 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043588766298896693, + "loss": 1.528, + "theoretical_loss": 3.9865495718096584, + "tokens_seen": 451936256 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004358776328986961, + "loss": 1.5792, + "theoretical_loss": 3.9864866757314017, + "tokens_seen": 452001792 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004358676028084253, + "loss": 1.4855, + "theoretical_loss": 3.9864237913248157, + "tokens_seen": 452067328 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004358575727181545, + "loss": 1.5538, + "theoretical_loss": 3.9863609185860422, + "tokens_seen": 452132864 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 173483, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.549603819847107, + "objective/train/theoretical_loss": 3.9862980575112257, + "objective/train/tokens_used": 452198400, + "theoretical_loss": 3.9862980575112257, + "tokens_seen": 452198400 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043584754262788366, + "loss": 1.5817, + "theoretical_loss": 3.9862980575112257, + "tokens_seen": 452198400 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043583751253761284, + "loss": 1.5065, + "theoretical_loss": 3.986235208096513, + "tokens_seen": 452263936 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043582748244734207, + "loss": 1.6657, + "theoretical_loss": 3.9861723703380516, + "tokens_seen": 452329472 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004358174523570712, + "loss": 1.5958, + "theoretical_loss": 3.986109544231992, + "tokens_seen": 452395008 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043580742226680043, + "loss": 1.4715, + "theoretical_loss": 3.9860467297744844, + "tokens_seen": 452460544 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043579739217652956, + "loss": 1.5981, + "theoretical_loss": 3.985983926961684, + "tokens_seen": 452526080 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004357873620862588, + "loss": 1.5383, + "theoretical_loss": 3.985921135789745, + "tokens_seen": 452591616 + }, + { + "epoch": 0.14, + "learning_rate": 0.000435777331995988, + "loss": 1.4226, + "theoretical_loss": 3.9858583562548247, + "tokens_seen": 452657152 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043576730190571716, + "loss": 1.5658, + "theoretical_loss": 3.9857955883530827, + "tokens_seen": 452722688 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043575727181544634, + "loss": 1.5127, + "theoretical_loss": 3.985732832080679, + "tokens_seen": 452788224 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004357472417251755, + "loss": 1.3967, + "theoretical_loss": 3.985670087433777, + "tokens_seen": 452853760 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004357372116349047, + "loss": 1.5148, + "theoretical_loss": 3.9856073544085406, + "tokens_seen": 452919296 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043572718154463394, + "loss": 1.4664, + "theoretical_loss": 3.9855446330011364, + "tokens_seen": 452984832 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043571715145436306, + "loss": 1.421, + "theoretical_loss": 3.9854819232077316, + "tokens_seen": 453050368 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004357071213640923, + "loss": 1.4311, + "theoretical_loss": 3.985419225024498, + "tokens_seen": 453115904 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004356970912738215, + "loss": 1.63, + "theoretical_loss": 3.985356538447605, + "tokens_seen": 453181440 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043568706118355066, + "loss": 1.5447, + "theoretical_loss": 3.9852938634732276, + "tokens_seen": 453246976 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043567703109327984, + "loss": 1.5874, + "theoretical_loss": 3.985231200097541, + "tokens_seen": 453312512 + }, + { + "epoch": 0.14, + "learning_rate": 0.000435667001003009, + "loss": 1.4824, + "theoretical_loss": 3.9851685483167225, + "tokens_seen": 453378048 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004356569709127382, + "loss": 1.5362, + "theoretical_loss": 3.9851059081269504, + "tokens_seen": 453443584 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043564694082246744, + "loss": 1.4772, + "theoretical_loss": 3.9850432795244055, + "tokens_seen": 453509120 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043563691073219657, + "loss": 1.5072, + "theoretical_loss": 3.9849806625052713, + "tokens_seen": 453574656 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004356268806419258, + "loss": 1.4174, + "theoretical_loss": 3.984918057065731, + "tokens_seen": 453640192 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043561685055165493, + "loss": 1.6109, + "theoretical_loss": 3.984855463201972, + "tokens_seen": 453705728 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043560682046138417, + "loss": 1.5751, + "theoretical_loss": 3.984792880910182, + "tokens_seen": 453771264 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043559679037111335, + "loss": 1.552, + "theoretical_loss": 3.98473031018655, + "tokens_seen": 453836800 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043558676028084253, + "loss": 1.5684, + "theoretical_loss": 3.984667751027268, + "tokens_seen": 453902336 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004355767301905717, + "loss": 1.4277, + "theoretical_loss": 3.98460520342853, + "tokens_seen": 453967872 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004355667001003009, + "loss": 1.5813, + "theoretical_loss": 3.984542667386531, + "tokens_seen": 454033408 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043555667001003007, + "loss": 1.5327, + "theoretical_loss": 3.9844801428974685, + "tokens_seen": 454098944 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004355466399197593, + "loss": 1.5212, + "theoretical_loss": 3.98441762995754, + "tokens_seen": 454164480 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043553660982948843, + "loss": 1.5342, + "theoretical_loss": 3.984355128562947, + "tokens_seen": 454230016 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043552657973921767, + "loss": 1.5635, + "theoretical_loss": 3.984292638709892, + "tokens_seen": 454295552 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043551654964894685, + "loss": 1.4791, + "theoretical_loss": 3.9842301603945787, + "tokens_seen": 454361088 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043550651955867603, + "loss": 1.4171, + "theoretical_loss": 3.9841676936132133, + "tokens_seen": 454426624 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004354964894684052, + "loss": 1.4737, + "theoretical_loss": 3.984105238362005, + "tokens_seen": 454492160 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004354864593781344, + "loss": 1.5635, + "theoretical_loss": 3.984042794637161, + "tokens_seen": 454557696 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004354764292878636, + "loss": 1.5556, + "theoretical_loss": 3.9839803624348944, + "tokens_seen": 454623232 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004354663991975928, + "loss": 1.4827, + "theoretical_loss": 3.983917941751418, + "tokens_seen": 454688768 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043545636910732194, + "loss": 1.5087, + "theoretical_loss": 3.983855532582947, + "tokens_seen": 454754304 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043544633901705117, + "loss": 1.5537, + "theoretical_loss": 3.983793134925697, + "tokens_seen": 454819840 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004354363089267803, + "loss": 1.4712, + "theoretical_loss": 3.9837307487758884, + "tokens_seen": 454885376 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043542627883650953, + "loss": 1.6322, + "theoretical_loss": 3.9836683741297403, + "tokens_seen": 454950912 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043541624874623877, + "loss": 1.5608, + "theoretical_loss": 3.9836060109834746, + "tokens_seen": 455016448 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004354062186559679, + "loss": 1.5637, + "theoretical_loss": 3.983543659333316, + "tokens_seen": 455081984 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043539618856569713, + "loss": 1.6357, + "theoretical_loss": 3.9834813191754908, + "tokens_seen": 455147520 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004353861584754263, + "loss": 1.5014, + "theoretical_loss": 3.983418990506225, + "tokens_seen": 455213056 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004353761283851555, + "loss": 1.6461, + "theoretical_loss": 3.9833566733217483, + "tokens_seen": 455278592 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004353660982948847, + "loss": 1.5396, + "theoretical_loss": 3.983294367618292, + "tokens_seen": 455344128 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043535606820461386, + "loss": 1.6089, + "theoretical_loss": 3.983232073392089, + "tokens_seen": 455409664 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 174995, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.669136881828308, + "objective/train/theoretical_loss": 3.983169790639374, + "objective/train/tokens_used": 455475200, + "theoretical_loss": 3.983169790639374, + "tokens_seen": 455475200 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043534603811434304, + "loss": 1.553, + "theoretical_loss": 3.983169790639374, + "tokens_seen": 455475200 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043533600802407227, + "loss": 1.5684, + "theoretical_loss": 3.983107519356383, + "tokens_seen": 455540736 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004353259779338014, + "loss": 1.5761, + "theoretical_loss": 3.9830452595393546, + "tokens_seen": 455606272 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043531594784353063, + "loss": 1.6226, + "theoretical_loss": 3.9829830111845284, + "tokens_seen": 455671808 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043530591775325976, + "loss": 1.6003, + "theoretical_loss": 3.9829207742881456, + "tokens_seen": 455737344 + }, + { + "epoch": 0.14, + "learning_rate": 0.000435295887662989, + "loss": 1.6784, + "theoretical_loss": 3.982858548846451, + "tokens_seen": 455802880 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004352858575727182, + "loss": 1.482, + "theoretical_loss": 3.982796334855689, + "tokens_seen": 455868416 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043527582748244736, + "loss": 1.4729, + "theoretical_loss": 3.9827341323121064, + "tokens_seen": 455933952 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043526579739217654, + "loss": 1.4271, + "theoretical_loss": 3.9826719412119527, + "tokens_seen": 455999488 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004352557673019057, + "loss": 1.4985, + "theoretical_loss": 3.982609761551478, + "tokens_seen": 456065024 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004352457372116349, + "loss": 1.5594, + "theoretical_loss": 3.982547593326934, + "tokens_seen": 456130560 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043523570712136414, + "loss": 1.6796, + "theoretical_loss": 3.982485436534576, + "tokens_seen": 456196096 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043522567703109326, + "loss": 1.5871, + "theoretical_loss": 3.98242329117066, + "tokens_seen": 456261632 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004352156469408225, + "loss": 1.4636, + "theoretical_loss": 3.982361157231442, + "tokens_seen": 456327168 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004352056168505517, + "loss": 1.5395, + "theoretical_loss": 3.982299034713183, + "tokens_seen": 456392704 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043519558676028086, + "loss": 1.5786, + "theoretical_loss": 3.9822369236121427, + "tokens_seen": 456458240 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043518555667001004, + "loss": 1.4979, + "theoretical_loss": 3.982174823924586, + "tokens_seen": 456523776 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004351755265797392, + "loss": 1.5331, + "theoretical_loss": 3.9821127356467754, + "tokens_seen": 456589312 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004351654964894684, + "loss": 1.5346, + "theoretical_loss": 3.9820506587749787, + "tokens_seen": 456654848 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043515546639919764, + "loss": 1.4626, + "theoretical_loss": 3.9819885933054637, + "tokens_seen": 456720384 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043514543630892677, + "loss": 1.4789, + "theoretical_loss": 3.9819265392345, + "tokens_seen": 456785920 + }, + { + "epoch": 0.14, + "learning_rate": 0.000435135406218656, + "loss": 1.4382, + "theoretical_loss": 3.9818644965583596, + "tokens_seen": 456851456 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043512537612838513, + "loss": 1.5065, + "theoretical_loss": 3.981802465273316, + "tokens_seen": 456916992 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043511534603811437, + "loss": 1.5549, + "theoretical_loss": 3.9817404453756446, + "tokens_seen": 456982528 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043510531594784355, + "loss": 1.5629, + "theoretical_loss": 3.981678436861622, + "tokens_seen": 457048064 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043509528585757273, + "loss": 1.6096, + "theoretical_loss": 3.981616439727527, + "tokens_seen": 457113600 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004350852557673019, + "loss": 1.4931, + "theoretical_loss": 3.9815544539696397, + "tokens_seen": 457179136 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004350752256770311, + "loss": 1.5747, + "theoretical_loss": 3.981492479584243, + "tokens_seen": 457244672 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043506519558676027, + "loss": 1.5772, + "theoretical_loss": 3.9814305165676203, + "tokens_seen": 457310208 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004350551654964895, + "loss": 1.5657, + "theoretical_loss": 3.9813685649160577, + "tokens_seen": 457375744 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043504513540621863, + "loss": 1.6038, + "theoretical_loss": 3.981306624625842, + "tokens_seen": 457441280 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043503510531594787, + "loss": 1.5299, + "theoretical_loss": 3.9812446956932632, + "tokens_seen": 457506816 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043502507522567705, + "loss": 1.5873, + "theoretical_loss": 3.9811827781146123, + "tokens_seen": 457572352 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043501504513540623, + "loss": 1.549, + "theoretical_loss": 3.981120871886181, + "tokens_seen": 457637888 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004350050150451354, + "loss": 1.6297, + "theoretical_loss": 3.9810589770042637, + "tokens_seen": 457703424 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004349949849548646, + "loss": 1.5712, + "theoretical_loss": 3.9809970934651577, + "tokens_seen": 457768960 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004349849548645938, + "loss": 1.5578, + "theoretical_loss": 3.9809352212651596, + "tokens_seen": 457834496 + }, + { + "epoch": 0.14, + "learning_rate": 0.000434974924774323, + "loss": 1.5702, + "theoretical_loss": 3.9808733604005706, + "tokens_seen": 457900032 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043496489468405214, + "loss": 1.4503, + "theoretical_loss": 3.9808115108676905, + "tokens_seen": 457965568 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043495486459378137, + "loss": 1.4266, + "theoretical_loss": 3.9807496726628235, + "tokens_seen": 458031104 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004349448345035105, + "loss": 1.387, + "theoretical_loss": 3.9806878457822736, + "tokens_seen": 458096640 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043493480441323973, + "loss": 1.4492, + "theoretical_loss": 3.9806260302223477, + "tokens_seen": 458162176 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004349247743229689, + "loss": 1.5814, + "theoretical_loss": 3.9805642259793546, + "tokens_seen": 458227712 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004349147442326981, + "loss": 1.5598, + "theoretical_loss": 3.9805024330496037, + "tokens_seen": 458293248 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004349047141424273, + "loss": 1.5469, + "theoretical_loss": 3.980440651429407, + "tokens_seen": 458358784 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004348946840521565, + "loss": 1.5737, + "theoretical_loss": 3.9803788811150778, + "tokens_seen": 458424320 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043488465396188564, + "loss": 1.5037, + "theoretical_loss": 3.980317122102931, + "tokens_seen": 458489856 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004348746238716149, + "loss": 1.4991, + "theoretical_loss": 3.9802553743892846, + "tokens_seen": 458555392 + }, + { + "epoch": 0.14, + "learning_rate": 0.000434864593781344, + "loss": 1.4496, + "theoretical_loss": 3.9801936379704568, + "tokens_seen": 458620928 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043485456369107324, + "loss": 1.6024, + "theoretical_loss": 3.9801319128427677, + "tokens_seen": 458686464 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 176189, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4482204914093018, + "objective/train/theoretical_loss": 3.9800701990025393, + "objective/train/tokens_used": 458752000, + "theoretical_loss": 3.9800701990025393, + "tokens_seen": 458752000 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004348445336008024, + "loss": 1.5582, + "theoretical_loss": 3.9800701990025393, + "tokens_seen": 458752000 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004348345035105316, + "loss": 1.5864, + "theoretical_loss": 3.9800084964460964, + "tokens_seen": 458817536 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004348244734202608, + "loss": 1.6071, + "theoretical_loss": 3.9799468051697633, + "tokens_seen": 458883072 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043481444332998996, + "loss": 1.5699, + "theoretical_loss": 3.9798851251698686, + "tokens_seen": 458948608 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043480441323971914, + "loss": 1.5193, + "theoretical_loss": 3.97982345644274, + "tokens_seen": 459014144 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004347943831494484, + "loss": 1.5116, + "theoretical_loss": 3.9797617989847094, + "tokens_seen": 459079680 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004347843530591775, + "loss": 1.4833, + "theoretical_loss": 3.979700152792108, + "tokens_seen": 459145216 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043477432296890674, + "loss": 1.5518, + "theoretical_loss": 3.9796385178612717, + "tokens_seen": 459210752 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043476429287863587, + "loss": 1.5603, + "theoretical_loss": 3.9795768941885346, + "tokens_seen": 459276288 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004347542627883651, + "loss": 1.5787, + "theoretical_loss": 3.9795152817702357, + "tokens_seen": 459341824 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004347442326980943, + "loss": 1.4756, + "theoretical_loss": 3.979453680602713, + "tokens_seen": 459407360 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043473420260782346, + "loss": 1.5748, + "theoretical_loss": 3.9793920906823086, + "tokens_seen": 459472896 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043472417251755265, + "loss": 1.5224, + "theoretical_loss": 3.979330512005365, + "tokens_seen": 459538432 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004347141424272819, + "loss": 1.5581, + "theoretical_loss": 3.9792689445682257, + "tokens_seen": 459603968 + }, + { + "epoch": 0.14, + "learning_rate": 0.000434704112337011, + "loss": 1.5806, + "theoretical_loss": 3.9792073883672385, + "tokens_seen": 459669504 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043469408224674024, + "loss": 1.601, + "theoretical_loss": 3.9791458433987503, + "tokens_seen": 459735040 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004346840521564694, + "loss": 1.6386, + "theoretical_loss": 3.97908430965911, + "tokens_seen": 459800576 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004346740220661986, + "loss": 1.5818, + "theoretical_loss": 3.979022787144671, + "tokens_seen": 459866112 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043466399197592784, + "loss": 1.5079, + "theoretical_loss": 3.9789612758517836, + "tokens_seen": 459931648 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043465396188565697, + "loss": 1.5486, + "theoretical_loss": 3.9788997757768043, + "tokens_seen": 459997184 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004346439317953862, + "loss": 1.638, + "theoretical_loss": 3.9788382869160897, + "tokens_seen": 460062720 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043463390170511533, + "loss": 1.5189, + "theoretical_loss": 3.9787768092659963, + "tokens_seen": 460128256 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043462387161484457, + "loss": 1.482, + "theoretical_loss": 3.978715342822885, + "tokens_seen": 460193792 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043461384152457375, + "loss": 1.5828, + "theoretical_loss": 3.9786538875831168, + "tokens_seen": 460259328 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043460381143430293, + "loss": 1.549, + "theoretical_loss": 3.9785924435430555, + "tokens_seen": 460324864 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004345937813440321, + "loss": 1.5804, + "theoretical_loss": 3.978531010699066, + "tokens_seen": 460390400 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004345837512537613, + "loss": 1.5695, + "theoretical_loss": 3.9784695890475144, + "tokens_seen": 460455936 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043457372116349047, + "loss": 1.6548, + "theoretical_loss": 3.9784081785847687, + "tokens_seen": 460521472 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004345636910732197, + "loss": 1.6533, + "theoretical_loss": 3.9783467793071994, + "tokens_seen": 460587008 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043455366098294883, + "loss": 1.4967, + "theoretical_loss": 3.978285391211178, + "tokens_seen": 460652544 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043454363089267807, + "loss": 1.5202, + "theoretical_loss": 3.9782240142930787, + "tokens_seen": 460718080 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043453360080240725, + "loss": 1.4996, + "theoretical_loss": 3.978162648549275, + "tokens_seen": 460783616 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043452357071213643, + "loss": 1.517, + "theoretical_loss": 3.9781012939761453, + "tokens_seen": 460849152 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004345135406218656, + "loss": 1.5928, + "theoretical_loss": 3.978039950570067, + "tokens_seen": 460914688 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004345035105315948, + "loss": 1.5339, + "theoretical_loss": 3.9779786183274197, + "tokens_seen": 460980224 + }, + { + "epoch": 0.14, + "learning_rate": 0.000434493480441324, + "loss": 1.5277, + "theoretical_loss": 3.977917297244587, + "tokens_seen": 461045760 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004344834503510532, + "loss": 1.5761, + "theoretical_loss": 3.9778559873179504, + "tokens_seen": 461111296 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043447342026078234, + "loss": 1.5656, + "theoretical_loss": 3.977794688543897, + "tokens_seen": 461176832 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043446339017051157, + "loss": 1.5189, + "theoretical_loss": 3.9777334009188117, + "tokens_seen": 461242368 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004344533600802407, + "loss": 1.5673, + "theoretical_loss": 3.9776721244390845, + "tokens_seen": 461307904 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043444332998996993, + "loss": 1.4248, + "theoretical_loss": 3.9776108591011052, + "tokens_seen": 461373440 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004344332998996991, + "loss": 1.5, + "theoretical_loss": 3.9775496049012657, + "tokens_seen": 461438976 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004344232698094283, + "loss": 1.4501, + "theoretical_loss": 3.9774883618359596, + "tokens_seen": 461504512 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004344132397191575, + "loss": 1.677, + "theoretical_loss": 3.9774271299015824, + "tokens_seen": 461570048 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004344032096288867, + "loss": 1.5459, + "theoretical_loss": 3.977365909094531, + "tokens_seen": 461635584 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043439317953861584, + "loss": 1.4996, + "theoretical_loss": 3.977304699411204, + "tokens_seen": 461701120 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004343831494483451, + "loss": 1.6178, + "theoretical_loss": 3.977243500848001, + "tokens_seen": 461766656 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004343731193580742, + "loss": 1.5752, + "theoretical_loss": 3.977182313401325, + "tokens_seen": 461832192 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043436308926780344, + "loss": 1.5685, + "theoretical_loss": 3.977121137067579, + "tokens_seen": 461897728 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004343530591775326, + "loss": 1.5843, + "theoretical_loss": 3.9770599718431687, + "tokens_seen": 461963264 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 177354, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5109574794769287, + "objective/train/theoretical_loss": 3.9769988177245015, + "objective/train/tokens_used": 462028800, + "theoretical_loss": 3.9769988177245015, + "tokens_seen": 462028800 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004343430290872618, + "loss": 1.4412, + "theoretical_loss": 3.9769988177245015, + "tokens_seen": 462028800 + }, + { + "epoch": 0.14, + "learning_rate": 0.000434332998996991, + "loss": 1.5953, + "theoretical_loss": 3.976937674707985, + "tokens_seen": 462094336 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043432296890672016, + "loss": 1.5661, + "theoretical_loss": 3.9768765427900306, + "tokens_seen": 462159872 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043431293881644934, + "loss": 1.676, + "theoretical_loss": 3.97681542196705, + "tokens_seen": 462225408 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004343029087261786, + "loss": 1.4685, + "theoretical_loss": 3.976754312235456, + "tokens_seen": 462290944 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004342928786359077, + "loss": 1.5091, + "theoretical_loss": 3.9766932135916653, + "tokens_seen": 462356480 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043428284854563694, + "loss": 1.5708, + "theoretical_loss": 3.9766321260320945, + "tokens_seen": 462422016 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043427281845536607, + "loss": 1.5451, + "theoretical_loss": 3.976571049553162, + "tokens_seen": 462487552 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004342627883650953, + "loss": 1.6525, + "theoretical_loss": 3.9765099841512885, + "tokens_seen": 462553088 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004342527582748245, + "loss": 1.618, + "theoretical_loss": 3.976448929822895, + "tokens_seen": 462618624 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043424272818455367, + "loss": 1.5661, + "theoretical_loss": 3.9763878865644067, + "tokens_seen": 462684160 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043423269809428285, + "loss": 1.5151, + "theoretical_loss": 3.9763268543722483, + "tokens_seen": 462749696 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004342226680040121, + "loss": 1.4199, + "theoretical_loss": 3.976265833242846, + "tokens_seen": 462815232 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004342126379137412, + "loss": 1.5561, + "theoretical_loss": 3.9762048231726297, + "tokens_seen": 462880768 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043420260782347044, + "loss": 1.5311, + "theoretical_loss": 3.9761438241580294, + "tokens_seen": 462946304 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043419257773319957, + "loss": 1.6086, + "theoretical_loss": 3.976082836195477, + "tokens_seen": 463011840 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004341825476429288, + "loss": 1.5401, + "theoretical_loss": 3.976021859281406, + "tokens_seen": 463077376 + }, + { + "epoch": 0.14, + "learning_rate": 0.000434172517552658, + "loss": 1.5323, + "theoretical_loss": 3.975960893412251, + "tokens_seen": 463142912 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043416248746238717, + "loss": 1.5306, + "theoretical_loss": 3.9758999385844502, + "tokens_seen": 463208448 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043415245737211635, + "loss": 1.5876, + "theoretical_loss": 3.975838994794441, + "tokens_seen": 463273984 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043414242728184553, + "loss": 1.6105, + "theoretical_loss": 3.975778062038665, + "tokens_seen": 463339520 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004341323971915747, + "loss": 1.5832, + "theoretical_loss": 3.975717140313564, + "tokens_seen": 463405056 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043412236710130395, + "loss": 1.4546, + "theoretical_loss": 3.9756562296155797, + "tokens_seen": 463470592 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004341123370110331, + "loss": 1.6095, + "theoretical_loss": 3.975595329941159, + "tokens_seen": 463536128 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004341023069207623, + "loss": 1.6681, + "theoretical_loss": 3.975534441286748, + "tokens_seen": 463601664 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043409227683049144, + "loss": 1.4648, + "theoretical_loss": 3.975473563648796, + "tokens_seen": 463667200 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043408224674022067, + "loss": 1.5359, + "theoretical_loss": 3.975412697023753, + "tokens_seen": 463732736 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043407221664994985, + "loss": 1.5796, + "theoretical_loss": 3.9753518414080693, + "tokens_seen": 463798272 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043406218655967903, + "loss": 1.7493, + "theoretical_loss": 3.9752909967982, + "tokens_seen": 463863808 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004340521564694082, + "loss": 1.5764, + "theoretical_loss": 3.9752301631906, + "tokens_seen": 463929344 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043404212637913745, + "loss": 1.5355, + "theoretical_loss": 3.9751693405817248, + "tokens_seen": 463994880 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004340320962888666, + "loss": 1.6096, + "theoretical_loss": 3.9751085289680343, + "tokens_seen": 464060416 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004340220661985958, + "loss": 1.5136, + "theoretical_loss": 3.975047728345987, + "tokens_seen": 464125952 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043401203610832494, + "loss": 1.5469, + "theoretical_loss": 3.9749869387120462, + "tokens_seen": 464191488 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004340020060180542, + "loss": 1.5233, + "theoretical_loss": 3.974926160062674, + "tokens_seen": 464257024 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043399197592778336, + "loss": 1.5705, + "theoretical_loss": 3.9748653923943356, + "tokens_seen": 464322560 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043398194583751254, + "loss": 1.6226, + "theoretical_loss": 3.9748046357034976, + "tokens_seen": 464388096 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004339719157472417, + "loss": 1.6281, + "theoretical_loss": 3.9747438899866276, + "tokens_seen": 464453632 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004339618856569709, + "loss": 1.5722, + "theoretical_loss": 3.9746831552401964, + "tokens_seen": 464519168 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004339518555667001, + "loss": 1.5509, + "theoretical_loss": 3.9746224314606753, + "tokens_seen": 464584704 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004339418254764293, + "loss": 1.5297, + "theoretical_loss": 3.9745617186445372, + "tokens_seen": 464650240 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004339317953861585, + "loss": 1.5083, + "theoretical_loss": 3.974501016788256, + "tokens_seen": 464715776 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004339217652958877, + "loss": 1.5678, + "theoretical_loss": 3.9744403258883096, + "tokens_seen": 464781312 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004339117352056169, + "loss": 1.6076, + "theoretical_loss": 3.974379645941174, + "tokens_seen": 464846848 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043390170511534604, + "loss": 1.5336, + "theoretical_loss": 3.974318976943331, + "tokens_seen": 464912384 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004338916750250753, + "loss": 1.6515, + "theoretical_loss": 3.974258318891261, + "tokens_seen": 464977920 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004338816449348044, + "loss": 1.5905, + "theoretical_loss": 3.974197671781446, + "tokens_seen": 465043456 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043387161484453364, + "loss": 1.7324, + "theoretical_loss": 3.9741370356103713, + "tokens_seen": 465108992 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004338615847542628, + "loss": 1.7269, + "theoretical_loss": 3.9740764103745225, + "tokens_seen": 465174528 + }, + { + "epoch": 0.14, + "learning_rate": 0.000433851554663992, + "loss": 1.6044, + "theoretical_loss": 3.974015796070388, + "tokens_seen": 465240064 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 178655, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5899386405944824, + "objective/train/theoretical_loss": 3.973955192694457, + "objective/train/tokens_used": 465305600, + "theoretical_loss": 3.973955192694457, + "tokens_seen": 465305600 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004338415245737212, + "loss": 1.6659, + "theoretical_loss": 3.973955192694457, + "tokens_seen": 465305600 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043383149448345036, + "loss": 1.5209, + "theoretical_loss": 3.97389460024322, + "tokens_seen": 465371136 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043382146439317954, + "loss": 1.6666, + "theoretical_loss": 3.9738340187131698, + "tokens_seen": 465436672 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004338114343029088, + "loss": 1.5894, + "theoretical_loss": 3.973773448100801, + "tokens_seen": 465502208 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004338014042126379, + "loss": 1.4923, + "theoretical_loss": 3.973712888402609, + "tokens_seen": 465567744 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043379137412236714, + "loss": 1.6537, + "theoretical_loss": 3.9736523396150916, + "tokens_seen": 465633280 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043378134403209627, + "loss": 1.6573, + "theoretical_loss": 3.973591801734748, + "tokens_seen": 465698816 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004337713139418255, + "loss": 1.5634, + "theoretical_loss": 3.9735312747580775, + "tokens_seen": 465764352 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004337612838515547, + "loss": 1.5535, + "theoretical_loss": 3.973470758681584, + "tokens_seen": 465829888 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043375125376128387, + "loss": 1.5232, + "theoretical_loss": 3.9734102535017715, + "tokens_seen": 465895424 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043374122367101305, + "loss": 1.6938, + "theoretical_loss": 3.9733497592151443, + "tokens_seen": 465960960 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004337311935807423, + "loss": 1.5574, + "theoretical_loss": 3.9732892758182103, + "tokens_seen": 466026496 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004337211634904714, + "loss": 1.5633, + "theoretical_loss": 3.973228803307478, + "tokens_seen": 466092032 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043371113340020064, + "loss": 1.5282, + "theoretical_loss": 3.9731683416794583, + "tokens_seen": 466157568 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043370110330992977, + "loss": 1.5784, + "theoretical_loss": 3.9731078909306623, + "tokens_seen": 466223104 + }, + { + "epoch": 0.14, + "learning_rate": 0.000433691073219659, + "loss": 1.5039, + "theoretical_loss": 3.9730474510576044, + "tokens_seen": 466288640 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004336810431293882, + "loss": 1.5392, + "theoretical_loss": 3.9729870220567993, + "tokens_seen": 466354176 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043367101303911737, + "loss": 1.6051, + "theoretical_loss": 3.9729266039247637, + "tokens_seen": 466419712 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043366098294884655, + "loss": 1.572, + "theoretical_loss": 3.9728661966580163, + "tokens_seen": 466485248 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043365095285857573, + "loss": 1.5432, + "theoretical_loss": 3.9728058002530773, + "tokens_seen": 466550784 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004336409227683049, + "loss": 1.6931, + "theoretical_loss": 3.972745414706468, + "tokens_seen": 466616320 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043363089267803415, + "loss": 1.5247, + "theoretical_loss": 3.972685040014711, + "tokens_seen": 466681856 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004336208625877633, + "loss": 1.4998, + "theoretical_loss": 3.9726246761743322, + "tokens_seen": 466747392 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004336108324974925, + "loss": 1.4691, + "theoretical_loss": 3.9725643231818575, + "tokens_seen": 466812928 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043360080240722164, + "loss": 1.5533, + "theoretical_loss": 3.9725039810338147, + "tokens_seen": 466878464 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043359077231695087, + "loss": 1.4918, + "theoretical_loss": 3.972443649726734, + "tokens_seen": 466944000 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043358074222668005, + "loss": 1.5591, + "theoretical_loss": 3.9723833292571458, + "tokens_seen": 467009536 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043357071213640923, + "loss": 1.5815, + "theoretical_loss": 3.9723230196215837, + "tokens_seen": 467075072 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004335606820461384, + "loss": 1.5164, + "theoretical_loss": 3.9722627208165817, + "tokens_seen": 467140608 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043355065195586765, + "loss": 1.5285, + "theoretical_loss": 3.9722024328386754, + "tokens_seen": 467206144 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004335406218655968, + "loss": 1.4729, + "theoretical_loss": 3.9721421556844034, + "tokens_seen": 467271680 + }, + { + "epoch": 0.14, + "learning_rate": 0.000433530591775326, + "loss": 1.6019, + "theoretical_loss": 3.972081889350304, + "tokens_seen": 467337216 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043352056168505514, + "loss": 1.5426, + "theoretical_loss": 3.9720216338329184, + "tokens_seen": 467402752 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004335105315947844, + "loss": 1.4894, + "theoretical_loss": 3.971961389128789, + "tokens_seen": 467468288 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043350050150451356, + "loss": 1.5987, + "theoretical_loss": 3.97190115523446, + "tokens_seen": 467533824 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043349047141424274, + "loss": 1.4169, + "theoretical_loss": 3.9718409321464754, + "tokens_seen": 467599360 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004334804413239719, + "loss": 1.4994, + "theoretical_loss": 3.9717807198613837, + "tokens_seen": 467664896 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004334704112337011, + "loss": 1.5346, + "theoretical_loss": 3.971720518375734, + "tokens_seen": 467730432 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004334603811434303, + "loss": 1.5637, + "theoretical_loss": 3.971660327686075, + "tokens_seen": 467795968 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004334503510531595, + "loss": 1.6334, + "theoretical_loss": 3.97160014778896, + "tokens_seen": 467861504 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043344032096288864, + "loss": 1.5026, + "theoretical_loss": 3.9715399786809424, + "tokens_seen": 467927040 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004334302908726179, + "loss": 1.6779, + "theoretical_loss": 3.971479820358576, + "tokens_seen": 467992576 + }, + { + "epoch": 0.14, + "learning_rate": 0.000433420260782347, + "loss": 1.4739, + "theoretical_loss": 3.9714196728184197, + "tokens_seen": 468058112 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043341023069207624, + "loss": 1.5182, + "theoretical_loss": 3.971359536057029, + "tokens_seen": 468123648 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004334002006018054, + "loss": 1.4721, + "theoretical_loss": 3.9712994100709653, + "tokens_seen": 468189184 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004333901705115346, + "loss": 1.5223, + "theoretical_loss": 3.97123929485679, + "tokens_seen": 468254720 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004333801404212638, + "loss": 1.575, + "theoretical_loss": 3.971179190411066, + "tokens_seen": 468320256 + }, + { + "epoch": 0.14, + "learning_rate": 0.000433370110330993, + "loss": 1.5505, + "theoretical_loss": 3.9711190967303565, + "tokens_seen": 468385792 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043336008024072215, + "loss": 1.4993, + "theoretical_loss": 3.971059013811229, + "tokens_seen": 468451328 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004333500501504514, + "loss": 1.6762, + "theoretical_loss": 3.9709989416502514, + "tokens_seen": 468516864 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 179769, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4446120262145996, + "objective/train/theoretical_loss": 3.9709388802439918, + "objective/train/tokens_used": 468582400, + "theoretical_loss": 3.9709388802439918, + "tokens_seen": 468582400 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004333400200601805, + "loss": 1.4716, + "theoretical_loss": 3.9709388802439918, + "tokens_seen": 468582400 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043332998996990974, + "loss": 1.6917, + "theoretical_loss": 3.9708788295890223, + "tokens_seen": 468647936 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004333199598796389, + "loss": 1.4781, + "theoretical_loss": 3.9708187896819136, + "tokens_seen": 468713472 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004333099297893681, + "loss": 1.5414, + "theoretical_loss": 3.9707587605192414, + "tokens_seen": 468779008 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004332998996990973, + "loss": 1.5261, + "theoretical_loss": 3.9706987420975803, + "tokens_seen": 468844544 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043328986960882647, + "loss": 1.5007, + "theoretical_loss": 3.970638734413508, + "tokens_seen": 468910080 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043327983951855565, + "loss": 1.5859, + "theoretical_loss": 3.970578737463603, + "tokens_seen": 468975616 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004332698094282849, + "loss": 1.5541, + "theoretical_loss": 3.9705187512444446, + "tokens_seen": 469041152 + }, + { + "epoch": 0.14, + "learning_rate": 0.000433259779338014, + "loss": 1.5477, + "theoretical_loss": 3.9704587757526157, + "tokens_seen": 469106688 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043324974924774325, + "loss": 1.4476, + "theoretical_loss": 3.9703988109846997, + "tokens_seen": 469172224 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043323971915747243, + "loss": 1.5355, + "theoretical_loss": 3.970338856937281, + "tokens_seen": 469237760 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004332296890672016, + "loss": 1.5082, + "theoretical_loss": 3.970278913606947, + "tokens_seen": 469303296 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004332196589769308, + "loss": 1.5246, + "theoretical_loss": 3.9702189809902846, + "tokens_seen": 469368832 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043320962888665997, + "loss": 1.462, + "theoretical_loss": 3.9701590590838842, + "tokens_seen": 469434368 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043319959879638915, + "loss": 1.4728, + "theoretical_loss": 3.970099147884336, + "tokens_seen": 469499904 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004331895687061184, + "loss": 1.5587, + "theoretical_loss": 3.9700392473882347, + "tokens_seen": 469565440 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043317953861584757, + "loss": 1.5211, + "theoretical_loss": 3.969979357592173, + "tokens_seen": 469630976 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043316950852557675, + "loss": 1.4437, + "theoretical_loss": 3.969919478492747, + "tokens_seen": 469696512 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043315947843530593, + "loss": 1.465, + "theoretical_loss": 3.9698596100865546, + "tokens_seen": 469762048 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004331494483450351, + "loss": 1.5753, + "theoretical_loss": 3.9697997523701947, + "tokens_seen": 469827584 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043313941825476435, + "loss": 1.5785, + "theoretical_loss": 3.969739905340268, + "tokens_seen": 469893120 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004331293881644935, + "loss": 1.5554, + "theoretical_loss": 3.9696800689933758, + "tokens_seen": 469958656 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004331193580742227, + "loss": 1.492, + "theoretical_loss": 3.9696202433261227, + "tokens_seen": 470024192 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043310932798395184, + "loss": 1.5713, + "theoretical_loss": 3.9695604283351136, + "tokens_seen": 470089728 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043309929789368107, + "loss": 1.5119, + "theoretical_loss": 3.9695006240169546, + "tokens_seen": 470155264 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043308926780341025, + "loss": 1.5506, + "theoretical_loss": 3.9694408303682556, + "tokens_seen": 470220800 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043307923771313943, + "loss": 1.5301, + "theoretical_loss": 3.9693810473856246, + "tokens_seen": 470286336 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004330692076228686, + "loss": 1.4897, + "theoretical_loss": 3.969321275065674, + "tokens_seen": 470351872 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043305917753259785, + "loss": 1.5971, + "theoretical_loss": 3.969261513405017, + "tokens_seen": 470417408 + }, + { + "epoch": 0.14, + "learning_rate": 0.000433049147442327, + "loss": 1.4671, + "theoretical_loss": 3.9692017624002682, + "tokens_seen": 470482944 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004330391173520562, + "loss": 1.5924, + "theoretical_loss": 3.9691420220480422, + "tokens_seen": 470548480 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043302908726178534, + "loss": 1.5195, + "theoretical_loss": 3.9690822923449582, + "tokens_seen": 470614016 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004330190571715146, + "loss": 1.5873, + "theoretical_loss": 3.969022573287635, + "tokens_seen": 470679552 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043300902708124376, + "loss": 1.5812, + "theoretical_loss": 3.968962864872693, + "tokens_seen": 470745088 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043299899699097294, + "loss": 1.541, + "theoretical_loss": 3.9689031670967543, + "tokens_seen": 470810624 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004329889669007021, + "loss": 1.5471, + "theoretical_loss": 3.9688434799564427, + "tokens_seen": 470876160 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004329789368104313, + "loss": 1.5761, + "theoretical_loss": 3.9687838034483844, + "tokens_seen": 470941696 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004329689067201605, + "loss": 1.5033, + "theoretical_loss": 3.9687241375692053, + "tokens_seen": 471007232 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004329588766298897, + "loss": 1.5952, + "theoretical_loss": 3.9686644823155337, + "tokens_seen": 471072768 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043294884653961884, + "loss": 1.5081, + "theoretical_loss": 3.9686048376840004, + "tokens_seen": 471138304 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004329388164493481, + "loss": 1.6305, + "theoretical_loss": 3.968545203671236, + "tokens_seen": 471203840 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004329287863590772, + "loss": 1.5951, + "theoretical_loss": 3.9684855802738745, + "tokens_seen": 471269376 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043291875626880644, + "loss": 1.4433, + "theoretical_loss": 3.9684259674885496, + "tokens_seen": 471334912 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004329087261785356, + "loss": 1.4388, + "theoretical_loss": 3.9683663653118977, + "tokens_seen": 471400448 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004328986960882648, + "loss": 1.6332, + "theoretical_loss": 3.968306773740556, + "tokens_seen": 471465984 + }, + { + "epoch": 0.14, + "learning_rate": 0.000432888665997994, + "loss": 1.4628, + "theoretical_loss": 3.9682471927711642, + "tokens_seen": 471531520 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004328786359077232, + "loss": 1.5521, + "theoretical_loss": 3.9681876224003627, + "tokens_seen": 471597056 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043286860581745235, + "loss": 1.6, + "theoretical_loss": 3.9681280626247943, + "tokens_seen": 471662592 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004328585757271816, + "loss": 1.6666, + "theoretical_loss": 3.9680685134411022, + "tokens_seen": 471728128 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004328485456369107, + "loss": 1.5656, + "theoretical_loss": 3.9680089748459313, + "tokens_seen": 471793664 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 180997, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5616438388824463, + "objective/train/theoretical_loss": 3.967949446835929, + "objective/train/tokens_used": 471859200, + "theoretical_loss": 3.967949446835929, + "tokens_seen": 471859200 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043283851554663994, + "loss": 1.6053, + "theoretical_loss": 3.967949446835929, + "tokens_seen": 471859200 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004328284854563691, + "loss": 1.5159, + "theoretical_loss": 3.967889929407743, + "tokens_seen": 471924736 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004328184553660983, + "loss": 1.607, + "theoretical_loss": 3.9678304225580248, + "tokens_seen": 471990272 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004328084252758275, + "loss": 1.4277, + "theoretical_loss": 3.967770926283424, + "tokens_seen": 472055808 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043279839518555667, + "loss": 1.5748, + "theoretical_loss": 3.9677114405805938, + "tokens_seen": 472121344 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043278836509528585, + "loss": 1.5191, + "theoretical_loss": 3.967651965446189, + "tokens_seen": 472186880 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004327783350050151, + "loss": 1.563, + "theoretical_loss": 3.9675925008768655, + "tokens_seen": 472252416 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004327683049147442, + "loss": 1.4527, + "theoretical_loss": 3.9675330468692804, + "tokens_seen": 472317952 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043275827482447345, + "loss": 1.5815, + "theoretical_loss": 3.9674736034200935, + "tokens_seen": 472383488 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043274824473420263, + "loss": 1.5499, + "theoretical_loss": 3.967414170525965, + "tokens_seen": 472449024 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004327382146439318, + "loss": 1.4456, + "theoretical_loss": 3.9673547481835567, + "tokens_seen": 472514560 + }, + { + "epoch": 0.14, + "learning_rate": 0.000432728184553661, + "loss": 1.5386, + "theoretical_loss": 3.9672953363895322, + "tokens_seen": 472580096 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043271815446339017, + "loss": 1.4955, + "theoretical_loss": 3.9672359351405566, + "tokens_seen": 472645632 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043270812437311935, + "loss": 1.443, + "theoretical_loss": 3.967176544433296, + "tokens_seen": 472711168 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004326980942828486, + "loss": 1.5772, + "theoretical_loss": 3.96711716426442, + "tokens_seen": 472776704 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004326880641925777, + "loss": 1.5303, + "theoretical_loss": 3.9670577946305965, + "tokens_seen": 472842240 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043267803410230695, + "loss": 1.6113, + "theoretical_loss": 3.9669984355284975, + "tokens_seen": 472907776 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004326680040120361, + "loss": 1.4417, + "theoretical_loss": 3.9669390869547954, + "tokens_seen": 472973312 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004326579739217653, + "loss": 1.5714, + "theoretical_loss": 3.966879748906165, + "tokens_seen": 473038848 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004326479438314945, + "loss": 1.5682, + "theoretical_loss": 3.966820421379281, + "tokens_seen": 473104384 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004326379137412237, + "loss": 1.4373, + "theoretical_loss": 3.966761104370821, + "tokens_seen": 473169920 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043262788365095286, + "loss": 1.55, + "theoretical_loss": 3.9667017978774632, + "tokens_seen": 473235456 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043261785356068204, + "loss": 1.5831, + "theoretical_loss": 3.966642501895889, + "tokens_seen": 473300992 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004326078234704112, + "loss": 1.5335, + "theoretical_loss": 3.9665832164227792, + "tokens_seen": 473366528 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043259779338014045, + "loss": 1.6021, + "theoretical_loss": 3.966523941454817, + "tokens_seen": 473432064 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004325877632898696, + "loss": 1.5838, + "theoretical_loss": 3.9664646769886875, + "tokens_seen": 473497600 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004325777331995988, + "loss": 1.6354, + "theoretical_loss": 3.966405423021077, + "tokens_seen": 473563136 + }, + { + "epoch": 0.14, + "learning_rate": 0.000432567703109328, + "loss": 1.5096, + "theoretical_loss": 3.9663461795486725, + "tokens_seen": 473628672 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004325576730190572, + "loss": 1.44, + "theoretical_loss": 3.9662869465681636, + "tokens_seen": 473694208 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043254764292878636, + "loss": 1.5101, + "theoretical_loss": 3.9662277240762416, + "tokens_seen": 473759744 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043253761283851554, + "loss": 1.5413, + "theoretical_loss": 3.9661685120695975, + "tokens_seen": 473825280 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004325275827482447, + "loss": 1.5526, + "theoretical_loss": 3.9661093105449265, + "tokens_seen": 473890816 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043251755265797396, + "loss": 1.5723, + "theoretical_loss": 3.966050119498923, + "tokens_seen": 473956352 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004325075225677031, + "loss": 1.4924, + "theoretical_loss": 3.9659909389282832, + "tokens_seen": 474021888 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004324974924774323, + "loss": 1.5183, + "theoretical_loss": 3.9659317688297064, + "tokens_seen": 474087424 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043248746238716145, + "loss": 1.5946, + "theoretical_loss": 3.965872609199892, + "tokens_seen": 474152960 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004324774322968907, + "loss": 1.6389, + "theoretical_loss": 3.965813460035541, + "tokens_seen": 474218496 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043246740220661986, + "loss": 1.5321, + "theoretical_loss": 3.965754321333356, + "tokens_seen": 474284032 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043245737211634904, + "loss": 1.5557, + "theoretical_loss": 3.965695193090042, + "tokens_seen": 474349568 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004324473420260782, + "loss": 1.5846, + "theoretical_loss": 3.9656360753023034, + "tokens_seen": 474415104 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004324373119358074, + "loss": 1.5707, + "theoretical_loss": 3.9655769679668484, + "tokens_seen": 474480640 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043242728184553664, + "loss": 1.473, + "theoretical_loss": 3.965517871080386, + "tokens_seen": 474546176 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004324172517552658, + "loss": 1.4528, + "theoretical_loss": 3.965458784639625, + "tokens_seen": 474611712 + }, + { + "epoch": 0.14, + "learning_rate": 0.000432407221664995, + "loss": 1.5736, + "theoretical_loss": 3.965399708641278, + "tokens_seen": 474677248 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004323971915747242, + "loss": 1.51, + "theoretical_loss": 3.9653406430820586, + "tokens_seen": 474742784 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004323871614844534, + "loss": 1.5597, + "theoretical_loss": 3.9652815879586805, + "tokens_seen": 474808320 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043237713139418255, + "loss": 1.5126, + "theoretical_loss": 3.96522254326786, + "tokens_seen": 474873856 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004323671013039118, + "loss": 1.5743, + "theoretical_loss": 3.9651635090063153, + "tokens_seen": 474939392 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004323570712136409, + "loss": 1.6201, + "theoretical_loss": 3.965104485170765, + "tokens_seen": 475004928 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043234704112337014, + "loss": 1.5473, + "theoretical_loss": 3.9650454717579295, + "tokens_seen": 475070464 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 182209, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4865361452102661, + "objective/train/theoretical_loss": 3.9649864687645318, + "objective/train/tokens_used": 475136000, + "theoretical_loss": 3.9649864687645318, + "tokens_seen": 475136000 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004323370110330993, + "loss": 1.5437, + "theoretical_loss": 3.9649864687645318, + "tokens_seen": 475136000 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004323269809428285, + "loss": 1.6794, + "theoretical_loss": 3.9649274761872944, + "tokens_seen": 475201536 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004323169508525577, + "loss": 1.5952, + "theoretical_loss": 3.9648684940229426, + "tokens_seen": 475267072 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043230692076228687, + "loss": 1.5192, + "theoretical_loss": 3.9648095222682036, + "tokens_seen": 475332608 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043229689067201605, + "loss": 1.6322, + "theoretical_loss": 3.9647505609198044, + "tokens_seen": 475398144 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004322868605817453, + "loss": 1.464, + "theoretical_loss": 3.9646916099744756, + "tokens_seen": 475463680 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004322768304914744, + "loss": 1.5993, + "theoretical_loss": 3.9646326694289473, + "tokens_seen": 475529216 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043226680040120365, + "loss": 1.4451, + "theoretical_loss": 3.9645737392799516, + "tokens_seen": 475594752 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043225677031093283, + "loss": 1.5616, + "theoretical_loss": 3.964514819524224, + "tokens_seen": 475660288 + }, + { + "epoch": 0.14, + "learning_rate": 0.000432246740220662, + "loss": 1.5222, + "theoretical_loss": 3.964455910158498, + "tokens_seen": 475725824 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004322367101303912, + "loss": 1.4401, + "theoretical_loss": 3.9643970111795115, + "tokens_seen": 475791360 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043222668004012037, + "loss": 1.438, + "theoretical_loss": 3.964338122584003, + "tokens_seen": 475856896 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043221664994984955, + "loss": 1.5658, + "theoretical_loss": 3.964279244368712, + "tokens_seen": 475922432 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004322066198595788, + "loss": 1.5135, + "theoretical_loss": 3.9642203765303794, + "tokens_seen": 475987968 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004321965897693079, + "loss": 1.592, + "theoretical_loss": 3.9641615190657484, + "tokens_seen": 476053504 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043218655967903715, + "loss": 1.5492, + "theoretical_loss": 3.9641026719715633, + "tokens_seen": 476119040 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004321765295887663, + "loss": 1.6254, + "theoretical_loss": 3.9640438352445697, + "tokens_seen": 476184576 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004321664994984955, + "loss": 1.6769, + "theoretical_loss": 3.9639850088815143, + "tokens_seen": 476250112 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004321564694082247, + "loss": 1.4784, + "theoretical_loss": 3.9639261928791463, + "tokens_seen": 476315648 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004321464393179539, + "loss": 1.5351, + "theoretical_loss": 3.963867387234216, + "tokens_seen": 476381184 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043213640922768306, + "loss": 1.5823, + "theoretical_loss": 3.963808591943474, + "tokens_seen": 476446720 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043212637913741224, + "loss": 1.4922, + "theoretical_loss": 3.9637498070036745, + "tokens_seen": 476512256 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004321163490471414, + "loss": 1.5009, + "theoretical_loss": 3.963691032411571, + "tokens_seen": 476577792 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043210631895687065, + "loss": 1.6575, + "theoretical_loss": 3.9636322681639196, + "tokens_seen": 476643328 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004320962888665998, + "loss": 1.4753, + "theoretical_loss": 3.9635735142574786, + "tokens_seen": 476708864 + }, + { + "epoch": 0.14, + "learning_rate": 0.000432086258776329, + "loss": 1.5737, + "theoretical_loss": 3.963514770689006, + "tokens_seen": 476774400 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004320762286860582, + "loss": 1.5189, + "theoretical_loss": 3.9634560374552623, + "tokens_seen": 476839936 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004320661985957874, + "loss": 1.5216, + "theoretical_loss": 3.96339731455301, + "tokens_seen": 476905472 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043205616850551656, + "loss": 1.4635, + "theoretical_loss": 3.963338601979011, + "tokens_seen": 476971008 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043204613841524574, + "loss": 1.6631, + "theoretical_loss": 3.9632798997300314, + "tokens_seen": 477036544 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004320361083249749, + "loss": 1.5633, + "theoretical_loss": 3.963221207802836, + "tokens_seen": 477102080 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043202607823470416, + "loss": 1.4872, + "theoretical_loss": 3.9631625261941945, + "tokens_seen": 477167616 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004320160481444333, + "loss": 1.5669, + "theoretical_loss": 3.963103854900874, + "tokens_seen": 477233152 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004320060180541625, + "loss": 1.5287, + "theoretical_loss": 3.963045193919646, + "tokens_seen": 477298688 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043199598796389165, + "loss": 1.5472, + "theoretical_loss": 3.9629865432472817, + "tokens_seen": 477364224 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004319859578736209, + "loss": 1.4819, + "theoretical_loss": 3.9629279028805553, + "tokens_seen": 477429760 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043197592778335006, + "loss": 1.5694, + "theoretical_loss": 3.962869272816242, + "tokens_seen": 477495296 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043196589769307924, + "loss": 1.6357, + "theoretical_loss": 3.9628106530511173, + "tokens_seen": 477560832 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004319558676028084, + "loss": 1.6222, + "theoretical_loss": 3.9627520435819594, + "tokens_seen": 477626368 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004319458375125376, + "loss": 1.5236, + "theoretical_loss": 3.9626934444055477, + "tokens_seen": 477691904 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004319358074222668, + "loss": 1.5224, + "theoretical_loss": 3.962634855518663, + "tokens_seen": 477757440 + }, + { + "epoch": 0.14, + "learning_rate": 0.000431925777331996, + "loss": 1.4276, + "theoretical_loss": 3.962576276918087, + "tokens_seen": 477822976 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043191574724172515, + "loss": 1.6288, + "theoretical_loss": 3.962517708600603, + "tokens_seen": 477888512 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004319057171514544, + "loss": 1.6357, + "theoretical_loss": 3.9624591505629976, + "tokens_seen": 477954048 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043189568706118357, + "loss": 1.5189, + "theoretical_loss": 3.962400602802056, + "tokens_seen": 478019584 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043188565697091275, + "loss": 1.5031, + "theoretical_loss": 3.962342065314566, + "tokens_seen": 478085120 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043187562688064193, + "loss": 1.5451, + "theoretical_loss": 3.962283538097318, + "tokens_seen": 478150656 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004318655967903711, + "loss": 1.4954, + "theoretical_loss": 3.962225021147102, + "tokens_seen": 478216192 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004318555667001003, + "loss": 1.5969, + "theoretical_loss": 3.9621665144607103, + "tokens_seen": 478281728 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004318455366098295, + "loss": 1.4663, + "theoretical_loss": 3.9621080180349373, + "tokens_seen": 478347264 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 183452, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.491711139678955, + "objective/train/theoretical_loss": 3.9620495318665774, + "objective/train/tokens_used": 478412800, + "theoretical_loss": 3.9620495318665774, + "tokens_seen": 478412800 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043183550651955865, + "loss": 1.5592, + "theoretical_loss": 3.9620495318665774, + "tokens_seen": 478412800 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004318254764292879, + "loss": 1.5966, + "theoretical_loss": 3.9619910559524274, + "tokens_seen": 478478336 + }, + { + "epoch": 0.15, + "learning_rate": 0.000431815446339017, + "loss": 1.5328, + "theoretical_loss": 3.9619325902892855, + "tokens_seen": 478543872 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043180541624874625, + "loss": 1.7105, + "theoretical_loss": 3.961874134873951, + "tokens_seen": 478609408 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043179538615847543, + "loss": 1.5476, + "theoretical_loss": 3.9618156897032257, + "tokens_seen": 478674944 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004317853560682046, + "loss": 1.6357, + "theoretical_loss": 3.961757254773911, + "tokens_seen": 478740480 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004317753259779338, + "loss": 1.6976, + "theoretical_loss": 3.9616988300828098, + "tokens_seen": 478806016 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043176529588766303, + "loss": 1.559, + "theoretical_loss": 3.9616404156267286, + "tokens_seen": 478871552 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043175526579739215, + "loss": 1.6585, + "theoretical_loss": 3.961582011402475, + "tokens_seen": 478937088 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004317452357071214, + "loss": 1.5173, + "theoretical_loss": 3.9615236174068547, + "tokens_seen": 479002624 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004317352056168505, + "loss": 1.6051, + "theoretical_loss": 3.961465233636679, + "tokens_seen": 479068160 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043172517552657975, + "loss": 1.5832, + "theoretical_loss": 3.9614068600887578, + "tokens_seen": 479133696 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043171514543630893, + "loss": 1.671, + "theoretical_loss": 3.9613484967599044, + "tokens_seen": 479199232 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004317051153460381, + "loss": 1.5543, + "theoretical_loss": 3.961290143646932, + "tokens_seen": 479264768 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004316950852557673, + "loss": 1.5473, + "theoretical_loss": 3.961231800746656, + "tokens_seen": 479330304 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004316850551654965, + "loss": 1.552, + "theoretical_loss": 3.961173468055893, + "tokens_seen": 479395840 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004316750250752257, + "loss": 1.6315, + "theoretical_loss": 3.961115145571461, + "tokens_seen": 479461376 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004316649949849549, + "loss": 1.4653, + "theoretical_loss": 3.96105683329018, + "tokens_seen": 479526912 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004316549648946841, + "loss": 1.5615, + "theoretical_loss": 3.9609985312088707, + "tokens_seen": 479592448 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043164493480441326, + "loss": 1.5169, + "theoretical_loss": 3.960940239324355, + "tokens_seen": 479657984 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043163490471414244, + "loss": 1.7147, + "theoretical_loss": 3.960881957633457, + "tokens_seen": 479723520 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004316248746238716, + "loss": 1.5527, + "theoretical_loss": 3.9608236861330024, + "tokens_seen": 479789056 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043161484453360085, + "loss": 1.4977, + "theoretical_loss": 3.960765424819817, + "tokens_seen": 479854592 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043160481444333, + "loss": 1.6159, + "theoretical_loss": 3.9607071736907296, + "tokens_seen": 479920128 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004315947843530592, + "loss": 1.4505, + "theoretical_loss": 3.9606489327425694, + "tokens_seen": 479985664 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004315847542627884, + "loss": 1.5659, + "theoretical_loss": 3.960590701972167, + "tokens_seen": 480051200 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004315747241725176, + "loss": 1.4614, + "theoretical_loss": 3.960532481376355, + "tokens_seen": 480116736 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043156469408224676, + "loss": 1.5223, + "theoretical_loss": 3.960474270951967, + "tokens_seen": 480182272 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043155466399197594, + "loss": 1.4461, + "theoretical_loss": 3.9604160706958385, + "tokens_seen": 480247808 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004315446339017051, + "loss": 1.5396, + "theoretical_loss": 3.960357880604806, + "tokens_seen": 480313344 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043153460381143436, + "loss": 1.4979, + "theoretical_loss": 3.9602997006757072, + "tokens_seen": 480378880 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004315245737211635, + "loss": 1.5, + "theoretical_loss": 3.960241530905382, + "tokens_seen": 480444416 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004315145436308927, + "loss": 1.4218, + "theoretical_loss": 3.96018337129067, + "tokens_seen": 480509952 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043150451354062185, + "loss": 1.641, + "theoretical_loss": 3.960125221828415, + "tokens_seen": 480575488 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004314944834503511, + "loss": 1.5826, + "theoretical_loss": 3.9600670825154594, + "tokens_seen": 480641024 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043148445336008026, + "loss": 1.5425, + "theoretical_loss": 3.960008953348649, + "tokens_seen": 480706560 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043147442326980944, + "loss": 1.628, + "theoretical_loss": 3.9599508343248297, + "tokens_seen": 480772096 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004314643931795386, + "loss": 1.5243, + "theoretical_loss": 3.9598927254408505, + "tokens_seen": 480837632 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004314543630892678, + "loss": 1.4818, + "theoretical_loss": 3.9598346266935596, + "tokens_seen": 480903168 + }, + { + "epoch": 0.15, + "learning_rate": 0.000431444332998997, + "loss": 1.6176, + "theoretical_loss": 3.959776538079808, + "tokens_seen": 480968704 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004314343029087262, + "loss": 1.604, + "theoretical_loss": 3.9597184595964476, + "tokens_seen": 481034240 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043142427281845535, + "loss": 1.4591, + "theoretical_loss": 3.959660391240332, + "tokens_seen": 481099776 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004314142427281846, + "loss": 1.5295, + "theoretical_loss": 3.9596023330083163, + "tokens_seen": 481165312 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043140421263791377, + "loss": 1.5232, + "theoretical_loss": 3.9595442848972566, + "tokens_seen": 481230848 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043139418254764295, + "loss": 1.6393, + "theoretical_loss": 3.959486246904011, + "tokens_seen": 481296384 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043138415245737213, + "loss": 1.628, + "theoretical_loss": 3.959428219025438, + "tokens_seen": 481361920 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004313741223671013, + "loss": 1.5229, + "theoretical_loss": 3.9593702012583987, + "tokens_seen": 481427456 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004313640922768305, + "loss": 1.559, + "theoretical_loss": 3.9593121935997546, + "tokens_seen": 481492992 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004313540621865597, + "loss": 1.4607, + "theoretical_loss": 3.95925419604637, + "tokens_seen": 481558528 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043134403209628885, + "loss": 1.4885, + "theoretical_loss": 3.959196208595108, + "tokens_seen": 481624064 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 184594, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4575998783111572, + "objective/train/theoretical_loss": 3.959138231242836, + "objective/train/tokens_used": 481689600, + "theoretical_loss": 3.959138231242836, + "tokens_seen": 481689600 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004313340020060181, + "loss": 1.5137, + "theoretical_loss": 3.959138231242836, + "tokens_seen": 481689600 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004313239719157472, + "loss": 1.4492, + "theoretical_loss": 3.959080263986421, + "tokens_seen": 481755136 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043131394182547645, + "loss": 1.4539, + "theoretical_loss": 3.9590223068227317, + "tokens_seen": 481820672 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043130391173520563, + "loss": 1.4678, + "theoretical_loss": 3.9589643597486393, + "tokens_seen": 481886208 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004312938816449348, + "loss": 1.478, + "theoretical_loss": 3.9589064227610145, + "tokens_seen": 481951744 + }, + { + "epoch": 0.15, + "learning_rate": 0.000431283851554664, + "loss": 1.5106, + "theoretical_loss": 3.9588484958567314, + "tokens_seen": 482017280 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043127382146439323, + "loss": 1.4631, + "theoretical_loss": 3.958790579032663, + "tokens_seen": 482082816 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043126379137412236, + "loss": 1.5982, + "theoretical_loss": 3.958732672285687, + "tokens_seen": 482148352 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004312537612838516, + "loss": 1.5533, + "theoretical_loss": 3.95867477561268, + "tokens_seen": 482213888 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004312437311935807, + "loss": 1.4167, + "theoretical_loss": 3.95861688901052, + "tokens_seen": 482279424 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043123370110330995, + "loss": 1.5941, + "theoretical_loss": 3.958559012476088, + "tokens_seen": 482344960 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043122367101303913, + "loss": 1.5301, + "theoretical_loss": 3.958501146006265, + "tokens_seen": 482410496 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004312136409227683, + "loss": 1.4692, + "theoretical_loss": 3.958443289597934, + "tokens_seen": 482476032 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004312036108324975, + "loss": 1.6221, + "theoretical_loss": 3.958385443247979, + "tokens_seen": 482541568 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004311935807422267, + "loss": 1.646, + "theoretical_loss": 3.958327606953286, + "tokens_seen": 482607104 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043118355065195586, + "loss": 1.6047, + "theoretical_loss": 3.958269780710742, + "tokens_seen": 482672640 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004311735205616851, + "loss": 1.4657, + "theoretical_loss": 3.9582119645172344, + "tokens_seen": 482738176 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004311634904714142, + "loss": 1.6844, + "theoretical_loss": 3.9581541583696547, + "tokens_seen": 482803712 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043115346038114346, + "loss": 1.5857, + "theoretical_loss": 3.958096362264893, + "tokens_seen": 482869248 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004311434302908726, + "loss": 1.4913, + "theoretical_loss": 3.958038576199842, + "tokens_seen": 482934784 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004311334002006018, + "loss": 1.5891, + "theoretical_loss": 3.957980800171396, + "tokens_seen": 483000320 + }, + { + "epoch": 0.15, + "learning_rate": 0.000431123370110331, + "loss": 1.6145, + "theoretical_loss": 3.9579230341764493, + "tokens_seen": 483065856 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004311133400200602, + "loss": 1.4455, + "theoretical_loss": 3.9578652782118997, + "tokens_seen": 483131392 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043110330992978936, + "loss": 1.3935, + "theoretical_loss": 3.957807532274645, + "tokens_seen": 483196928 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004310932798395186, + "loss": 1.4665, + "theoretical_loss": 3.957749796361585, + "tokens_seen": 483262464 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004310832497492477, + "loss": 1.6195, + "theoretical_loss": 3.9576920704696192, + "tokens_seen": 483328000 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043107321965897696, + "loss": 1.5155, + "theoretical_loss": 3.957634354595651, + "tokens_seen": 483393536 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004310631895687061, + "loss": 1.5452, + "theoretical_loss": 3.957576648736584, + "tokens_seen": 483459072 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004310531594784353, + "loss": 1.6034, + "theoretical_loss": 3.9575189528893233, + "tokens_seen": 483524608 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004310431293881645, + "loss": 1.582, + "theoretical_loss": 3.957461267050774, + "tokens_seen": 483590144 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004310330992978937, + "loss": 1.609, + "theoretical_loss": 3.957403591217845, + "tokens_seen": 483655680 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043102306920762286, + "loss": 1.575, + "theoretical_loss": 3.9573459253874455, + "tokens_seen": 483721216 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043101303911735205, + "loss": 1.5502, + "theoretical_loss": 3.9572882695564853, + "tokens_seen": 483786752 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004310030090270812, + "loss": 1.5081, + "theoretical_loss": 3.9572306237218764, + "tokens_seen": 483852288 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043099297893681046, + "loss": 1.5166, + "theoretical_loss": 3.957172987880532, + "tokens_seen": 483917824 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004309829488465396, + "loss": 1.6386, + "theoretical_loss": 3.9571153620293673, + "tokens_seen": 483983360 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004309729187562688, + "loss": 1.5499, + "theoretical_loss": 3.9570577461652974, + "tokens_seen": 484048896 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043096288866599795, + "loss": 1.6093, + "theoretical_loss": 3.95700014028524, + "tokens_seen": 484114432 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004309528585757272, + "loss": 1.5825, + "theoretical_loss": 3.956942544386114, + "tokens_seen": 484179968 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043094282848545637, + "loss": 1.6742, + "theoretical_loss": 3.9568849584648396, + "tokens_seen": 484245504 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043093279839518555, + "loss": 1.505, + "theoretical_loss": 3.9568273825183375, + "tokens_seen": 484311040 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004309227683049148, + "loss": 1.621, + "theoretical_loss": 3.9567698165435305, + "tokens_seen": 484376576 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043091273821464397, + "loss": 1.5307, + "theoretical_loss": 3.9567122605373437, + "tokens_seen": 484442112 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043090270812437315, + "loss": 1.4967, + "theoretical_loss": 3.956654714496702, + "tokens_seen": 484507648 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043089267803410233, + "loss": 1.5865, + "theoretical_loss": 3.9565971784185328, + "tokens_seen": 484573184 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004308826479438315, + "loss": 1.4918, + "theoretical_loss": 3.9565396522997633, + "tokens_seen": 484638720 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004308726178535607, + "loss": 1.5323, + "theoretical_loss": 3.9564821361373244, + "tokens_seen": 484704256 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004308625877632899, + "loss": 1.5366, + "theoretical_loss": 3.956424629928146, + "tokens_seen": 484769792 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043085255767301905, + "loss": 1.533, + "theoretical_loss": 3.9563671336691613, + "tokens_seen": 484835328 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004308425275827483, + "loss": 1.4813, + "theoretical_loss": 3.956309647357303, + "tokens_seen": 484900864 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 185723, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.529571294784546, + "objective/train/theoretical_loss": 3.956252170989507, + "objective/train/tokens_used": 484966400, + "theoretical_loss": 3.956252170989507, + "tokens_seen": 484966400 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004308324974924774, + "loss": 1.5217, + "theoretical_loss": 3.956252170989507, + "tokens_seen": 484966400 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043082246740220665, + "loss": 1.4706, + "theoretical_loss": 3.9561947045627095, + "tokens_seen": 485031936 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043081243731193583, + "loss": 1.5435, + "theoretical_loss": 3.9561372480738486, + "tokens_seen": 485097472 + }, + { + "epoch": 0.15, + "learning_rate": 0.000430802407221665, + "loss": 1.5079, + "theoretical_loss": 3.956079801519863, + "tokens_seen": 485163008 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004307923771313942, + "loss": 1.4619, + "theoretical_loss": 3.9560223648976924, + "tokens_seen": 485228544 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043078234704112343, + "loss": 1.5752, + "theoretical_loss": 3.9559649382042803, + "tokens_seen": 485294080 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043077231695085256, + "loss": 1.5426, + "theoretical_loss": 3.955907521436569, + "tokens_seen": 485359616 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004307622868605818, + "loss": 1.4852, + "theoretical_loss": 3.9558501145915033, + "tokens_seen": 485425152 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004307522567703109, + "loss": 1.5856, + "theoretical_loss": 3.9557927176660286, + "tokens_seen": 485490688 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043074222668004015, + "loss": 1.4959, + "theoretical_loss": 3.9557353306570926, + "tokens_seen": 485556224 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043073219658976933, + "loss": 1.6191, + "theoretical_loss": 3.9556779535616444, + "tokens_seen": 485621760 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004307221664994985, + "loss": 1.5456, + "theoretical_loss": 3.955620586376633, + "tokens_seen": 485687296 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004307121364092277, + "loss": 1.4848, + "theoretical_loss": 3.9555632290990097, + "tokens_seen": 485752832 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004307021063189569, + "loss": 1.5854, + "theoretical_loss": 3.9555058817257285, + "tokens_seen": 485818368 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043069207622868606, + "loss": 1.5684, + "theoretical_loss": 3.9554485442537417, + "tokens_seen": 485883904 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004306820461384153, + "loss": 1.5159, + "theoretical_loss": 3.955391216680006, + "tokens_seen": 485949440 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004306720160481444, + "loss": 1.4625, + "theoretical_loss": 3.9553338990014772, + "tokens_seen": 486014976 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043066198595787366, + "loss": 1.4946, + "theoretical_loss": 3.9552765912151138, + "tokens_seen": 486080512 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004306519558676028, + "loss": 1.4616, + "theoretical_loss": 3.9552192933178745, + "tokens_seen": 486146048 + }, + { + "epoch": 0.15, + "learning_rate": 0.000430641925777332, + "loss": 1.5255, + "theoretical_loss": 3.9551620053067214, + "tokens_seen": 486211584 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004306318956870612, + "loss": 1.5555, + "theoretical_loss": 3.9551047271786155, + "tokens_seen": 486277120 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004306218655967904, + "loss": 1.4379, + "theoretical_loss": 3.9550474589305202, + "tokens_seen": 486342656 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043061183550651956, + "loss": 1.4852, + "theoretical_loss": 3.954990200559401, + "tokens_seen": 486408192 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004306018054162488, + "loss": 1.5302, + "theoretical_loss": 3.9549329520622236, + "tokens_seen": 486473728 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004305917753259779, + "loss": 1.5463, + "theoretical_loss": 3.954875713435955, + "tokens_seen": 486539264 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043058174523570716, + "loss": 1.4966, + "theoretical_loss": 3.9548184846775647, + "tokens_seen": 486604800 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004305717151454363, + "loss": 1.4498, + "theoretical_loss": 3.9547612657840228, + "tokens_seen": 486670336 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004305616850551655, + "loss": 1.5136, + "theoretical_loss": 3.9547040567523, + "tokens_seen": 486735872 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004305516549648947, + "loss": 1.5758, + "theoretical_loss": 3.9546468575793696, + "tokens_seen": 486801408 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004305416248746239, + "loss": 1.5366, + "theoretical_loss": 3.9545896682622064, + "tokens_seen": 486866944 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043053159478435306, + "loss": 1.5654, + "theoretical_loss": 3.954532488797785, + "tokens_seen": 486932480 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043052156469408225, + "loss": 1.5979, + "theoretical_loss": 3.954475319183082, + "tokens_seen": 486998016 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004305115346038114, + "loss": 1.5237, + "theoretical_loss": 3.9544181594150762, + "tokens_seen": 487063552 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043050150451354066, + "loss": 1.4391, + "theoretical_loss": 3.9543610094907464, + "tokens_seen": 487129088 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004304914744232698, + "loss": 1.5872, + "theoretical_loss": 3.9543038694070747, + "tokens_seen": 487194624 + }, + { + "epoch": 0.15, + "learning_rate": 0.000430481444332999, + "loss": 1.5235, + "theoretical_loss": 3.9542467391610416, + "tokens_seen": 487260160 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043047141424272815, + "loss": 1.4745, + "theoretical_loss": 3.9541896187496315, + "tokens_seen": 487325696 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004304613841524574, + "loss": 1.5553, + "theoretical_loss": 3.9541325081698293, + "tokens_seen": 487391232 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043045135406218657, + "loss": 1.6295, + "theoretical_loss": 3.954075407418621, + "tokens_seen": 487456768 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043044132397191575, + "loss": 1.6011, + "theoretical_loss": 3.954018316492994, + "tokens_seen": 487522304 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043043129388164493, + "loss": 1.6178, + "theoretical_loss": 3.9539612353899365, + "tokens_seen": 487587840 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043042126379137417, + "loss": 1.474, + "theoretical_loss": 3.9539041641064396, + "tokens_seen": 487653376 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004304112337011033, + "loss": 1.555, + "theoretical_loss": 3.953847102639494, + "tokens_seen": 487718912 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043040120361083253, + "loss": 1.582, + "theoretical_loss": 3.9537900509860933, + "tokens_seen": 487784448 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043039117352056165, + "loss": 1.5655, + "theoretical_loss": 3.953733009143231, + "tokens_seen": 487849984 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004303811434302909, + "loss": 1.4684, + "theoretical_loss": 3.9536759771079018, + "tokens_seen": 487915520 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043037111334002007, + "loss": 1.4985, + "theoretical_loss": 3.9536189548771037, + "tokens_seen": 487981056 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043036108324974925, + "loss": 1.441, + "theoretical_loss": 3.953561942447834, + "tokens_seen": 488046592 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043035105315947843, + "loss": 1.5162, + "theoretical_loss": 3.953504939817093, + "tokens_seen": 488112128 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004303410230692076, + "loss": 1.5409, + "theoretical_loss": 3.953447946981881, + "tokens_seen": 488177664 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 186992, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5685129165649414, + "objective/train/theoretical_loss": 3.9533909639391993, + "objective/train/tokens_used": 488243200, + "theoretical_loss": 3.9533909639391993, + "tokens_seen": 488243200 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004303309929789368, + "loss": 1.5384, + "theoretical_loss": 3.9533909639391993, + "tokens_seen": 488243200 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043032096288866603, + "loss": 1.4958, + "theoretical_loss": 3.953333990686051, + "tokens_seen": 488308736 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043031093279839516, + "loss": 1.517, + "theoretical_loss": 3.9532770272194426, + "tokens_seen": 488374272 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004303009027081244, + "loss": 1.5531, + "theoretical_loss": 3.9532200735363787, + "tokens_seen": 488439808 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004302908726178535, + "loss": 1.6255, + "theoretical_loss": 3.953163129633867, + "tokens_seen": 488505344 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043028084252758276, + "loss": 1.4148, + "theoretical_loss": 3.9531061955089157, + "tokens_seen": 488570880 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043027081243731194, + "loss": 1.4839, + "theoretical_loss": 3.9530492711585357, + "tokens_seen": 488636416 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004302607823470411, + "loss": 1.5484, + "theoretical_loss": 3.9529923565797374, + "tokens_seen": 488701952 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004302507522567703, + "loss": 1.4373, + "theoretical_loss": 3.952935451769533, + "tokens_seen": 488767488 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043024072216649953, + "loss": 1.5669, + "theoretical_loss": 3.9528785567249374, + "tokens_seen": 488833024 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043023069207622866, + "loss": 1.5682, + "theoretical_loss": 3.9528216714429654, + "tokens_seen": 488898560 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004302206619859579, + "loss": 1.4771, + "theoretical_loss": 3.9527647959206336, + "tokens_seen": 488964096 + }, + { + "epoch": 0.15, + "learning_rate": 0.000430210631895687, + "loss": 1.5179, + "theoretical_loss": 3.9527079301549595, + "tokens_seen": 489029632 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043020060180541626, + "loss": 1.4898, + "theoretical_loss": 3.9526510741429624, + "tokens_seen": 489095168 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043019057171514544, + "loss": 1.5534, + "theoretical_loss": 3.9525942278816633, + "tokens_seen": 489160704 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004301805416248746, + "loss": 1.6054, + "theoretical_loss": 3.952537391368083, + "tokens_seen": 489226240 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043017051153460386, + "loss": 1.5684, + "theoretical_loss": 3.9524805645992447, + "tokens_seen": 489291776 + }, + { + "epoch": 0.15, + "learning_rate": 0.000430160481444333, + "loss": 1.6397, + "theoretical_loss": 3.9524237475721726, + "tokens_seen": 489357312 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004301504513540622, + "loss": 1.6548, + "theoretical_loss": 3.952366940283894, + "tokens_seen": 489422848 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004301404212637914, + "loss": 1.5387, + "theoretical_loss": 3.952310142731434, + "tokens_seen": 489488384 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004301303911735206, + "loss": 1.4512, + "theoretical_loss": 3.9522533549118215, + "tokens_seen": 489553920 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043012036108324976, + "loss": 1.5929, + "theoretical_loss": 3.952196576822086, + "tokens_seen": 489619456 + }, + { + "epoch": 0.15, + "learning_rate": 0.000430110330992979, + "loss": 1.5294, + "theoretical_loss": 3.9521398084592585, + "tokens_seen": 489684992 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004301003009027081, + "loss": 1.5213, + "theoretical_loss": 3.952083049820372, + "tokens_seen": 489750528 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043009027081243736, + "loss": 1.4728, + "theoretical_loss": 3.9520263009024585, + "tokens_seen": 489816064 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004300802407221665, + "loss": 1.5708, + "theoretical_loss": 3.951969561702554, + "tokens_seen": 489881600 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004300702106318957, + "loss": 1.4932, + "theoretical_loss": 3.9519128322176935, + "tokens_seen": 489947136 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004300601805416249, + "loss": 1.5196, + "theoretical_loss": 3.951856112444915, + "tokens_seen": 490012672 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004300501504513541, + "loss": 1.5979, + "theoretical_loss": 3.951799402381258, + "tokens_seen": 490078208 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043004012036108326, + "loss": 1.3681, + "theoretical_loss": 3.951742702023761, + "tokens_seen": 490143744 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043003009027081245, + "loss": 1.5301, + "theoretical_loss": 3.951686011369466, + "tokens_seen": 490209280 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043002006018054163, + "loss": 1.6402, + "theoretical_loss": 3.9516293304154155, + "tokens_seen": 490274816 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043001003009027086, + "loss": 1.5882, + "theoretical_loss": 3.9515726591586544, + "tokens_seen": 490340352 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043, + "loss": 1.447, + "theoretical_loss": 3.9515159975962257, + "tokens_seen": 490405888 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004299899699097292, + "loss": 1.5797, + "theoretical_loss": 3.951459345725178, + "tokens_seen": 490471424 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042997993981945835, + "loss": 1.5229, + "theoretical_loss": 3.9514027035425583, + "tokens_seen": 490536960 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004299699097291876, + "loss": 1.4374, + "theoretical_loss": 3.951346071045415, + "tokens_seen": 490602496 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042995987963891677, + "loss": 1.5931, + "theoretical_loss": 3.951289448230799, + "tokens_seen": 490668032 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042994984954864595, + "loss": 1.5183, + "theoretical_loss": 3.9512328350957624, + "tokens_seen": 490733568 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042993981945837513, + "loss": 1.5574, + "theoretical_loss": 3.9511762316373575, + "tokens_seen": 490799104 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042992978936810437, + "loss": 1.5786, + "theoretical_loss": 3.951119637852639, + "tokens_seen": 490864640 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004299197592778335, + "loss": 1.4856, + "theoretical_loss": 3.951063053738662, + "tokens_seen": 490930176 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042990972918756273, + "loss": 1.5722, + "theoretical_loss": 3.951006479292484, + "tokens_seen": 490995712 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042989969909729185, + "loss": 1.6395, + "theoretical_loss": 3.9509499145111624, + "tokens_seen": 491061248 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004298896690070211, + "loss": 1.6112, + "theoretical_loss": 3.950893359391756, + "tokens_seen": 491126784 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042987963891675027, + "loss": 1.5859, + "theoretical_loss": 3.9508368139313275, + "tokens_seen": 491192320 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042986960882647945, + "loss": 1.59, + "theoretical_loss": 3.950780278126937, + "tokens_seen": 491257856 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042985957873620863, + "loss": 1.4891, + "theoretical_loss": 3.9507237519756484, + "tokens_seen": 491323392 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004298495486459378, + "loss": 1.3546, + "theoretical_loss": 3.9506672354745263, + "tokens_seen": 491388928 + }, + { + "epoch": 0.15, + "learning_rate": 0.000429839518555667, + "loss": 1.5268, + "theoretical_loss": 3.950610728620636, + "tokens_seen": 491454464 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 188044, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5719361305236816, + "objective/train/theoretical_loss": 3.950554231411046, + "objective/train/tokens_used": 491520000, + "theoretical_loss": 3.950554231411046, + "tokens_seen": 491520000 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042982948846539623, + "loss": 1.5094, + "theoretical_loss": 3.950554231411046, + "tokens_seen": 491520000 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042981945837512536, + "loss": 1.4953, + "theoretical_loss": 3.9504977438428233, + "tokens_seen": 491585536 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004298094282848546, + "loss": 1.5596, + "theoretical_loss": 3.950441265913038, + "tokens_seen": 491651072 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004297993981945837, + "loss": 1.4946, + "theoretical_loss": 3.950384797618761, + "tokens_seen": 491716608 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042978936810431296, + "loss": 1.4908, + "theoretical_loss": 3.950328338957064, + "tokens_seen": 491782144 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042977933801404214, + "loss": 1.5107, + "theoretical_loss": 3.950271889925022, + "tokens_seen": 491847680 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004297693079237713, + "loss": 1.4673, + "theoretical_loss": 3.9502154505197087, + "tokens_seen": 491913216 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004297592778335005, + "loss": 1.5683, + "theoretical_loss": 3.9501590207382, + "tokens_seen": 491978752 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042974924774322973, + "loss": 1.5783, + "theoretical_loss": 3.9501026005775732, + "tokens_seen": 492044288 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042973921765295886, + "loss": 1.6245, + "theoretical_loss": 3.9500461900349078, + "tokens_seen": 492109824 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004297291875626881, + "loss": 1.5327, + "theoretical_loss": 3.949989789107283, + "tokens_seen": 492175360 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004297191574724172, + "loss": 1.6717, + "theoretical_loss": 3.9499333977917805, + "tokens_seen": 492240896 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042970912738214646, + "loss": 1.5637, + "theoretical_loss": 3.949877016085482, + "tokens_seen": 492306432 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042969909729187564, + "loss": 1.4612, + "theoretical_loss": 3.949820643985472, + "tokens_seen": 492371968 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004296890672016048, + "loss": 1.5427, + "theoretical_loss": 3.9497642814888345, + "tokens_seen": 492437504 + }, + { + "epoch": 0.15, + "learning_rate": 0.000429679037111334, + "loss": 1.5934, + "theoretical_loss": 3.949707928592657, + "tokens_seen": 492503040 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004296690070210632, + "loss": 1.514, + "theoretical_loss": 3.949651585294026, + "tokens_seen": 492568576 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042965897693079236, + "loss": 1.5077, + "theoretical_loss": 3.9495952515900306, + "tokens_seen": 492634112 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004296489468405216, + "loss": 1.5718, + "theoretical_loss": 3.9495389274777613, + "tokens_seen": 492699648 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004296389167502507, + "loss": 1.4902, + "theoretical_loss": 3.9494826129543084, + "tokens_seen": 492765184 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042962888665997996, + "loss": 1.5636, + "theoretical_loss": 3.9494263080167658, + "tokens_seen": 492830720 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004296188565697091, + "loss": 1.5331, + "theoretical_loss": 3.949370012662227, + "tokens_seen": 492896256 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004296088264794383, + "loss": 1.5929, + "theoretical_loss": 3.9493137268877865, + "tokens_seen": 492961792 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004295987963891675, + "loss": 1.5688, + "theoretical_loss": 3.9492574506905416, + "tokens_seen": 493027328 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004295887662988967, + "loss": 1.4179, + "theoretical_loss": 3.949201184067589, + "tokens_seen": 493092864 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042957873620862587, + "loss": 1.5757, + "theoretical_loss": 3.949144927016029, + "tokens_seen": 493158400 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004295687061183551, + "loss": 1.4611, + "theoretical_loss": 3.9490886795329603, + "tokens_seen": 493223936 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042955867602808423, + "loss": 1.5117, + "theoretical_loss": 3.9490324416154854, + "tokens_seen": 493289472 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042954864593781347, + "loss": 1.5496, + "theoretical_loss": 3.9489762132607065, + "tokens_seen": 493355008 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004295386158475426, + "loss": 1.5777, + "theoretical_loss": 3.948919994465728, + "tokens_seen": 493420544 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042952858575727183, + "loss": 1.4946, + "theoretical_loss": 3.9488637852276547, + "tokens_seen": 493486080 + }, + { + "epoch": 0.15, + "learning_rate": 0.000429518555667001, + "loss": 1.5524, + "theoretical_loss": 3.9488075855435936, + "tokens_seen": 493551616 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004295085255767302, + "loss": 1.5323, + "theoretical_loss": 3.9487513954106523, + "tokens_seen": 493617152 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042949849548645937, + "loss": 1.5827, + "theoretical_loss": 3.94869521482594, + "tokens_seen": 493682688 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042948846539618855, + "loss": 1.5876, + "theoretical_loss": 3.9486390437865664, + "tokens_seen": 493748224 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042947843530591773, + "loss": 1.5491, + "theoretical_loss": 3.948582882289644, + "tokens_seen": 493813760 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042946840521564697, + "loss": 1.525, + "theoretical_loss": 3.9485267303322846, + "tokens_seen": 493879296 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004294583751253761, + "loss": 1.4661, + "theoretical_loss": 3.9484705879116033, + "tokens_seen": 493944832 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042944834503510533, + "loss": 1.4414, + "theoretical_loss": 3.9484144550247144, + "tokens_seen": 494010368 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004294383149448345, + "loss": 1.5223, + "theoretical_loss": 3.948358331668735, + "tokens_seen": 494075904 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004294282848545637, + "loss": 1.5239, + "theoretical_loss": 3.9483022178407827, + "tokens_seen": 494141440 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042941825476429293, + "loss": 1.5068, + "theoretical_loss": 3.9482461135379774, + "tokens_seen": 494206976 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042940822467402206, + "loss": 1.5475, + "theoretical_loss": 3.9481900187574386, + "tokens_seen": 494272512 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004293981945837513, + "loss": 1.5818, + "theoretical_loss": 3.948133933496288, + "tokens_seen": 494338048 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042938816449348047, + "loss": 1.4622, + "theoretical_loss": 3.948077857751649, + "tokens_seen": 494403584 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042937813440320965, + "loss": 1.4642, + "theoretical_loss": 3.948021791520645, + "tokens_seen": 494469120 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042936810431293883, + "loss": 1.4333, + "theoretical_loss": 3.9479657348004014, + "tokens_seen": 494534656 + }, + { + "epoch": 0.15, + "learning_rate": 0.000429358074222668, + "loss": 1.6327, + "theoretical_loss": 3.9479096875880453, + "tokens_seen": 494600192 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004293480441323972, + "loss": 1.4009, + "theoretical_loss": 3.947853649880704, + "tokens_seen": 494665728 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042933801404212643, + "loss": 1.5626, + "theoretical_loss": 3.947797621675507, + "tokens_seen": 494731264 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 189198, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.456634759902954, + "objective/train/theoretical_loss": 3.947741602969585, + "objective/train/tokens_used": 494796800, + "theoretical_loss": 3.947741602969585, + "tokens_seen": 494796800 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042932798395185556, + "loss": 1.489, + "theoretical_loss": 3.947741602969585, + "tokens_seen": 494796800 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004293179538615848, + "loss": 1.6056, + "theoretical_loss": 3.9476855937600686, + "tokens_seen": 494862336 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004293079237713139, + "loss": 1.4814, + "theoretical_loss": 3.9476295940440913, + "tokens_seen": 494927872 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042929789368104316, + "loss": 1.6332, + "theoretical_loss": 3.9475736038187867, + "tokens_seen": 494993408 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042928786359077234, + "loss": 1.6519, + "theoretical_loss": 3.947517623081291, + "tokens_seen": 495058944 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004292778335005015, + "loss": 1.5952, + "theoretical_loss": 3.9474616518287395, + "tokens_seen": 495124480 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004292678034102307, + "loss": 1.5186, + "theoretical_loss": 3.9474056900582717, + "tokens_seen": 495190016 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042925777331995993, + "loss": 1.5758, + "theoretical_loss": 3.9473497377670244, + "tokens_seen": 495255552 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042924774322968906, + "loss": 1.4303, + "theoretical_loss": 3.94729379495214, + "tokens_seen": 495321088 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004292377131394183, + "loss": 1.5112, + "theoretical_loss": 3.947237861610759, + "tokens_seen": 495386624 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004292276830491474, + "loss": 1.5299, + "theoretical_loss": 3.9471819377400243, + "tokens_seen": 495452160 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042921765295887666, + "loss": 1.5083, + "theoretical_loss": 3.94712602333708, + "tokens_seen": 495517696 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042920762286860584, + "loss": 1.6455, + "theoretical_loss": 3.9470701183990715, + "tokens_seen": 495583232 + }, + { + "epoch": 0.15, + "learning_rate": 0.000429197592778335, + "loss": 1.4753, + "theoretical_loss": 3.947014222923145, + "tokens_seen": 495648768 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004291875626880642, + "loss": 1.4816, + "theoretical_loss": 3.946958336906449, + "tokens_seen": 495714304 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004291775325977934, + "loss": 1.4827, + "theoretical_loss": 3.9469024603461307, + "tokens_seen": 495779840 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042916750250752256, + "loss": 1.4703, + "theoretical_loss": 3.946846593239342, + "tokens_seen": 495845376 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004291574724172518, + "loss": 1.5564, + "theoretical_loss": 3.946790735583234, + "tokens_seen": 495910912 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004291474423269809, + "loss": 1.4951, + "theoretical_loss": 3.9467348873749586, + "tokens_seen": 495976448 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042913741223671016, + "loss": 1.4471, + "theoretical_loss": 3.9466790486116707, + "tokens_seen": 496041984 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004291273821464393, + "loss": 1.5253, + "theoretical_loss": 3.946623219290525, + "tokens_seen": 496107520 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004291173520561685, + "loss": 1.495, + "theoretical_loss": 3.9465673994086776, + "tokens_seen": 496173056 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004291073219658977, + "loss": 1.5005, + "theoretical_loss": 3.9465115889632862, + "tokens_seen": 496238592 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004290972918756269, + "loss": 1.6499, + "theoretical_loss": 3.9464557879515105, + "tokens_seen": 496304128 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042908726178535607, + "loss": 1.4214, + "theoretical_loss": 3.946399996370509, + "tokens_seen": 496369664 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004290772316950853, + "loss": 1.4612, + "theoretical_loss": 3.9463442142174445, + "tokens_seen": 496435200 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042906720160481443, + "loss": 1.5803, + "theoretical_loss": 3.9462884414894788, + "tokens_seen": 496500736 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042905717151454367, + "loss": 1.6216, + "theoretical_loss": 3.946232678183775, + "tokens_seen": 496566272 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004290471414242728, + "loss": 1.4221, + "theoretical_loss": 3.9461769242974998, + "tokens_seen": 496631808 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042903711133400203, + "loss": 1.551, + "theoretical_loss": 3.946121179827818, + "tokens_seen": 496697344 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004290270812437312, + "loss": 1.5776, + "theoretical_loss": 3.9460654447718975, + "tokens_seen": 496762880 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004290170511534604, + "loss": 1.55, + "theoretical_loss": 3.9460097191269066, + "tokens_seen": 496828416 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042900702106318957, + "loss": 1.5037, + "theoretical_loss": 3.9459540028900157, + "tokens_seen": 496893952 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042899699097291875, + "loss": 1.5311, + "theoretical_loss": 3.945898296058396, + "tokens_seen": 496959488 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042898696088264793, + "loss": 1.4029, + "theoretical_loss": 3.9458425986292194, + "tokens_seen": 497025024 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042897693079237717, + "loss": 1.5255, + "theoretical_loss": 3.9457869105996597, + "tokens_seen": 497090560 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004289669007021063, + "loss": 1.5632, + "theoretical_loss": 3.9457312319668914, + "tokens_seen": 497156096 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042895687061183553, + "loss": 1.4143, + "theoretical_loss": 3.9456755627280904, + "tokens_seen": 497221632 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004289468405215647, + "loss": 1.6069, + "theoretical_loss": 3.945619902880434, + "tokens_seen": 497287168 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004289368104312939, + "loss": 1.5307, + "theoretical_loss": 3.945564252421102, + "tokens_seen": 497352704 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004289267803410231, + "loss": 1.5571, + "theoretical_loss": 3.9455086113472717, + "tokens_seen": 497418240 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042891675025075226, + "loss": 1.5988, + "theoretical_loss": 3.9454529796561255, + "tokens_seen": 497483776 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042890672016048144, + "loss": 1.5568, + "theoretical_loss": 3.9453973573448455, + "tokens_seen": 497549312 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042889669007021067, + "loss": 1.4988, + "theoretical_loss": 3.945341744410615, + "tokens_seen": 497614848 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004288866599799398, + "loss": 1.5285, + "theoretical_loss": 3.9452861408506172, + "tokens_seen": 497680384 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042887662988966903, + "loss": 1.5539, + "theoretical_loss": 3.945230546662039, + "tokens_seen": 497745920 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042886659979939816, + "loss": 1.5564, + "theoretical_loss": 3.945174961842068, + "tokens_seen": 497811456 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004288565697091274, + "loss": 1.5393, + "theoretical_loss": 3.9451193863878906, + "tokens_seen": 497876992 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004288465396188566, + "loss": 1.4425, + "theoretical_loss": 3.945063820296698, + "tokens_seen": 497942528 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042883650952858576, + "loss": 1.4416, + "theoretical_loss": 3.9450082635656796, + "tokens_seen": 498008064 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 190526, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5104748010635376, + "objective/train/theoretical_loss": 3.9449527161920273, + "objective/train/tokens_used": 498073600, + "theoretical_loss": 3.9449527161920273, + "tokens_seen": 498073600 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042882647943831494, + "loss": 1.4944, + "theoretical_loss": 3.9449527161920273, + "tokens_seen": 498073600 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004288164493480441, + "loss": 1.4751, + "theoretical_loss": 3.944897178172935, + "tokens_seen": 498139136 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004288064192577733, + "loss": 1.4287, + "theoretical_loss": 3.9448416495055962, + "tokens_seen": 498204672 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042879638916750254, + "loss": 1.4626, + "theoretical_loss": 3.9447861301872065, + "tokens_seen": 498270208 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042878635907723166, + "loss": 1.565, + "theoretical_loss": 3.944730620214962, + "tokens_seen": 498335744 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004287763289869609, + "loss": 1.4718, + "theoretical_loss": 3.944675119586061, + "tokens_seen": 498401280 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004287662988966901, + "loss": 1.5671, + "theoretical_loss": 3.944619628297704, + "tokens_seen": 498466816 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042875626880641926, + "loss": 1.5782, + "theoretical_loss": 3.944564146347089, + "tokens_seen": 498532352 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042874623871614844, + "loss": 1.507, + "theoretical_loss": 3.9445086737314186, + "tokens_seen": 498597888 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004287362086258776, + "loss": 1.5432, + "theoretical_loss": 3.9444532104478958, + "tokens_seen": 498663424 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004287261785356068, + "loss": 1.4779, + "theoretical_loss": 3.9443977564937236, + "tokens_seen": 498728960 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042871614844533604, + "loss": 1.5422, + "theoretical_loss": 3.9443423118661074, + "tokens_seen": 498794496 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042870611835506517, + "loss": 1.4958, + "theoretical_loss": 3.944286876562254, + "tokens_seen": 498860032 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004286960882647944, + "loss": 1.4761, + "theoretical_loss": 3.9442314505793705, + "tokens_seen": 498925568 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042868605817452353, + "loss": 1.5499, + "theoretical_loss": 3.944176033914666, + "tokens_seen": 498991104 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042867602808425276, + "loss": 1.484, + "theoretical_loss": 3.9441206265653497, + "tokens_seen": 499056640 + }, + { + "epoch": 0.15, + "learning_rate": 0.000428665997993982, + "loss": 1.5674, + "theoretical_loss": 3.9440652285286335, + "tokens_seen": 499122176 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004286559679037111, + "loss": 1.4706, + "theoretical_loss": 3.9440098398017294, + "tokens_seen": 499187712 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042864593781344036, + "loss": 1.4965, + "theoretical_loss": 3.9439544603818506, + "tokens_seen": 499253248 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004286359077231695, + "loss": 1.508, + "theoretical_loss": 3.9438990902662128, + "tokens_seen": 499318784 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004286258776328987, + "loss": 1.5418, + "theoretical_loss": 3.9438437294520305, + "tokens_seen": 499384320 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004286158475426279, + "loss": 1.5202, + "theoretical_loss": 3.943788377936522, + "tokens_seen": 499449856 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004286058174523571, + "loss": 1.4409, + "theoretical_loss": 3.943733035716905, + "tokens_seen": 499515392 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042859578736208627, + "loss": 1.6078, + "theoretical_loss": 3.9436777027903993, + "tokens_seen": 499580928 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004285857572718155, + "loss": 1.5218, + "theoretical_loss": 3.9436223791542258, + "tokens_seen": 499646464 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042857572718154463, + "loss": 1.3979, + "theoretical_loss": 3.943567064805606, + "tokens_seen": 499712000 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042856569709127387, + "loss": 1.483, + "theoretical_loss": 3.943511759741763, + "tokens_seen": 499777536 + }, + { + "epoch": 0.15, + "learning_rate": 0.000428555667001003, + "loss": 1.5711, + "theoretical_loss": 3.9434564639599214, + "tokens_seen": 499843072 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042854563691073223, + "loss": 1.4471, + "theoretical_loss": 3.9434011774573063, + "tokens_seen": 499908608 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004285356068204614, + "loss": 1.5177, + "theoretical_loss": 3.9433459002311446, + "tokens_seen": 499974144 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004285255767301906, + "loss": 1.5787, + "theoretical_loss": 3.9432906322786643, + "tokens_seen": 500039680 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042851554663991977, + "loss": 1.5198, + "theoretical_loss": 3.943235373597094, + "tokens_seen": 500105216 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042850551654964895, + "loss": 1.4214, + "theoretical_loss": 3.943180124183665, + "tokens_seen": 500170752 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042849548645937813, + "loss": 1.5242, + "theoretical_loss": 3.9431248840356075, + "tokens_seen": 500236288 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042848545636910737, + "loss": 1.5235, + "theoretical_loss": 3.943069653150155, + "tokens_seen": 500301824 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004284754262788365, + "loss": 1.5057, + "theoretical_loss": 3.9430144315245412, + "tokens_seen": 500367360 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042846539618856573, + "loss": 1.4871, + "theoretical_loss": 3.942959219156, + "tokens_seen": 500432896 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004284553660982949, + "loss": 1.5205, + "theoretical_loss": 3.942904016041769, + "tokens_seen": 500498432 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004284453360080241, + "loss": 1.5161, + "theoretical_loss": 3.9428488221790854, + "tokens_seen": 500563968 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004284353059177533, + "loss": 1.462, + "theoretical_loss": 3.9427936375651877, + "tokens_seen": 500629504 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042842527582748246, + "loss": 1.4635, + "theoretical_loss": 3.9427384621973154, + "tokens_seen": 500695040 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042841524573721164, + "loss": 1.5904, + "theoretical_loss": 3.9426832960727087, + "tokens_seen": 500760576 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042840521564694087, + "loss": 1.4321, + "theoretical_loss": 3.9426281391886113, + "tokens_seen": 500826112 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042839518555667, + "loss": 1.6003, + "theoretical_loss": 3.9425729915422654, + "tokens_seen": 500891648 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042838515546639923, + "loss": 1.5999, + "theoretical_loss": 3.942517853130916, + "tokens_seen": 500957184 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042837512537612836, + "loss": 1.5573, + "theoretical_loss": 3.9424627239518086, + "tokens_seen": 501022720 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004283650952858576, + "loss": 1.5021, + "theoretical_loss": 3.9424076040021903, + "tokens_seen": 501088256 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004283550651955868, + "loss": 1.4853, + "theoretical_loss": 3.9423524932793086, + "tokens_seen": 501153792 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042834503510531596, + "loss": 1.5495, + "theoretical_loss": 3.9422973917804134, + "tokens_seen": 501219328 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042833500501504514, + "loss": 1.4942, + "theoretical_loss": 3.9422422995027544, + "tokens_seen": 501284864 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 191761, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4874608516693115, + "objective/train/theoretical_loss": 3.9421872164435836, + "objective/train/tokens_used": 501350400, + "theoretical_loss": 3.9421872164435836, + "tokens_seen": 501350400 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004283249749247743, + "loss": 1.4673, + "theoretical_loss": 3.9421872164435836, + "tokens_seen": 501350400 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004283149448345035, + "loss": 1.4192, + "theoretical_loss": 3.942132142600154, + "tokens_seen": 501415936 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042830491474423274, + "loss": 1.5013, + "theoretical_loss": 3.942077077969719, + "tokens_seen": 501481472 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042829488465396186, + "loss": 1.4666, + "theoretical_loss": 3.9420220225495335, + "tokens_seen": 501547008 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004282848545636911, + "loss": 1.5143, + "theoretical_loss": 3.9419669763368548, + "tokens_seen": 501612544 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004282748244734203, + "loss": 1.5658, + "theoretical_loss": 3.9419119393289397, + "tokens_seen": 501678080 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042826479438314946, + "loss": 1.5847, + "theoretical_loss": 3.941856911523047, + "tokens_seen": 501743616 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042825476429287864, + "loss": 1.5976, + "theoretical_loss": 3.9418018929164362, + "tokens_seen": 501809152 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004282447342026078, + "loss": 1.4842, + "theoretical_loss": 3.9417468835063687, + "tokens_seen": 501874688 + }, + { + "epoch": 0.15, + "learning_rate": 0.000428234704112337, + "loss": 1.555, + "theoretical_loss": 3.9416918832901064, + "tokens_seen": 501940224 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042822467402206624, + "loss": 1.5839, + "theoretical_loss": 3.9416368922649134, + "tokens_seen": 502005760 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042821464393179537, + "loss": 1.6551, + "theoretical_loss": 3.9415819104280527, + "tokens_seen": 502071296 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004282046138415246, + "loss": 1.5104, + "theoretical_loss": 3.941526937776791, + "tokens_seen": 502136832 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042819458375125373, + "loss": 1.3889, + "theoretical_loss": 3.941471974308395, + "tokens_seen": 502202368 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042818455366098296, + "loss": 1.4883, + "theoretical_loss": 3.9414170200201326, + "tokens_seen": 502267904 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042817452357071215, + "loss": 1.617, + "theoretical_loss": 3.9413620749092733, + "tokens_seen": 502333440 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004281644934804413, + "loss": 1.4119, + "theoretical_loss": 3.9413071389730874, + "tokens_seen": 502398976 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004281544633901705, + "loss": 1.5485, + "theoretical_loss": 3.941252212208846, + "tokens_seen": 502464512 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004281444332998997, + "loss": 1.4655, + "theoretical_loss": 3.941197294613822, + "tokens_seen": 502530048 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042813440320962887, + "loss": 1.5093, + "theoretical_loss": 3.9411423861852892, + "tokens_seen": 502595584 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004281243731193581, + "loss": 1.5298, + "theoretical_loss": 3.941087486920523, + "tokens_seen": 502661120 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042811434302908723, + "loss": 1.5706, + "theoretical_loss": 3.9410325968167994, + "tokens_seen": 502726656 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042810431293881647, + "loss": 1.4345, + "theoretical_loss": 3.9409777158713952, + "tokens_seen": 502792192 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042809428284854565, + "loss": 1.4047, + "theoretical_loss": 3.9409228440815895, + "tokens_seen": 502857728 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042808425275827483, + "loss": 1.5669, + "theoretical_loss": 3.9408679814446623, + "tokens_seen": 502923264 + }, + { + "epoch": 0.15, + "learning_rate": 0.000428074222668004, + "loss": 1.3521, + "theoretical_loss": 3.9408131279578935, + "tokens_seen": 502988800 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004280641925777332, + "loss": 1.5385, + "theoretical_loss": 3.9407582836185657, + "tokens_seen": 503054336 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004280541624874624, + "loss": 1.5178, + "theoretical_loss": 3.940703448423962, + "tokens_seen": 503119872 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004280441323971916, + "loss": 1.5277, + "theoretical_loss": 3.9406486223713664, + "tokens_seen": 503185408 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042803410230692074, + "loss": 1.489, + "theoretical_loss": 3.940593805458065, + "tokens_seen": 503250944 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042802407221664997, + "loss": 1.6075, + "theoretical_loss": 3.940538997681344, + "tokens_seen": 503316480 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004280140421263791, + "loss": 1.4608, + "theoretical_loss": 3.9404841990384916, + "tokens_seen": 503382016 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042800401203610833, + "loss": 1.5282, + "theoretical_loss": 3.940429409526796, + "tokens_seen": 503447552 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004279939819458375, + "loss": 1.5333, + "theoretical_loss": 3.940374629143548, + "tokens_seen": 503513088 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004279839518555667, + "loss": 1.5705, + "theoretical_loss": 3.940319857886039, + "tokens_seen": 503578624 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004279739217652959, + "loss": 1.5461, + "theoretical_loss": 3.9402650957515606, + "tokens_seen": 503644160 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004279638916750251, + "loss": 1.5051, + "theoretical_loss": 3.940210342737407, + "tokens_seen": 503709696 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042795386158475424, + "loss": 1.5685, + "theoretical_loss": 3.940155598840873, + "tokens_seen": 503775232 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004279438314944835, + "loss": 1.4628, + "theoretical_loss": 3.9401008640592545, + "tokens_seen": 503840768 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004279338014042126, + "loss": 1.5061, + "theoretical_loss": 3.940046138389848, + "tokens_seen": 503906304 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042792377131394184, + "loss": 1.5224, + "theoretical_loss": 3.9399914218299523, + "tokens_seen": 503971840 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042791374122367107, + "loss": 1.5465, + "theoretical_loss": 3.9399367143768664, + "tokens_seen": 504037376 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004279037111334002, + "loss": 1.4507, + "theoretical_loss": 3.939882016027891, + "tokens_seen": 504102912 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042789368104312943, + "loss": 1.5567, + "theoretical_loss": 3.939827326780328, + "tokens_seen": 504168448 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042788365095285856, + "loss": 1.4564, + "theoretical_loss": 3.9397726466314795, + "tokens_seen": 504233984 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004278736208625878, + "loss": 1.3785, + "theoretical_loss": 3.93971797557865, + "tokens_seen": 504299520 + }, + { + "epoch": 0.15, + "learning_rate": 0.000427863590772317, + "loss": 1.5324, + "theoretical_loss": 3.939663313619145, + "tokens_seen": 504365056 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042785356068204616, + "loss": 1.4675, + "theoretical_loss": 3.9396086607502694, + "tokens_seen": 504430592 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042784353059177534, + "loss": 1.5151, + "theoretical_loss": 3.939554016969332, + "tokens_seen": 504496128 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004278335005015045, + "loss": 1.5745, + "theoretical_loss": 3.939499382273641, + "tokens_seen": 504561664 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 192915, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4823546409606934, + "objective/train/theoretical_loss": 3.939444756660505, + "objective/train/tokens_used": 504627200, + "theoretical_loss": 3.939444756660505, + "tokens_seen": 504627200 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004278234704112337, + "loss": 1.4905, + "theoretical_loss": 3.939444756660505, + "tokens_seen": 504627200 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042781344032096294, + "loss": 1.3943, + "theoretical_loss": 3.9393901401272364, + "tokens_seen": 504692736 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042780341023069206, + "loss": 1.4171, + "theoretical_loss": 3.939335532671146, + "tokens_seen": 504758272 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004277933801404213, + "loss": 1.4767, + "theoretical_loss": 3.939280934289548, + "tokens_seen": 504823808 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004277833500501505, + "loss": 1.4041, + "theoretical_loss": 3.9392263449797564, + "tokens_seen": 504889344 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042777331995987966, + "loss": 1.4685, + "theoretical_loss": 3.939171764739086, + "tokens_seen": 504954880 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042776328986960884, + "loss": 1.5057, + "theoretical_loss": 3.9391171935648535, + "tokens_seen": 505020416 + }, + { + "epoch": 0.15, + "learning_rate": 0.000427753259779338, + "loss": 1.5298, + "theoretical_loss": 3.9390626314543766, + "tokens_seen": 505085952 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004277432296890672, + "loss": 1.4244, + "theoretical_loss": 3.939008078404975, + "tokens_seen": 505151488 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042773319959879644, + "loss": 1.5736, + "theoretical_loss": 3.938953534413968, + "tokens_seen": 505217024 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042772316950852557, + "loss": 1.5263, + "theoretical_loss": 3.9388989994786767, + "tokens_seen": 505282560 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004277131394182548, + "loss": 1.5656, + "theoretical_loss": 3.938844473596423, + "tokens_seen": 505348096 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042770310932798393, + "loss": 1.5451, + "theoretical_loss": 3.9387899567645315, + "tokens_seen": 505413632 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042769307923771316, + "loss": 1.4815, + "theoretical_loss": 3.9387354489803252, + "tokens_seen": 505479168 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042768304914744235, + "loss": 1.4683, + "theoretical_loss": 3.938680950241131, + "tokens_seen": 505544704 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042767301905717153, + "loss": 1.5588, + "theoretical_loss": 3.938626460544275, + "tokens_seen": 505610240 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004276629889669007, + "loss": 1.602, + "theoretical_loss": 3.9385719798870857, + "tokens_seen": 505675776 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004276529588766299, + "loss": 1.3982, + "theoretical_loss": 3.9385175082668926, + "tokens_seen": 505741312 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042764292878635907, + "loss": 1.5214, + "theoretical_loss": 3.9384630456810243, + "tokens_seen": 505806848 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004276328986960883, + "loss": 1.5226, + "theoretical_loss": 3.9384085921268133, + "tokens_seen": 505872384 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042762286860581743, + "loss": 1.4783, + "theoretical_loss": 3.9383541476015926, + "tokens_seen": 505937920 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042761283851554667, + "loss": 1.5071, + "theoretical_loss": 3.9382997121026944, + "tokens_seen": 506003456 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042760280842527585, + "loss": 1.5019, + "theoretical_loss": 3.938245285627455, + "tokens_seen": 506068992 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042759277833500503, + "loss": 1.4816, + "theoretical_loss": 3.9381908681732085, + "tokens_seen": 506134528 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004275827482447342, + "loss": 1.5018, + "theoretical_loss": 3.9381364597372936, + "tokens_seen": 506200064 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004275727181544634, + "loss": 1.5401, + "theoretical_loss": 3.938082060317048, + "tokens_seen": 506265600 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004275626880641926, + "loss": 1.5707, + "theoretical_loss": 3.9380276699098102, + "tokens_seen": 506331136 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004275526579739218, + "loss": 1.4647, + "theoretical_loss": 3.9379732885129215, + "tokens_seen": 506396672 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042754262788365094, + "loss": 1.4488, + "theoretical_loss": 3.9379189161237234, + "tokens_seen": 506462208 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042753259779338017, + "loss": 1.5315, + "theoretical_loss": 3.9378645527395575, + "tokens_seen": 506527744 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004275225677031093, + "loss": 1.526, + "theoretical_loss": 3.937810198357769, + "tokens_seen": 506593280 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042751253761283853, + "loss": 1.5015, + "theoretical_loss": 3.9377558529757026, + "tokens_seen": 506658816 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004275025075225677, + "loss": 1.4812, + "theoretical_loss": 3.937701516590703, + "tokens_seen": 506724352 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004274924774322969, + "loss": 1.5207, + "theoretical_loss": 3.9376471892001197, + "tokens_seen": 506789888 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004274824473420261, + "loss": 1.532, + "theoretical_loss": 3.9375928708012986, + "tokens_seen": 506855424 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004274724172517553, + "loss": 1.5504, + "theoretical_loss": 3.9375385613915905, + "tokens_seen": 506920960 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042746238716148444, + "loss": 1.4715, + "theoretical_loss": 3.937484260968346, + "tokens_seen": 506986496 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004274523570712137, + "loss": 1.5269, + "theoretical_loss": 3.9374299695289157, + "tokens_seen": 507052032 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004274423269809428, + "loss": 1.5456, + "theoretical_loss": 3.9373756870706536, + "tokens_seen": 507117568 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042743229689067204, + "loss": 1.6107, + "theoretical_loss": 3.937321413590913, + "tokens_seen": 507183104 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004274222668004012, + "loss": 1.5128, + "theoretical_loss": 3.9372671490870492, + "tokens_seen": 507248640 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004274122367101304, + "loss": 1.3963, + "theoretical_loss": 3.937212893556418, + "tokens_seen": 507314176 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004274022066198596, + "loss": 1.5733, + "theoretical_loss": 3.9371586469963775, + "tokens_seen": 507379712 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042739217652958876, + "loss": 1.3881, + "theoretical_loss": 3.937104409404285, + "tokens_seen": 507445248 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042738214643931794, + "loss": 1.4711, + "theoretical_loss": 3.9370501807775007, + "tokens_seen": 507510784 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004273721163490472, + "loss": 1.4827, + "theoretical_loss": 3.936995961113385, + "tokens_seen": 507576320 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004273620862587763, + "loss": 1.372, + "theoretical_loss": 3.9369417504093, + "tokens_seen": 507641856 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042735205616850554, + "loss": 1.4085, + "theoretical_loss": 3.936887548662608, + "tokens_seen": 507707392 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042734202607823467, + "loss": 1.4199, + "theoretical_loss": 3.9368333558706734, + "tokens_seen": 507772928 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004273319959879639, + "loss": 1.4539, + "theoretical_loss": 3.936779172030861, + "tokens_seen": 507838464 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 194112, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5227311849594116, + "objective/train/theoretical_loss": 3.936724997140538, + "objective/train/tokens_used": 507904000, + "theoretical_loss": 3.936724997140538, + "tokens_seen": 507904000 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004273219658976931, + "loss": 1.5712, + "theoretical_loss": 3.936724997140538, + "tokens_seen": 507904000 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042731193580742226, + "loss": 1.501, + "theoretical_loss": 3.9366708311970706, + "tokens_seen": 507969536 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042730190571715145, + "loss": 1.5546, + "theoretical_loss": 3.9366166741978272, + "tokens_seen": 508035072 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004272918756268807, + "loss": 1.5818, + "theoretical_loss": 3.936562526140178, + "tokens_seen": 508100608 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004272818455366098, + "loss": 1.5193, + "theoretical_loss": 3.936508387021494, + "tokens_seen": 508166144 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042727181544633904, + "loss": 1.4691, + "theoretical_loss": 3.9364542568391454, + "tokens_seen": 508231680 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042726178535606817, + "loss": 1.4639, + "theoretical_loss": 3.9364001355905067, + "tokens_seen": 508297216 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004272517552657974, + "loss": 1.4531, + "theoretical_loss": 3.936346023272952, + "tokens_seen": 508362752 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004272417251755266, + "loss": 1.4026, + "theoretical_loss": 3.936291919883855, + "tokens_seen": 508428288 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042723169508525577, + "loss": 1.5577, + "theoretical_loss": 3.9362378254205925, + "tokens_seen": 508493824 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042722166499498495, + "loss": 1.4961, + "theoretical_loss": 3.936183739880543, + "tokens_seen": 508559360 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042721163490471413, + "loss": 1.5727, + "theoretical_loss": 3.9361296632610827, + "tokens_seen": 508624896 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004272016048144433, + "loss": 1.5452, + "theoretical_loss": 3.9360755955595934, + "tokens_seen": 508690432 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042719157472417255, + "loss": 1.5037, + "theoretical_loss": 3.936021536773455, + "tokens_seen": 508755968 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004271815446339017, + "loss": 1.5072, + "theoretical_loss": 3.9359674869000485, + "tokens_seen": 508821504 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004271715145436309, + "loss": 1.4427, + "theoretical_loss": 3.9359134459367575, + "tokens_seen": 508887040 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004271614844533601, + "loss": 1.4752, + "theoretical_loss": 3.9358594138809657, + "tokens_seen": 508952576 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042715145436308927, + "loss": 1.598, + "theoretical_loss": 3.9358053907300588, + "tokens_seen": 509018112 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004271414242728185, + "loss": 1.4259, + "theoretical_loss": 3.935751376481422, + "tokens_seen": 509083648 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042713139418254763, + "loss": 1.4807, + "theoretical_loss": 3.9356973711324432, + "tokens_seen": 509149184 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042712136409227687, + "loss": 1.4751, + "theoretical_loss": 3.9356433746805113, + "tokens_seen": 509214720 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042711133400200605, + "loss": 1.5598, + "theoretical_loss": 3.9355893871230143, + "tokens_seen": 509280256 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042710130391173523, + "loss": 1.5758, + "theoretical_loss": 3.935535408457344, + "tokens_seen": 509345792 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004270912738214644, + "loss": 1.5298, + "theoretical_loss": 3.9354814386808923, + "tokens_seen": 509411328 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004270812437311936, + "loss": 1.4981, + "theoretical_loss": 3.935427477791051, + "tokens_seen": 509476864 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004270712136409228, + "loss": 1.3967, + "theoretical_loss": 3.9353735257852147, + "tokens_seen": 509542400 + }, + { + "epoch": 0.15, + "learning_rate": 0.000427061183550652, + "loss": 1.5931, + "theoretical_loss": 3.935319582660778, + "tokens_seen": 509607936 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042705115346038114, + "loss": 1.5288, + "theoretical_loss": 3.935265648415138, + "tokens_seen": 509673472 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042704112337011037, + "loss": 1.4967, + "theoretical_loss": 3.9352117230456902, + "tokens_seen": 509739008 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004270310932798395, + "loss": 1.5754, + "theoretical_loss": 3.9351578065498343, + "tokens_seen": 509804544 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042702106318956873, + "loss": 1.513, + "theoretical_loss": 3.935103898924969, + "tokens_seen": 509870080 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004270110330992979, + "loss": 1.463, + "theoretical_loss": 3.935050000168495, + "tokens_seen": 509935616 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004270010030090271, + "loss": 1.4096, + "theoretical_loss": 3.9349961102778135, + "tokens_seen": 510001152 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004269909729187563, + "loss": 1.569, + "theoretical_loss": 3.934942229250328, + "tokens_seen": 510066688 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004269809428284855, + "loss": 1.472, + "theoretical_loss": 3.934888357083442, + "tokens_seen": 510132224 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042697091273821464, + "loss": 1.6128, + "theoretical_loss": 3.9348344937745594, + "tokens_seen": 510197760 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004269608826479439, + "loss": 1.4777, + "theoretical_loss": 3.9347806393210876, + "tokens_seen": 510263296 + }, + { + "epoch": 0.15, + "learning_rate": 0.000426950852557673, + "loss": 1.5684, + "theoretical_loss": 3.934726793720433, + "tokens_seen": 510328832 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042694082246740224, + "loss": 1.5274, + "theoretical_loss": 3.9346729569700036, + "tokens_seen": 510394368 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004269307923771314, + "loss": 1.5371, + "theoretical_loss": 3.934619129067209, + "tokens_seen": 510459904 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004269207622868606, + "loss": 1.5791, + "theoretical_loss": 3.934565310009459, + "tokens_seen": 510525440 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004269107321965898, + "loss": 1.6458, + "theoretical_loss": 3.9345114997941657, + "tokens_seen": 510590976 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042690070210631896, + "loss": 1.5554, + "theoretical_loss": 3.934457698418741, + "tokens_seen": 510656512 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042689067201604814, + "loss": 1.4997, + "theoretical_loss": 3.9344039058805986, + "tokens_seen": 510722048 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004268806419257774, + "loss": 1.5599, + "theoretical_loss": 3.934350122177154, + "tokens_seen": 510787584 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004268706118355065, + "loss": 1.5411, + "theoretical_loss": 3.9342963473058212, + "tokens_seen": 510853120 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042686058174523574, + "loss": 1.5288, + "theoretical_loss": 3.9342425812640194, + "tokens_seen": 510918656 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042685055165496487, + "loss": 1.5184, + "theoretical_loss": 3.9341888240491647, + "tokens_seen": 510984192 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004268405215646941, + "loss": 1.5502, + "theoretical_loss": 3.9341350756586766, + "tokens_seen": 511049728 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004268304914744233, + "loss": 1.5303, + "theoretical_loss": 3.934081336089976, + "tokens_seen": 511115264 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 195253, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4593555927276611, + "objective/train/theoretical_loss": 3.934027605340483, + "objective/train/tokens_used": 511180800, + "theoretical_loss": 3.934027605340483, + "tokens_seen": 511180800 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042682046138415246, + "loss": 1.5055, + "theoretical_loss": 3.934027605340483, + "tokens_seen": 511180800 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042681043129388165, + "loss": 1.6836, + "theoretical_loss": 3.93397388340762, + "tokens_seen": 511246336 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004268004012036109, + "loss": 1.4685, + "theoretical_loss": 3.933920170288811, + "tokens_seen": 511311872 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042679037111334, + "loss": 1.5587, + "theoretical_loss": 3.93386646598148, + "tokens_seen": 511377408 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042678034102306924, + "loss": 1.4923, + "theoretical_loss": 3.9338127704830534, + "tokens_seen": 511442944 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042677031093279837, + "loss": 1.5772, + "theoretical_loss": 3.933759083790957, + "tokens_seen": 511508480 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004267602808425276, + "loss": 1.4762, + "theoretical_loss": 3.933705405902619, + "tokens_seen": 511574016 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004267502507522568, + "loss": 1.5105, + "theoretical_loss": 3.9336517368154666, + "tokens_seen": 511639552 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042674022066198597, + "loss": 1.3989, + "theoretical_loss": 3.9335980765269323, + "tokens_seen": 511705088 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042673019057171515, + "loss": 1.5604, + "theoretical_loss": 3.9335444250344445, + "tokens_seen": 511770624 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042672016048144433, + "loss": 1.5309, + "theoretical_loss": 3.933490782335437, + "tokens_seen": 511836160 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004267101303911735, + "loss": 1.4216, + "theoretical_loss": 3.933437148427342, + "tokens_seen": 511901696 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042670010030090275, + "loss": 1.4768, + "theoretical_loss": 3.933383523307594, + "tokens_seen": 511967232 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004266900702106319, + "loss": 1.4837, + "theoretical_loss": 3.9333299069736283, + "tokens_seen": 512032768 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004266800401203611, + "loss": 1.5246, + "theoretical_loss": 3.9332762994228814, + "tokens_seen": 512098304 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042667001003009024, + "loss": 1.6142, + "theoretical_loss": 3.93322270065279, + "tokens_seen": 512163840 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042665997993981947, + "loss": 1.4755, + "theoretical_loss": 3.933169110660793, + "tokens_seen": 512229376 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042664994984954865, + "loss": 1.5796, + "theoretical_loss": 3.93311552944433, + "tokens_seen": 512294912 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042663991975927783, + "loss": 1.5177, + "theoretical_loss": 3.9330619570008416, + "tokens_seen": 512360448 + }, + { + "epoch": 0.16, + "learning_rate": 0.000426629889669007, + "loss": 1.5352, + "theoretical_loss": 3.9330083933277695, + "tokens_seen": 512425984 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042661985957873625, + "loss": 1.5298, + "theoretical_loss": 3.9329548384225568, + "tokens_seen": 512491520 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004266098294884654, + "loss": 1.5674, + "theoretical_loss": 3.932901292282647, + "tokens_seen": 512557056 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004265997993981946, + "loss": 1.5072, + "theoretical_loss": 3.932847754905485, + "tokens_seen": 512622592 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042658976930792374, + "loss": 1.4795, + "theoretical_loss": 3.9327942262885167, + "tokens_seen": 512688128 + }, + { + "epoch": 0.16, + "learning_rate": 0.000426579739217653, + "loss": 1.5045, + "theoretical_loss": 3.9327407064291897, + "tokens_seen": 512753664 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042656970912738216, + "loss": 1.618, + "theoretical_loss": 3.9326871953249514, + "tokens_seen": 512819200 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042655967903711134, + "loss": 1.5326, + "theoretical_loss": 3.9326336929732513, + "tokens_seen": 512884736 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004265496489468405, + "loss": 1.5076, + "theoretical_loss": 3.9325801993715404, + "tokens_seen": 512950272 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004265396188565697, + "loss": 1.5831, + "theoretical_loss": 3.9325267145172687, + "tokens_seen": 513015808 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004265295887662989, + "loss": 1.5168, + "theoretical_loss": 3.932473238407889, + "tokens_seen": 513081344 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004265195586760281, + "loss": 1.4922, + "theoretical_loss": 3.932419771040856, + "tokens_seen": 513146880 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042650952858575724, + "loss": 1.6087, + "theoretical_loss": 3.932366312413623, + "tokens_seen": 513212416 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004264994984954865, + "loss": 1.4908, + "theoretical_loss": 3.932312862523646, + "tokens_seen": 513277952 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004264894684052156, + "loss": 1.4728, + "theoretical_loss": 3.9322594213683812, + "tokens_seen": 513343488 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042647943831494484, + "loss": 1.4714, + "theoretical_loss": 3.9322059889452867, + "tokens_seen": 513409024 + }, + { + "epoch": 0.16, + "learning_rate": 0.000426469408224674, + "loss": 1.649, + "theoretical_loss": 3.9321525652518217, + "tokens_seen": 513474560 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004264593781344032, + "loss": 1.5268, + "theoretical_loss": 3.9320991502854454, + "tokens_seen": 513540096 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004264493480441324, + "loss": 1.4668, + "theoretical_loss": 3.932045744043619, + "tokens_seen": 513605632 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004264393179538616, + "loss": 1.4959, + "theoretical_loss": 3.9319923465238054, + "tokens_seen": 513671168 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042642928786359075, + "loss": 1.4545, + "theoretical_loss": 3.9319389577234665, + "tokens_seen": 513736704 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042641925777332, + "loss": 1.5544, + "theoretical_loss": 3.931885577640066, + "tokens_seen": 513802240 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042640922768304916, + "loss": 1.4341, + "theoretical_loss": 3.9318322062710704, + "tokens_seen": 513867776 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042639919759277834, + "loss": 1.5489, + "theoretical_loss": 3.931778843613945, + "tokens_seen": 513933312 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004263891675025076, + "loss": 1.4416, + "theoretical_loss": 3.9317254896661575, + "tokens_seen": 513998848 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004263791374122367, + "loss": 1.5614, + "theoretical_loss": 3.9316721444251765, + "tokens_seen": 514064384 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042636910732196594, + "loss": 1.4638, + "theoretical_loss": 3.9316188078884706, + "tokens_seen": 514129920 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042635907723169507, + "loss": 1.4252, + "theoretical_loss": 3.931565480053511, + "tokens_seen": 514195456 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004263490471414243, + "loss": 1.4686, + "theoretical_loss": 3.9315121609177694, + "tokens_seen": 514260992 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004263390170511535, + "loss": 1.5655, + "theoretical_loss": 3.931458850478718, + "tokens_seen": 514326528 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042632898696088266, + "loss": 1.4751, + "theoretical_loss": 3.9314055487338297, + "tokens_seen": 514392064 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 196592, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5901634693145752, + "objective/train/theoretical_loss": 3.93135225568058, + "objective/train/tokens_used": 514457600, + "theoretical_loss": 3.93135225568058, + "tokens_seen": 514457600 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042631895687061185, + "loss": 1.502, + "theoretical_loss": 3.93135225568058, + "tokens_seen": 514457600 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004263089267803411, + "loss": 1.5449, + "theoretical_loss": 3.9312989713164455, + "tokens_seen": 514523136 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004262988966900702, + "loss": 1.4704, + "theoretical_loss": 3.9312456956389017, + "tokens_seen": 514588672 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042628886659979944, + "loss": 1.4812, + "theoretical_loss": 3.931192428645427, + "tokens_seen": 514654208 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042627883650952857, + "loss": 1.4678, + "theoretical_loss": 3.9311391703334997, + "tokens_seen": 514719744 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004262688064192578, + "loss": 1.5393, + "theoretical_loss": 3.9310859207006006, + "tokens_seen": 514785280 + }, + { + "epoch": 0.16, + "learning_rate": 0.000426258776328987, + "loss": 1.4917, + "theoretical_loss": 3.9310326797442103, + "tokens_seen": 514850816 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042624874623871617, + "loss": 1.6516, + "theoretical_loss": 3.930979447461811, + "tokens_seen": 514916352 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042623871614844535, + "loss": 1.5391, + "theoretical_loss": 3.9309262238508857, + "tokens_seen": 514981888 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042622868605817453, + "loss": 1.5419, + "theoretical_loss": 3.930873008908919, + "tokens_seen": 515047424 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004262186559679037, + "loss": 1.4612, + "theoretical_loss": 3.9308198026333954, + "tokens_seen": 515112960 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042620862587763295, + "loss": 1.5219, + "theoretical_loss": 3.9307666050218018, + "tokens_seen": 515178496 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004261985957873621, + "loss": 1.4317, + "theoretical_loss": 3.930713416071625, + "tokens_seen": 515244032 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004261885656970913, + "loss": 1.4941, + "theoretical_loss": 3.930660235780354, + "tokens_seen": 515309568 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042617853560682044, + "loss": 1.4587, + "theoretical_loss": 3.9306070641454776, + "tokens_seen": 515375104 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042616850551654967, + "loss": 1.5595, + "theoretical_loss": 3.9305539011644863, + "tokens_seen": 515440640 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042615847542627885, + "loss": 1.4999, + "theoretical_loss": 3.9305007468348725, + "tokens_seen": 515506176 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042614844533600803, + "loss": 1.4414, + "theoretical_loss": 3.9304476011541274, + "tokens_seen": 515571712 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004261384152457372, + "loss": 1.4386, + "theoretical_loss": 3.930394464119746, + "tokens_seen": 515637248 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042612838515546645, + "loss": 1.5464, + "theoretical_loss": 3.9303413357292216, + "tokens_seen": 515702784 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004261183550651956, + "loss": 1.4699, + "theoretical_loss": 3.930288215980051, + "tokens_seen": 515768320 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004261083249749248, + "loss": 1.488, + "theoretical_loss": 3.9302351048697304, + "tokens_seen": 515833856 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042609829488465394, + "loss": 1.4153, + "theoretical_loss": 3.9301820023957577, + "tokens_seen": 515899392 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004260882647943832, + "loss": 1.4982, + "theoretical_loss": 3.930128908555632, + "tokens_seen": 515964928 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042607823470411236, + "loss": 1.5638, + "theoretical_loss": 3.9300758233468525, + "tokens_seen": 516030464 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042606820461384154, + "loss": 1.4649, + "theoretical_loss": 3.9300227467669204, + "tokens_seen": 516096000 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004260581745235707, + "loss": 1.4503, + "theoretical_loss": 3.929969678813338, + "tokens_seen": 516161536 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004260481444332999, + "loss": 1.5293, + "theoretical_loss": 3.929916619483608, + "tokens_seen": 516227072 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004260381143430291, + "loss": 1.4858, + "theoretical_loss": 3.9298635687752337, + "tokens_seen": 516292608 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004260280842527583, + "loss": 1.5637, + "theoretical_loss": 3.9298105266857215, + "tokens_seen": 516358144 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042601805416248744, + "loss": 1.5212, + "theoretical_loss": 3.929757493212577, + "tokens_seen": 516423680 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004260080240722167, + "loss": 1.54, + "theoretical_loss": 3.929704468353307, + "tokens_seen": 516489216 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004259979939819458, + "loss": 1.4927, + "theoretical_loss": 3.9296514521054196, + "tokens_seen": 516554752 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042598796389167504, + "loss": 1.4532, + "theoretical_loss": 3.929598444466425, + "tokens_seen": 516620288 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004259779338014042, + "loss": 1.4546, + "theoretical_loss": 3.9295454454338326, + "tokens_seen": 516685824 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004259679037111334, + "loss": 1.4772, + "theoretical_loss": 3.9294924550051533, + "tokens_seen": 516751360 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004259578736208626, + "loss": 1.5706, + "theoretical_loss": 3.9294394731779008, + "tokens_seen": 516816896 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004259478435305918, + "loss": 1.4949, + "theoretical_loss": 3.929386499949587, + "tokens_seen": 516882432 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042593781344032095, + "loss": 1.5605, + "theoretical_loss": 3.929333535317727, + "tokens_seen": 516947968 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004259277833500502, + "loss": 1.5408, + "theoretical_loss": 3.929280579279837, + "tokens_seen": 517013504 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004259177532597793, + "loss": 1.427, + "theoretical_loss": 3.929227631833432, + "tokens_seen": 517079040 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042590772316950854, + "loss": 1.418, + "theoretical_loss": 3.92917469297603, + "tokens_seen": 517144576 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004258976930792377, + "loss": 1.4281, + "theoretical_loss": 3.9291217627051496, + "tokens_seen": 517210112 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004258876629889669, + "loss": 1.4643, + "theoretical_loss": 3.9290688410183106, + "tokens_seen": 517275648 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004258776328986961, + "loss": 1.5894, + "theoretical_loss": 3.9290159279130332, + "tokens_seen": 517341184 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042586760280842527, + "loss": 1.5006, + "theoretical_loss": 3.928963023386839, + "tokens_seen": 517406720 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042585757271815445, + "loss": 1.4711, + "theoretical_loss": 3.9289101274372515, + "tokens_seen": 517472256 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004258475426278837, + "loss": 1.5211, + "theoretical_loss": 3.9288572400617934, + "tokens_seen": 517537792 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004258375125376128, + "loss": 1.5372, + "theoretical_loss": 3.9288043612579893, + "tokens_seen": 517603328 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042582748244734205, + "loss": 1.6018, + "theoretical_loss": 3.928751491023366, + "tokens_seen": 517668864 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 197659, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4554855823516846, + "objective/train/theoretical_loss": 3.9286986293554493, + "objective/train/tokens_used": 517734400, + "theoretical_loss": 3.9286986293554493, + "tokens_seen": 517734400 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004258174523570712, + "loss": 1.4081, + "theoretical_loss": 3.9286986293554493, + "tokens_seen": 517734400 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004258074222668004, + "loss": 1.4548, + "theoretical_loss": 3.9286457762517673, + "tokens_seen": 517799936 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004257973921765296, + "loss": 1.3949, + "theoretical_loss": 3.928592931709849, + "tokens_seen": 517865472 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042578736208625877, + "loss": 1.4637, + "theoretical_loss": 3.9285400957272234, + "tokens_seen": 517931008 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042577733199598795, + "loss": 1.5984, + "theoretical_loss": 3.9284872683014225, + "tokens_seen": 517996544 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004257673019057172, + "loss": 1.4683, + "theoretical_loss": 3.9284344494299774, + "tokens_seen": 518062080 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004257572718154463, + "loss": 1.4684, + "theoretical_loss": 3.9283816391104214, + "tokens_seen": 518127616 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042574724172517555, + "loss": 1.4911, + "theoretical_loss": 3.9283288373402883, + "tokens_seen": 518193152 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004257372116349047, + "loss": 1.6233, + "theoretical_loss": 3.9282760441171134, + "tokens_seen": 518258688 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004257271815446339, + "loss": 1.4392, + "theoretical_loss": 3.9282232594384316, + "tokens_seen": 518324224 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004257171514543631, + "loss": 1.4973, + "theoretical_loss": 3.9281704833017814, + "tokens_seen": 518389760 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004257071213640923, + "loss": 1.4824, + "theoretical_loss": 3.9281177157047003, + "tokens_seen": 518455296 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042569709127382145, + "loss": 1.5437, + "theoretical_loss": 3.928064956644726, + "tokens_seen": 518520832 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042568706118355064, + "loss": 1.4387, + "theoretical_loss": 3.9280122061194005, + "tokens_seen": 518586368 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004256770310932798, + "loss": 1.5591, + "theoretical_loss": 3.927959464126264, + "tokens_seen": 518651904 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042566700100300905, + "loss": 1.5489, + "theoretical_loss": 3.9279067306628583, + "tokens_seen": 518717440 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042565697091273823, + "loss": 1.4501, + "theoretical_loss": 3.9278540057267266, + "tokens_seen": 518782976 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004256469408224674, + "loss": 1.5934, + "theoretical_loss": 3.927801289315414, + "tokens_seen": 518848512 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042563691073219665, + "loss": 1.5345, + "theoretical_loss": 3.9277485814264645, + "tokens_seen": 518914048 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004256268806419258, + "loss": 1.4957, + "theoretical_loss": 3.9276958820574253, + "tokens_seen": 518979584 + }, + { + "epoch": 0.16, + "learning_rate": 0.000425616850551655, + "loss": 1.4707, + "theoretical_loss": 3.9276431912058425, + "tokens_seen": 519045120 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042560682046138414, + "loss": 1.6433, + "theoretical_loss": 3.9275905088692653, + "tokens_seen": 519110656 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004255967903711134, + "loss": 1.5104, + "theoretical_loss": 3.927537835045242, + "tokens_seen": 519176192 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042558676028084256, + "loss": 1.506, + "theoretical_loss": 3.927485169731323, + "tokens_seen": 519241728 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042557673019057174, + "loss": 1.5949, + "theoretical_loss": 3.9274325129250602, + "tokens_seen": 519307264 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004255667001003009, + "loss": 1.5525, + "theoretical_loss": 3.9273798646240055, + "tokens_seen": 519372800 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004255566700100301, + "loss": 1.5656, + "theoretical_loss": 3.9273272248257123, + "tokens_seen": 519438336 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004255466399197593, + "loss": 1.5671, + "theoretical_loss": 3.9272745935277347, + "tokens_seen": 519503872 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004255366098294885, + "loss": 1.4817, + "theoretical_loss": 3.927221970727628, + "tokens_seen": 519569408 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042552657973921764, + "loss": 1.5624, + "theoretical_loss": 3.927169356422948, + "tokens_seen": 519634944 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004255165496489469, + "loss": 1.4809, + "theoretical_loss": 3.9271167506112534, + "tokens_seen": 519700480 + }, + { + "epoch": 0.16, + "learning_rate": 0.000425506519558676, + "loss": 1.5486, + "theoretical_loss": 3.927064153290101, + "tokens_seen": 519766016 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042549648946840524, + "loss": 1.4574, + "theoretical_loss": 3.9270115644570516, + "tokens_seen": 519831552 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004254864593781344, + "loss": 1.5713, + "theoretical_loss": 3.9269589841096644, + "tokens_seen": 519897088 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004254764292878636, + "loss": 1.5627, + "theoretical_loss": 3.9269064122455006, + "tokens_seen": 519962624 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004254663991975928, + "loss": 1.5062, + "theoretical_loss": 3.926853848862124, + "tokens_seen": 520028160 + }, + { + "epoch": 0.16, + "learning_rate": 0.000425456369107322, + "loss": 1.4747, + "theoretical_loss": 3.9268012939570967, + "tokens_seen": 520093696 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042544633901705115, + "loss": 1.4546, + "theoretical_loss": 3.9267487475279834, + "tokens_seen": 520159232 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004254363089267804, + "loss": 1.5151, + "theoretical_loss": 3.9266962095723494, + "tokens_seen": 520224768 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004254262788365095, + "loss": 1.4006, + "theoretical_loss": 3.9266436800877615, + "tokens_seen": 520290304 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042541624874623874, + "loss": 1.6012, + "theoretical_loss": 3.926591159071787, + "tokens_seen": 520355840 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004254062186559679, + "loss": 1.5015, + "theoretical_loss": 3.9265386465219936, + "tokens_seen": 520421376 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004253961885656971, + "loss": 1.5719, + "theoretical_loss": 3.9264861424359516, + "tokens_seen": 520486912 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004253861584754263, + "loss": 1.5124, + "theoretical_loss": 3.9264336468112315, + "tokens_seen": 520552448 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042537612838515547, + "loss": 1.5749, + "theoretical_loss": 3.9263811596454037, + "tokens_seen": 520617984 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042536609829488465, + "loss": 1.5098, + "theoretical_loss": 3.9263286809360407, + "tokens_seen": 520683520 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004253560682046139, + "loss": 1.5393, + "theoretical_loss": 3.9262762106807174, + "tokens_seen": 520749056 + }, + { + "epoch": 0.16, + "learning_rate": 0.000425346038114343, + "loss": 1.5208, + "theoretical_loss": 3.926223748877007, + "tokens_seen": 520814592 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042533600802407225, + "loss": 1.6173, + "theoretical_loss": 3.926171295522485, + "tokens_seen": 520880128 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004253259779338014, + "loss": 1.6087, + "theoretical_loss": 3.9261188506147278, + "tokens_seen": 520945664 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 198696, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.442735195159912, + "objective/train/theoretical_loss": 3.926066414151313, + "objective/train/tokens_used": 521011200, + "theoretical_loss": 3.926066414151313, + "tokens_seen": 521011200 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004253159478435306, + "loss": 1.4762, + "theoretical_loss": 3.926066414151313, + "tokens_seen": 521011200 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004253059177532598, + "loss": 1.572, + "theoretical_loss": 3.926013986129819, + "tokens_seen": 521076736 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042529588766298897, + "loss": 1.556, + "theoretical_loss": 3.9259615665478256, + "tokens_seen": 521142272 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042528585757271815, + "loss": 1.4641, + "theoretical_loss": 3.9259091554029126, + "tokens_seen": 521207808 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004252758274824474, + "loss": 1.5823, + "theoretical_loss": 3.9258567526926615, + "tokens_seen": 521273344 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004252657973921765, + "loss": 1.6287, + "theoretical_loss": 3.9258043584146556, + "tokens_seen": 521338880 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042525576730190575, + "loss": 1.5084, + "theoretical_loss": 3.925751972566477, + "tokens_seen": 521404416 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004252457372116349, + "loss": 1.5618, + "theoretical_loss": 3.9256995951457103, + "tokens_seen": 521469952 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004252357071213641, + "loss": 1.3916, + "theoretical_loss": 3.925647226149942, + "tokens_seen": 521535488 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004252256770310933, + "loss": 1.5173, + "theoretical_loss": 3.925594865576758, + "tokens_seen": 521601024 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004252156469408225, + "loss": 1.4533, + "theoretical_loss": 3.9255425134237445, + "tokens_seen": 521666560 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042520561685055165, + "loss": 1.5423, + "theoretical_loss": 3.9254901696884916, + "tokens_seen": 521732096 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042519558676028084, + "loss": 1.4404, + "theoretical_loss": 3.925437834368588, + "tokens_seen": 521797632 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042518555667001, + "loss": 1.5339, + "theoretical_loss": 3.925385507461624, + "tokens_seen": 521863168 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042517552657973925, + "loss": 1.5372, + "theoretical_loss": 3.9253331889651912, + "tokens_seen": 521928704 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004251654964894684, + "loss": 1.5577, + "theoretical_loss": 3.9252808788768814, + "tokens_seen": 521994240 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004251554663991976, + "loss": 1.5194, + "theoretical_loss": 3.9252285771942876, + "tokens_seen": 522059776 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004251454363089268, + "loss": 1.5289, + "theoretical_loss": 3.9251762839150057, + "tokens_seen": 522125312 + }, + { + "epoch": 0.16, + "learning_rate": 0.000425135406218656, + "loss": 1.5882, + "theoretical_loss": 3.9251239990366305, + "tokens_seen": 522190848 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042512537612838516, + "loss": 1.6474, + "theoretical_loss": 3.925071722556757, + "tokens_seen": 522256384 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042511534603811434, + "loss": 1.4551, + "theoretical_loss": 3.9250194544729844, + "tokens_seen": 522321920 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004251053159478435, + "loss": 1.4804, + "theoretical_loss": 3.9249671947829095, + "tokens_seen": 522387456 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042509528585757276, + "loss": 1.5403, + "theoretical_loss": 3.924914943484133, + "tokens_seen": 522452992 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004250852557673019, + "loss": 1.6487, + "theoretical_loss": 3.9248627005742533, + "tokens_seen": 522518528 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004250752256770311, + "loss": 1.5392, + "theoretical_loss": 3.9248104660508734, + "tokens_seen": 522584064 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042506519558676024, + "loss": 1.4447, + "theoretical_loss": 3.9247582399115943, + "tokens_seen": 522649600 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004250551654964895, + "loss": 1.4903, + "theoretical_loss": 3.92470602215402, + "tokens_seen": 522715136 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042504513540621866, + "loss": 1.4989, + "theoretical_loss": 3.9246538127757544, + "tokens_seen": 522780672 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042503510531594784, + "loss": 1.5642, + "theoretical_loss": 3.9246016117744036, + "tokens_seen": 522846208 + }, + { + "epoch": 0.16, + "learning_rate": 0.000425025075225677, + "loss": 1.3979, + "theoretical_loss": 3.924549419147572, + "tokens_seen": 522911744 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004250150451354062, + "loss": 1.4154, + "theoretical_loss": 3.9244972348928684, + "tokens_seen": 522977280 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004250050150451354, + "loss": 1.4899, + "theoretical_loss": 3.9244450590079003, + "tokens_seen": 523042816 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004249949849548646, + "loss": 1.554, + "theoretical_loss": 3.9243928914902764, + "tokens_seen": 523108352 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042498495486459375, + "loss": 1.3936, + "theoretical_loss": 3.924340732337608, + "tokens_seen": 523173888 + }, + { + "epoch": 0.16, + "learning_rate": 0.000424974924774323, + "loss": 1.436, + "theoretical_loss": 3.924288581547505, + "tokens_seen": 523239424 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042496489468405216, + "loss": 1.5982, + "theoretical_loss": 3.9242364391175797, + "tokens_seen": 523304960 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042495486459378135, + "loss": 1.5834, + "theoretical_loss": 3.924184305045446, + "tokens_seen": 523370496 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004249448345035105, + "loss": 1.4681, + "theoretical_loss": 3.9241321793287174, + "tokens_seen": 523436032 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004249348044132397, + "loss": 1.5166, + "theoretical_loss": 3.9240800619650082, + "tokens_seen": 523501568 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042492477432296894, + "loss": 1.5132, + "theoretical_loss": 3.9240279529519357, + "tokens_seen": 523567104 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004249147442326981, + "loss": 1.5602, + "theoretical_loss": 3.9239758522871164, + "tokens_seen": 523632640 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004249047141424273, + "loss": 1.5315, + "theoretical_loss": 3.923923759968168, + "tokens_seen": 523698176 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004248946840521565, + "loss": 1.5189, + "theoretical_loss": 3.9238716759927095, + "tokens_seen": 523763712 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042488465396188567, + "loss": 1.5222, + "theoretical_loss": 3.923819600358361, + "tokens_seen": 523829248 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042487462387161485, + "loss": 1.5498, + "theoretical_loss": 3.923767533062743, + "tokens_seen": 523894784 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004248645937813441, + "loss": 1.5617, + "theoretical_loss": 3.9237154741034783, + "tokens_seen": 523960320 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004248545636910732, + "loss": 1.4739, + "theoretical_loss": 3.923663423478189, + "tokens_seen": 524025856 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042484453360080245, + "loss": 1.5221, + "theoretical_loss": 3.9236113811844984, + "tokens_seen": 524091392 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004248345035105316, + "loss": 1.4357, + "theoretical_loss": 3.9235593472200323, + "tokens_seen": 524156928 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004248244734202608, + "loss": 1.4496, + "theoretical_loss": 3.9235073215824166, + "tokens_seen": 524222464 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 199786, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5273642539978027, + "objective/train/theoretical_loss": 3.9234553042692766, + "objective/train/tokens_used": 524288000, + "theoretical_loss": 3.9234553042692766, + "tokens_seen": 524288000 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042481444332999, + "loss": 1.4739, + "theoretical_loss": 3.9234553042692766, + "tokens_seen": 524288000 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042480441323971917, + "loss": 1.5026, + "theoretical_loss": 3.9234032952782414, + "tokens_seen": 524353536 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042479438314944835, + "loss": 1.5239, + "theoretical_loss": 3.923351294606939, + "tokens_seen": 524419072 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004247843530591776, + "loss": 1.4751, + "theoretical_loss": 3.923299302253, + "tokens_seen": 524484608 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004247743229689067, + "loss": 1.4935, + "theoretical_loss": 3.923247318214054, + "tokens_seen": 524550144 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042476429287863595, + "loss": 1.5888, + "theoretical_loss": 3.923195342487733, + "tokens_seen": 524615680 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004247542627883651, + "loss": 1.5427, + "theoretical_loss": 3.9231433750716693, + "tokens_seen": 524681216 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004247442326980943, + "loss": 1.5096, + "theoretical_loss": 3.9230914159634964, + "tokens_seen": 524746752 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004247342026078235, + "loss": 1.5641, + "theoretical_loss": 3.9230394651608496, + "tokens_seen": 524812288 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004247241725175527, + "loss": 1.4846, + "theoretical_loss": 3.922987522661363, + "tokens_seen": 524877824 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042471414242728186, + "loss": 1.5305, + "theoretical_loss": 3.922935588462675, + "tokens_seen": 524943360 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042470411233701104, + "loss": 1.51, + "theoretical_loss": 3.9228836625624206, + "tokens_seen": 525008896 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004246940822467402, + "loss": 1.5637, + "theoretical_loss": 3.9228317449582404, + "tokens_seen": 525074432 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042468405215646945, + "loss": 1.5033, + "theoretical_loss": 3.9227798356477726, + "tokens_seen": 525139968 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004246740220661986, + "loss": 1.5283, + "theoretical_loss": 3.922727934628658, + "tokens_seen": 525205504 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004246639919759278, + "loss": 1.5751, + "theoretical_loss": 3.922676041898537, + "tokens_seen": 525271040 + }, + { + "epoch": 0.16, + "learning_rate": 0.000424653961885657, + "loss": 1.5069, + "theoretical_loss": 3.9226241574550533, + "tokens_seen": 525336576 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004246439317953862, + "loss": 1.5008, + "theoretical_loss": 3.9225722812958486, + "tokens_seen": 525402112 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042463390170511536, + "loss": 1.5869, + "theoretical_loss": 3.922520413418568, + "tokens_seen": 525467648 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042462387161484454, + "loss": 1.5199, + "theoretical_loss": 3.922468553820856, + "tokens_seen": 525533184 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004246138415245737, + "loss": 1.5072, + "theoretical_loss": 3.9224167025003593, + "tokens_seen": 525598720 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042460381143430296, + "loss": 1.5462, + "theoretical_loss": 3.922364859454725, + "tokens_seen": 525664256 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004245937813440321, + "loss": 1.6239, + "theoretical_loss": 3.9223130246816007, + "tokens_seen": 525729792 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004245837512537613, + "loss": 1.5485, + "theoretical_loss": 3.9222611981786355, + "tokens_seen": 525795328 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042457372116349044, + "loss": 1.4372, + "theoretical_loss": 3.92220937994348, + "tokens_seen": 525860864 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004245636910732197, + "loss": 1.4662, + "theoretical_loss": 3.9221575699737836, + "tokens_seen": 525926400 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042455366098294886, + "loss": 1.4682, + "theoretical_loss": 3.9221057682672, + "tokens_seen": 525991936 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042454363089267804, + "loss": 1.591, + "theoretical_loss": 3.92205397482138, + "tokens_seen": 526057472 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004245336008024072, + "loss": 1.5611, + "theoretical_loss": 3.9220021896339796, + "tokens_seen": 526123008 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004245235707121364, + "loss": 1.5543, + "theoretical_loss": 3.921950412702652, + "tokens_seen": 526188544 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004245135406218656, + "loss": 1.5037, + "theoretical_loss": 3.9218986440250534, + "tokens_seen": 526254080 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004245035105315948, + "loss": 1.4294, + "theoretical_loss": 3.9218468835988403, + "tokens_seen": 526319616 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042449348044132395, + "loss": 1.5513, + "theoretical_loss": 3.9217951314216704, + "tokens_seen": 526385152 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004244834503510532, + "loss": 1.569, + "theoretical_loss": 3.9217433874912024, + "tokens_seen": 526450688 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042447342026078236, + "loss": 1.5711, + "theoretical_loss": 3.921691651805096, + "tokens_seen": 526516224 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042446339017051155, + "loss": 1.5585, + "theoretical_loss": 3.9216399243610116, + "tokens_seen": 526581760 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004244533600802407, + "loss": 1.5758, + "theoretical_loss": 3.9215882051566098, + "tokens_seen": 526647296 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004244433299899699, + "loss": 1.5665, + "theoretical_loss": 3.921536494189554, + "tokens_seen": 526712832 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004244332998996991, + "loss": 1.4972, + "theoretical_loss": 3.921484791457507, + "tokens_seen": 526778368 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004244232698094283, + "loss": 1.5476, + "theoretical_loss": 3.921433096958134, + "tokens_seen": 526843904 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042441323971915745, + "loss": 1.494, + "theoretical_loss": 3.921381410689099, + "tokens_seen": 526909440 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004244032096288867, + "loss": 1.5273, + "theoretical_loss": 3.9213297326480694, + "tokens_seen": 526974976 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004243931795386158, + "loss": 1.5188, + "theoretical_loss": 3.9212780628327115, + "tokens_seen": 527040512 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042438314944834505, + "loss": 1.4967, + "theoretical_loss": 3.921226401240693, + "tokens_seen": 527106048 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042437311935807423, + "loss": 1.609, + "theoretical_loss": 3.9211747478696846, + "tokens_seen": 527171584 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004243630892678034, + "loss": 1.5232, + "theoretical_loss": 3.9211231027173548, + "tokens_seen": 527237120 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004243530591775326, + "loss": 1.488, + "theoretical_loss": 3.921071465781375, + "tokens_seen": 527302656 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004243430290872618, + "loss": 1.5724, + "theoretical_loss": 3.9210198370594176, + "tokens_seen": 527368192 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042433299899699095, + "loss": 1.5059, + "theoretical_loss": 3.920968216549155, + "tokens_seen": 527433728 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004243229689067202, + "loss": 1.6036, + "theoretical_loss": 3.9209166042482613, + "tokens_seen": 527499264 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 201277, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4557223320007324, + "objective/train/theoretical_loss": 3.9208650001544108, + "objective/train/tokens_used": 527564800, + "theoretical_loss": 3.9208650001544108, + "tokens_seen": 527564800 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004243129388164493, + "loss": 1.4411, + "theoretical_loss": 3.9208650001544108, + "tokens_seen": 527564800 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042430290872617855, + "loss": 1.4983, + "theoretical_loss": 3.920813404265279, + "tokens_seen": 527630336 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042429287863590773, + "loss": 1.6434, + "theoretical_loss": 3.920761816578543, + "tokens_seen": 527695872 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004242828485456369, + "loss": 1.4975, + "theoretical_loss": 3.920710237091881, + "tokens_seen": 527761408 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004242728184553661, + "loss": 1.4716, + "theoretical_loss": 3.9206586658029714, + "tokens_seen": 527826944 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004242627883650953, + "loss": 1.5009, + "theoretical_loss": 3.920607102709493, + "tokens_seen": 527892480 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042425275827482446, + "loss": 1.5362, + "theoretical_loss": 3.9205555478091254, + "tokens_seen": 527958016 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004242427281845537, + "loss": 1.4566, + "theoretical_loss": 3.920504001099552, + "tokens_seen": 528023552 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004242326980942828, + "loss": 1.5851, + "theoretical_loss": 3.920452462578454, + "tokens_seen": 528089088 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042422266800401206, + "loss": 1.486, + "theoretical_loss": 3.920400932243515, + "tokens_seen": 528154624 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004242126379137412, + "loss": 1.6139, + "theoretical_loss": 3.920349410092419, + "tokens_seen": 528220160 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004242026078234704, + "loss": 1.51, + "theoretical_loss": 3.920297896122851, + "tokens_seen": 528285696 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004241925777331996, + "loss": 1.5174, + "theoretical_loss": 3.9202463903324976, + "tokens_seen": 528351232 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004241825476429288, + "loss": 1.4677, + "theoretical_loss": 3.920194892719045, + "tokens_seen": 528416768 + }, + { + "epoch": 0.16, + "learning_rate": 0.000424172517552658, + "loss": 1.4437, + "theoretical_loss": 3.920143403280182, + "tokens_seen": 528482304 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004241624874623872, + "loss": 1.4809, + "theoretical_loss": 3.9200919220135972, + "tokens_seen": 528547840 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004241524573721164, + "loss": 1.5899, + "theoretical_loss": 3.920040448916981, + "tokens_seen": 528613376 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042414242728184556, + "loss": 1.5172, + "theoretical_loss": 3.919988983988023, + "tokens_seen": 528678912 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042413239719157474, + "loss": 1.5246, + "theoretical_loss": 3.919937527224416, + "tokens_seen": 528744448 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004241223671013039, + "loss": 1.537, + "theoretical_loss": 3.9198860786238523, + "tokens_seen": 528809984 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042411233701103316, + "loss": 1.4987, + "theoretical_loss": 3.919834638184025, + "tokens_seen": 528875520 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004241023069207623, + "loss": 1.4668, + "theoretical_loss": 3.91978320590263, + "tokens_seen": 528941056 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004240922768304915, + "loss": 1.4715, + "theoretical_loss": 3.919731781777361, + "tokens_seen": 529006592 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042408224674022065, + "loss": 1.5755, + "theoretical_loss": 3.919680365805916, + "tokens_seen": 529072128 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004240722166499499, + "loss": 1.4192, + "theoretical_loss": 3.9196289579859913, + "tokens_seen": 529137664 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042406218655967906, + "loss": 1.4242, + "theoretical_loss": 3.919577558315286, + "tokens_seen": 529203200 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042405215646940824, + "loss": 1.4693, + "theoretical_loss": 3.9195261667914982, + "tokens_seen": 529268736 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004240421263791374, + "loss": 1.5686, + "theoretical_loss": 3.91947478341233, + "tokens_seen": 529334272 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004240320962888666, + "loss": 1.5783, + "theoretical_loss": 3.9194234081754797, + "tokens_seen": 529399808 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004240220661985958, + "loss": 1.4702, + "theoretical_loss": 3.9193720410786517, + "tokens_seen": 529465344 + }, + { + "epoch": 0.16, + "learning_rate": 0.000424012036108325, + "loss": 1.5346, + "theoretical_loss": 3.9193206821195483, + "tokens_seen": 529530880 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042400200601805415, + "loss": 1.558, + "theoretical_loss": 3.9192693312958733, + "tokens_seen": 529596416 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004239919759277834, + "loss": 1.5344, + "theoretical_loss": 3.919217988605331, + "tokens_seen": 529661952 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042398194583751256, + "loss": 1.4785, + "theoretical_loss": 3.919166654045628, + "tokens_seen": 529727488 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042397191574724175, + "loss": 1.3798, + "theoretical_loss": 3.919115327614471, + "tokens_seen": 529793024 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004239618856569709, + "loss": 1.5909, + "theoretical_loss": 3.9190640093095666, + "tokens_seen": 529858560 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004239518555667001, + "loss": 1.549, + "theoretical_loss": 3.9190126991286243, + "tokens_seen": 529924096 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004239418254764293, + "loss": 1.6064, + "theoretical_loss": 3.918961397069353, + "tokens_seen": 529989632 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004239317953861585, + "loss": 1.5319, + "theoretical_loss": 3.9189101031294635, + "tokens_seen": 530055168 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042392176529588765, + "loss": 1.4792, + "theoretical_loss": 3.918858817306667, + "tokens_seen": 530120704 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004239117352056169, + "loss": 1.447, + "theoretical_loss": 3.918807539598676, + "tokens_seen": 530186240 + }, + { + "epoch": 0.16, + "learning_rate": 0.000423901705115346, + "loss": 1.4905, + "theoretical_loss": 3.918756270003203, + "tokens_seen": 530251776 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042389167502507525, + "loss": 1.5663, + "theoretical_loss": 3.9187050085179633, + "tokens_seen": 530317312 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042388164493480443, + "loss": 1.6205, + "theoretical_loss": 3.918653755140671, + "tokens_seen": 530382848 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004238716148445336, + "loss": 1.4307, + "theoretical_loss": 3.918602509869042, + "tokens_seen": 530448384 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004238615847542628, + "loss": 1.5433, + "theoretical_loss": 3.9185512727007934, + "tokens_seen": 530513920 + }, + { + "epoch": 0.16, + "learning_rate": 0.000423851554663992, + "loss": 1.5306, + "theoretical_loss": 3.9185000436336432, + "tokens_seen": 530579456 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042384152457372115, + "loss": 1.5691, + "theoretical_loss": 3.9184488226653102, + "tokens_seen": 530644992 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004238314944834504, + "loss": 1.6719, + "theoretical_loss": 3.918397609793514, + "tokens_seen": 530710528 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004238214643931795, + "loss": 1.5838, + "theoretical_loss": 3.9183464050159746, + "tokens_seen": 530776064 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 202413, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5526180267333984, + "objective/train/theoretical_loss": 3.918295208330415, + "objective/train/tokens_used": 530841600, + "theoretical_loss": 3.918295208330415, + "tokens_seen": 530841600 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042381143430290875, + "loss": 1.5261, + "theoretical_loss": 3.918295208330415, + "tokens_seen": 530841600 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042380140421263793, + "loss": 1.4567, + "theoretical_loss": 3.918244019734556, + "tokens_seen": 530907136 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004237913741223671, + "loss": 1.5934, + "theoretical_loss": 3.9181928392261214, + "tokens_seen": 530972672 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004237813440320963, + "loss": 1.5703, + "theoretical_loss": 3.9181416668028355, + "tokens_seen": 531038208 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004237713139418255, + "loss": 1.4585, + "theoretical_loss": 3.918090502462425, + "tokens_seen": 531103744 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042376128385155466, + "loss": 1.5743, + "theoretical_loss": 3.9180393462026135, + "tokens_seen": 531169280 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004237512537612839, + "loss": 1.5467, + "theoretical_loss": 3.9179881980211295, + "tokens_seen": 531234816 + }, + { + "epoch": 0.16, + "learning_rate": 0.000423741223671013, + "loss": 1.4757, + "theoretical_loss": 3.917937057915701, + "tokens_seen": 531300352 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042373119358074226, + "loss": 1.6083, + "theoretical_loss": 3.9178859258840566, + "tokens_seen": 531365888 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004237211634904714, + "loss": 1.4592, + "theoretical_loss": 3.9178348019239264, + "tokens_seen": 531431424 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004237111334002006, + "loss": 1.5188, + "theoretical_loss": 3.91778368603304, + "tokens_seen": 531496960 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004237011033099298, + "loss": 1.4858, + "theoretical_loss": 3.9177325782091303, + "tokens_seen": 531562496 + }, + { + "epoch": 0.16, + "learning_rate": 0.000423691073219659, + "loss": 1.5155, + "theoretical_loss": 3.9176814784499303, + "tokens_seen": 531628032 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042368104312938816, + "loss": 1.4892, + "theoretical_loss": 3.9176303867531717, + "tokens_seen": 531693568 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004236710130391174, + "loss": 1.4433, + "theoretical_loss": 3.9175793031165895, + "tokens_seen": 531759104 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004236609829488465, + "loss": 1.4753, + "theoretical_loss": 3.91752822753792, + "tokens_seen": 531824640 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042365095285857576, + "loss": 1.5678, + "theoretical_loss": 3.917477160014898, + "tokens_seen": 531890176 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004236409227683049, + "loss": 1.4697, + "theoretical_loss": 3.9174261005452617, + "tokens_seen": 531955712 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004236308926780341, + "loss": 1.516, + "theoretical_loss": 3.917375049126749, + "tokens_seen": 532021248 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004236208625877633, + "loss": 1.4389, + "theoretical_loss": 3.9173240057570986, + "tokens_seen": 532086784 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004236108324974925, + "loss": 1.5993, + "theoretical_loss": 3.91727297043405, + "tokens_seen": 532152320 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042360080240722166, + "loss": 1.4961, + "theoretical_loss": 3.917221943155345, + "tokens_seen": 532217856 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042359077231695085, + "loss": 1.3967, + "theoretical_loss": 3.9171709239187247, + "tokens_seen": 532283392 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042358074222668, + "loss": 1.5302, + "theoretical_loss": 3.9171199127219314, + "tokens_seen": 532348928 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042357071213640926, + "loss": 1.3868, + "theoretical_loss": 3.917068909562709, + "tokens_seen": 532414464 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004235606820461384, + "loss": 1.44, + "theoretical_loss": 3.917017914438802, + "tokens_seen": 532480000 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004235506519558676, + "loss": 1.5442, + "theoretical_loss": 3.9169669273479553, + "tokens_seen": 532545536 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042354062186559675, + "loss": 1.5329, + "theoretical_loss": 3.916915948287916, + "tokens_seen": 532611072 + }, + { + "epoch": 0.16, + "learning_rate": 0.000423530591775326, + "loss": 1.46, + "theoretical_loss": 3.91686497725643, + "tokens_seen": 532676608 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042352056168505517, + "loss": 1.5378, + "theoretical_loss": 3.916814014251247, + "tokens_seen": 532742144 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042351053159478435, + "loss": 1.5131, + "theoretical_loss": 3.9167630592701146, + "tokens_seen": 532807680 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042350050150451353, + "loss": 1.5287, + "theoretical_loss": 3.9167121123107833, + "tokens_seen": 532873216 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042349047141424276, + "loss": 1.4339, + "theoretical_loss": 3.9166611733710037, + "tokens_seen": 532938752 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004234804413239719, + "loss": 1.5737, + "theoretical_loss": 3.9166102424485274, + "tokens_seen": 533004288 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042347041123370113, + "loss": 1.4426, + "theoretical_loss": 3.9165593195411077, + "tokens_seen": 533069824 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042346038114343025, + "loss": 1.5476, + "theoretical_loss": 3.916508404646497, + "tokens_seen": 533135360 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004234503510531595, + "loss": 1.6234, + "theoretical_loss": 3.9164574977624507, + "tokens_seen": 533200896 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042344032096288867, + "loss": 1.6352, + "theoretical_loss": 3.916406598886724, + "tokens_seen": 533266432 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042343029087261785, + "loss": 1.5529, + "theoretical_loss": 3.916355708017073, + "tokens_seen": 533331968 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004234202607823471, + "loss": 1.5066, + "theoretical_loss": 3.916304825151254, + "tokens_seen": 533397504 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004234102306920762, + "loss": 1.5032, + "theoretical_loss": 3.916253950287026, + "tokens_seen": 533463040 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042340020060180545, + "loss": 1.5233, + "theoretical_loss": 3.916203083422148, + "tokens_seen": 533528576 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042339017051153463, + "loss": 1.4852, + "theoretical_loss": 3.91615222455438, + "tokens_seen": 533594112 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004233801404212638, + "loss": 1.6301, + "theoretical_loss": 3.9161013736814816, + "tokens_seen": 533659648 + }, + { + "epoch": 0.16, + "learning_rate": 0.000423370110330993, + "loss": 1.5245, + "theoretical_loss": 3.916050530801215, + "tokens_seen": 533725184 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004233600802407222, + "loss": 1.5944, + "theoretical_loss": 3.915999695911343, + "tokens_seen": 533790720 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042335005015045135, + "loss": 1.6311, + "theoretical_loss": 3.91594886900963, + "tokens_seen": 533856256 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004233400200601806, + "loss": 1.5985, + "theoretical_loss": 3.9158980500938383, + "tokens_seen": 533921792 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004233299899699097, + "loss": 1.5164, + "theoretical_loss": 3.9158472391617343, + "tokens_seen": 533987328 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042331995987963895, + "loss": 1.5052, + "theoretical_loss": 3.9157964362110844, + "tokens_seen": 534052864 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 203492, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4729149341583252, + "objective/train/theoretical_loss": 3.9157456412396554, + "objective/train/tokens_used": 534118400, + "theoretical_loss": 3.9157456412396554, + "tokens_seen": 534118400 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042330992978936813, + "loss": 1.5061, + "theoretical_loss": 3.9157456412396554, + "tokens_seen": 534118400 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004232998996990973, + "loss": 1.5349, + "theoretical_loss": 3.9156948542452144, + "tokens_seen": 534183936 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004232898696088265, + "loss": 1.5815, + "theoretical_loss": 3.915644075225532, + "tokens_seen": 534249472 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004232798395185557, + "loss": 1.6169, + "theoretical_loss": 3.915593304178376, + "tokens_seen": 534315008 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042326980942828486, + "loss": 1.5698, + "theoretical_loss": 3.9155425411015186, + "tokens_seen": 534380544 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004232597793380141, + "loss": 1.5265, + "theoretical_loss": 3.9154917859927307, + "tokens_seen": 534446080 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004232497492477432, + "loss": 1.5257, + "theoretical_loss": 3.9154410388497842, + "tokens_seen": 534511616 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042323971915747246, + "loss": 1.4685, + "theoretical_loss": 3.9153902996704533, + "tokens_seen": 534577152 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004232296890672016, + "loss": 1.5636, + "theoretical_loss": 3.915339568452512, + "tokens_seen": 534642688 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004232196589769308, + "loss": 1.619, + "theoretical_loss": 3.9152888451937358, + "tokens_seen": 534708224 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042320962888666, + "loss": 1.5349, + "theoretical_loss": 3.9152381298919, + "tokens_seen": 534773760 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004231995987963892, + "loss": 1.5007, + "theoretical_loss": 3.915187422544782, + "tokens_seen": 534839296 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042318956870611836, + "loss": 1.4008, + "theoretical_loss": 3.915136723150159, + "tokens_seen": 534904832 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004231795386158476, + "loss": 1.5589, + "theoretical_loss": 3.91508603170581, + "tokens_seen": 534970368 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004231695085255767, + "loss": 1.6502, + "theoretical_loss": 3.915035348209514, + "tokens_seen": 535035904 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042315947843530596, + "loss": 1.5433, + "theoretical_loss": 3.9149846726590534, + "tokens_seen": 535101440 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004231494483450351, + "loss": 1.5343, + "theoretical_loss": 3.914934005052208, + "tokens_seen": 535166976 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004231394182547643, + "loss": 1.4994, + "theoretical_loss": 3.91488334538676, + "tokens_seen": 535232512 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004231293881644935, + "loss": 1.4366, + "theoretical_loss": 3.914832693660493, + "tokens_seen": 535298048 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004231193580742227, + "loss": 1.5868, + "theoretical_loss": 3.9147820498711914, + "tokens_seen": 535363584 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042310932798395186, + "loss": 1.5228, + "theoretical_loss": 3.91473141401664, + "tokens_seen": 535429120 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042309929789368105, + "loss": 1.4324, + "theoretical_loss": 3.9146807860946238, + "tokens_seen": 535494656 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004230892678034102, + "loss": 1.4914, + "theoretical_loss": 3.91463016610293, + "tokens_seen": 535560192 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042307923771313946, + "loss": 1.5585, + "theoretical_loss": 3.9145795540393467, + "tokens_seen": 535625728 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004230692076228686, + "loss": 1.4541, + "theoretical_loss": 3.9145289499016616, + "tokens_seen": 535691264 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004230591775325978, + "loss": 1.5567, + "theoretical_loss": 3.9144783536876644, + "tokens_seen": 535756800 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042304914744232695, + "loss": 1.5512, + "theoretical_loss": 3.9144277653951454, + "tokens_seen": 535822336 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004230391173520562, + "loss": 1.407, + "theoretical_loss": 3.914377185021896, + "tokens_seen": 535887872 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042302908726178537, + "loss": 1.6305, + "theoretical_loss": 3.914326612565708, + "tokens_seen": 535953408 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042301905717151455, + "loss": 1.5104, + "theoretical_loss": 3.914276048024374, + "tokens_seen": 536018944 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042300902708124373, + "loss": 1.4638, + "theoretical_loss": 3.9142254913956886, + "tokens_seen": 536084480 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042299899699097297, + "loss": 1.5106, + "theoretical_loss": 3.914174942677446, + "tokens_seen": 536150016 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004229889669007021, + "loss": 1.5401, + "theoretical_loss": 3.9141244018674413, + "tokens_seen": 536215552 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042297893681043133, + "loss": 1.489, + "theoretical_loss": 3.9140738689634715, + "tokens_seen": 536281088 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042296890672016045, + "loss": 1.5474, + "theoretical_loss": 3.914023343963334, + "tokens_seen": 536346624 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004229588766298897, + "loss": 1.4965, + "theoretical_loss": 3.9139728268648266, + "tokens_seen": 536412160 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042294884653961887, + "loss": 1.5028, + "theoretical_loss": 3.913922317665749, + "tokens_seen": 536477696 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042293881644934805, + "loss": 1.4563, + "theoretical_loss": 3.9138718163639012, + "tokens_seen": 536543232 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042292878635907723, + "loss": 1.5677, + "theoretical_loss": 3.9138213229570837, + "tokens_seen": 536608768 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004229187562688064, + "loss": 1.6301, + "theoretical_loss": 3.9137708374430984, + "tokens_seen": 536674304 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004229087261785356, + "loss": 1.6535, + "theoretical_loss": 3.913720359819748, + "tokens_seen": 536739840 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042289869608826483, + "loss": 1.4253, + "theoretical_loss": 3.913669890084835, + "tokens_seen": 536805376 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042288866599799396, + "loss": 1.4477, + "theoretical_loss": 3.9136194282361654, + "tokens_seen": 536870912 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004228786359077232, + "loss": 1.4737, + "theoretical_loss": 3.913568974271543, + "tokens_seen": 536936448 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004228686058174523, + "loss": 1.493, + "theoretical_loss": 3.913518528188775, + "tokens_seen": 537001984 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042285857572718155, + "loss": 1.5444, + "theoretical_loss": 3.9134680899856686, + "tokens_seen": 537067520 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042284854563691074, + "loss": 1.5113, + "theoretical_loss": 3.913417659660031, + "tokens_seen": 537133056 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004228385155466399, + "loss": 1.6145, + "theoretical_loss": 3.913367237209671, + "tokens_seen": 537198592 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004228284854563691, + "loss": 1.5406, + "theoretical_loss": 3.9133168226323987, + "tokens_seen": 537264128 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042281845536609833, + "loss": 1.4855, + "theoretical_loss": 3.913266415926024, + "tokens_seen": 537329664 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 204661, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3257076740264893, + "objective/train/theoretical_loss": 3.9132160170883585, + "objective/train/tokens_used": 537395200, + "theoretical_loss": 3.9132160170883585, + "tokens_seen": 537395200 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042280842527582746, + "loss": 1.5311, + "theoretical_loss": 3.9132160170883585, + "tokens_seen": 537395200 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004227983951855567, + "loss": 1.517, + "theoretical_loss": 3.913165626117215, + "tokens_seen": 537460736 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004227883650952858, + "loss": 1.4876, + "theoretical_loss": 3.913115243010407, + "tokens_seen": 537526272 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042277833500501506, + "loss": 1.4436, + "theoretical_loss": 3.9130648677657476, + "tokens_seen": 537591808 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042276830491474424, + "loss": 1.5651, + "theoretical_loss": 3.9130145003810517, + "tokens_seen": 537657344 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004227582748244734, + "loss": 1.5737, + "theoretical_loss": 3.912964140854135, + "tokens_seen": 537722880 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004227482447342026, + "loss": 1.5552, + "theoretical_loss": 3.9129137891828147, + "tokens_seen": 537788416 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004227382146439318, + "loss": 1.6388, + "theoretical_loss": 3.9128634453649083, + "tokens_seen": 537853952 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042272818455366096, + "loss": 1.411, + "theoretical_loss": 3.9128131093982343, + "tokens_seen": 537919488 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004227181544633902, + "loss": 1.5989, + "theoretical_loss": 3.9127627812806116, + "tokens_seen": 537985024 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004227081243731193, + "loss": 1.4766, + "theoretical_loss": 3.91271246100986, + "tokens_seen": 538050560 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042269809428284856, + "loss": 1.5058, + "theoretical_loss": 3.9126621485838013, + "tokens_seen": 538116096 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004226880641925777, + "loss": 1.4914, + "theoretical_loss": 3.9126118440002573, + "tokens_seen": 538181632 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004226780341023069, + "loss": 1.5288, + "theoretical_loss": 3.9125615472570505, + "tokens_seen": 538247168 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042266800401203616, + "loss": 1.5033, + "theoretical_loss": 3.912511258352004, + "tokens_seen": 538312704 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004226579739217653, + "loss": 1.491, + "theoretical_loss": 3.9124609772829437, + "tokens_seen": 538378240 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004226479438314945, + "loss": 1.4799, + "theoretical_loss": 3.9124107040476934, + "tokens_seen": 538443776 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004226379137412237, + "loss": 1.5714, + "theoretical_loss": 3.91236043864408, + "tokens_seen": 538509312 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004226278836509529, + "loss": 1.5668, + "theoretical_loss": 3.912310181069931, + "tokens_seen": 538574848 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042261785356068206, + "loss": 1.5292, + "theoretical_loss": 3.9122599313230735, + "tokens_seen": 538640384 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042260782347041125, + "loss": 1.5138, + "theoretical_loss": 3.9122096894013376, + "tokens_seen": 538705920 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004225977933801404, + "loss": 1.5086, + "theoretical_loss": 3.9121594553025516, + "tokens_seen": 538771456 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042258776328986966, + "loss": 1.5439, + "theoretical_loss": 3.912109229024547, + "tokens_seen": 538836992 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004225777331995988, + "loss": 1.5046, + "theoretical_loss": 3.9120590105651547, + "tokens_seen": 538902528 + }, + { + "epoch": 0.16, + "learning_rate": 0.000422567703109328, + "loss": 1.4944, + "theoretical_loss": 3.9120087999222073, + "tokens_seen": 538968064 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042255767301905715, + "loss": 1.5146, + "theoretical_loss": 3.9119585970935375, + "tokens_seen": 539033600 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004225476429287864, + "loss": 1.526, + "theoretical_loss": 3.91190840207698, + "tokens_seen": 539099136 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042253761283851557, + "loss": 1.4881, + "theoretical_loss": 3.9118582148703696, + "tokens_seen": 539164672 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042252758274824475, + "loss": 1.4367, + "theoretical_loss": 3.9118080354715414, + "tokens_seen": 539230208 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042251755265797393, + "loss": 1.5314, + "theoretical_loss": 3.911757863878333, + "tokens_seen": 539295744 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042250752256770317, + "loss": 1.355, + "theoretical_loss": 3.9117077000885807, + "tokens_seen": 539361280 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004224974924774323, + "loss": 1.5874, + "theoretical_loss": 3.911657544100124, + "tokens_seen": 539426816 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042248746238716153, + "loss": 1.5304, + "theoretical_loss": 3.911607395910801, + "tokens_seen": 539492352 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042247743229689065, + "loss": 1.6421, + "theoretical_loss": 3.911557255518453, + "tokens_seen": 539557888 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004224674022066199, + "loss": 1.5027, + "theoretical_loss": 3.9115071229209195, + "tokens_seen": 539623424 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042245737211634907, + "loss": 1.4853, + "theoretical_loss": 3.9114569981160434, + "tokens_seen": 539688960 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042244734202607825, + "loss": 1.4467, + "theoretical_loss": 3.9114068811016662, + "tokens_seen": 539754496 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042243731193580743, + "loss": 1.5414, + "theoretical_loss": 3.9113567718756332, + "tokens_seen": 539820032 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004224272818455366, + "loss": 1.5318, + "theoretical_loss": 3.9113066704357866, + "tokens_seen": 539885568 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004224172517552658, + "loss": 1.5551, + "theoretical_loss": 3.9112565767799734, + "tokens_seen": 539951104 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042240722166499503, + "loss": 1.5985, + "theoretical_loss": 3.9112064909060393, + "tokens_seen": 540016640 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042239719157472416, + "loss": 1.5389, + "theoretical_loss": 3.91115641281183, + "tokens_seen": 540082176 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004223871614844534, + "loss": 1.4999, + "theoretical_loss": 3.911106342495195, + "tokens_seen": 540147712 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004223771313941825, + "loss": 1.4635, + "theoretical_loss": 3.911056279953981, + "tokens_seen": 540213248 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042236710130391176, + "loss": 1.5536, + "theoretical_loss": 3.91100622518604, + "tokens_seen": 540278784 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042235707121364094, + "loss": 1.4954, + "theoretical_loss": 3.9109561781892204, + "tokens_seen": 540344320 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004223470411233701, + "loss": 1.5746, + "theoretical_loss": 3.9109061389613737, + "tokens_seen": 540409856 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004223370110330993, + "loss": 1.5127, + "theoretical_loss": 3.9108561075003525, + "tokens_seen": 540475392 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042232698094282853, + "loss": 1.549, + "theoretical_loss": 3.9108060838040095, + "tokens_seen": 540540928 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042231695085255766, + "loss": 1.4941, + "theoretical_loss": 3.9107560678701985, + "tokens_seen": 540606464 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 205878, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5906744003295898, + "objective/train/theoretical_loss": 3.910706059696774, + "objective/train/tokens_used": 540672000, + "theoretical_loss": 3.910706059696774, + "tokens_seen": 540672000 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004223069207622869, + "loss": 1.5469, + "theoretical_loss": 3.910706059696774, + "tokens_seen": 540672000 + }, + { + "epoch": 0.16, + "learning_rate": 0.000422296890672016, + "loss": 1.49, + "theoretical_loss": 3.910656059281591, + "tokens_seen": 540737536 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042228686058174526, + "loss": 1.4587, + "theoretical_loss": 3.910606066622507, + "tokens_seen": 540803072 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042227683049147444, + "loss": 1.6862, + "theoretical_loss": 3.9105560817173783, + "tokens_seen": 540868608 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004222668004012036, + "loss": 1.4768, + "theoretical_loss": 3.910506104564063, + "tokens_seen": 540934144 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004222567703109328, + "loss": 1.5955, + "theoretical_loss": 3.9104561351604206, + "tokens_seen": 540999680 + }, + { + "epoch": 0.16, + "learning_rate": 0.000422246740220662, + "loss": 1.5579, + "theoretical_loss": 3.9104061735043105, + "tokens_seen": 541065216 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042223671013039116, + "loss": 1.5867, + "theoretical_loss": 3.9103562195935933, + "tokens_seen": 541130752 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004222266800401204, + "loss": 1.5684, + "theoretical_loss": 3.91030627342613, + "tokens_seen": 541196288 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004222166499498495, + "loss": 1.4885, + "theoretical_loss": 3.910256334999783, + "tokens_seen": 541261824 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042220661985957876, + "loss": 1.4582, + "theoretical_loss": 3.910206404312416, + "tokens_seen": 541327360 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004221965897693079, + "loss": 1.4842, + "theoretical_loss": 3.9101564813618923, + "tokens_seen": 541392896 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004221865596790371, + "loss": 1.473, + "theoretical_loss": 3.9101065661460774, + "tokens_seen": 541458432 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004221765295887663, + "loss": 1.4676, + "theoretical_loss": 3.9100566586628362, + "tokens_seen": 541523968 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004221664994984955, + "loss": 1.5545, + "theoretical_loss": 3.9100067589100362, + "tokens_seen": 541589504 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042215646940822467, + "loss": 1.6311, + "theoretical_loss": 3.909956866885544, + "tokens_seen": 541655040 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004221464393179539, + "loss": 1.452, + "theoretical_loss": 3.909906982587228, + "tokens_seen": 541720576 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042213640922768303, + "loss": 1.4722, + "theoretical_loss": 3.909857106012957, + "tokens_seen": 541786112 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042212637913741226, + "loss": 1.4463, + "theoretical_loss": 3.9098072371606016, + "tokens_seen": 541851648 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004221163490471414, + "loss": 1.479, + "theoretical_loss": 3.9097573760280326, + "tokens_seen": 541917184 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004221063189568706, + "loss": 1.4978, + "theoretical_loss": 3.90970752261312, + "tokens_seen": 541982720 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004220962888665998, + "loss": 1.5736, + "theoretical_loss": 3.9096576769137386, + "tokens_seen": 542048256 + }, + { + "epoch": 0.16, + "learning_rate": 0.000422086258776329, + "loss": 1.5714, + "theoretical_loss": 3.90960783892776, + "tokens_seen": 542113792 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042207622868605817, + "loss": 1.4324, + "theoretical_loss": 3.9095580086530592, + "tokens_seen": 542179328 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042206619859578735, + "loss": 1.5211, + "theoretical_loss": 3.90950818608751, + "tokens_seen": 542244864 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042205616850551653, + "loss": 1.4893, + "theoretical_loss": 3.9094583712289896, + "tokens_seen": 542310400 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042204613841524577, + "loss": 1.5049, + "theoretical_loss": 3.909408564075374, + "tokens_seen": 542375936 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004220361083249749, + "loss": 1.5895, + "theoretical_loss": 3.9093587646245407, + "tokens_seen": 542441472 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042202607823470413, + "loss": 1.4932, + "theoretical_loss": 3.9093089728743684, + "tokens_seen": 542507008 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004220160481444333, + "loss": 1.5405, + "theoretical_loss": 3.9092591888227357, + "tokens_seen": 542572544 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004220060180541625, + "loss": 1.5055, + "theoretical_loss": 3.909209412467523, + "tokens_seen": 542638080 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004219959879638917, + "loss": 1.4428, + "theoretical_loss": 3.9091596438066114, + "tokens_seen": 542703616 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042198595787362085, + "loss": 1.5746, + "theoretical_loss": 3.909109882837882, + "tokens_seen": 542769152 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042197592778335004, + "loss": 1.5455, + "theoretical_loss": 3.9090601295592178, + "tokens_seen": 542834688 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042196589769307927, + "loss": 1.5461, + "theoretical_loss": 3.9090103839685013, + "tokens_seen": 542900224 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004219558676028084, + "loss": 1.4434, + "theoretical_loss": 3.908960646063618, + "tokens_seen": 542965760 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042194583751253763, + "loss": 1.4686, + "theoretical_loss": 3.908910915842452, + "tokens_seen": 543031296 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042193580742226676, + "loss": 1.5159, + "theoretical_loss": 3.90886119330289, + "tokens_seen": 543096832 + }, + { + "epoch": 0.16, + "learning_rate": 0.000421925777331996, + "loss": 1.6537, + "theoretical_loss": 3.908811478442818, + "tokens_seen": 543162368 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042191574724172523, + "loss": 1.6233, + "theoretical_loss": 3.908761771260124, + "tokens_seen": 543227904 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042190571715145436, + "loss": 1.4781, + "theoretical_loss": 3.908712071752696, + "tokens_seen": 543293440 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004218956870611836, + "loss": 1.5067, + "theoretical_loss": 3.908662379918424, + "tokens_seen": 543358976 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004218856569709127, + "loss": 1.6113, + "theoretical_loss": 3.908612695755197, + "tokens_seen": 543424512 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042187562688064196, + "loss": 1.499, + "theoretical_loss": 3.9085630192609067, + "tokens_seen": 543490048 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042186559679037114, + "loss": 1.5223, + "theoretical_loss": 3.908513350433444, + "tokens_seen": 543555584 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004218555667001003, + "loss": 1.5841, + "theoretical_loss": 3.908463689270703, + "tokens_seen": 543621120 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004218455366098295, + "loss": 1.5641, + "theoretical_loss": 3.9084140357705754, + "tokens_seen": 543686656 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042183550651955873, + "loss": 1.5155, + "theoretical_loss": 3.9083643899309566, + "tokens_seen": 543752192 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042182547642928786, + "loss": 1.6255, + "theoretical_loss": 3.9083147517497414, + "tokens_seen": 543817728 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004218154463390171, + "loss": 1.5083, + "theoretical_loss": 3.908265121224825, + "tokens_seen": 543883264 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 207102, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5554263591766357, + "objective/train/theoretical_loss": 3.908215498354105, + "objective/train/tokens_used": 543948800, + "theoretical_loss": 3.908215498354105, + "tokens_seen": 543948800 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004218054162487462, + "loss": 1.5293, + "theoretical_loss": 3.908215498354105, + "tokens_seen": 543948800 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042179538615847546, + "loss": 1.6625, + "theoretical_loss": 3.9081658831354784, + "tokens_seen": 544014336 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042178535606820464, + "loss": 1.5106, + "theoretical_loss": 3.908116275566844, + "tokens_seen": 544079872 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004217753259779338, + "loss": 1.5095, + "theoretical_loss": 3.9080666756461007, + "tokens_seen": 544145408 + }, + { + "epoch": 0.16, + "learning_rate": 0.000421765295887663, + "loss": 1.4574, + "theoretical_loss": 3.908017083371149, + "tokens_seen": 544210944 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004217552657973922, + "loss": 1.5465, + "theoretical_loss": 3.9079674987398896, + "tokens_seen": 544276480 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042174523570712136, + "loss": 1.4554, + "theoretical_loss": 3.907917921750224, + "tokens_seen": 544342016 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004217352056168506, + "loss": 1.5767, + "theoretical_loss": 3.9078683524000546, + "tokens_seen": 544407552 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004217251755265797, + "loss": 1.5141, + "theoretical_loss": 3.907818790687285, + "tokens_seen": 544473088 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042171514543630896, + "loss": 1.5037, + "theoretical_loss": 3.90776923660982, + "tokens_seen": 544538624 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004217051153460381, + "loss": 1.4746, + "theoretical_loss": 3.907719690165564, + "tokens_seen": 544604160 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004216950852557673, + "loss": 1.4643, + "theoretical_loss": 3.9076701513524226, + "tokens_seen": 544669696 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004216850551654965, + "loss": 1.5227, + "theoretical_loss": 3.9076206201683026, + "tokens_seen": 544735232 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004216750250752257, + "loss": 1.5413, + "theoretical_loss": 3.907571096611112, + "tokens_seen": 544800768 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042166499498495487, + "loss": 1.5265, + "theoretical_loss": 3.9075215806787584, + "tokens_seen": 544866304 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004216549648946841, + "loss": 1.4879, + "theoretical_loss": 3.907472072369152, + "tokens_seen": 544931840 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042164493480441323, + "loss": 1.5324, + "theoretical_loss": 3.907422571680202, + "tokens_seen": 544997376 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042163490471414246, + "loss": 1.6126, + "theoretical_loss": 3.907373078609819, + "tokens_seen": 545062912 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004216248746238716, + "loss": 1.5368, + "theoretical_loss": 3.9073235931559154, + "tokens_seen": 545128448 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004216148445336008, + "loss": 1.5401, + "theoretical_loss": 3.907274115316403, + "tokens_seen": 545193984 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042160481444333, + "loss": 1.4964, + "theoretical_loss": 3.907224645089195, + "tokens_seen": 545259520 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004215947843530592, + "loss": 1.4565, + "theoretical_loss": 3.907175182472206, + "tokens_seen": 545325056 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042158475426278837, + "loss": 1.5007, + "theoretical_loss": 3.9071257274633506, + "tokens_seen": 545390592 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042157472417251755, + "loss": 1.5247, + "theoretical_loss": 3.9070762800605454, + "tokens_seen": 545456128 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042156469408224673, + "loss": 1.5719, + "theoretical_loss": 3.9070268402617057, + "tokens_seen": 545521664 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042155466399197597, + "loss": 1.5683, + "theoretical_loss": 3.9069774080647495, + "tokens_seen": 545587200 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004215446339017051, + "loss": 1.5341, + "theoretical_loss": 3.9069279834675945, + "tokens_seen": 545652736 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042153460381143433, + "loss": 1.4581, + "theoretical_loss": 3.9068785664681602, + "tokens_seen": 545718272 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004215245737211635, + "loss": 1.4634, + "theoretical_loss": 3.906829157064367, + "tokens_seen": 545783808 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004215145436308927, + "loss": 1.5774, + "theoretical_loss": 3.906779755254134, + "tokens_seen": 545849344 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004215045135406219, + "loss": 1.558, + "theoretical_loss": 3.9067303610353847, + "tokens_seen": 545914880 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042149448345035105, + "loss": 1.3913, + "theoretical_loss": 3.9066809744060396, + "tokens_seen": 545980416 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042148445336008024, + "loss": 1.506, + "theoretical_loss": 3.9066315953640225, + "tokens_seen": 546045952 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042147442326980947, + "loss": 1.5149, + "theoretical_loss": 3.9065822239072574, + "tokens_seen": 546111488 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004214643931795386, + "loss": 1.5711, + "theoretical_loss": 3.9065328600336695, + "tokens_seen": 546177024 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042145436308926783, + "loss": 1.5251, + "theoretical_loss": 3.9064835037411836, + "tokens_seen": 546242560 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042144433299899696, + "loss": 1.5654, + "theoretical_loss": 3.9064341550277266, + "tokens_seen": 546308096 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004214343029087262, + "loss": 1.5334, + "theoretical_loss": 3.9063848138912247, + "tokens_seen": 546373632 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004214242728184554, + "loss": 1.4665, + "theoretical_loss": 3.9063354803296075, + "tokens_seen": 546439168 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042141424272818456, + "loss": 1.601, + "theoretical_loss": 3.906286154340802, + "tokens_seen": 546504704 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042140421263791374, + "loss": 1.5163, + "theoretical_loss": 3.9062368359227397, + "tokens_seen": 546570240 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004213941825476429, + "loss": 1.5741, + "theoretical_loss": 3.90618752507335, + "tokens_seen": 546635776 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004213841524573721, + "loss": 1.6265, + "theoretical_loss": 3.906138221790565, + "tokens_seen": 546701312 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042137412236710134, + "loss": 1.5843, + "theoretical_loss": 3.9060889260723153, + "tokens_seen": 546766848 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042136409227683046, + "loss": 1.5042, + "theoretical_loss": 3.906039637916535, + "tokens_seen": 546832384 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004213540621865597, + "loss": 1.505, + "theoretical_loss": 3.9059903573211576, + "tokens_seen": 546897920 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004213440320962889, + "loss": 1.5573, + "theoretical_loss": 3.905941084284117, + "tokens_seen": 546963456 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042133400200601806, + "loss": 1.5331, + "theoretical_loss": 3.9058918188033496, + "tokens_seen": 547028992 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042132397191574724, + "loss": 1.4566, + "theoretical_loss": 3.905842560876791, + "tokens_seen": 547094528 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004213139418254764, + "loss": 1.3923, + "theoretical_loss": 3.9057933105023777, + "tokens_seen": 547160064 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 208235, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.380865216255188, + "objective/train/theoretical_loss": 3.9057440676780484, + "objective/train/tokens_used": 547225600, + "theoretical_loss": 3.9057440676780484, + "tokens_seen": 547225600 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004213039117352056, + "loss": 1.4683, + "theoretical_loss": 3.9057440676780484, + "tokens_seen": 547225600 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042129388164493484, + "loss": 1.5381, + "theoretical_loss": 3.905694832401741, + "tokens_seen": 547291136 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042128385155466397, + "loss": 1.5316, + "theoretical_loss": 3.905645604671395, + "tokens_seen": 547356672 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004212738214643932, + "loss": 1.6149, + "theoretical_loss": 3.9055963844849506, + "tokens_seen": 547422208 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042126379137412233, + "loss": 1.5528, + "theoretical_loss": 3.9055471718403494, + "tokens_seen": 547487744 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042125376128385156, + "loss": 1.484, + "theoretical_loss": 3.905497966735532, + "tokens_seen": 547553280 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042124373119358075, + "loss": 1.5074, + "theoretical_loss": 3.9054487691684425, + "tokens_seen": 547618816 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004212337011033099, + "loss": 1.4105, + "theoretical_loss": 3.905399579137023, + "tokens_seen": 547684352 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004212236710130391, + "loss": 1.4884, + "theoretical_loss": 3.9053503966392187, + "tokens_seen": 547749888 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004212136409227683, + "loss": 1.575, + "theoretical_loss": 3.905301221672974, + "tokens_seen": 547815424 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042120361083249747, + "loss": 1.439, + "theoretical_loss": 3.9052520542362346, + "tokens_seen": 547880960 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004211935807422267, + "loss": 1.4202, + "theoretical_loss": 3.9052028943269477, + "tokens_seen": 547946496 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042118355065195583, + "loss": 1.5545, + "theoretical_loss": 3.905153741943061, + "tokens_seen": 548012032 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042117352056168507, + "loss": 1.479, + "theoretical_loss": 3.905104597082522, + "tokens_seen": 548077568 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004211634904714143, + "loss": 1.4875, + "theoretical_loss": 3.9050554597432807, + "tokens_seen": 548143104 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042115346038114343, + "loss": 1.5335, + "theoretical_loss": 3.905006329923286, + "tokens_seen": 548208640 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042114343029087266, + "loss": 1.4877, + "theoretical_loss": 3.904957207620489, + "tokens_seen": 548274176 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004211334002006018, + "loss": 1.5809, + "theoretical_loss": 3.904908092832841, + "tokens_seen": 548339712 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042112337011033103, + "loss": 1.5197, + "theoretical_loss": 3.904858985558295, + "tokens_seen": 548405248 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004211133400200602, + "loss": 1.4835, + "theoretical_loss": 3.904809885794803, + "tokens_seen": 548470784 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004211033099297894, + "loss": 1.524, + "theoretical_loss": 3.9047607935403197, + "tokens_seen": 548536320 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042109327983951857, + "loss": 1.4268, + "theoretical_loss": 3.9047117087927994, + "tokens_seen": 548601856 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042108324974924775, + "loss": 1.6444, + "theoretical_loss": 3.904662631550198, + "tokens_seen": 548667392 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042107321965897693, + "loss": 1.5784, + "theoretical_loss": 3.904613561810471, + "tokens_seen": 548732928 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042106318956870617, + "loss": 1.5606, + "theoretical_loss": 3.9045644995715767, + "tokens_seen": 548798464 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004210531594784353, + "loss": 1.5268, + "theoretical_loss": 3.9045154448314716, + "tokens_seen": 548864000 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042104312938816453, + "loss": 1.4718, + "theoretical_loss": 3.9044663975881155, + "tokens_seen": 548929536 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004210330992978937, + "loss": 1.5451, + "theoretical_loss": 3.9044173578394674, + "tokens_seen": 548995072 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004210230692076229, + "loss": 1.4392, + "theoretical_loss": 3.904368325583488, + "tokens_seen": 549060608 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004210130391173521, + "loss": 1.4643, + "theoretical_loss": 3.9043193008181376, + "tokens_seen": 549126144 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042100300902708125, + "loss": 1.5177, + "theoretical_loss": 3.9042702835413787, + "tokens_seen": 549191680 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042099297893681044, + "loss": 1.5251, + "theoretical_loss": 3.9042212737511743, + "tokens_seen": 549257216 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042098294884653967, + "loss": 1.4831, + "theoretical_loss": 3.904172271445487, + "tokens_seen": 549322752 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004209729187562688, + "loss": 1.472, + "theoretical_loss": 3.9041232766222813, + "tokens_seen": 549388288 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042096288866599803, + "loss": 1.5471, + "theoretical_loss": 3.9040742892795226, + "tokens_seen": 549453824 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042095285857572716, + "loss": 1.5193, + "theoretical_loss": 3.9040253094151773, + "tokens_seen": 549519360 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004209428284854564, + "loss": 1.5046, + "theoretical_loss": 3.9039763370272107, + "tokens_seen": 549584896 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004209327983951856, + "loss": 1.4369, + "theoretical_loss": 3.903927372113592, + "tokens_seen": 549650432 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042092276830491476, + "loss": 1.5394, + "theoretical_loss": 3.9038784146722874, + "tokens_seen": 549715968 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042091273821464394, + "loss": 1.5281, + "theoretical_loss": 3.9038294647012672, + "tokens_seen": 549781504 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004209027081243731, + "loss": 1.4378, + "theoretical_loss": 3.9037805221985016, + "tokens_seen": 549847040 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004208926780341023, + "loss": 1.5435, + "theoretical_loss": 3.90373158716196, + "tokens_seen": 549912576 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042088264794383154, + "loss": 1.6108, + "theoretical_loss": 3.903682659589615, + "tokens_seen": 549978112 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042087261785356066, + "loss": 1.4775, + "theoretical_loss": 3.903633739479438, + "tokens_seen": 550043648 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004208625877632899, + "loss": 1.4998, + "theoretical_loss": 3.9035848268294027, + "tokens_seen": 550109184 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004208525576730191, + "loss": 1.4522, + "theoretical_loss": 3.9035359216374825, + "tokens_seen": 550174720 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042084252758274826, + "loss": 1.3701, + "theoretical_loss": 3.9034870239016524, + "tokens_seen": 550240256 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042083249749247744, + "loss": 1.3904, + "theoretical_loss": 3.903438133619887, + "tokens_seen": 550305792 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004208224674022066, + "loss": 1.5489, + "theoretical_loss": 3.9033892507901635, + "tokens_seen": 550371328 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004208124373119358, + "loss": 1.4393, + "theoretical_loss": 3.9033403754104583, + "tokens_seen": 550436864 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 209442, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.495992660522461, + "objective/train/theoretical_loss": 3.9032915074787486, + "objective/train/tokens_used": 550502400, + "theoretical_loss": 3.9032915074787486, + "tokens_seen": 550502400 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042080240722166504, + "loss": 1.5283, + "theoretical_loss": 3.9032915074787486, + "tokens_seen": 550502400 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042079237713139417, + "loss": 1.5425, + "theoretical_loss": 3.903242646993014, + "tokens_seen": 550567936 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004207823470411234, + "loss": 1.4709, + "theoretical_loss": 3.9031937939512336, + "tokens_seen": 550633472 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042077231695085253, + "loss": 1.5132, + "theoretical_loss": 3.903144948351387, + "tokens_seen": 550699008 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042076228686058176, + "loss": 1.5187, + "theoretical_loss": 3.9030961101914556, + "tokens_seen": 550764544 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042075225677031095, + "loss": 1.4936, + "theoretical_loss": 3.903047279469421, + "tokens_seen": 550830080 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004207422266800401, + "loss": 1.5911, + "theoretical_loss": 3.9029984561832656, + "tokens_seen": 550895616 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004207321965897693, + "loss": 1.6871, + "theoretical_loss": 3.9029496403309727, + "tokens_seen": 550961152 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004207221664994985, + "loss": 1.4044, + "theoretical_loss": 3.902900831910527, + "tokens_seen": 551026688 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042071213640922767, + "loss": 1.5889, + "theoretical_loss": 3.902852030919912, + "tokens_seen": 551092224 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004207021063189569, + "loss": 1.5274, + "theoretical_loss": 3.902803237357115, + "tokens_seen": 551157760 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042069207622868603, + "loss": 1.4009, + "theoretical_loss": 3.9027544512201207, + "tokens_seen": 551223296 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042068204613841527, + "loss": 1.5514, + "theoretical_loss": 3.902705672506918, + "tokens_seen": 551288832 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042067201604814445, + "loss": 1.5837, + "theoretical_loss": 3.9026569012154932, + "tokens_seen": 551354368 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042066198595787363, + "loss": 1.4148, + "theoretical_loss": 3.9026081373438366, + "tokens_seen": 551419904 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004206519558676028, + "loss": 1.5137, + "theoretical_loss": 3.9025593808899366, + "tokens_seen": 551485440 + }, + { + "epoch": 0.17, + "learning_rate": 0.000420641925777332, + "loss": 1.4664, + "theoretical_loss": 3.9025106318517846, + "tokens_seen": 551550976 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004206318956870612, + "loss": 1.6103, + "theoretical_loss": 3.902461890227371, + "tokens_seen": 551616512 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004206218655967904, + "loss": 1.5906, + "theoretical_loss": 3.902413156014688, + "tokens_seen": 551682048 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042061183550651954, + "loss": 1.5158, + "theoretical_loss": 3.902364429211728, + "tokens_seen": 551747584 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042060180541624877, + "loss": 1.5293, + "theoretical_loss": 3.9023157098164853, + "tokens_seen": 551813120 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004205917753259779, + "loss": 1.5598, + "theoretical_loss": 3.902266997826953, + "tokens_seen": 551878656 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042058174523570713, + "loss": 1.4665, + "theoretical_loss": 3.902218293241127, + "tokens_seen": 551944192 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004205717151454363, + "loss": 1.4775, + "theoretical_loss": 3.9021695960570026, + "tokens_seen": 552009728 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004205616850551655, + "loss": 1.5925, + "theoretical_loss": 3.902120906272577, + "tokens_seen": 552075264 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004205516549648947, + "loss": 1.4407, + "theoretical_loss": 3.9020722238858467, + "tokens_seen": 552140800 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004205416248746239, + "loss": 1.5064, + "theoretical_loss": 3.902023548894811, + "tokens_seen": 552206336 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042053159478435304, + "loss": 1.5655, + "theoretical_loss": 3.901974881297468, + "tokens_seen": 552271872 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004205215646940823, + "loss": 1.4468, + "theoretical_loss": 3.9019262210918173, + "tokens_seen": 552337408 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004205115346038114, + "loss": 1.5782, + "theoretical_loss": 3.90187756827586, + "tokens_seen": 552402944 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042050150451354064, + "loss": 1.4932, + "theoretical_loss": 3.9018289228475975, + "tokens_seen": 552468480 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004204914744232698, + "loss": 1.5204, + "theoretical_loss": 3.9017802848050307, + "tokens_seen": 552534016 + }, + { + "epoch": 0.17, + "learning_rate": 0.000420481444332999, + "loss": 1.4518, + "theoretical_loss": 3.9017316541461637, + "tokens_seen": 552599552 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004204714142427282, + "loss": 1.4562, + "theoretical_loss": 3.9016830308689996, + "tokens_seen": 552665088 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042046138415245736, + "loss": 1.467, + "theoretical_loss": 3.901634414971543, + "tokens_seen": 552730624 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042045135406218654, + "loss": 1.5539, + "theoretical_loss": 3.901585806451799, + "tokens_seen": 552796160 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004204413239719158, + "loss": 1.4841, + "theoretical_loss": 3.901537205307773, + "tokens_seen": 552861696 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004204312938816449, + "loss": 1.3881, + "theoretical_loss": 3.9014886115374723, + "tokens_seen": 552927232 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042042126379137414, + "loss": 1.5495, + "theoretical_loss": 3.9014400251389043, + "tokens_seen": 552992768 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004204112337011033, + "loss": 1.6577, + "theoretical_loss": 3.901391446110077, + "tokens_seen": 553058304 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004204012036108325, + "loss": 1.4967, + "theoretical_loss": 3.901342874449, + "tokens_seen": 553123840 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042039117352056174, + "loss": 1.4881, + "theoretical_loss": 3.9012943101536823, + "tokens_seen": 553189376 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042038114343029086, + "loss": 1.4975, + "theoretical_loss": 3.901245753222135, + "tokens_seen": 553254912 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004203711133400201, + "loss": 1.5273, + "theoretical_loss": 3.90119720365237, + "tokens_seen": 553320448 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004203610832497493, + "loss": 1.5724, + "theoretical_loss": 3.9011486614423982, + "tokens_seen": 553385984 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042035105315947846, + "loss": 1.575, + "theoretical_loss": 3.901100126590233, + "tokens_seen": 553451520 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042034102306920764, + "loss": 1.5101, + "theoretical_loss": 3.901051599093889, + "tokens_seen": 553517056 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004203309929789368, + "loss": 1.5405, + "theoretical_loss": 3.901003078951379, + "tokens_seen": 553582592 + }, + { + "epoch": 0.17, + "learning_rate": 0.000420320962888666, + "loss": 1.5595, + "theoretical_loss": 3.9009545661607192, + "tokens_seen": 553648128 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042031093279839524, + "loss": 1.5251, + "theoretical_loss": 3.900906060719925, + "tokens_seen": 553713664 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 210755, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6109105348587036, + "objective/train/theoretical_loss": 3.9008575626270146, + "objective/train/tokens_used": 553779200, + "theoretical_loss": 3.9008575626270146, + "tokens_seen": 553779200 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042030090270812437, + "loss": 1.5804, + "theoretical_loss": 3.9008575626270146, + "tokens_seen": 553779200 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004202908726178536, + "loss": 1.5611, + "theoretical_loss": 3.900809071880004, + "tokens_seen": 553844736 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042028084252758273, + "loss": 1.5501, + "theoretical_loss": 3.900760588476912, + "tokens_seen": 553910272 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042027081243731196, + "loss": 1.5819, + "theoretical_loss": 3.900712112415758, + "tokens_seen": 553975808 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042026078234704115, + "loss": 1.4877, + "theoretical_loss": 3.900663643694561, + "tokens_seen": 554041344 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004202507522567703, + "loss": 1.6396, + "theoretical_loss": 3.900615182311342, + "tokens_seen": 554106880 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004202407221664995, + "loss": 1.5147, + "theoretical_loss": 3.900566728264123, + "tokens_seen": 554172416 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004202306920762287, + "loss": 1.5238, + "theoretical_loss": 3.9005182815509256, + "tokens_seen": 554237952 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042022066198595787, + "loss": 1.4852, + "theoretical_loss": 3.9004698421697723, + "tokens_seen": 554303488 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004202106318956871, + "loss": 1.5277, + "theoretical_loss": 3.900421410118688, + "tokens_seen": 554369024 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042020060180541623, + "loss": 1.4922, + "theoretical_loss": 3.9003729853956957, + "tokens_seen": 554434560 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042019057171514547, + "loss": 1.4854, + "theoretical_loss": 3.9003245679988217, + "tokens_seen": 554500096 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042018054162487465, + "loss": 1.4742, + "theoretical_loss": 3.9002761579260907, + "tokens_seen": 554565632 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042017051153460383, + "loss": 1.5631, + "theoretical_loss": 3.9002277551755307, + "tokens_seen": 554631168 + }, + { + "epoch": 0.17, + "learning_rate": 0.000420160481444333, + "loss": 1.5757, + "theoretical_loss": 3.9001793597451684, + "tokens_seen": 554696704 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004201504513540622, + "loss": 1.3057, + "theoretical_loss": 3.900130971633033, + "tokens_seen": 554762240 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004201404212637914, + "loss": 1.3806, + "theoretical_loss": 3.9000825908371524, + "tokens_seen": 554827776 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004201303911735206, + "loss": 1.5162, + "theoretical_loss": 3.9000342173555573, + "tokens_seen": 554893312 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042012036108324974, + "loss": 1.453, + "theoretical_loss": 3.899985851186277, + "tokens_seen": 554958848 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042011033099297897, + "loss": 1.5477, + "theoretical_loss": 3.8999374923273447, + "tokens_seen": 555024384 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004201003009027081, + "loss": 1.446, + "theoretical_loss": 3.899889140776791, + "tokens_seen": 555089920 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042009027081243733, + "loss": 1.4539, + "theoretical_loss": 3.8998407965326494, + "tokens_seen": 555155456 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004200802407221665, + "loss": 1.3861, + "theoretical_loss": 3.899792459592953, + "tokens_seen": 555220992 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004200702106318957, + "loss": 1.4862, + "theoretical_loss": 3.899744129955737, + "tokens_seen": 555286528 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004200601805416249, + "loss": 1.4514, + "theoretical_loss": 3.899695807619035, + "tokens_seen": 555352064 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004200501504513541, + "loss": 1.4402, + "theoretical_loss": 3.899647492580885, + "tokens_seen": 555417600 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042004012036108324, + "loss": 1.4226, + "theoretical_loss": 3.899599184839322, + "tokens_seen": 555483136 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004200300902708125, + "loss": 1.4831, + "theoretical_loss": 3.899550884392384, + "tokens_seen": 555548672 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004200200601805416, + "loss": 1.4504, + "theoretical_loss": 3.8995025912381087, + "tokens_seen": 555614208 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042001003009027084, + "loss": 1.5528, + "theoretical_loss": 3.899454305374536, + "tokens_seen": 555679744 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042, + "loss": 1.645, + "theoretical_loss": 3.899406026799704, + "tokens_seen": 555745280 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004199899699097292, + "loss": 1.4379, + "theoretical_loss": 3.899357755511655, + "tokens_seen": 555810816 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004199799398194584, + "loss": 1.505, + "theoretical_loss": 3.899309491508429, + "tokens_seen": 555876352 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041996990972918756, + "loss": 1.6008, + "theoretical_loss": 3.899261234788068, + "tokens_seen": 555941888 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041995987963891674, + "loss": 1.4045, + "theoretical_loss": 3.8992129853486155, + "tokens_seen": 556007424 + }, + { + "epoch": 0.17, + "learning_rate": 0.000419949849548646, + "loss": 1.4209, + "theoretical_loss": 3.8991647431881145, + "tokens_seen": 556072960 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004199398194583751, + "loss": 1.4824, + "theoretical_loss": 3.899116508304609, + "tokens_seen": 556138496 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041992978936810434, + "loss": 1.5952, + "theoretical_loss": 3.899068280696144, + "tokens_seen": 556204032 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041991975927783347, + "loss": 1.5, + "theoretical_loss": 3.8990200603607654, + "tokens_seen": 556269568 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004199097291875627, + "loss": 1.5783, + "theoretical_loss": 3.8989718472965196, + "tokens_seen": 556335104 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004198996990972919, + "loss": 1.5356, + "theoretical_loss": 3.898923641501454, + "tokens_seen": 556400640 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041988966900702106, + "loss": 1.5216, + "theoretical_loss": 3.898875442973617, + "tokens_seen": 556466176 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041987963891675025, + "loss": 1.5128, + "theoretical_loss": 3.8988272517110563, + "tokens_seen": 556531712 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004198696088264795, + "loss": 1.5042, + "theoretical_loss": 3.898779067711822, + "tokens_seen": 556597248 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004198595787362086, + "loss": 1.574, + "theoretical_loss": 3.8987308909739644, + "tokens_seen": 556662784 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041984954864593784, + "loss": 1.4573, + "theoretical_loss": 3.898682721495535, + "tokens_seen": 556728320 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041983951855566697, + "loss": 1.5314, + "theoretical_loss": 3.8986345592745844, + "tokens_seen": 556793856 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004198294884653962, + "loss": 1.458, + "theoretical_loss": 3.898586404309166, + "tokens_seen": 556859392 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004198194583751254, + "loss": 1.2861, + "theoretical_loss": 3.8985382565973326, + "tokens_seen": 556924928 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041980942828485457, + "loss": 1.5385, + "theoretical_loss": 3.898490116137139, + "tokens_seen": 556990464 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 211950, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.583616018295288, + "objective/train/theoretical_loss": 3.898441982926639, + "objective/train/tokens_used": 557056000, + "theoretical_loss": 3.898441982926639, + "tokens_seen": 557056000 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041979939819458375, + "loss": 1.556, + "theoretical_loss": 3.898441982926639, + "tokens_seen": 557056000 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041978936810431293, + "loss": 1.4789, + "theoretical_loss": 3.8983938569638883, + "tokens_seen": 557121536 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004197793380140421, + "loss": 1.5111, + "theoretical_loss": 3.8983457382469435, + "tokens_seen": 557187072 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041976930792377135, + "loss": 1.5581, + "theoretical_loss": 3.8982976267738616, + "tokens_seen": 557252608 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041975927783350047, + "loss": 1.4506, + "theoretical_loss": 3.8982495225427005, + "tokens_seen": 557318144 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004197492477432297, + "loss": 1.5215, + "theoretical_loss": 3.8982014255515183, + "tokens_seen": 557383680 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041973921765295883, + "loss": 1.4698, + "theoretical_loss": 3.8981533357983746, + "tokens_seen": 557449216 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041972918756268807, + "loss": 1.4863, + "theoretical_loss": 3.8981052532813294, + "tokens_seen": 557514752 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041971915747241725, + "loss": 1.4912, + "theoretical_loss": 3.898057177998443, + "tokens_seen": 557580288 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041970912738214643, + "loss": 1.5483, + "theoretical_loss": 3.898009109947777, + "tokens_seen": 557645824 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004196990972918756, + "loss": 1.5763, + "theoretical_loss": 3.897961049127395, + "tokens_seen": 557711360 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041968906720160485, + "loss": 1.5349, + "theoretical_loss": 3.8979129955353584, + "tokens_seen": 557776896 + }, + { + "epoch": 0.17, + "learning_rate": 0.000419679037111334, + "loss": 1.6247, + "theoretical_loss": 3.8978649491697315, + "tokens_seen": 557842432 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004196690070210632, + "loss": 1.4761, + "theoretical_loss": 3.8978169100285784, + "tokens_seen": 557907968 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004196589769307924, + "loss": 1.506, + "theoretical_loss": 3.8977688781099653, + "tokens_seen": 557973504 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004196489468405216, + "loss": 1.4843, + "theoretical_loss": 3.8977208534119576, + "tokens_seen": 558039040 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004196389167502508, + "loss": 1.4606, + "theoretical_loss": 3.8976728359326223, + "tokens_seen": 558104576 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041962888665997994, + "loss": 1.5304, + "theoretical_loss": 3.897624825670026, + "tokens_seen": 558170112 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041961885656970917, + "loss": 1.5975, + "theoretical_loss": 3.8975768226222383, + "tokens_seen": 558235648 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004196088264794383, + "loss": 1.4744, + "theoretical_loss": 3.897528826787327, + "tokens_seen": 558301184 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041959879638916753, + "loss": 1.5038, + "theoretical_loss": 3.8974808381633625, + "tokens_seen": 558366720 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004195887662988967, + "loss": 1.4525, + "theoretical_loss": 3.8974328567484147, + "tokens_seen": 558432256 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004195787362086259, + "loss": 1.5449, + "theoretical_loss": 3.8973848825405555, + "tokens_seen": 558497792 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004195687061183551, + "loss": 1.4952, + "theoretical_loss": 3.8973369155378563, + "tokens_seen": 558563328 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004195586760280843, + "loss": 1.4974, + "theoretical_loss": 3.89728895573839, + "tokens_seen": 558628864 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041954864593781344, + "loss": 1.6135, + "theoretical_loss": 3.8972410031402296, + "tokens_seen": 558694400 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004195386158475427, + "loss": 1.5255, + "theoretical_loss": 3.8971930577414495, + "tokens_seen": 558759936 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004195285857572718, + "loss": 1.5008, + "theoretical_loss": 3.8971451195401254, + "tokens_seen": 558825472 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041951855566700104, + "loss": 1.5584, + "theoretical_loss": 3.897097188534332, + "tokens_seen": 558891008 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004195085255767302, + "loss": 1.5382, + "theoretical_loss": 3.8970492647221464, + "tokens_seen": 558956544 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004194984954864594, + "loss": 1.4753, + "theoretical_loss": 3.897001348101645, + "tokens_seen": 559022080 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004194884653961886, + "loss": 1.5929, + "theoretical_loss": 3.8969534386709053, + "tokens_seen": 559087616 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041947843530591776, + "loss": 1.5455, + "theoretical_loss": 3.896905536428007, + "tokens_seen": 559153152 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041946840521564694, + "loss": 1.4819, + "theoretical_loss": 3.8968576413710294, + "tokens_seen": 559218688 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004194583751253762, + "loss": 1.4647, + "theoretical_loss": 3.8968097534980517, + "tokens_seen": 559284224 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004194483450351053, + "loss": 1.5633, + "theoretical_loss": 3.8967618728071556, + "tokens_seen": 559349760 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041943831494483454, + "loss": 1.5448, + "theoretical_loss": 3.896713999296422, + "tokens_seen": 559415296 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041942828485456367, + "loss": 1.4689, + "theoretical_loss": 3.8966661329639334, + "tokens_seen": 559480832 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004194182547642929, + "loss": 1.6063, + "theoretical_loss": 3.896618273807773, + "tokens_seen": 559546368 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004194082246740221, + "loss": 1.5302, + "theoretical_loss": 3.8965704218260244, + "tokens_seen": 559611904 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041939819458375126, + "loss": 1.5085, + "theoretical_loss": 3.896522577016772, + "tokens_seen": 559677440 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041938816449348045, + "loss": 1.4744, + "theoretical_loss": 3.896474739378102, + "tokens_seen": 559742976 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004193781344032097, + "loss": 1.5524, + "theoretical_loss": 3.896426908908099, + "tokens_seen": 559808512 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004193681043129388, + "loss": 1.5314, + "theoretical_loss": 3.89637908560485, + "tokens_seen": 559874048 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041935807422266804, + "loss": 1.4485, + "theoretical_loss": 3.896331269466443, + "tokens_seen": 559939584 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041934804413239717, + "loss": 1.4603, + "theoretical_loss": 3.8962834604909657, + "tokens_seen": 560005120 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004193380140421264, + "loss": 1.5443, + "theoretical_loss": 3.8962356586765075, + "tokens_seen": 560070656 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004193279839518556, + "loss": 1.475, + "theoretical_loss": 3.896187864021157, + "tokens_seen": 560136192 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041931795386158477, + "loss": 1.5223, + "theoretical_loss": 3.8961400765230065, + "tokens_seen": 560201728 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041930792377131395, + "loss": 1.5176, + "theoretical_loss": 3.8960922961801447, + "tokens_seen": 560267264 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 213150, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5923283100128174, + "objective/train/theoretical_loss": 3.8960445229906657, + "objective/train/tokens_used": 560332800, + "theoretical_loss": 3.8960445229906657, + "tokens_seen": 560332800 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041929789368104313, + "loss": 1.4737, + "theoretical_loss": 3.8960445229906657, + "tokens_seen": 560332800 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004192878635907723, + "loss": 1.3979, + "theoretical_loss": 3.895996756952661, + "tokens_seen": 560398336 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041927783350050155, + "loss": 1.5611, + "theoretical_loss": 3.895948998064223, + "tokens_seen": 560463872 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041926780341023067, + "loss": 1.4983, + "theoretical_loss": 3.8959012463234473, + "tokens_seen": 560529408 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004192577733199599, + "loss": 1.4323, + "theoretical_loss": 3.8958535017284284, + "tokens_seen": 560594944 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041924774322968904, + "loss": 1.4833, + "theoretical_loss": 3.895805764277261, + "tokens_seen": 560660480 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041923771313941827, + "loss": 1.5569, + "theoretical_loss": 3.895758033968042, + "tokens_seen": 560726016 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041922768304914745, + "loss": 1.5602, + "theoretical_loss": 3.895710310798868, + "tokens_seen": 560791552 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041921765295887663, + "loss": 1.5358, + "theoretical_loss": 3.895662594767837, + "tokens_seen": 560857088 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004192076228686058, + "loss": 1.5032, + "theoretical_loss": 3.895614885873047, + "tokens_seen": 560922624 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041919759277833505, + "loss": 1.5411, + "theoretical_loss": 3.895567184112598, + "tokens_seen": 560988160 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004191875626880642, + "loss": 1.5944, + "theoretical_loss": 3.8955194894845886, + "tokens_seen": 561053696 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004191775325977934, + "loss": 1.4914, + "theoretical_loss": 3.89547180198712, + "tokens_seen": 561119232 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041916750250752254, + "loss": 1.4918, + "theoretical_loss": 3.8954241216182943, + "tokens_seen": 561184768 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004191574724172518, + "loss": 1.5863, + "theoretical_loss": 3.8953764483762123, + "tokens_seen": 561250304 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041914744232698095, + "loss": 1.4605, + "theoretical_loss": 3.8953287822589777, + "tokens_seen": 561315840 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041913741223671014, + "loss": 1.6026, + "theoretical_loss": 3.8952811232646933, + "tokens_seen": 561381376 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004191273821464393, + "loss": 1.4684, + "theoretical_loss": 3.895233471391464, + "tokens_seen": 561446912 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004191173520561685, + "loss": 1.5329, + "theoretical_loss": 3.895185826637394, + "tokens_seen": 561512448 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004191073219658977, + "loss": 1.6282, + "theoretical_loss": 3.89513818900059, + "tokens_seen": 561577984 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004190972918756269, + "loss": 1.5635, + "theoretical_loss": 3.8950905584791577, + "tokens_seen": 561643520 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041908726178535604, + "loss": 1.4635, + "theoretical_loss": 3.8950429350712037, + "tokens_seen": 561709056 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004190772316950853, + "loss": 1.5043, + "theoretical_loss": 3.8949953187748374, + "tokens_seen": 561774592 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004190672016048144, + "loss": 1.4603, + "theoretical_loss": 3.894947709588166, + "tokens_seen": 561840128 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041905717151454364, + "loss": 1.6065, + "theoretical_loss": 3.8949001075093, + "tokens_seen": 561905664 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004190471414242728, + "loss": 1.5808, + "theoretical_loss": 3.8948525125363487, + "tokens_seen": 561971200 + }, + { + "epoch": 0.17, + "learning_rate": 0.000419037111334002, + "loss": 1.5328, + "theoretical_loss": 3.8948049246674223, + "tokens_seen": 562036736 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004190270812437312, + "loss": 1.5137, + "theoretical_loss": 3.894757343900633, + "tokens_seen": 562102272 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004190170511534604, + "loss": 1.4664, + "theoretical_loss": 3.894709770234093, + "tokens_seen": 562167808 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041900702106318954, + "loss": 1.4793, + "theoretical_loss": 3.8946622036659155, + "tokens_seen": 562233344 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004189969909729188, + "loss": 1.497, + "theoretical_loss": 3.8946146441942138, + "tokens_seen": 562298880 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004189869608826479, + "loss": 1.5493, + "theoretical_loss": 3.894567091817102, + "tokens_seen": 562364416 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041897693079237714, + "loss": 1.4733, + "theoretical_loss": 3.8945195465326954, + "tokens_seen": 562429952 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004189669007021063, + "loss": 1.4817, + "theoretical_loss": 3.8944720083391093, + "tokens_seen": 562495488 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004189568706118355, + "loss": 1.4265, + "theoretical_loss": 3.8944244772344616, + "tokens_seen": 562561024 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004189468405215647, + "loss": 1.5693, + "theoretical_loss": 3.894376953216868, + "tokens_seen": 562626560 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041893681043129387, + "loss": 1.4413, + "theoretical_loss": 3.8943294362844476, + "tokens_seen": 562692096 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041892678034102305, + "loss": 1.4788, + "theoretical_loss": 3.8942819264353186, + "tokens_seen": 562757632 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004189167502507523, + "loss": 1.5145, + "theoretical_loss": 3.8942344236676005, + "tokens_seen": 562823168 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041890672016048146, + "loss": 1.5501, + "theoretical_loss": 3.894186927979413, + "tokens_seen": 562888704 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041889669007021065, + "loss": 1.6281, + "theoretical_loss": 3.8941394393688777, + "tokens_seen": 562954240 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004188866599799399, + "loss": 1.4883, + "theoretical_loss": 3.8940919578341155, + "tokens_seen": 563019776 + }, + { + "epoch": 0.17, + "learning_rate": 0.000418876629889669, + "loss": 1.5709, + "theoretical_loss": 3.894044483373249, + "tokens_seen": 563085312 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041886659979939824, + "loss": 1.5987, + "theoretical_loss": 3.893997015984401, + "tokens_seen": 563150848 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041885656970912737, + "loss": 1.4491, + "theoretical_loss": 3.893949555665696, + "tokens_seen": 563216384 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004188465396188566, + "loss": 1.553, + "theoretical_loss": 3.893902102415257, + "tokens_seen": 563281920 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004188365095285858, + "loss": 1.4305, + "theoretical_loss": 3.89385465623121, + "tokens_seen": 563347456 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041882647943831497, + "loss": 1.516, + "theoretical_loss": 3.8938072171116804, + "tokens_seen": 563412992 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041881644934804415, + "loss": 1.4014, + "theoretical_loss": 3.893759785054796, + "tokens_seen": 563478528 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041880641925777333, + "loss": 1.6814, + "theoretical_loss": 3.8937123600586823, + "tokens_seen": 563544064 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 214661, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3954126834869385, + "objective/train/theoretical_loss": 3.8936649421214686, + "objective/train/tokens_used": 563609600, + "theoretical_loss": 3.8936649421214686, + "tokens_seen": 563609600 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004187963891675025, + "loss": 1.468, + "theoretical_loss": 3.8936649421214686, + "tokens_seen": 563609600 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041878635907723175, + "loss": 1.453, + "theoretical_loss": 3.8936175312412833, + "tokens_seen": 563675136 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004187763289869609, + "loss": 1.5178, + "theoretical_loss": 3.8935701274162553, + "tokens_seen": 563740672 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004187662988966901, + "loss": 1.3852, + "theoretical_loss": 3.8935227306445155, + "tokens_seen": 563806208 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041875626880641924, + "loss": 1.556, + "theoretical_loss": 3.8934753409241942, + "tokens_seen": 563871744 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041874623871614847, + "loss": 1.5049, + "theoretical_loss": 3.8934279582534232, + "tokens_seen": 563937280 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041873620862587765, + "loss": 1.4927, + "theoretical_loss": 3.8933805826303347, + "tokens_seen": 564002816 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041872617853560683, + "loss": 1.4125, + "theoretical_loss": 3.893333214053062, + "tokens_seen": 564068352 + }, + { + "epoch": 0.17, + "learning_rate": 0.000418716148445336, + "loss": 1.5351, + "theoretical_loss": 3.8932858525197385, + "tokens_seen": 564133888 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041870611835506525, + "loss": 1.5417, + "theoretical_loss": 3.8932384980284986, + "tokens_seen": 564199424 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004186960882647944, + "loss": 1.559, + "theoretical_loss": 3.8931911505774766, + "tokens_seen": 564264960 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004186860581745236, + "loss": 1.4349, + "theoretical_loss": 3.89314381016481, + "tokens_seen": 564330496 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041867602808425274, + "loss": 1.5366, + "theoretical_loss": 3.8930964767886347, + "tokens_seen": 564396032 + }, + { + "epoch": 0.17, + "learning_rate": 0.000418665997993982, + "loss": 1.4721, + "theoretical_loss": 3.893049150447087, + "tokens_seen": 564461568 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041865596790371115, + "loss": 1.4791, + "theoretical_loss": 3.8930018311383066, + "tokens_seen": 564527104 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041864593781344034, + "loss": 1.5024, + "theoretical_loss": 3.8929545188604306, + "tokens_seen": 564592640 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004186359077231695, + "loss": 1.5426, + "theoretical_loss": 3.8929072136115987, + "tokens_seen": 564658176 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004186258776328987, + "loss": 1.5549, + "theoretical_loss": 3.8928599153899515, + "tokens_seen": 564723712 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004186158475426279, + "loss": 1.4654, + "theoretical_loss": 3.8928126241936294, + "tokens_seen": 564789248 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004186058174523571, + "loss": 1.486, + "theoretical_loss": 3.8927653400207745, + "tokens_seen": 564854784 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041859578736208624, + "loss": 1.6182, + "theoretical_loss": 3.892718062869528, + "tokens_seen": 564920320 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004185857572718155, + "loss": 1.4939, + "theoretical_loss": 3.8926707927380337, + "tokens_seen": 564985856 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004185757271815446, + "loss": 1.4621, + "theoretical_loss": 3.892623529624435, + "tokens_seen": 565051392 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041856569709127384, + "loss": 1.4667, + "theoretical_loss": 3.892576273526876, + "tokens_seen": 565116928 + }, + { + "epoch": 0.17, + "learning_rate": 0.000418555667001003, + "loss": 1.5293, + "theoretical_loss": 3.892529024443502, + "tokens_seen": 565182464 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004185456369107322, + "loss": 1.3692, + "theoretical_loss": 3.892481782372458, + "tokens_seen": 565248000 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004185356068204614, + "loss": 1.5302, + "theoretical_loss": 3.8924345473118915, + "tokens_seen": 565313536 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004185255767301906, + "loss": 1.4073, + "theoretical_loss": 3.892387319259949, + "tokens_seen": 565379072 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041851554663991974, + "loss": 1.5572, + "theoretical_loss": 3.8923400982147784, + "tokens_seen": 565444608 + }, + { + "epoch": 0.17, + "learning_rate": 0.000418505516549649, + "loss": 1.5163, + "theoretical_loss": 3.8922928841745286, + "tokens_seen": 565510144 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004184954864593781, + "loss": 1.495, + "theoretical_loss": 3.892245677137349, + "tokens_seen": 565575680 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041848545636910734, + "loss": 1.5016, + "theoretical_loss": 3.8921984771013887, + "tokens_seen": 565641216 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004184754262788365, + "loss": 1.3407, + "theoretical_loss": 3.8921512840647985, + "tokens_seen": 565706752 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004184653961885657, + "loss": 1.5518, + "theoretical_loss": 3.892104098025731, + "tokens_seen": 565772288 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004184553660982949, + "loss": 1.4848, + "theoretical_loss": 3.892056918982337, + "tokens_seen": 565837824 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041844533600802407, + "loss": 1.4779, + "theoretical_loss": 3.8920097469327697, + "tokens_seen": 565903360 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041843530591775325, + "loss": 1.5019, + "theoretical_loss": 3.8919625818751826, + "tokens_seen": 565968896 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004184252758274825, + "loss": 1.4977, + "theoretical_loss": 3.8919154238077294, + "tokens_seen": 566034432 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004184152457372116, + "loss": 1.4953, + "theoretical_loss": 3.8918682727285665, + "tokens_seen": 566099968 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041840521564694085, + "loss": 1.4517, + "theoretical_loss": 3.8918211286358475, + "tokens_seen": 566165504 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041839518555666997, + "loss": 1.4931, + "theoretical_loss": 3.89177399152773, + "tokens_seen": 566231040 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004183851554663992, + "loss": 1.3653, + "theoretical_loss": 3.89172686140237, + "tokens_seen": 566296576 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004183751253761284, + "loss": 1.5509, + "theoretical_loss": 3.8916797382579262, + "tokens_seen": 566362112 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041836509528585757, + "loss": 1.5462, + "theoretical_loss": 3.8916326220925566, + "tokens_seen": 566427648 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041835506519558675, + "loss": 1.5091, + "theoretical_loss": 3.8915855129044195, + "tokens_seen": 566493184 + }, + { + "epoch": 0.17, + "learning_rate": 0.000418345035105316, + "loss": 1.4681, + "theoretical_loss": 3.891538410691676, + "tokens_seen": 566558720 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004183350050150451, + "loss": 1.4474, + "theoretical_loss": 3.8914913154524857, + "tokens_seen": 566624256 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041832497492477435, + "loss": 1.4462, + "theoretical_loss": 3.89144422718501, + "tokens_seen": 566689792 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004183149448345035, + "loss": 1.4376, + "theoretical_loss": 3.891397145887411, + "tokens_seen": 566755328 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004183049147442327, + "loss": 1.4929, + "theoretical_loss": 3.8913500715578504, + "tokens_seen": 566820864 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 215957, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5541365146636963, + "objective/train/theoretical_loss": 3.8913030041944925, + "objective/train/tokens_used": 566886400, + "theoretical_loss": 3.8913030041944925, + "tokens_seen": 566886400 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004182948846539619, + "loss": 1.5459, + "theoretical_loss": 3.8913030041944925, + "tokens_seen": 566886400 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004182848545636911, + "loss": 1.507, + "theoretical_loss": 3.891255943795501, + "tokens_seen": 566951936 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041827482447342025, + "loss": 1.477, + "theoretical_loss": 3.8912088903590405, + "tokens_seen": 567017472 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041826479438314944, + "loss": 1.4886, + "theoretical_loss": 3.8911618438832756, + "tokens_seen": 567083008 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004182547642928786, + "loss": 1.4102, + "theoretical_loss": 3.8911148043663735, + "tokens_seen": 567148544 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041824473420260785, + "loss": 1.5189, + "theoretical_loss": 3.8910677718065, + "tokens_seen": 567214080 + }, + { + "epoch": 0.17, + "learning_rate": 0.000418234704112337, + "loss": 1.5171, + "theoretical_loss": 3.8910207462018236, + "tokens_seen": 567279616 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004182246740220662, + "loss": 1.4867, + "theoretical_loss": 3.890973727550511, + "tokens_seen": 567345152 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004182146439317954, + "loss": 1.5154, + "theoretical_loss": 3.890926715850733, + "tokens_seen": 567410688 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004182046138415246, + "loss": 1.4282, + "theoretical_loss": 3.890879711100657, + "tokens_seen": 567476224 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041819458375125376, + "loss": 1.5294, + "theoretical_loss": 3.8908327132984546, + "tokens_seen": 567541760 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041818455366098294, + "loss": 1.5247, + "theoretical_loss": 3.890785722442296, + "tokens_seen": 567607296 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004181745235707121, + "loss": 1.4591, + "theoretical_loss": 3.890738738530353, + "tokens_seen": 567672832 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041816449348044136, + "loss": 1.5239, + "theoretical_loss": 3.890691761560798, + "tokens_seen": 567738368 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041815446339017054, + "loss": 1.5275, + "theoretical_loss": 3.8906447915318036, + "tokens_seen": 567803904 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004181444332998997, + "loss": 1.4716, + "theoretical_loss": 3.8905978284415443, + "tokens_seen": 567869440 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004181344032096289, + "loss": 1.5225, + "theoretical_loss": 3.890550872288194, + "tokens_seen": 567934976 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004181243731193581, + "loss": 1.4243, + "theoretical_loss": 3.890503923069927, + "tokens_seen": 568000512 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004181143430290873, + "loss": 1.6317, + "theoretical_loss": 3.8904569807849203, + "tokens_seen": 568066048 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041810431293881644, + "loss": 1.4858, + "theoretical_loss": 3.8904100454313495, + "tokens_seen": 568131584 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004180942828485457, + "loss": 1.5065, + "theoretical_loss": 3.890363117007392, + "tokens_seen": 568197120 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004180842527582748, + "loss": 1.4958, + "theoretical_loss": 3.890316195511226, + "tokens_seen": 568262656 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041807422266800404, + "loss": 1.5383, + "theoretical_loss": 3.8902692809410295, + "tokens_seen": 568328192 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004180641925777332, + "loss": 1.484, + "theoretical_loss": 3.8902223732949817, + "tokens_seen": 568393728 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004180541624874624, + "loss": 1.6036, + "theoretical_loss": 3.8901754725712627, + "tokens_seen": 568459264 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004180441323971916, + "loss": 1.5281, + "theoretical_loss": 3.890128578768053, + "tokens_seen": 568524800 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004180341023069208, + "loss": 1.4585, + "theoretical_loss": 3.8900816918835335, + "tokens_seen": 568590336 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041802407221664994, + "loss": 1.509, + "theoretical_loss": 3.8900348119158865, + "tokens_seen": 568655872 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004180140421263792, + "loss": 1.5471, + "theoretical_loss": 3.8899879388632947, + "tokens_seen": 568721408 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004180040120361083, + "loss": 1.5409, + "theoretical_loss": 3.889941072723942, + "tokens_seen": 568786944 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041799398194583754, + "loss": 1.4951, + "theoretical_loss": 3.8898942134960106, + "tokens_seen": 568852480 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004179839518555667, + "loss": 1.4214, + "theoretical_loss": 3.8898473611776865, + "tokens_seen": 568918016 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004179739217652959, + "loss": 1.5771, + "theoretical_loss": 3.8898005157671554, + "tokens_seen": 568983552 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004179638916750251, + "loss": 1.4536, + "theoretical_loss": 3.889753677262603, + "tokens_seen": 569049088 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041795386158475427, + "loss": 1.5454, + "theoretical_loss": 3.8897068456622153, + "tokens_seen": 569114624 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041794383149448345, + "loss": 1.547, + "theoretical_loss": 3.8896600209641807, + "tokens_seen": 569180160 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004179338014042127, + "loss": 1.4778, + "theoretical_loss": 3.889613203166687, + "tokens_seen": 569245696 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004179237713139418, + "loss": 1.4943, + "theoretical_loss": 3.8895663922679224, + "tokens_seen": 569311232 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041791374122367105, + "loss": 1.5062, + "theoretical_loss": 3.8895195882660776, + "tokens_seen": 569376768 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041790371113340017, + "loss": 1.467, + "theoretical_loss": 3.8894727911593416, + "tokens_seen": 569442304 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004178936810431294, + "loss": 1.3901, + "theoretical_loss": 3.889426000945906, + "tokens_seen": 569507840 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004178836509528586, + "loss": 1.486, + "theoretical_loss": 3.889379217623962, + "tokens_seen": 569573376 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041787362086258777, + "loss": 1.5832, + "theoretical_loss": 3.8893324411917014, + "tokens_seen": 569638912 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041786359077231695, + "loss": 1.427, + "theoretical_loss": 3.8892856716473183, + "tokens_seen": 569704448 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004178535606820462, + "loss": 1.4611, + "theoretical_loss": 3.8892389089890056, + "tokens_seen": 569769984 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004178435305917753, + "loss": 1.4643, + "theoretical_loss": 3.8891921532149576, + "tokens_seen": 569835520 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041783350050150455, + "loss": 1.4647, + "theoretical_loss": 3.8891454043233686, + "tokens_seen": 569901056 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004178234704112337, + "loss": 1.4983, + "theoretical_loss": 3.8890986623124353, + "tokens_seen": 569966592 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004178134403209629, + "loss": 1.4725, + "theoretical_loss": 3.889051927180353, + "tokens_seen": 570032128 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004178034102306921, + "loss": 1.5551, + "theoretical_loss": 3.88900519892532, + "tokens_seen": 570097664 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 217201, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4359532594680786, + "objective/train/theoretical_loss": 3.8889584775455326, + "objective/train/tokens_used": 570163200, + "theoretical_loss": 3.8889584775455326, + "tokens_seen": 570163200 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004177933801404213, + "loss": 1.437, + "theoretical_loss": 3.8889584775455326, + "tokens_seen": 570163200 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041778335005015045, + "loss": 1.5197, + "theoretical_loss": 3.8889117630391894, + "tokens_seen": 570228736 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041777331995987964, + "loss": 1.3705, + "theoretical_loss": 3.888865055404491, + "tokens_seen": 570294272 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004177632898696088, + "loss": 1.488, + "theoretical_loss": 3.8888183546396347, + "tokens_seen": 570359808 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041775325977933805, + "loss": 1.508, + "theoretical_loss": 3.8887716607428224, + "tokens_seen": 570425344 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004177432296890672, + "loss": 1.4542, + "theoretical_loss": 3.8887249737122547, + "tokens_seen": 570490880 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004177331995987964, + "loss": 1.4601, + "theoretical_loss": 3.8886782935461333, + "tokens_seen": 570556416 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004177231695085256, + "loss": 1.5246, + "theoretical_loss": 3.888631620242661, + "tokens_seen": 570621952 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004177131394182548, + "loss": 1.4964, + "theoretical_loss": 3.8885849538000397, + "tokens_seen": 570687488 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041770310932798396, + "loss": 1.4411, + "theoretical_loss": 3.8885382942164752, + "tokens_seen": 570753024 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041769307923771314, + "loss": 1.4917, + "theoretical_loss": 3.8884916414901705, + "tokens_seen": 570818560 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004176830491474423, + "loss": 1.4606, + "theoretical_loss": 3.8884449956193308, + "tokens_seen": 570884096 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041767301905717156, + "loss": 1.4858, + "theoretical_loss": 3.8883983566021625, + "tokens_seen": 570949632 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004176629889669007, + "loss": 1.4849, + "theoretical_loss": 3.8883517244368715, + "tokens_seen": 571015168 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004176529588766299, + "loss": 1.5329, + "theoretical_loss": 3.8883050991216654, + "tokens_seen": 571080704 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041764292878635904, + "loss": 1.441, + "theoretical_loss": 3.888258480654751, + "tokens_seen": 571146240 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004176328986960883, + "loss": 1.4612, + "theoretical_loss": 3.888211869034339, + "tokens_seen": 571211776 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041762286860581746, + "loss": 1.4745, + "theoretical_loss": 3.888165264258636, + "tokens_seen": 571277312 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041761283851554664, + "loss": 1.4425, + "theoretical_loss": 3.888118666325853, + "tokens_seen": 571342848 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004176028084252758, + "loss": 1.4912, + "theoretical_loss": 3.888072075234201, + "tokens_seen": 571408384 + }, + { + "epoch": 0.17, + "learning_rate": 0.000417592778335005, + "loss": 1.4067, + "theoretical_loss": 3.8880254909818905, + "tokens_seen": 571473920 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004175827482447342, + "loss": 1.5797, + "theoretical_loss": 3.8879789135671334, + "tokens_seen": 571539456 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004175727181544634, + "loss": 1.5444, + "theoretical_loss": 3.8879323429881425, + "tokens_seen": 571604992 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041756268806419255, + "loss": 1.5755, + "theoretical_loss": 3.887885779243131, + "tokens_seen": 571670528 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004175526579739218, + "loss": 1.3898, + "theoretical_loss": 3.887839222330313, + "tokens_seen": 571736064 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041754262788365096, + "loss": 1.3589, + "theoretical_loss": 3.887792672247902, + "tokens_seen": 571801600 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041753259779338015, + "loss": 1.4151, + "theoretical_loss": 3.887746128994114, + "tokens_seen": 571867136 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004175225677031093, + "loss": 1.4723, + "theoretical_loss": 3.8876995925671656, + "tokens_seen": 571932672 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004175125376128385, + "loss": 1.523, + "theoretical_loss": 3.8876530629652724, + "tokens_seen": 571998208 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004175025075225677, + "loss": 1.5791, + "theoretical_loss": 3.8876065401866513, + "tokens_seen": 572063744 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004174924774322969, + "loss": 1.493, + "theoretical_loss": 3.887560024229521, + "tokens_seen": 572129280 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041748244734202605, + "loss": 1.5018, + "theoretical_loss": 3.8875135150921, + "tokens_seen": 572194816 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004174724172517553, + "loss": 1.4614, + "theoretical_loss": 3.887467012772607, + "tokens_seen": 572260352 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004174623871614844, + "loss": 1.6011, + "theoretical_loss": 3.8874205172692626, + "tokens_seen": 572325888 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041745235707121365, + "loss": 1.533, + "theoretical_loss": 3.887374028580287, + "tokens_seen": 572391424 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041744232698094283, + "loss": 1.4781, + "theoretical_loss": 3.8873275467039017, + "tokens_seen": 572456960 + }, + { + "epoch": 0.17, + "learning_rate": 0.000417432296890672, + "loss": 1.5729, + "theoretical_loss": 3.887281071638328, + "tokens_seen": 572522496 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004174222668004012, + "loss": 1.6581, + "theoretical_loss": 3.8872346033817893, + "tokens_seen": 572588032 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041741223671013037, + "loss": 1.5188, + "theoretical_loss": 3.887188141932508, + "tokens_seen": 572653568 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004174022066198596, + "loss": 1.5031, + "theoretical_loss": 3.8871416872887083, + "tokens_seen": 572719104 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004173921765295888, + "loss": 1.5479, + "theoretical_loss": 3.8870952394486156, + "tokens_seen": 572784640 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041738214643931797, + "loss": 1.5067, + "theoretical_loss": 3.8870487984104543, + "tokens_seen": 572850176 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041737211634904715, + "loss": 1.5309, + "theoretical_loss": 3.8870023641724503, + "tokens_seen": 572915712 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004173620862587764, + "loss": 1.5956, + "theoretical_loss": 3.886955936732831, + "tokens_seen": 572981248 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004173520561685055, + "loss": 1.5342, + "theoretical_loss": 3.886909516089822, + "tokens_seen": 573046784 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041734202607823475, + "loss": 1.5782, + "theoretical_loss": 3.8868631022416533, + "tokens_seen": 573112320 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004173319959879639, + "loss": 1.6705, + "theoretical_loss": 3.886816695186552, + "tokens_seen": 573177856 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004173219658976931, + "loss": 1.4683, + "theoretical_loss": 3.886770294922748, + "tokens_seen": 573243392 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004173119358074223, + "loss": 1.4606, + "theoretical_loss": 3.8867239014484705, + "tokens_seen": 573308928 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004173019057171515, + "loss": 1.5772, + "theoretical_loss": 3.886677514761951, + "tokens_seen": 573374464 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 218333, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5484840869903564, + "objective/train/theoretical_loss": 3.8866311348614198, + "objective/train/tokens_used": 573440000, + "theoretical_loss": 3.8866311348614198, + "tokens_seen": 573440000 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041729187562688065, + "loss": 1.5872, + "theoretical_loss": 3.8866311348614198, + "tokens_seen": 573440000 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041728184553660984, + "loss": 1.5052, + "theoretical_loss": 3.8865847617451097, + "tokens_seen": 573505536 + }, + { + "epoch": 0.17, + "learning_rate": 0.000417271815446339, + "loss": 1.5483, + "theoretical_loss": 3.8865383954112525, + "tokens_seen": 573571072 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041726178535606825, + "loss": 1.4474, + "theoretical_loss": 3.8864920358580823, + "tokens_seen": 573636608 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004172517552657974, + "loss": 1.5433, + "theoretical_loss": 3.8864456830838314, + "tokens_seen": 573702144 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004172417251755266, + "loss": 1.5146, + "theoretical_loss": 3.8863993370867362, + "tokens_seen": 573767680 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004172316950852558, + "loss": 1.4627, + "theoretical_loss": 3.886352997865031, + "tokens_seen": 573833216 + }, + { + "epoch": 0.17, + "learning_rate": 0.000417221664994985, + "loss": 1.5612, + "theoretical_loss": 3.8863066654169516, + "tokens_seen": 573898752 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041721163490471416, + "loss": 1.6233, + "theoretical_loss": 3.886260339740735, + "tokens_seen": 573964288 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041720160481444334, + "loss": 1.4166, + "theoretical_loss": 3.8862140208346174, + "tokens_seen": 574029824 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004171915747241725, + "loss": 1.5029, + "theoretical_loss": 3.8861677086968376, + "tokens_seen": 574095360 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041718154463390176, + "loss": 1.6154, + "theoretical_loss": 3.8861214033256344, + "tokens_seen": 574160896 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004171715145436309, + "loss": 1.4978, + "theoretical_loss": 3.8860751047192457, + "tokens_seen": 574226432 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004171614844533601, + "loss": 1.4522, + "theoretical_loss": 3.886028812875912, + "tokens_seen": 574291968 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041715145436308924, + "loss": 1.4943, + "theoretical_loss": 3.885982527793874, + "tokens_seen": 574357504 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004171414242728185, + "loss": 1.4159, + "theoretical_loss": 3.8859362494713725, + "tokens_seen": 574423040 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041713139418254766, + "loss": 1.5814, + "theoretical_loss": 3.8858899779066496, + "tokens_seen": 574488576 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041712136409227684, + "loss": 1.5173, + "theoretical_loss": 3.8858437130979473, + "tokens_seen": 574554112 + }, + { + "epoch": 0.17, + "learning_rate": 0.000417111334002006, + "loss": 1.5419, + "theoretical_loss": 3.8857974550435097, + "tokens_seen": 574619648 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004171013039117352, + "loss": 1.4787, + "theoretical_loss": 3.885751203741579, + "tokens_seen": 574685184 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004170912738214644, + "loss": 1.4738, + "theoretical_loss": 3.8857049591904014, + "tokens_seen": 574750720 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004170812437311936, + "loss": 1.4605, + "theoretical_loss": 3.8856587213882205, + "tokens_seen": 574816256 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041707121364092275, + "loss": 1.4323, + "theoretical_loss": 3.8856124903332825, + "tokens_seen": 574881792 + }, + { + "epoch": 0.17, + "learning_rate": 0.000417061183550652, + "loss": 1.3669, + "theoretical_loss": 3.885566266023834, + "tokens_seen": 574947328 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041705115346038116, + "loss": 1.5263, + "theoretical_loss": 3.8855200484581225, + "tokens_seen": 575012864 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041704112337011035, + "loss": 1.4771, + "theoretical_loss": 3.885473837634395, + "tokens_seen": 575078400 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004170310932798395, + "loss": 1.4881, + "theoretical_loss": 3.8854276335508997, + "tokens_seen": 575143936 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004170210631895687, + "loss": 1.6159, + "theoretical_loss": 3.885381436205886, + "tokens_seen": 575209472 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004170110330992979, + "loss": 1.4656, + "theoretical_loss": 3.885335245597604, + "tokens_seen": 575275008 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004170010030090271, + "loss": 1.4854, + "theoretical_loss": 3.885289061724303, + "tokens_seen": 575340544 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041699097291875625, + "loss": 1.4582, + "theoretical_loss": 3.885242884584235, + "tokens_seen": 575406080 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004169809428284855, + "loss": 1.4518, + "theoretical_loss": 3.8851967141756507, + "tokens_seen": 575471616 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004169709127382146, + "loss": 1.5543, + "theoretical_loss": 3.885150550496803, + "tokens_seen": 575537152 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041696088264794385, + "loss": 1.509, + "theoretical_loss": 3.8851043935459453, + "tokens_seen": 575602688 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041695085255767303, + "loss": 1.4775, + "theoretical_loss": 3.88505824332133, + "tokens_seen": 575668224 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004169408224674022, + "loss": 1.5493, + "theoretical_loss": 3.8850120998212123, + "tokens_seen": 575733760 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004169307923771314, + "loss": 1.5621, + "theoretical_loss": 3.884965963043847, + "tokens_seen": 575799296 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004169207622868606, + "loss": 1.5847, + "theoretical_loss": 3.884919832987489, + "tokens_seen": 575864832 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041691073219658975, + "loss": 1.3966, + "theoretical_loss": 3.884873709650395, + "tokens_seen": 575930368 + }, + { + "epoch": 0.17, + "learning_rate": 0.000416900702106319, + "loss": 1.4946, + "theoretical_loss": 3.8848275930308223, + "tokens_seen": 575995904 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004168906720160481, + "loss": 1.5262, + "theoretical_loss": 3.884781483127027, + "tokens_seen": 576061440 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041688064192577735, + "loss": 1.5753, + "theoretical_loss": 3.8847353799372684, + "tokens_seen": 576126976 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041687061183550653, + "loss": 1.4722, + "theoretical_loss": 3.8846892834598052, + "tokens_seen": 576192512 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004168605817452357, + "loss": 1.5674, + "theoretical_loss": 3.884643193692896, + "tokens_seen": 576258048 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004168505516549649, + "loss": 1.4964, + "theoretical_loss": 3.884597110634802, + "tokens_seen": 576323584 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004168405215646941, + "loss": 1.4192, + "theoretical_loss": 3.884551034283784, + "tokens_seen": 576389120 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041683049147442326, + "loss": 1.4457, + "theoretical_loss": 3.8845049646381025, + "tokens_seen": 576454656 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004168204613841525, + "loss": 1.442, + "theoretical_loss": 3.8844589016960196, + "tokens_seen": 576520192 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004168104312938816, + "loss": 1.4439, + "theoretical_loss": 3.884412845455799, + "tokens_seen": 576585728 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041680040120361085, + "loss": 1.4807, + "theoretical_loss": 3.8843667959157027, + "tokens_seen": 576651264 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 219460, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4630844593048096, + "objective/train/theoretical_loss": 3.884320753073996, + "objective/train/tokens_used": 576716800, + "theoretical_loss": 3.884320753073996, + "tokens_seen": 576716800 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041679037111334, + "loss": 1.5228, + "theoretical_loss": 3.884320753073996, + "tokens_seen": 576716800 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004167803410230692, + "loss": 1.4981, + "theoretical_loss": 3.884274716928942, + "tokens_seen": 576782336 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004167703109327984, + "loss": 1.4326, + "theoretical_loss": 3.8842286874788075, + "tokens_seen": 576847872 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004167602808425276, + "loss": 1.5159, + "theoretical_loss": 3.8841826647218576, + "tokens_seen": 576913408 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041675025075225676, + "loss": 1.3919, + "theoretical_loss": 3.8841366486563595, + "tokens_seen": 576978944 + }, + { + "epoch": 0.17, + "learning_rate": 0.000416740220661986, + "loss": 1.5088, + "theoretical_loss": 3.8840906392805796, + "tokens_seen": 577044480 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004167301905717151, + "loss": 1.4123, + "theoretical_loss": 3.8840446365927863, + "tokens_seen": 577110016 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041672016048144436, + "loss": 1.5019, + "theoretical_loss": 3.8839986405912477, + "tokens_seen": 577175552 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004167101303911735, + "loss": 1.48, + "theoretical_loss": 3.8839526512742335, + "tokens_seen": 577241088 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004167001003009027, + "loss": 1.4297, + "theoretical_loss": 3.883906668640013, + "tokens_seen": 577306624 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004166900702106319, + "loss": 1.6309, + "theoretical_loss": 3.8838606926868575, + "tokens_seen": 577372160 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004166800401203611, + "loss": 1.4186, + "theoretical_loss": 3.8838147234130367, + "tokens_seen": 577437696 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041667001003009026, + "loss": 1.5054, + "theoretical_loss": 3.8837687608168237, + "tokens_seen": 577503232 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041665997993981944, + "loss": 1.5964, + "theoretical_loss": 3.8837228048964896, + "tokens_seen": 577568768 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004166499498495487, + "loss": 1.3754, + "theoretical_loss": 3.8836768556503083, + "tokens_seen": 577634304 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041663991975927786, + "loss": 1.5428, + "theoretical_loss": 3.883630913076553, + "tokens_seen": 577699840 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041662988966900704, + "loss": 1.5068, + "theoretical_loss": 3.883584977173499, + "tokens_seen": 577765376 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004166198595787362, + "loss": 1.5781, + "theoretical_loss": 3.8835390479394203, + "tokens_seen": 577830912 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004166098294884654, + "loss": 1.469, + "theoretical_loss": 3.883493125372592, + "tokens_seen": 577896448 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004165997993981946, + "loss": 1.5036, + "theoretical_loss": 3.8834472094712913, + "tokens_seen": 577961984 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004165897693079238, + "loss": 1.518, + "theoretical_loss": 3.8834013002337953, + "tokens_seen": 578027520 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041657973921765295, + "loss": 1.5288, + "theoretical_loss": 3.883355397658381, + "tokens_seen": 578093056 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004165697091273822, + "loss": 1.5531, + "theoretical_loss": 3.8833095017433257, + "tokens_seen": 578158592 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041655967903711136, + "loss": 1.5392, + "theoretical_loss": 3.8832636124869095, + "tokens_seen": 578224128 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041654964894684055, + "loss": 1.5489, + "theoretical_loss": 3.883217729887411, + "tokens_seen": 578289664 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004165396188565697, + "loss": 1.5104, + "theoretical_loss": 3.883171853943111, + "tokens_seen": 578355200 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004165295887662989, + "loss": 1.4584, + "theoretical_loss": 3.8831259846522896, + "tokens_seen": 578420736 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004165195586760281, + "loss": 1.5219, + "theoretical_loss": 3.8830801220132285, + "tokens_seen": 578486272 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004165095285857573, + "loss": 1.4799, + "theoretical_loss": 3.883034266024209, + "tokens_seen": 578551808 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041649949849548645, + "loss": 1.419, + "theoretical_loss": 3.8829884166835154, + "tokens_seen": 578617344 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004164894684052157, + "loss": 1.4606, + "theoretical_loss": 3.882942573989429, + "tokens_seen": 578682880 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004164794383149448, + "loss": 1.4959, + "theoretical_loss": 3.882896737940235, + "tokens_seen": 578748416 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041646940822467405, + "loss": 1.5075, + "theoretical_loss": 3.882850908534217, + "tokens_seen": 578813952 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041645937813440323, + "loss": 1.5042, + "theoretical_loss": 3.882805085769661, + "tokens_seen": 578879488 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004164493480441324, + "loss": 1.5032, + "theoretical_loss": 3.8827592696448523, + "tokens_seen": 578945024 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004164393179538616, + "loss": 1.534, + "theoretical_loss": 3.8827134601580777, + "tokens_seen": 579010560 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004164292878635908, + "loss": 1.3927, + "theoretical_loss": 3.8826676573076235, + "tokens_seen": 579076096 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041641925777331995, + "loss": 1.6232, + "theoretical_loss": 3.8826218610917786, + "tokens_seen": 579141632 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004164092276830492, + "loss": 1.5238, + "theoretical_loss": 3.8825760715088307, + "tokens_seen": 579207168 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004163991975927783, + "loss": 1.5267, + "theoretical_loss": 3.882530288557069, + "tokens_seen": 579272704 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041638916750250755, + "loss": 1.4945, + "theoretical_loss": 3.8824845122347824, + "tokens_seen": 579338240 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041637913741223673, + "loss": 1.4877, + "theoretical_loss": 3.8824387425402618, + "tokens_seen": 579403776 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004163691073219659, + "loss": 1.5264, + "theoretical_loss": 3.8823929794717986, + "tokens_seen": 579469312 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004163590772316951, + "loss": 1.5837, + "theoretical_loss": 3.882347223027683, + "tokens_seen": 579534848 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004163490471414243, + "loss": 1.4229, + "theoretical_loss": 3.882301473206209, + "tokens_seen": 579600384 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041633901705115346, + "loss": 1.5884, + "theoretical_loss": 3.8822557300056673, + "tokens_seen": 579665920 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004163289869608827, + "loss": 1.453, + "theoretical_loss": 3.882209993424353, + "tokens_seen": 579731456 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004163189568706118, + "loss": 1.5291, + "theoretical_loss": 3.8821642634605587, + "tokens_seen": 579796992 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041630892678034105, + "loss": 1.4391, + "theoretical_loss": 3.8821185401125806, + "tokens_seen": 579862528 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004162988966900702, + "loss": 1.4584, + "theoretical_loss": 3.8820728233787136, + "tokens_seen": 579928064 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 220781, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.511155605316162, + "objective/train/theoretical_loss": 3.8820271132572524, + "objective/train/tokens_used": 579993600, + "theoretical_loss": 3.8820271132572524, + "tokens_seen": 579993600 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004162888665997994, + "loss": 1.5095, + "theoretical_loss": 3.8820271132572524, + "tokens_seen": 579993600 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004162788365095286, + "loss": 1.4841, + "theoretical_loss": 3.881981409746495, + "tokens_seen": 580059136 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004162688064192578, + "loss": 1.4085, + "theoretical_loss": 3.8819357128447383, + "tokens_seen": 580124672 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041625877632898696, + "loss": 1.4248, + "theoretical_loss": 3.88189002255028, + "tokens_seen": 580190208 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004162487462387162, + "loss": 1.5334, + "theoretical_loss": 3.8818443388614186, + "tokens_seen": 580255744 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004162387161484453, + "loss": 1.3916, + "theoretical_loss": 3.881798661776453, + "tokens_seen": 580321280 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041622868605817456, + "loss": 1.4411, + "theoretical_loss": 3.8817529912936832, + "tokens_seen": 580386816 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004162186559679037, + "loss": 1.4772, + "theoretical_loss": 3.88170732741141, + "tokens_seen": 580452352 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004162086258776329, + "loss": 1.5859, + "theoretical_loss": 3.8816616701279334, + "tokens_seen": 580517888 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004161985957873621, + "loss": 1.5349, + "theoretical_loss": 3.881616019441556, + "tokens_seen": 580583424 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004161885656970913, + "loss": 1.5494, + "theoretical_loss": 3.881570375350579, + "tokens_seen": 580648960 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041617853560682046, + "loss": 1.5674, + "theoretical_loss": 3.881524737853306, + "tokens_seen": 580714496 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041616850551654964, + "loss": 1.6244, + "theoretical_loss": 3.881479106948041, + "tokens_seen": 580780032 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004161584754262788, + "loss": 1.5609, + "theoretical_loss": 3.8814334826330876, + "tokens_seen": 580845568 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041614844533600806, + "loss": 1.4415, + "theoretical_loss": 3.88138786490675, + "tokens_seen": 580911104 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004161384152457372, + "loss": 1.4113, + "theoretical_loss": 3.8813422537673343, + "tokens_seen": 580976640 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004161283851554664, + "loss": 1.5656, + "theoretical_loss": 3.881296649213146, + "tokens_seen": 581042176 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041611835506519555, + "loss": 1.5739, + "theoretical_loss": 3.8812510512424927, + "tokens_seen": 581107712 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004161083249749248, + "loss": 1.4328, + "theoretical_loss": 3.881205459853681, + "tokens_seen": 581173248 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041609829488465397, + "loss": 1.4307, + "theoretical_loss": 3.8811598750450185, + "tokens_seen": 581238784 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041608826479438315, + "loss": 1.5257, + "theoretical_loss": 3.8811142968148142, + "tokens_seen": 581304320 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041607823470411233, + "loss": 1.5629, + "theoretical_loss": 3.881068725161377, + "tokens_seen": 581369856 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041606820461384156, + "loss": 1.5102, + "theoretical_loss": 3.881023160083018, + "tokens_seen": 581435392 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004160581745235707, + "loss": 1.5472, + "theoretical_loss": 3.8809776015780453, + "tokens_seen": 581500928 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004160481444332999, + "loss": 1.4645, + "theoretical_loss": 3.880932049644771, + "tokens_seen": 581566464 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041603811434302905, + "loss": 1.4478, + "theoretical_loss": 3.8808865042815075, + "tokens_seen": 581632000 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004160280842527583, + "loss": 1.4392, + "theoretical_loss": 3.880840965486566, + "tokens_seen": 581697536 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041601805416248747, + "loss": 1.5968, + "theoretical_loss": 3.8807954332582595, + "tokens_seen": 581763072 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041600802407221665, + "loss": 1.4625, + "theoretical_loss": 3.880749907594902, + "tokens_seen": 581828608 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041599799398194583, + "loss": 1.4454, + "theoretical_loss": 3.8807043884948085, + "tokens_seen": 581894144 + }, + { + "epoch": 0.18, + "learning_rate": 0.000415987963891675, + "loss": 1.5345, + "theoretical_loss": 3.880658875956292, + "tokens_seen": 581959680 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004159779338014042, + "loss": 1.5115, + "theoretical_loss": 3.8806133699776684, + "tokens_seen": 582025216 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041596790371113343, + "loss": 1.465, + "theoretical_loss": 3.880567870557254, + "tokens_seen": 582090752 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041595787362086256, + "loss": 1.5194, + "theoretical_loss": 3.8805223776933655, + "tokens_seen": 582156288 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004159478435305918, + "loss": 1.5244, + "theoretical_loss": 3.88047689138432, + "tokens_seen": 582221824 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004159378134403209, + "loss": 1.4592, + "theoretical_loss": 3.880431411628436, + "tokens_seen": 582287360 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041592778335005015, + "loss": 1.5827, + "theoretical_loss": 3.8803859384240305, + "tokens_seen": 582352896 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041591775325977934, + "loss": 1.539, + "theoretical_loss": 3.880340471769424, + "tokens_seen": 582418432 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004159077231695085, + "loss": 1.4045, + "theoretical_loss": 3.880295011662936, + "tokens_seen": 582483968 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041589769307923775, + "loss": 1.4968, + "theoretical_loss": 3.8802495581028857, + "tokens_seen": 582549504 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041588766298896693, + "loss": 1.4536, + "theoretical_loss": 3.880204111087596, + "tokens_seen": 582615040 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004158776328986961, + "loss": 1.5017, + "theoretical_loss": 3.880158670615387, + "tokens_seen": 582680576 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004158676028084253, + "loss": 1.571, + "theoretical_loss": 3.880113236684582, + "tokens_seen": 582746112 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004158575727181545, + "loss": 1.4222, + "theoretical_loss": 3.880067809293503, + "tokens_seen": 582811648 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041584754262788366, + "loss": 1.4039, + "theoretical_loss": 3.880022388440474, + "tokens_seen": 582877184 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004158375125376129, + "loss": 1.5708, + "theoretical_loss": 3.879976974123818, + "tokens_seen": 582942720 + }, + { + "epoch": 0.18, + "learning_rate": 0.000415827482447342, + "loss": 1.4566, + "theoretical_loss": 3.8799315663418614, + "tokens_seen": 583008256 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041581745235707126, + "loss": 1.4948, + "theoretical_loss": 3.8798861650929277, + "tokens_seen": 583073792 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004158074222668004, + "loss": 1.5536, + "theoretical_loss": 3.8798407703753446, + "tokens_seen": 583139328 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004157973921765296, + "loss": 1.4779, + "theoretical_loss": 3.879795382187438, + "tokens_seen": 583204864 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 221910, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4498310089111328, + "objective/train/theoretical_loss": 3.879750000527534, + "objective/train/tokens_used": 583270400, + "theoretical_loss": 3.879750000527534, + "tokens_seen": 583270400 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004157873620862588, + "loss": 1.4094, + "theoretical_loss": 3.879750000527534, + "tokens_seen": 583270400 + }, + { + "epoch": 0.18, + "learning_rate": 0.000415777331995988, + "loss": 1.4799, + "theoretical_loss": 3.8797046253939618, + "tokens_seen": 583335936 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041576730190571716, + "loss": 1.5837, + "theoretical_loss": 3.8796592567850485, + "tokens_seen": 583401472 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004157572718154464, + "loss": 1.4669, + "theoretical_loss": 3.8796138946991245, + "tokens_seen": 583467008 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004157472417251755, + "loss": 1.4236, + "theoretical_loss": 3.8795685391345183, + "tokens_seen": 583532544 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041573721163490476, + "loss": 1.4968, + "theoretical_loss": 3.879523190089561, + "tokens_seen": 583598080 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004157271815446339, + "loss": 1.5271, + "theoretical_loss": 3.8794778475625833, + "tokens_seen": 583663616 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004157171514543631, + "loss": 1.4803, + "theoretical_loss": 3.879432511551916, + "tokens_seen": 583729152 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004157071213640923, + "loss": 1.5194, + "theoretical_loss": 3.8793871820558916, + "tokens_seen": 583794688 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004156970912738215, + "loss": 1.5101, + "theoretical_loss": 3.8793418590728423, + "tokens_seen": 583860224 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041568706118355066, + "loss": 1.5805, + "theoretical_loss": 3.879296542601103, + "tokens_seen": 583925760 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041567703109327984, + "loss": 1.4707, + "theoretical_loss": 3.879251232639006, + "tokens_seen": 583991296 + }, + { + "epoch": 0.18, + "learning_rate": 0.000415667001003009, + "loss": 1.4035, + "theoretical_loss": 3.879205929184886, + "tokens_seen": 584056832 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041565697091273826, + "loss": 1.5497, + "theoretical_loss": 3.8791606322370793, + "tokens_seen": 584122368 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004156469408224674, + "loss": 1.3884, + "theoretical_loss": 3.8791153417939204, + "tokens_seen": 584187904 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004156369107321966, + "loss": 1.3936, + "theoretical_loss": 3.8790700578537463, + "tokens_seen": 584253440 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041562688064192575, + "loss": 1.5798, + "theoretical_loss": 3.879024780414894, + "tokens_seen": 584318976 + }, + { + "epoch": 0.18, + "learning_rate": 0.000415616850551655, + "loss": 1.4799, + "theoretical_loss": 3.878979509475701, + "tokens_seen": 584384512 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041560682046138417, + "loss": 1.509, + "theoretical_loss": 3.8789342450345052, + "tokens_seen": 584450048 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041559679037111335, + "loss": 1.3779, + "theoretical_loss": 3.8788889870896455, + "tokens_seen": 584515584 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041558676028084253, + "loss": 1.5239, + "theoretical_loss": 3.878843735639462, + "tokens_seen": 584581120 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041557673019057176, + "loss": 1.4999, + "theoretical_loss": 3.8787984906822945, + "tokens_seen": 584646656 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004155667001003009, + "loss": 1.5552, + "theoretical_loss": 3.878753252216483, + "tokens_seen": 584712192 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004155566700100301, + "loss": 1.4315, + "theoretical_loss": 3.8787080202403694, + "tokens_seen": 584777728 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041554663991975925, + "loss": 1.6537, + "theoretical_loss": 3.8786627947522954, + "tokens_seen": 584843264 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004155366098294885, + "loss": 1.4182, + "theoretical_loss": 3.878617575750603, + "tokens_seen": 584908800 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041552657973921767, + "loss": 1.456, + "theoretical_loss": 3.8785723632336366, + "tokens_seen": 584974336 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041551654964894685, + "loss": 1.4683, + "theoretical_loss": 3.8785271571997386, + "tokens_seen": 585039872 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041550651955867603, + "loss": 1.4842, + "theoretical_loss": 3.878481957647254, + "tokens_seen": 585105408 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004154964894684052, + "loss": 1.4383, + "theoretical_loss": 3.878436764574527, + "tokens_seen": 585170944 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004154864593781344, + "loss": 1.5491, + "theoretical_loss": 3.8783915779799036, + "tokens_seen": 585236480 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041547642928786363, + "loss": 1.5273, + "theoretical_loss": 3.8783463978617303, + "tokens_seen": 585302016 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041546639919759276, + "loss": 1.5236, + "theoretical_loss": 3.878301224218353, + "tokens_seen": 585367552 + }, + { + "epoch": 0.18, + "learning_rate": 0.000415456369107322, + "loss": 1.517, + "theoretical_loss": 3.8782560570481204, + "tokens_seen": 585433088 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004154463390170511, + "loss": 1.5933, + "theoretical_loss": 3.8782108963493793, + "tokens_seen": 585498624 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041543630892678035, + "loss": 1.4788, + "theoretical_loss": 3.878165742120478, + "tokens_seen": 585564160 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041542627883650954, + "loss": 1.5547, + "theoretical_loss": 3.8781205943597667, + "tokens_seen": 585629696 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004154162487462387, + "loss": 1.4798, + "theoretical_loss": 3.8780754530655948, + "tokens_seen": 585695232 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004154062186559679, + "loss": 1.4997, + "theoretical_loss": 3.8780303182363123, + "tokens_seen": 585760768 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041539618856569713, + "loss": 1.4199, + "theoretical_loss": 3.877985189870271, + "tokens_seen": 585826304 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041538615847542626, + "loss": 1.4929, + "theoretical_loss": 3.877940067965821, + "tokens_seen": 585891840 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004153761283851555, + "loss": 1.457, + "theoretical_loss": 3.8778949525213164, + "tokens_seen": 585957376 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004153660982948846, + "loss": 1.3519, + "theoretical_loss": 3.8778498435351088, + "tokens_seen": 586022912 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041535606820461386, + "loss": 1.45, + "theoretical_loss": 3.8778047410055514, + "tokens_seen": 586088448 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041534603811434304, + "loss": 1.5672, + "theoretical_loss": 3.8777596449309994, + "tokens_seen": 586153984 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004153360080240722, + "loss": 1.5455, + "theoretical_loss": 3.8777145553098062, + "tokens_seen": 586219520 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004153259779338014, + "loss": 1.4223, + "theoretical_loss": 3.8776694721403278, + "tokens_seen": 586285056 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004153159478435306, + "loss": 1.5397, + "theoretical_loss": 3.877624395420919, + "tokens_seen": 586350592 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041530591775325976, + "loss": 1.4323, + "theoretical_loss": 3.8775793251499375, + "tokens_seen": 586416128 + }, + { + "epoch": 0.18, + "learning_rate": 0.000415295887662989, + "loss": 1.643, + "theoretical_loss": 3.8775342613257395, + "tokens_seen": 586481664 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 223107, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.384760856628418, + "objective/train/theoretical_loss": 3.8774892039466833, + "objective/train/tokens_used": 586547200, + "theoretical_loss": 3.8774892039466833, + "tokens_seen": 586547200 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004152858575727181, + "loss": 1.4601, + "theoretical_loss": 3.8774892039466833, + "tokens_seen": 586547200 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041527582748244736, + "loss": 1.46, + "theoretical_loss": 3.877444153011126, + "tokens_seen": 586612736 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004152657973921765, + "loss": 1.4217, + "theoretical_loss": 3.8773991085174284, + "tokens_seen": 586678272 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004152557673019057, + "loss": 1.4837, + "theoretical_loss": 3.8773540704639475, + "tokens_seen": 586743808 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004152457372116349, + "loss": 1.4365, + "theoretical_loss": 3.8773090388490448, + "tokens_seen": 586809344 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004152357071213641, + "loss": 1.5306, + "theoretical_loss": 3.8772640136710805, + "tokens_seen": 586874880 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041522567703109327, + "loss": 1.5934, + "theoretical_loss": 3.877218994928416, + "tokens_seen": 586940416 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004152156469408225, + "loss": 1.4256, + "theoretical_loss": 3.8771739826194134, + "tokens_seen": 587005952 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041520561685055163, + "loss": 1.4641, + "theoretical_loss": 3.8771289767424344, + "tokens_seen": 587071488 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041519558676028086, + "loss": 1.5691, + "theoretical_loss": 3.8770839772958428, + "tokens_seen": 587137024 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041518555667001, + "loss": 1.4213, + "theoretical_loss": 3.877038984278002, + "tokens_seen": 587202560 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004151755265797392, + "loss": 1.5534, + "theoretical_loss": 3.8769939976872756, + "tokens_seen": 587268096 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004151654964894684, + "loss": 1.5, + "theoretical_loss": 3.8769490175220294, + "tokens_seen": 587333632 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004151554663991976, + "loss": 1.5215, + "theoretical_loss": 3.876904043780628, + "tokens_seen": 587399168 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004151454363089268, + "loss": 1.5802, + "theoretical_loss": 3.8768590764614377, + "tokens_seen": 587464704 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041513540621865595, + "loss": 1.4874, + "theoretical_loss": 3.8768141155628255, + "tokens_seen": 587530240 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004151253761283852, + "loss": 1.441, + "theoretical_loss": 3.8767691610831587, + "tokens_seen": 587595776 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041511534603811437, + "loss": 1.5481, + "theoretical_loss": 3.876724213020804, + "tokens_seen": 587661312 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041510531594784355, + "loss": 1.6291, + "theoretical_loss": 3.876679271374131, + "tokens_seen": 587726848 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041509528585757273, + "loss": 1.5277, + "theoretical_loss": 3.8766343361415085, + "tokens_seen": 587792384 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041508525576730196, + "loss": 1.3834, + "theoretical_loss": 3.876589407321305, + "tokens_seen": 587857920 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004150752256770311, + "loss": 1.4229, + "theoretical_loss": 3.8765444849118924, + "tokens_seen": 587923456 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004150651955867603, + "loss": 1.5305, + "theoretical_loss": 3.8764995689116404, + "tokens_seen": 587988992 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041505516549648945, + "loss": 1.4696, + "theoretical_loss": 3.8764546593189206, + "tokens_seen": 588054528 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004150451354062187, + "loss": 1.6174, + "theoretical_loss": 3.876409756132105, + "tokens_seen": 588120064 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041503510531594787, + "loss": 1.5629, + "theoretical_loss": 3.8763648593495663, + "tokens_seen": 588185600 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041502507522567705, + "loss": 1.497, + "theoretical_loss": 3.8763199689696783, + "tokens_seen": 588251136 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041501504513540623, + "loss": 1.4688, + "theoretical_loss": 3.876275084990813, + "tokens_seen": 588316672 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004150050150451354, + "loss": 1.5598, + "theoretical_loss": 3.876230207411347, + "tokens_seen": 588382208 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004149949849548646, + "loss": 1.5631, + "theoretical_loss": 3.8761853362296534, + "tokens_seen": 588447744 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041498495486459383, + "loss": 1.4433, + "theoretical_loss": 3.8761404714441086, + "tokens_seen": 588513280 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041497492477432296, + "loss": 1.5159, + "theoretical_loss": 3.8760956130530886, + "tokens_seen": 588578816 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004149648946840522, + "loss": 1.4364, + "theoretical_loss": 3.8760507610549704, + "tokens_seen": 588644352 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004149548645937813, + "loss": 1.5767, + "theoretical_loss": 3.876005915448131, + "tokens_seen": 588709888 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041494483450351055, + "loss": 1.5738, + "theoretical_loss": 3.8759610762309484, + "tokens_seen": 588775424 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041493480441323974, + "loss": 1.508, + "theoretical_loss": 3.875916243401801, + "tokens_seen": 588840960 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004149247743229689, + "loss": 1.4552, + "theoretical_loss": 3.8758714169590682, + "tokens_seen": 588906496 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004149147442326981, + "loss": 1.463, + "theoretical_loss": 3.87582659690113, + "tokens_seen": 588972032 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041490471414242733, + "loss": 1.5567, + "theoretical_loss": 3.875781783226366, + "tokens_seen": 589037568 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041489468405215646, + "loss": 1.4926, + "theoretical_loss": 3.8757369759331572, + "tokens_seen": 589103104 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004148846539618857, + "loss": 1.4762, + "theoretical_loss": 3.8756921750198856, + "tokens_seen": 589168640 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004148746238716148, + "loss": 1.5207, + "theoretical_loss": 3.8756473804849323, + "tokens_seen": 589234176 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041486459378134406, + "loss": 1.472, + "theoretical_loss": 3.8756025923266813, + "tokens_seen": 589299712 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041485456369107324, + "loss": 1.459, + "theoretical_loss": 3.8755578105435147, + "tokens_seen": 589365248 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004148445336008024, + "loss": 1.5546, + "theoretical_loss": 3.8755130351338165, + "tokens_seen": 589430784 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004148345035105316, + "loss": 1.4863, + "theoretical_loss": 3.8754682660959716, + "tokens_seen": 589496320 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004148244734202608, + "loss": 1.4848, + "theoretical_loss": 3.8754235034283653, + "tokens_seen": 589561856 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041481444332998996, + "loss": 1.4533, + "theoretical_loss": 3.875378747129382, + "tokens_seen": 589627392 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004148044132397192, + "loss": 1.508, + "theoretical_loss": 3.8753339971974086, + "tokens_seen": 589692928 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004147943831494483, + "loss": 1.4999, + "theoretical_loss": 3.8752892536308323, + "tokens_seen": 589758464 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 224258, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3773882389068604, + "objective/train/theoretical_loss": 3.87524451642804, + "objective/train/tokens_used": 589824000, + "theoretical_loss": 3.87524451642804, + "tokens_seen": 589824000 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041478435305917756, + "loss": 1.427, + "theoretical_loss": 3.87524451642804, + "tokens_seen": 589824000 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004147743229689067, + "loss": 1.4824, + "theoretical_loss": 3.8751997855874194, + "tokens_seen": 589889536 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004147642928786359, + "loss": 1.5382, + "theoretical_loss": 3.8751550611073595, + "tokens_seen": 589955072 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004147542627883651, + "loss": 1.5691, + "theoretical_loss": 3.875110342986249, + "tokens_seen": 590020608 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004147442326980943, + "loss": 1.5666, + "theoretical_loss": 3.875065631222478, + "tokens_seen": 590086144 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041473420260782347, + "loss": 1.4829, + "theoretical_loss": 3.875020925814437, + "tokens_seen": 590151680 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004147241725175527, + "loss": 1.5435, + "theoretical_loss": 3.8749762267605163, + "tokens_seen": 590217216 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041471414242728183, + "loss": 1.6688, + "theoretical_loss": 3.8749315340591073, + "tokens_seen": 590282752 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041470411233701106, + "loss": 1.5698, + "theoretical_loss": 3.874886847708603, + "tokens_seen": 590348288 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004146940822467402, + "loss": 1.5918, + "theoretical_loss": 3.8748421677073956, + "tokens_seen": 590413824 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004146840521564694, + "loss": 1.4598, + "theoretical_loss": 3.874797494053878, + "tokens_seen": 590479360 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004146740220661986, + "loss": 1.5514, + "theoretical_loss": 3.8747528267464437, + "tokens_seen": 590544896 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004146639919759278, + "loss": 1.5916, + "theoretical_loss": 3.8747081657834874, + "tokens_seen": 590610432 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041465396188565697, + "loss": 1.5733, + "theoretical_loss": 3.874663511163405, + "tokens_seen": 590675968 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041464393179538615, + "loss": 1.5967, + "theoretical_loss": 3.8746188628845912, + "tokens_seen": 590741504 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041463390170511533, + "loss": 1.4742, + "theoretical_loss": 3.8745742209454423, + "tokens_seen": 590807040 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041462387161484457, + "loss": 1.4816, + "theoretical_loss": 3.874529585344355, + "tokens_seen": 590872576 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004146138415245737, + "loss": 1.5075, + "theoretical_loss": 3.874484956079727, + "tokens_seen": 590938112 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041460381143430293, + "loss": 1.4935, + "theoretical_loss": 3.8744403331499555, + "tokens_seen": 591003648 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004145937813440321, + "loss": 1.4544, + "theoretical_loss": 3.8743957165534395, + "tokens_seen": 591069184 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004145837512537613, + "loss": 1.471, + "theoretical_loss": 3.8743511062885783, + "tokens_seen": 591134720 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004145737211634905, + "loss": 1.4172, + "theoretical_loss": 3.874306502353771, + "tokens_seen": 591200256 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041456369107321965, + "loss": 1.5015, + "theoretical_loss": 3.8742619047474176, + "tokens_seen": 591265792 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041455366098294884, + "loss": 1.4873, + "theoretical_loss": 3.87421731346792, + "tokens_seen": 591331328 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041454363089267807, + "loss": 1.323, + "theoretical_loss": 3.874172728513678, + "tokens_seen": 591396864 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004145336008024072, + "loss": 1.4538, + "theoretical_loss": 3.874128149883095, + "tokens_seen": 591462400 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041452357071213643, + "loss": 1.4549, + "theoretical_loss": 3.8740835775745732, + "tokens_seen": 591527936 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041451354062186556, + "loss": 1.5516, + "theoretical_loss": 3.8740390115865155, + "tokens_seen": 591593472 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004145035105315948, + "loss": 1.5457, + "theoretical_loss": 3.873994451917326, + "tokens_seen": 591659008 + }, + { + "epoch": 0.18, + "learning_rate": 0.000414493480441324, + "loss": 1.4702, + "theoretical_loss": 3.8739498985654084, + "tokens_seen": 591724544 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041448345035105316, + "loss": 1.4666, + "theoretical_loss": 3.873905351529168, + "tokens_seen": 591790080 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041447342026078234, + "loss": 1.4826, + "theoretical_loss": 3.8738608108070096, + "tokens_seen": 591855616 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004144633901705115, + "loss": 1.5147, + "theoretical_loss": 3.8738162763973407, + "tokens_seen": 591921152 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004144533600802407, + "loss": 1.419, + "theoretical_loss": 3.8737717482985667, + "tokens_seen": 591986688 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041444332998996994, + "loss": 1.4346, + "theoretical_loss": 3.8737272265090947, + "tokens_seen": 592052224 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041443329989969906, + "loss": 1.4588, + "theoretical_loss": 3.8736827110273335, + "tokens_seen": 592117760 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004144232698094283, + "loss": 1.5815, + "theoretical_loss": 3.8736382018516906, + "tokens_seen": 592183296 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041441323971915753, + "loss": 1.5333, + "theoretical_loss": 3.8735936989805744, + "tokens_seen": 592248832 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041440320962888666, + "loss": 1.5148, + "theoretical_loss": 3.873549202412396, + "tokens_seen": 592314368 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004143931795386159, + "loss": 1.4262, + "theoretical_loss": 3.8735047121455644, + "tokens_seen": 592379904 + }, + { + "epoch": 0.18, + "learning_rate": 0.000414383149448345, + "loss": 1.4958, + "theoretical_loss": 3.873460228178491, + "tokens_seen": 592445440 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041437311935807426, + "loss": 1.5973, + "theoretical_loss": 3.8734157505095856, + "tokens_seen": 592510976 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041436308926780344, + "loss": 1.4764, + "theoretical_loss": 3.8733712791372614, + "tokens_seen": 592576512 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004143530591775326, + "loss": 1.3932, + "theoretical_loss": 3.8733268140599306, + "tokens_seen": 592642048 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004143430290872618, + "loss": 1.5277, + "theoretical_loss": 3.8732823552760056, + "tokens_seen": 592707584 + }, + { + "epoch": 0.18, + "learning_rate": 0.000414332998996991, + "loss": 1.4903, + "theoretical_loss": 3.8732379027839, + "tokens_seen": 592773120 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041432296890672016, + "loss": 1.5271, + "theoretical_loss": 3.8731934565820287, + "tokens_seen": 592838656 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004143129388164494, + "loss": 1.5268, + "theoretical_loss": 3.8731490166688056, + "tokens_seen": 592904192 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004143029087261785, + "loss": 1.4811, + "theoretical_loss": 3.873104583042646, + "tokens_seen": 592969728 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041429287863590776, + "loss": 1.642, + "theoretical_loss": 3.8730601557019657, + "tokens_seen": 593035264 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 225581, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5393953323364258, + "objective/train/theoretical_loss": 3.873015734645182, + "objective/train/tokens_used": 593100800, + "theoretical_loss": 3.873015734645182, + "tokens_seen": 593100800 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004142828485456369, + "loss": 1.5067, + "theoretical_loss": 3.873015734645182, + "tokens_seen": 593100800 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004142728184553661, + "loss": 1.5658, + "theoretical_loss": 3.872971319870711, + "tokens_seen": 593166336 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004142627883650953, + "loss": 1.5321, + "theoretical_loss": 3.872926911376971, + "tokens_seen": 593231872 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004142527582748245, + "loss": 1.6531, + "theoretical_loss": 3.8728825091623786, + "tokens_seen": 593297408 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041424272818455367, + "loss": 1.4322, + "theoretical_loss": 3.872838113225354, + "tokens_seen": 593362944 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004142326980942829, + "loss": 1.5073, + "theoretical_loss": 3.8727937235643157, + "tokens_seen": 593428480 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041422266800401203, + "loss": 1.5421, + "theoretical_loss": 3.872749340177685, + "tokens_seen": 593494016 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041421263791374126, + "loss": 1.5989, + "theoretical_loss": 3.87270496306388, + "tokens_seen": 593559552 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004142026078234704, + "loss": 1.4628, + "theoretical_loss": 3.8726605922213233, + "tokens_seen": 593625088 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004141925777331996, + "loss": 1.4604, + "theoretical_loss": 3.872616227648436, + "tokens_seen": 593690624 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004141825476429288, + "loss": 1.4113, + "theoretical_loss": 3.8725718693436404, + "tokens_seen": 593756160 + }, + { + "epoch": 0.18, + "learning_rate": 0.000414172517552658, + "loss": 1.4722, + "theoretical_loss": 3.8725275173053593, + "tokens_seen": 593821696 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041416248746238717, + "loss": 1.5004, + "theoretical_loss": 3.872483171532015, + "tokens_seen": 593887232 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041415245737211635, + "loss": 1.5061, + "theoretical_loss": 3.8724388320220333, + "tokens_seen": 593952768 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041414242728184553, + "loss": 1.5012, + "theoretical_loss": 3.872394498773837, + "tokens_seen": 594018304 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041413239719157477, + "loss": 1.4644, + "theoretical_loss": 3.8723501717858513, + "tokens_seen": 594083840 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004141223671013039, + "loss": 1.4669, + "theoretical_loss": 3.872305851056502, + "tokens_seen": 594149376 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041411233701103313, + "loss": 1.4482, + "theoretical_loss": 3.8722615365842157, + "tokens_seen": 594214912 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004141023069207623, + "loss": 1.448, + "theoretical_loss": 3.8722172283674183, + "tokens_seen": 594280448 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004140922768304915, + "loss": 1.5318, + "theoretical_loss": 3.8721729264045375, + "tokens_seen": 594345984 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004140822467402207, + "loss": 1.6096, + "theoretical_loss": 3.872128630694001, + "tokens_seen": 594411520 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041407221664994985, + "loss": 1.5428, + "theoretical_loss": 3.8720843412342374, + "tokens_seen": 594477056 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041406218655967904, + "loss": 1.3736, + "theoretical_loss": 3.872040058023676, + "tokens_seen": 594542592 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041405215646940827, + "loss": 1.5429, + "theoretical_loss": 3.871995781060745, + "tokens_seen": 594608128 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004140421263791374, + "loss": 1.4023, + "theoretical_loss": 3.8719515103438757, + "tokens_seen": 594673664 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041403209628886663, + "loss": 1.4503, + "theoretical_loss": 3.871907245871499, + "tokens_seen": 594739200 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041402206619859576, + "loss": 1.4476, + "theoretical_loss": 3.871862987642045, + "tokens_seen": 594804736 + }, + { + "epoch": 0.18, + "learning_rate": 0.000414012036108325, + "loss": 1.5285, + "theoretical_loss": 3.871818735653947, + "tokens_seen": 594870272 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004140020060180542, + "loss": 1.5299, + "theoretical_loss": 3.8717744899056354, + "tokens_seen": 594935808 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041399197592778336, + "loss": 1.5582, + "theoretical_loss": 3.8717302503955446, + "tokens_seen": 595001344 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041398194583751254, + "loss": 1.5144, + "theoretical_loss": 3.8716860171221086, + "tokens_seen": 595066880 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004139719157472417, + "loss": 1.4946, + "theoretical_loss": 3.8716417900837596, + "tokens_seen": 595132416 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004139618856569709, + "loss": 1.4463, + "theoretical_loss": 3.8715975692789337, + "tokens_seen": 595197952 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041395185556670014, + "loss": 1.5055, + "theoretical_loss": 3.871553354706066, + "tokens_seen": 595263488 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041394182547642926, + "loss": 1.4823, + "theoretical_loss": 3.871509146363592, + "tokens_seen": 595329024 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004139317953861585, + "loss": 1.5404, + "theoretical_loss": 3.8714649442499476, + "tokens_seen": 595394560 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004139217652958877, + "loss": 1.5338, + "theoretical_loss": 3.8714207483635703, + "tokens_seen": 595460096 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041391173520561686, + "loss": 1.6063, + "theoretical_loss": 3.8713765587028974, + "tokens_seen": 595525632 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041390170511534604, + "loss": 1.5, + "theoretical_loss": 3.871332375266367, + "tokens_seen": 595591168 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004138916750250752, + "loss": 1.4201, + "theoretical_loss": 3.871288198052418, + "tokens_seen": 595656704 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004138816449348044, + "loss": 1.4602, + "theoretical_loss": 3.871244027059489, + "tokens_seen": 595722240 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041387161484453364, + "loss": 1.5021, + "theoretical_loss": 3.8711998622860193, + "tokens_seen": 595787776 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041386158475426277, + "loss": 1.4299, + "theoretical_loss": 3.8711557037304507, + "tokens_seen": 595853312 + }, + { + "epoch": 0.18, + "learning_rate": 0.000413851554663992, + "loss": 1.4441, + "theoretical_loss": 3.8711115513912224, + "tokens_seen": 595918848 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041384152457372113, + "loss": 1.5939, + "theoretical_loss": 3.8710674052667766, + "tokens_seen": 595984384 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041383149448345036, + "loss": 1.5101, + "theoretical_loss": 3.8710232653555554, + "tokens_seen": 596049920 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041382146439317954, + "loss": 1.5915, + "theoretical_loss": 3.870979131656001, + "tokens_seen": 596115456 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004138114343029087, + "loss": 1.5006, + "theoretical_loss": 3.870935004166557, + "tokens_seen": 596180992 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004138014042126379, + "loss": 1.4793, + "theoretical_loss": 3.870890882885667, + "tokens_seen": 596246528 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004137913741223671, + "loss": 1.5159, + "theoretical_loss": 3.8708467678117744, + "tokens_seen": 596312064 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 227080, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5293521881103516, + "objective/train/theoretical_loss": 3.870802658943324, + "objective/train/tokens_used": 596377600, + "theoretical_loss": 3.870802658943324, + "tokens_seen": 596377600 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041378134403209627, + "loss": 1.531, + "theoretical_loss": 3.870802658943324, + "tokens_seen": 596377600 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004137713139418255, + "loss": 1.4697, + "theoretical_loss": 3.8707585562787625, + "tokens_seen": 596443136 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041376128385155463, + "loss": 1.4779, + "theoretical_loss": 3.8707144598165346, + "tokens_seen": 596508672 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041375125376128387, + "loss": 1.5669, + "theoretical_loss": 3.8706703695550875, + "tokens_seen": 596574208 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041374122367101305, + "loss": 1.5561, + "theoretical_loss": 3.8706262854928672, + "tokens_seen": 596639744 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041373119358074223, + "loss": 1.4663, + "theoretical_loss": 3.8705822076283223, + "tokens_seen": 596705280 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004137211634904714, + "loss": 1.5521, + "theoretical_loss": 3.870538135959901, + "tokens_seen": 596770816 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004137111334002006, + "loss": 1.4379, + "theoretical_loss": 3.8704940704860507, + "tokens_seen": 596836352 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041370110330992977, + "loss": 1.4917, + "theoretical_loss": 3.8704500112052216, + "tokens_seen": 596901888 + }, + { + "epoch": 0.18, + "learning_rate": 0.000413691073219659, + "loss": 1.5656, + "theoretical_loss": 3.870405958115864, + "tokens_seen": 596967424 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041368104312938813, + "loss": 1.5278, + "theoretical_loss": 3.870361911216427, + "tokens_seen": 597032960 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041367101303911737, + "loss": 1.5693, + "theoretical_loss": 3.8703178705053625, + "tokens_seen": 597098496 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041366098294884655, + "loss": 1.5551, + "theoretical_loss": 3.8702738359811217, + "tokens_seen": 597164032 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041365095285857573, + "loss": 1.4591, + "theoretical_loss": 3.8702298076421564, + "tokens_seen": 597229568 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041364092276830497, + "loss": 1.4249, + "theoretical_loss": 3.87018578548692, + "tokens_seen": 597295104 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004136308926780341, + "loss": 1.5279, + "theoretical_loss": 3.8701417695138645, + "tokens_seen": 597360640 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041362086258776333, + "loss": 1.5214, + "theoretical_loss": 3.8700977597214443, + "tokens_seen": 597426176 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004136108324974925, + "loss": 1.4813, + "theoretical_loss": 3.870053756108114, + "tokens_seen": 597491712 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004136008024072217, + "loss": 1.5501, + "theoretical_loss": 3.8700097586723277, + "tokens_seen": 597557248 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004135907723169509, + "loss": 1.4747, + "theoretical_loss": 3.869965767412541, + "tokens_seen": 597622784 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041358074222668005, + "loss": 1.6024, + "theoretical_loss": 3.8699217823272103, + "tokens_seen": 597688320 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041357071213640924, + "loss": 1.5424, + "theoretical_loss": 3.8698778034147914, + "tokens_seen": 597753856 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041356068204613847, + "loss": 1.4724, + "theoretical_loss": 3.869833830673742, + "tokens_seen": 597819392 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004135506519558676, + "loss": 1.4655, + "theoretical_loss": 3.8697898641025192, + "tokens_seen": 597884928 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041354062186559683, + "loss": 1.5786, + "theoretical_loss": 3.8697459036995814, + "tokens_seen": 597950464 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041353059177532596, + "loss": 1.5357, + "theoretical_loss": 3.869701949463387, + "tokens_seen": 598016000 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004135205616850552, + "loss": 1.3477, + "theoretical_loss": 3.869658001392396, + "tokens_seen": 598081536 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004135105315947844, + "loss": 1.4285, + "theoretical_loss": 3.869614059485068, + "tokens_seen": 598147072 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041350050150451356, + "loss": 1.4174, + "theoretical_loss": 3.8695701237398623, + "tokens_seen": 598212608 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041349047141424274, + "loss": 1.3404, + "theoretical_loss": 3.869526194155241, + "tokens_seen": 598278144 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004134804413239719, + "loss": 1.4516, + "theoretical_loss": 3.869482270729665, + "tokens_seen": 598343680 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004134704112337011, + "loss": 1.4249, + "theoretical_loss": 3.8694383534615966, + "tokens_seen": 598409216 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041346038114343034, + "loss": 1.5518, + "theoretical_loss": 3.8693944423494986, + "tokens_seen": 598474752 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041345035105315946, + "loss": 1.5515, + "theoretical_loss": 3.869350537391834, + "tokens_seen": 598540288 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004134403209628887, + "loss": 1.4626, + "theoretical_loss": 3.869306638587066, + "tokens_seen": 598605824 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004134302908726179, + "loss": 1.5709, + "theoretical_loss": 3.8692627459336597, + "tokens_seen": 598671360 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041342026078234706, + "loss": 1.5358, + "theoretical_loss": 3.8692188594300787, + "tokens_seen": 598736896 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041341023069207624, + "loss": 1.5526, + "theoretical_loss": 3.8691749790747894, + "tokens_seen": 598802432 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004134002006018054, + "loss": 1.5204, + "theoretical_loss": 3.8691311048662573, + "tokens_seen": 598867968 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004133901705115346, + "loss": 1.5882, + "theoretical_loss": 3.869087236802949, + "tokens_seen": 598933504 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041338014042126384, + "loss": 1.4599, + "theoretical_loss": 3.869043374883331, + "tokens_seen": 598999040 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041337011033099297, + "loss": 1.482, + "theoretical_loss": 3.8689995191058717, + "tokens_seen": 599064576 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004133600802407222, + "loss": 1.4594, + "theoretical_loss": 3.868955669469039, + "tokens_seen": 599130112 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041335005015045133, + "loss": 1.497, + "theoretical_loss": 3.868911825971301, + "tokens_seen": 599195648 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041334002006018056, + "loss": 1.508, + "theoretical_loss": 3.868867988611127, + "tokens_seen": 599261184 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041332998996990975, + "loss": 1.4582, + "theoretical_loss": 3.868824157386987, + "tokens_seen": 599326720 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004133199598796389, + "loss": 1.4696, + "theoretical_loss": 3.8687803322973506, + "tokens_seen": 599392256 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004133099297893681, + "loss": 1.5181, + "theoretical_loss": 3.86873651334069, + "tokens_seen": 599457792 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004132998996990973, + "loss": 1.5085, + "theoretical_loss": 3.8686927005154756, + "tokens_seen": 599523328 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041328986960882647, + "loss": 1.6037, + "theoretical_loss": 3.8686488938201795, + "tokens_seen": 599588864 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 228205, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4352452754974365, + "objective/train/theoretical_loss": 3.868605093253274, + "objective/train/tokens_used": 599654400, + "theoretical_loss": 3.868605093253274, + "tokens_seen": 599654400 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004132798395185557, + "loss": 1.4503, + "theoretical_loss": 3.868605093253274, + "tokens_seen": 599654400 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041326980942828483, + "loss": 1.4702, + "theoretical_loss": 3.8685612988132334, + "tokens_seen": 599719936 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041325977933801407, + "loss": 1.6362, + "theoretical_loss": 3.8685175104985294, + "tokens_seen": 599785472 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041324974924774325, + "loss": 1.6019, + "theoretical_loss": 3.868473728307637, + "tokens_seen": 599851008 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041323971915747243, + "loss": 1.4776, + "theoretical_loss": 3.868429952239031, + "tokens_seen": 599916544 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004132296890672016, + "loss": 1.4838, + "theoretical_loss": 3.8683861822911867, + "tokens_seen": 599982080 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004132196589769308, + "loss": 1.5197, + "theoretical_loss": 3.8683424184625794, + "tokens_seen": 600047616 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041320962888665997, + "loss": 1.4899, + "theoretical_loss": 3.8682986607516856, + "tokens_seen": 600113152 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004131995987963892, + "loss": 1.5843, + "theoretical_loss": 3.8682549091569824, + "tokens_seen": 600178688 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041318956870611833, + "loss": 1.5442, + "theoretical_loss": 3.868211163676947, + "tokens_seen": 600244224 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041317953861584757, + "loss": 1.487, + "theoretical_loss": 3.868167424310057, + "tokens_seen": 600309760 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004131695085255767, + "loss": 1.4906, + "theoretical_loss": 3.868123691054792, + "tokens_seen": 600375296 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041315947843530593, + "loss": 1.4688, + "theoretical_loss": 3.8680799639096297, + "tokens_seen": 600440832 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004131494483450351, + "loss": 1.4756, + "theoretical_loss": 3.8680362428730506, + "tokens_seen": 600506368 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004131394182547643, + "loss": 1.5073, + "theoretical_loss": 3.8679925279435343, + "tokens_seen": 600571904 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004131293881644935, + "loss": 1.5003, + "theoretical_loss": 3.8679488191195617, + "tokens_seen": 600637440 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004131193580742227, + "loss": 1.4205, + "theoretical_loss": 3.8679051163996143, + "tokens_seen": 600702976 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041310932798395184, + "loss": 1.46, + "theoretical_loss": 3.867861419782173, + "tokens_seen": 600768512 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004130992978936811, + "loss": 1.4062, + "theoretical_loss": 3.867817729265721, + "tokens_seen": 600834048 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004130892678034102, + "loss": 1.5368, + "theoretical_loss": 3.8677740448487414, + "tokens_seen": 600899584 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041307923771313944, + "loss": 1.4181, + "theoretical_loss": 3.867730366529716, + "tokens_seen": 600965120 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004130692076228686, + "loss": 1.5556, + "theoretical_loss": 3.86768669430713, + "tokens_seen": 601030656 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004130591775325978, + "loss": 1.4351, + "theoretical_loss": 3.867643028179468, + "tokens_seen": 601096192 + }, + { + "epoch": 0.18, + "learning_rate": 0.000413049147442327, + "loss": 1.5366, + "theoretical_loss": 3.867599368145214, + "tokens_seen": 601161728 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041303911735205616, + "loss": 1.3528, + "theoretical_loss": 3.8675557142028545, + "tokens_seen": 601227264 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041302908726178534, + "loss": 1.5907, + "theoretical_loss": 3.8675120663508755, + "tokens_seen": 601292800 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004130190571715146, + "loss": 1.6228, + "theoretical_loss": 3.867468424587763, + "tokens_seen": 601358336 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004130090270812437, + "loss": 1.5003, + "theoretical_loss": 3.867424788912005, + "tokens_seen": 601423872 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041299899699097294, + "loss": 1.4884, + "theoretical_loss": 3.8673811593220884, + "tokens_seen": 601489408 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041298896690070207, + "loss": 1.6127, + "theoretical_loss": 3.8673375358165014, + "tokens_seen": 601554944 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004129789368104313, + "loss": 1.5327, + "theoretical_loss": 3.867293918393734, + "tokens_seen": 601620480 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004129689067201605, + "loss": 1.5762, + "theoretical_loss": 3.8672503070522746, + "tokens_seen": 601686016 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041295887662988966, + "loss": 1.4653, + "theoretical_loss": 3.8672067017906135, + "tokens_seen": 601751552 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041294884653961884, + "loss": 1.5365, + "theoretical_loss": 3.8671631026072406, + "tokens_seen": 601817088 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004129388164493481, + "loss": 1.4848, + "theoretical_loss": 3.8671195095006468, + "tokens_seen": 601882624 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004129287863590772, + "loss": 1.4374, + "theoretical_loss": 3.8670759224693247, + "tokens_seen": 601948160 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041291875626880644, + "loss": 1.3587, + "theoretical_loss": 3.867032341511765, + "tokens_seen": 602013696 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004129087261785356, + "loss": 1.496, + "theoretical_loss": 3.866988766626461, + "tokens_seen": 602079232 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004128986960882648, + "loss": 1.4673, + "theoretical_loss": 3.866945197811906, + "tokens_seen": 602144768 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041288866599799404, + "loss": 1.5547, + "theoretical_loss": 3.866901635066593, + "tokens_seen": 602210304 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041287863590772317, + "loss": 1.435, + "theoretical_loss": 3.8668580783890167, + "tokens_seen": 602275840 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004128686058174524, + "loss": 1.5551, + "theoretical_loss": 3.866814527777671, + "tokens_seen": 602341376 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041285857572718153, + "loss": 1.4346, + "theoretical_loss": 3.8667709832310524, + "tokens_seen": 602406912 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041284854563691076, + "loss": 1.4926, + "theoretical_loss": 3.866727444747656, + "tokens_seen": 602472448 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041283851554663995, + "loss": 1.5024, + "theoretical_loss": 3.8666839123259784, + "tokens_seen": 602537984 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004128284854563691, + "loss": 1.5546, + "theoretical_loss": 3.8666403859645158, + "tokens_seen": 602603520 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004128184553660983, + "loss": 1.5123, + "theoretical_loss": 3.8665968656617666, + "tokens_seen": 602669056 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004128084252758275, + "loss": 1.4942, + "theoretical_loss": 3.8665533514162282, + "tokens_seen": 602734592 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041279839518555667, + "loss": 1.4678, + "theoretical_loss": 3.8665098432263987, + "tokens_seen": 602800128 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004127883650952859, + "loss": 1.4371, + "theoretical_loss": 3.866466341090778, + "tokens_seen": 602865664 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 229402, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.449509620666504, + "objective/train/theoretical_loss": 3.866422845007865, + "objective/train/tokens_used": 602931200, + "theoretical_loss": 3.866422845007865, + "tokens_seen": 602931200 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041277833500501503, + "loss": 1.4831, + "theoretical_loss": 3.866422845007865, + "tokens_seen": 602931200 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041276830491474427, + "loss": 1.458, + "theoretical_loss": 3.86637935497616, + "tokens_seen": 602996736 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041275827482447345, + "loss": 1.5643, + "theoretical_loss": 3.8663358709941633, + "tokens_seen": 603062272 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041274824473420263, + "loss": 1.4161, + "theoretical_loss": 3.8662923930603768, + "tokens_seen": 603127808 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004127382146439318, + "loss": 1.5412, + "theoretical_loss": 3.8662489211733018, + "tokens_seen": 603193344 + }, + { + "epoch": 0.18, + "learning_rate": 0.000412728184553661, + "loss": 1.4535, + "theoretical_loss": 3.8662054553314404, + "tokens_seen": 603258880 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041271815446339017, + "loss": 1.443, + "theoretical_loss": 3.866161995533295, + "tokens_seen": 603324416 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004127081243731194, + "loss": 1.3476, + "theoretical_loss": 3.86611854177737, + "tokens_seen": 603389952 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041269809428284854, + "loss": 1.4434, + "theoretical_loss": 3.866075094062168, + "tokens_seen": 603455488 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041268806419257777, + "loss": 1.4759, + "theoretical_loss": 3.8660316523861944, + "tokens_seen": 603521024 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004126780341023069, + "loss": 1.5885, + "theoretical_loss": 3.8659882167479527, + "tokens_seen": 603586560 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041266800401203613, + "loss": 1.4871, + "theoretical_loss": 3.86594478714595, + "tokens_seen": 603652096 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004126579739217653, + "loss": 1.5406, + "theoretical_loss": 3.8659013635786907, + "tokens_seen": 603717632 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004126479438314945, + "loss": 1.5587, + "theoretical_loss": 3.8658579460446827, + "tokens_seen": 603783168 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004126379137412237, + "loss": 1.4307, + "theoretical_loss": 3.865814534542432, + "tokens_seen": 603848704 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004126278836509529, + "loss": 1.4443, + "theoretical_loss": 3.8657711290704473, + "tokens_seen": 603914240 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041261785356068204, + "loss": 1.4068, + "theoretical_loss": 3.8657277296272348, + "tokens_seen": 603979776 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004126078234704113, + "loss": 1.5047, + "theoretical_loss": 3.865684336211305, + "tokens_seen": 604045312 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004125977933801404, + "loss": 1.5551, + "theoretical_loss": 3.8656409488211656, + "tokens_seen": 604110848 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041258776328986964, + "loss": 1.4335, + "theoretical_loss": 3.8655975674553273, + "tokens_seen": 604176384 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004125777331995988, + "loss": 1.3817, + "theoretical_loss": 3.8655541921123, + "tokens_seen": 604241920 + }, + { + "epoch": 0.18, + "learning_rate": 0.000412567703109328, + "loss": 1.4057, + "theoretical_loss": 3.8655108227905943, + "tokens_seen": 604307456 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004125576730190572, + "loss": 1.5198, + "theoretical_loss": 3.8654674594887215, + "tokens_seen": 604372992 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041254764292878636, + "loss": 1.4446, + "theoretical_loss": 3.8654241022051927, + "tokens_seen": 604438528 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041253761283851554, + "loss": 1.4948, + "theoretical_loss": 3.865380750938522, + "tokens_seen": 604504064 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004125275827482448, + "loss": 1.5283, + "theoretical_loss": 3.86533740568722, + "tokens_seen": 604569600 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004125175526579739, + "loss": 1.467, + "theoretical_loss": 3.8652940664498026, + "tokens_seen": 604635136 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041250752256770314, + "loss": 1.4948, + "theoretical_loss": 3.865250733224781, + "tokens_seen": 604700672 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041249749247743227, + "loss": 1.5835, + "theoretical_loss": 3.865207406010672, + "tokens_seen": 604766208 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004124874623871615, + "loss": 1.6045, + "theoretical_loss": 3.865164084805989, + "tokens_seen": 604831744 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004124774322968907, + "loss": 1.4384, + "theoretical_loss": 3.865120769609248, + "tokens_seen": 604897280 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041246740220661986, + "loss": 1.2854, + "theoretical_loss": 3.8650774604189655, + "tokens_seen": 604962816 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041245737211634904, + "loss": 1.4222, + "theoretical_loss": 3.8650341572336573, + "tokens_seen": 605028352 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004124473420260783, + "loss": 1.451, + "theoretical_loss": 3.8649908600518406, + "tokens_seen": 605093888 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004124373119358074, + "loss": 1.5067, + "theoretical_loss": 3.8649475688720334, + "tokens_seen": 605159424 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041242728184553664, + "loss": 1.3186, + "theoretical_loss": 3.864904283692754, + "tokens_seen": 605224960 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041241725175526577, + "loss": 1.4625, + "theoretical_loss": 3.8648610045125196, + "tokens_seen": 605290496 + }, + { + "epoch": 0.18, + "learning_rate": 0.000412407221664995, + "loss": 1.478, + "theoretical_loss": 3.8648177313298513, + "tokens_seen": 605356032 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004123971915747242, + "loss": 1.4973, + "theoretical_loss": 3.8647744641432675, + "tokens_seen": 605421568 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041238716148445337, + "loss": 1.5071, + "theoretical_loss": 3.8647312029512886, + "tokens_seen": 605487104 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041237713139418255, + "loss": 1.414, + "theoretical_loss": 3.864687947752436, + "tokens_seen": 605552640 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041236710130391173, + "loss": 1.5018, + "theoretical_loss": 3.8646446985452307, + "tokens_seen": 605618176 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004123570712136409, + "loss": 1.5712, + "theoretical_loss": 3.864601455328194, + "tokens_seen": 605683712 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041234704112337015, + "loss": 1.4748, + "theoretical_loss": 3.864558218099849, + "tokens_seen": 605749248 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041233701103309927, + "loss": 1.5032, + "theoretical_loss": 3.864514986858718, + "tokens_seen": 605814784 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004123269809428285, + "loss": 1.5493, + "theoretical_loss": 3.8644717616033244, + "tokens_seen": 605880320 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041231695085255763, + "loss": 1.5043, + "theoretical_loss": 3.8644285423321927, + "tokens_seen": 605945856 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041230692076228687, + "loss": 1.4045, + "theoretical_loss": 3.864385329043847, + "tokens_seen": 606011392 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041229689067201605, + "loss": 1.4542, + "theoretical_loss": 3.8643421217368115, + "tokens_seen": 606076928 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041228686058174523, + "loss": 1.3896, + "theoretical_loss": 3.8642989204096123, + "tokens_seen": 606142464 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 230572, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5414414405822754, + "objective/train/theoretical_loss": 3.8642557250607754, + "objective/train/tokens_used": 606208000, + "theoretical_loss": 3.8642557250607754, + "tokens_seen": 606208000 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004122768304914744, + "loss": 1.5026, + "theoretical_loss": 3.8642557250607754, + "tokens_seen": 606208000 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041226680040120365, + "loss": 1.5326, + "theoretical_loss": 3.8642125356888277, + "tokens_seen": 606273536 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004122567703109328, + "loss": 1.4342, + "theoretical_loss": 3.8641693522922953, + "tokens_seen": 606339072 + }, + { + "epoch": 0.18, + "learning_rate": 0.000412246740220662, + "loss": 1.5456, + "theoretical_loss": 3.8641261748697064, + "tokens_seen": 606404608 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041223671013039114, + "loss": 1.5595, + "theoretical_loss": 3.8640830034195894, + "tokens_seen": 606470144 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004122266800401204, + "loss": 1.5477, + "theoretical_loss": 3.864039837940472, + "tokens_seen": 606535680 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041221664994984955, + "loss": 1.525, + "theoretical_loss": 3.863996678430884, + "tokens_seen": 606601216 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041220661985957874, + "loss": 1.4298, + "theoretical_loss": 3.8639535248893555, + "tokens_seen": 606666752 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004121965897693079, + "loss": 1.4894, + "theoretical_loss": 3.863910377314415, + "tokens_seen": 606732288 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004121865596790371, + "loss": 1.4618, + "theoretical_loss": 3.863867235704595, + "tokens_seen": 606797824 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004121765295887663, + "loss": 1.4489, + "theoretical_loss": 3.8638241000584257, + "tokens_seen": 606863360 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004121664994984955, + "loss": 1.4237, + "theoretical_loss": 3.863780970374439, + "tokens_seen": 606928896 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004121564694082247, + "loss": 1.5016, + "theoretical_loss": 3.863737846651167, + "tokens_seen": 606994432 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004121464393179539, + "loss": 1.5674, + "theoretical_loss": 3.8636947288871424, + "tokens_seen": 607059968 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004121364092276831, + "loss": 1.4889, + "theoretical_loss": 3.863651617080899, + "tokens_seen": 607125504 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041212637913741224, + "loss": 1.5151, + "theoretical_loss": 3.8636085112309715, + "tokens_seen": 607191040 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004121163490471415, + "loss": 1.4181, + "theoretical_loss": 3.863565411335892, + "tokens_seen": 607256576 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004121063189568706, + "loss": 1.5466, + "theoretical_loss": 3.8635223173941964, + "tokens_seen": 607322112 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041209628886659984, + "loss": 1.5453, + "theoretical_loss": 3.8634792294044202, + "tokens_seen": 607387648 + }, + { + "epoch": 0.18, + "learning_rate": 0.000412086258776329, + "loss": 1.3683, + "theoretical_loss": 3.8634361473650993, + "tokens_seen": 607453184 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004120762286860582, + "loss": 1.5054, + "theoretical_loss": 3.8633930712747704, + "tokens_seen": 607518720 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004120661985957874, + "loss": 1.4194, + "theoretical_loss": 3.8633500011319692, + "tokens_seen": 607584256 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041205616850551656, + "loss": 1.4262, + "theoretical_loss": 3.8633069369352344, + "tokens_seen": 607649792 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041204613841524574, + "loss": 1.4893, + "theoretical_loss": 3.8632638786831035, + "tokens_seen": 607715328 + }, + { + "epoch": 0.18, + "learning_rate": 0.000412036108324975, + "loss": 1.588, + "theoretical_loss": 3.863220826374115, + "tokens_seen": 607780864 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004120260782347041, + "loss": 1.4474, + "theoretical_loss": 3.863177780006808, + "tokens_seen": 607846400 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041201604814443334, + "loss": 1.4918, + "theoretical_loss": 3.8631347395797215, + "tokens_seen": 607911936 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041200601805416247, + "loss": 1.5631, + "theoretical_loss": 3.8630917050913958, + "tokens_seen": 607977472 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004119959879638917, + "loss": 1.4705, + "theoretical_loss": 3.863048676540372, + "tokens_seen": 608043008 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004119859578736209, + "loss": 1.5076, + "theoretical_loss": 3.8630056539251902, + "tokens_seen": 608108544 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041197592778335006, + "loss": 1.5231, + "theoretical_loss": 3.862962637244393, + "tokens_seen": 608174080 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041196589769307924, + "loss": 1.434, + "theoretical_loss": 3.862919626496521, + "tokens_seen": 608239616 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004119558676028085, + "loss": 1.4665, + "theoretical_loss": 3.862876621680119, + "tokens_seen": 608305152 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004119458375125376, + "loss": 1.548, + "theoretical_loss": 3.862833622793728, + "tokens_seen": 608370688 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041193580742226684, + "loss": 1.4756, + "theoretical_loss": 3.8627906298358923, + "tokens_seen": 608436224 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041192577733199597, + "loss": 1.476, + "theoretical_loss": 3.862747642805157, + "tokens_seen": 608501760 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004119157472417252, + "loss": 1.4895, + "theoretical_loss": 3.8627046617000653, + "tokens_seen": 608567296 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004119057171514544, + "loss": 1.4914, + "theoretical_loss": 3.862661686519163, + "tokens_seen": 608632832 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041189568706118357, + "loss": 1.5832, + "theoretical_loss": 3.862618717260996, + "tokens_seen": 608698368 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041188565697091275, + "loss": 1.5492, + "theoretical_loss": 3.8625757539241103, + "tokens_seen": 608763904 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041187562688064193, + "loss": 1.4916, + "theoretical_loss": 3.862532796507053, + "tokens_seen": 608829440 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004118655967903711, + "loss": 1.5425, + "theoretical_loss": 3.8624898450083704, + "tokens_seen": 608894976 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041185556670010035, + "loss": 1.4695, + "theoretical_loss": 3.862446899426611, + "tokens_seen": 608960512 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041184553660982947, + "loss": 1.5217, + "theoretical_loss": 3.8624039597603232, + "tokens_seen": 609026048 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004118355065195587, + "loss": 1.5008, + "theoretical_loss": 3.862361026008055, + "tokens_seen": 609091584 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041182547642928783, + "loss": 1.5804, + "theoretical_loss": 3.8623180981683563, + "tokens_seen": 609157120 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041181544633901707, + "loss": 1.4606, + "theoretical_loss": 3.8622751762397765, + "tokens_seen": 609222656 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041180541624874625, + "loss": 1.4551, + "theoretical_loss": 3.862232260220866, + "tokens_seen": 609288192 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041179538615847543, + "loss": 1.5427, + "theoretical_loss": 3.8621893501101763, + "tokens_seen": 609353728 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004117853560682046, + "loss": 1.5027, + "theoretical_loss": 3.8621464459062573, + "tokens_seen": 609419264 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 231726, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3254683017730713, + "objective/train/theoretical_loss": 3.8621035476076617, + "objective/train/tokens_used": 609484800, + "theoretical_loss": 3.8621035476076617, + "tokens_seen": 609484800 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041177532597793385, + "loss": 1.4026, + "theoretical_loss": 3.8621035476076617, + "tokens_seen": 609484800 + }, + { + "epoch": 0.18, + "learning_rate": 0.000411765295887663, + "loss": 1.4927, + "theoretical_loss": 3.862060655212942, + "tokens_seen": 609550336 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004117552657973922, + "loss": 1.4354, + "theoretical_loss": 3.8620177687206514, + "tokens_seen": 609615872 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041174523570712134, + "loss": 1.4802, + "theoretical_loss": 3.861974888129342, + "tokens_seen": 609681408 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004117352056168506, + "loss": 1.6179, + "theoretical_loss": 3.8619320134375688, + "tokens_seen": 609746944 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041172517552657975, + "loss": 1.4807, + "theoretical_loss": 3.861889144643885, + "tokens_seen": 609812480 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041171514543630894, + "loss": 1.5085, + "theoretical_loss": 3.861846281746847, + "tokens_seen": 609878016 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004117051153460381, + "loss": 1.573, + "theoretical_loss": 3.861803424745009, + "tokens_seen": 609943552 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004116950852557673, + "loss": 1.4817, + "theoretical_loss": 3.8617605736369276, + "tokens_seen": 610009088 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004116850551654965, + "loss": 1.4127, + "theoretical_loss": 3.8617177284211595, + "tokens_seen": 610074624 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004116750250752257, + "loss": 1.5195, + "theoretical_loss": 3.8616748890962604, + "tokens_seen": 610140160 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041166499498495484, + "loss": 1.6785, + "theoretical_loss": 3.8616320556607886, + "tokens_seen": 610205696 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004116549648946841, + "loss": 1.4154, + "theoretical_loss": 3.861589228113302, + "tokens_seen": 610271232 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004116449348044132, + "loss": 1.5135, + "theoretical_loss": 3.8615464064523586, + "tokens_seen": 610336768 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041163490471414244, + "loss": 1.5421, + "theoretical_loss": 3.861503590676518, + "tokens_seen": 610402304 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004116248746238716, + "loss": 1.4651, + "theoretical_loss": 3.861460780784339, + "tokens_seen": 610467840 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004116148445336008, + "loss": 1.4739, + "theoretical_loss": 3.8614179767743826, + "tokens_seen": 610533376 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041160481444333, + "loss": 1.5628, + "theoretical_loss": 3.861375178645208, + "tokens_seen": 610598912 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004115947843530592, + "loss": 1.5974, + "theoretical_loss": 3.861332386395377, + "tokens_seen": 610664448 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041158475426278834, + "loss": 1.4539, + "theoretical_loss": 3.8612896000234516, + "tokens_seen": 610729984 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004115747241725176, + "loss": 1.4836, + "theoretical_loss": 3.861246819527992, + "tokens_seen": 610795520 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004115646940822467, + "loss": 1.4622, + "theoretical_loss": 3.861204044907563, + "tokens_seen": 610861056 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041155466399197594, + "loss": 1.4409, + "theoretical_loss": 3.8611612761607255, + "tokens_seen": 610926592 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004115446339017051, + "loss": 1.4452, + "theoretical_loss": 3.861118513286044, + "tokens_seen": 610992128 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004115346038114343, + "loss": 1.5284, + "theoretical_loss": 3.8610757562820828, + "tokens_seen": 611057664 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004115245737211635, + "loss": 1.5725, + "theoretical_loss": 3.861033005147406, + "tokens_seen": 611123200 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041151454363089267, + "loss": 1.3967, + "theoretical_loss": 3.8609902598805785, + "tokens_seen": 611188736 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041150451354062185, + "loss": 1.4686, + "theoretical_loss": 3.860947520480167, + "tokens_seen": 611254272 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004114944834503511, + "loss": 1.4307, + "theoretical_loss": 3.8609047869447357, + "tokens_seen": 611319808 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004114844533600802, + "loss": 1.5499, + "theoretical_loss": 3.8608620592728524, + "tokens_seen": 611385344 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041147442326980944, + "loss": 1.5288, + "theoretical_loss": 3.860819337463084, + "tokens_seen": 611450880 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041146439317953857, + "loss": 1.4788, + "theoretical_loss": 3.860776621513998, + "tokens_seen": 611516416 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004114543630892678, + "loss": 1.5319, + "theoretical_loss": 3.8607339114241617, + "tokens_seen": 611581952 + }, + { + "epoch": 0.19, + "learning_rate": 0.000411444332998997, + "loss": 1.4335, + "theoretical_loss": 3.860691207192145, + "tokens_seen": 611647488 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041143430290872617, + "loss": 1.5872, + "theoretical_loss": 3.860648508816516, + "tokens_seen": 611713024 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041142427281845535, + "loss": 1.4658, + "theoretical_loss": 3.8606058162958443, + "tokens_seen": 611778560 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004114142427281846, + "loss": 1.5506, + "theoretical_loss": 3.8605631296287006, + "tokens_seen": 611844096 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041140421263791377, + "loss": 1.4243, + "theoretical_loss": 3.860520448813655, + "tokens_seen": 611909632 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041139418254764295, + "loss": 1.5462, + "theoretical_loss": 3.8604777738492784, + "tokens_seen": 611975168 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041138415245737213, + "loss": 1.5161, + "theoretical_loss": 3.860435104734143, + "tokens_seen": 612040704 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004113741223671013, + "loss": 1.4715, + "theoretical_loss": 3.86039244146682, + "tokens_seen": 612106240 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041136409227683055, + "loss": 1.4288, + "theoretical_loss": 3.860349784045883, + "tokens_seen": 612171776 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041135406218655967, + "loss": 1.5871, + "theoretical_loss": 3.8603071324699045, + "tokens_seen": 612237312 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004113440320962889, + "loss": 1.3931, + "theoretical_loss": 3.8602644867374583, + "tokens_seen": 612302848 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041133400200601803, + "loss": 1.4394, + "theoretical_loss": 3.860221846847118, + "tokens_seen": 612368384 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041132397191574727, + "loss": 1.4312, + "theoretical_loss": 3.8601792127974583, + "tokens_seen": 612433920 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041131394182547645, + "loss": 1.373, + "theoretical_loss": 3.860136584587055, + "tokens_seen": 612499456 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041130391173520563, + "loss": 1.4697, + "theoretical_loss": 3.860093962214483, + "tokens_seen": 612564992 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004112938816449348, + "loss": 1.4604, + "theoretical_loss": 3.860051345678319, + "tokens_seen": 612630528 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041128385155466405, + "loss": 1.4877, + "theoretical_loss": 3.860008734977139, + "tokens_seen": 612696064 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 232921, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.506666898727417, + "objective/train/theoretical_loss": 3.8599661301095196, + "objective/train/tokens_used": 612761600, + "theoretical_loss": 3.8599661301095196, + "tokens_seen": 612761600 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004112738214643932, + "loss": 1.472, + "theoretical_loss": 3.8599661301095196, + "tokens_seen": 612761600 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004112637913741224, + "loss": 1.4204, + "theoretical_loss": 3.85992353107404, + "tokens_seen": 612827136 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041125376128385154, + "loss": 1.4136, + "theoretical_loss": 3.8598809378692764, + "tokens_seen": 612892672 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004112437311935808, + "loss": 1.4008, + "theoretical_loss": 3.859838350493809, + "tokens_seen": 612958208 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041123370110330995, + "loss": 1.6061, + "theoretical_loss": 3.8597957689462157, + "tokens_seen": 613023744 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041122367101303914, + "loss": 1.5227, + "theoretical_loss": 3.859753193225077, + "tokens_seen": 613089280 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004112136409227683, + "loss": 1.3553, + "theoretical_loss": 3.859710623328972, + "tokens_seen": 613154816 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004112036108324975, + "loss": 1.5657, + "theoretical_loss": 3.8596680592564816, + "tokens_seen": 613220352 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004111935807422267, + "loss": 1.4969, + "theoretical_loss": 3.859625501006188, + "tokens_seen": 613285888 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004111835506519559, + "loss": 1.5124, + "theoretical_loss": 3.859582948576671, + "tokens_seen": 613351424 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041117352056168504, + "loss": 1.4203, + "theoretical_loss": 3.8595404019665143, + "tokens_seen": 613416960 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004111634904714143, + "loss": 1.5657, + "theoretical_loss": 3.859497861174299, + "tokens_seen": 613482496 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004111534603811434, + "loss": 1.5326, + "theoretical_loss": 3.859455326198609, + "tokens_seen": 613548032 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041114343029087264, + "loss": 1.4375, + "theoretical_loss": 3.8594127970380274, + "tokens_seen": 613613568 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004111334002006018, + "loss": 1.4268, + "theoretical_loss": 3.859370273691139, + "tokens_seen": 613679104 + }, + { + "epoch": 0.19, + "learning_rate": 0.000411123370110331, + "loss": 1.4655, + "theoretical_loss": 3.8593277561565276, + "tokens_seen": 613744640 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004111133400200602, + "loss": 1.498, + "theoretical_loss": 3.8592852444327788, + "tokens_seen": 613810176 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004111033099297894, + "loss": 1.5886, + "theoretical_loss": 3.8592427385184775, + "tokens_seen": 613875712 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041109327983951854, + "loss": 1.5456, + "theoretical_loss": 3.85920023841221, + "tokens_seen": 613941248 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004110832497492478, + "loss": 1.5082, + "theoretical_loss": 3.859157744112563, + "tokens_seen": 614006784 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004110732196589769, + "loss": 1.558, + "theoretical_loss": 3.859115255618123, + "tokens_seen": 614072320 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041106318956870614, + "loss": 1.3945, + "theoretical_loss": 3.8590727729274787, + "tokens_seen": 614137856 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004110531594784353, + "loss": 1.4713, + "theoretical_loss": 3.859030296039217, + "tokens_seen": 614203392 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004110431293881645, + "loss": 1.4194, + "theoretical_loss": 3.8589878249519263, + "tokens_seen": 614268928 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004110330992978937, + "loss": 1.567, + "theoretical_loss": 3.8589453596641965, + "tokens_seen": 614334464 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041102306920762287, + "loss": 1.4927, + "theoretical_loss": 3.858902900174617, + "tokens_seen": 614400000 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041101303911735205, + "loss": 1.54, + "theoretical_loss": 3.858860446481777, + "tokens_seen": 614465536 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004110030090270813, + "loss": 1.4383, + "theoretical_loss": 3.8588179985842674, + "tokens_seen": 614531072 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004109929789368104, + "loss": 1.5149, + "theoretical_loss": 3.8587755564806794, + "tokens_seen": 614596608 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041098294884653965, + "loss": 1.5245, + "theoretical_loss": 3.8587331201696045, + "tokens_seen": 614662144 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041097291875626877, + "loss": 1.4419, + "theoretical_loss": 3.8586906896496336, + "tokens_seen": 614727680 + }, + { + "epoch": 0.19, + "learning_rate": 0.000410962888665998, + "loss": 1.5098, + "theoretical_loss": 3.8586482649193607, + "tokens_seen": 614793216 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004109528585757272, + "loss": 1.4721, + "theoretical_loss": 3.8586058459773778, + "tokens_seen": 614858752 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041094282848545637, + "loss": 1.5029, + "theoretical_loss": 3.8585634328222778, + "tokens_seen": 614924288 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041093279839518555, + "loss": 1.413, + "theoretical_loss": 3.8585210254526565, + "tokens_seen": 614989824 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004109227683049148, + "loss": 1.4726, + "theoretical_loss": 3.8584786238671063, + "tokens_seen": 615055360 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004109127382146439, + "loss": 1.4732, + "theoretical_loss": 3.8584362280642233, + "tokens_seen": 615120896 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041090270812437315, + "loss": 1.5304, + "theoretical_loss": 3.858393838042603, + "tokens_seen": 615186432 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004108926780341023, + "loss": 1.5056, + "theoretical_loss": 3.8583514538008403, + "tokens_seen": 615251968 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004108826479438315, + "loss": 1.5035, + "theoretical_loss": 3.858309075337532, + "tokens_seen": 615317504 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004108726178535607, + "loss": 1.5486, + "theoretical_loss": 3.8582667026512754, + "tokens_seen": 615383040 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041086258776328987, + "loss": 1.5159, + "theoretical_loss": 3.858224335740668, + "tokens_seen": 615448576 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041085255767301905, + "loss": 1.3648, + "theoretical_loss": 3.8581819746043067, + "tokens_seen": 615514112 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041084252758274823, + "loss": 1.5361, + "theoretical_loss": 3.8581396192407906, + "tokens_seen": 615579648 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004108324974924774, + "loss": 1.5035, + "theoretical_loss": 3.8580972696487184, + "tokens_seen": 615645184 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041082246740220665, + "loss": 1.4359, + "theoretical_loss": 3.858054925826689, + "tokens_seen": 615710720 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004108124373119358, + "loss": 1.4497, + "theoretical_loss": 3.8580125877733025, + "tokens_seen": 615776256 + }, + { + "epoch": 0.19, + "learning_rate": 0.000410802407221665, + "loss": 1.3818, + "theoretical_loss": 3.8579702554871593, + "tokens_seen": 615841792 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004107923771313942, + "loss": 1.4287, + "theoretical_loss": 3.8579279289668604, + "tokens_seen": 615907328 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004107823470411234, + "loss": 1.551, + "theoretical_loss": 3.8578856082110065, + "tokens_seen": 615972864 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 233987, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3074796199798584, + "objective/train/theoretical_loss": 3.8578432932182, + "objective/train/tokens_used": 616038400, + "theoretical_loss": 3.8578432932182, + "tokens_seen": 616038400 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041077231695085256, + "loss": 1.3496, + "theoretical_loss": 3.8578432932182, + "tokens_seen": 616038400 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041076228686058174, + "loss": 1.5319, + "theoretical_loss": 3.8578009839870426, + "tokens_seen": 616103936 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004107522567703109, + "loss": 1.4565, + "theoretical_loss": 3.8577586805161372, + "tokens_seen": 616169472 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041074222668004015, + "loss": 1.3608, + "theoretical_loss": 3.857716382804087, + "tokens_seen": 616235008 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004107321965897693, + "loss": 1.5311, + "theoretical_loss": 3.8576740908494958, + "tokens_seen": 616300544 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004107221664994985, + "loss": 1.5219, + "theoretical_loss": 3.8576318046509677, + "tokens_seen": 616366080 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041071213640922764, + "loss": 1.5011, + "theoretical_loss": 3.857589524207108, + "tokens_seen": 616431616 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004107021063189569, + "loss": 1.4386, + "theoretical_loss": 3.8575472495165206, + "tokens_seen": 616497152 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041069207622868606, + "loss": 1.3978, + "theoretical_loss": 3.8575049805778123, + "tokens_seen": 616562688 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041068204613841524, + "loss": 1.4565, + "theoretical_loss": 3.8574627173895886, + "tokens_seen": 616628224 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004106720160481444, + "loss": 1.3889, + "theoretical_loss": 3.8574204599504567, + "tokens_seen": 616693760 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004106619859578736, + "loss": 1.5355, + "theoretical_loss": 3.857378208259023, + "tokens_seen": 616759296 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041065195586760284, + "loss": 1.4113, + "theoretical_loss": 3.8573359623138956, + "tokens_seen": 616824832 + }, + { + "epoch": 0.19, + "learning_rate": 0.000410641925777332, + "loss": 1.2923, + "theoretical_loss": 3.857293722113683, + "tokens_seen": 616890368 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004106318956870612, + "loss": 1.413, + "theoretical_loss": 3.8572514876569923, + "tokens_seen": 616955904 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004106218655967904, + "loss": 1.5144, + "theoretical_loss": 3.857209258942434, + "tokens_seen": 617021440 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004106118355065196, + "loss": 1.4694, + "theoretical_loss": 3.857167035968617, + "tokens_seen": 617086976 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041060180541624874, + "loss": 1.4555, + "theoretical_loss": 3.8571248187341522, + "tokens_seen": 617152512 + }, + { + "epoch": 0.19, + "learning_rate": 0.000410591775325978, + "loss": 1.3948, + "theoretical_loss": 3.8570826072376487, + "tokens_seen": 617218048 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004105817452357071, + "loss": 1.5027, + "theoretical_loss": 3.8570404014777178, + "tokens_seen": 617283584 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041057171514543634, + "loss": 1.4631, + "theoretical_loss": 3.8569982014529716, + "tokens_seen": 617349120 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004105616850551655, + "loss": 1.3521, + "theoretical_loss": 3.8569560071620224, + "tokens_seen": 617414656 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004105516549648947, + "loss": 1.4272, + "theoretical_loss": 3.8569138186034815, + "tokens_seen": 617480192 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004105416248746239, + "loss": 1.4925, + "theoretical_loss": 3.8568716357759625, + "tokens_seen": 617545728 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041053159478435307, + "loss": 1.4301, + "theoretical_loss": 3.856829458678079, + "tokens_seen": 617611264 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041052156469408225, + "loss": 1.5398, + "theoretical_loss": 3.856787287308444, + "tokens_seen": 617676800 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004105115346038115, + "loss": 1.4601, + "theoretical_loss": 3.8567451216656723, + "tokens_seen": 617742336 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004105015045135406, + "loss": 1.3819, + "theoretical_loss": 3.8567029617483795, + "tokens_seen": 617807872 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041049147442326985, + "loss": 1.5051, + "theoretical_loss": 3.8566608075551803, + "tokens_seen": 617873408 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041048144433299897, + "loss": 1.4892, + "theoretical_loss": 3.8566186590846905, + "tokens_seen": 617938944 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004104714142427282, + "loss": 1.4536, + "theoretical_loss": 3.856576516335526, + "tokens_seen": 618004480 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004104613841524574, + "loss": 1.4862, + "theoretical_loss": 3.8565343793063045, + "tokens_seen": 618070016 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041045135406218657, + "loss": 1.3756, + "theoretical_loss": 3.856492247995643, + "tokens_seen": 618135552 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041044132397191575, + "loss": 1.4358, + "theoretical_loss": 3.856450122402159, + "tokens_seen": 618201088 + }, + { + "epoch": 0.19, + "learning_rate": 0.000410431293881645, + "loss": 1.4864, + "theoretical_loss": 3.8564080025244705, + "tokens_seen": 618266624 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004104212637913741, + "loss": 1.4668, + "theoretical_loss": 3.8563658883611964, + "tokens_seen": 618332160 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041041123370110335, + "loss": 1.3744, + "theoretical_loss": 3.856323779910956, + "tokens_seen": 618397696 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004104012036108325, + "loss": 1.4153, + "theoretical_loss": 3.856281677172369, + "tokens_seen": 618463232 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004103911735205617, + "loss": 1.4411, + "theoretical_loss": 3.856239580144056, + "tokens_seen": 618528768 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004103811434302909, + "loss": 1.3415, + "theoretical_loss": 3.8561974888246366, + "tokens_seen": 618594304 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041037111334002007, + "loss": 1.5545, + "theoretical_loss": 3.8561554032127323, + "tokens_seen": 618659840 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041036108324974925, + "loss": 1.4406, + "theoretical_loss": 3.856113323306965, + "tokens_seen": 618725376 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041035105315947844, + "loss": 1.4048, + "theoretical_loss": 3.8560712491059563, + "tokens_seen": 618790912 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004103410230692076, + "loss": 1.3042, + "theoretical_loss": 3.8560291806083296, + "tokens_seen": 618856448 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041033099297893685, + "loss": 1.4467, + "theoretical_loss": 3.855987117812707, + "tokens_seen": 618921984 + }, + { + "epoch": 0.19, + "learning_rate": 0.000410320962888666, + "loss": 1.5048, + "theoretical_loss": 3.8559450607177124, + "tokens_seen": 618987520 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004103109327983952, + "loss": 1.5052, + "theoretical_loss": 3.8559030093219695, + "tokens_seen": 619053056 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004103009027081244, + "loss": 1.4833, + "theoretical_loss": 3.855860963624103, + "tokens_seen": 619118592 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004102908726178536, + "loss": 1.5203, + "theoretical_loss": 3.855818923622738, + "tokens_seen": 619184128 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041028084252758276, + "loss": 1.465, + "theoretical_loss": 3.855776889316499, + "tokens_seen": 619249664 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 235166, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5961979627609253, + "objective/train/theoretical_loss": 3.855734860704014, + "objective/train/tokens_used": 619315200, + "theoretical_loss": 3.855734860704014, + "tokens_seen": 619315200 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041027081243731194, + "loss": 1.5119, + "theoretical_loss": 3.855734860704014, + "tokens_seen": 619315200 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004102607823470411, + "loss": 1.5081, + "theoretical_loss": 3.8556928377839066, + "tokens_seen": 619380736 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041025075225677035, + "loss": 1.5186, + "theoretical_loss": 3.8556508205548057, + "tokens_seen": 619446272 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004102407221664995, + "loss": 1.5074, + "theoretical_loss": 3.8556088090153375, + "tokens_seen": 619511808 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004102306920762287, + "loss": 1.4367, + "theoretical_loss": 3.8555668031641304, + "tokens_seen": 619577344 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041022066198595784, + "loss": 1.4652, + "theoretical_loss": 3.8555248029998124, + "tokens_seen": 619642880 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004102106318956871, + "loss": 1.377, + "theoretical_loss": 3.855482808521012, + "tokens_seen": 619708416 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041020060180541626, + "loss": 1.4303, + "theoretical_loss": 3.855440819726359, + "tokens_seen": 619773952 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041019057171514544, + "loss": 1.4252, + "theoretical_loss": 3.8553988366144827, + "tokens_seen": 619839488 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004101805416248746, + "loss": 1.5029, + "theoretical_loss": 3.8553568591840133, + "tokens_seen": 619905024 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004101705115346038, + "loss": 1.4177, + "theoretical_loss": 3.8553148874335816, + "tokens_seen": 619970560 + }, + { + "epoch": 0.19, + "learning_rate": 0.000410160481444333, + "loss": 1.4116, + "theoretical_loss": 3.855272921361819, + "tokens_seen": 620036096 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004101504513540622, + "loss": 1.5267, + "theoretical_loss": 3.855230960967356, + "tokens_seen": 620101632 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041014042126379135, + "loss": 1.4423, + "theoretical_loss": 3.8551890062488265, + "tokens_seen": 620167168 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004101303911735206, + "loss": 1.3925, + "theoretical_loss": 3.855147057204861, + "tokens_seen": 620232704 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041012036108324976, + "loss": 1.3591, + "theoretical_loss": 3.8551051138340937, + "tokens_seen": 620298240 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041011033099297894, + "loss": 1.5528, + "theoretical_loss": 3.8550631761351575, + "tokens_seen": 620363776 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004101003009027081, + "loss": 1.3734, + "theoretical_loss": 3.8550212441066876, + "tokens_seen": 620429312 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004100902708124373, + "loss": 1.399, + "theoretical_loss": 3.8549793177473166, + "tokens_seen": 620494848 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004100802407221665, + "loss": 1.4965, + "theoretical_loss": 3.8549373970556813, + "tokens_seen": 620560384 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004100702106318957, + "loss": 1.4139, + "theoretical_loss": 3.8548954820304155, + "tokens_seen": 620625920 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041006018054162485, + "loss": 1.4062, + "theoretical_loss": 3.854853572670156, + "tokens_seen": 620691456 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004100501504513541, + "loss": 1.5595, + "theoretical_loss": 3.8548116689735386, + "tokens_seen": 620756992 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004100401203610832, + "loss": 1.3795, + "theoretical_loss": 3.8547697709392006, + "tokens_seen": 620822528 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041003009027081245, + "loss": 1.5526, + "theoretical_loss": 3.854727878565779, + "tokens_seen": 620888064 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041002006018054163, + "loss": 1.4224, + "theoretical_loss": 3.8546859918519107, + "tokens_seen": 620953600 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004100100300902708, + "loss": 1.4348, + "theoretical_loss": 3.8546441107962357, + "tokens_seen": 621019136 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041, + "loss": 1.5238, + "theoretical_loss": 3.854602235397391, + "tokens_seen": 621084672 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040998996990972917, + "loss": 1.4929, + "theoretical_loss": 3.8545603656540175, + "tokens_seen": 621150208 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040997993981945835, + "loss": 1.4082, + "theoretical_loss": 3.854518501564753, + "tokens_seen": 621215744 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004099699097291876, + "loss": 1.4778, + "theoretical_loss": 3.8544766431282387, + "tokens_seen": 621281280 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004099598796389167, + "loss": 1.43, + "theoretical_loss": 3.854434790343115, + "tokens_seen": 621346816 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040994984954864595, + "loss": 1.4417, + "theoretical_loss": 3.8543929432080226, + "tokens_seen": 621412352 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040993981945837513, + "loss": 1.4485, + "theoretical_loss": 3.8543511017216034, + "tokens_seen": 621477888 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004099297893681043, + "loss": 1.4459, + "theoretical_loss": 3.8543092658825, + "tokens_seen": 621543424 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004099197592778335, + "loss": 1.4803, + "theoretical_loss": 3.8542674356893527, + "tokens_seen": 621608960 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004099097291875627, + "loss": 1.4502, + "theoretical_loss": 3.8542256111408064, + "tokens_seen": 621674496 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004098996990972919, + "loss": 1.5046, + "theoretical_loss": 3.854183792235504, + "tokens_seen": 621740032 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004098896690070211, + "loss": 1.4261, + "theoretical_loss": 3.8541419789720894, + "tokens_seen": 621805568 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004098796389167503, + "loss": 1.3877, + "theoretical_loss": 3.854100171349206, + "tokens_seen": 621871104 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040986960882647945, + "loss": 1.5503, + "theoretical_loss": 3.8540583693655, + "tokens_seen": 621936640 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040985957873620864, + "loss": 1.4914, + "theoretical_loss": 3.854016573019616, + "tokens_seen": 622002176 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004098495486459378, + "loss": 1.5011, + "theoretical_loss": 3.8539747823101997, + "tokens_seen": 622067712 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040983951855566705, + "loss": 1.3899, + "theoretical_loss": 3.853932997235897, + "tokens_seen": 622133248 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004098294884653962, + "loss": 1.5081, + "theoretical_loss": 3.8538912177953555, + "tokens_seen": 622198784 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004098194583751254, + "loss": 1.4918, + "theoretical_loss": 3.8538494439872215, + "tokens_seen": 622264320 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004098094282848546, + "loss": 1.4663, + "theoretical_loss": 3.8538076758101427, + "tokens_seen": 622329856 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004097993981945838, + "loss": 1.4653, + "theoretical_loss": 3.8537659132627677, + "tokens_seen": 622395392 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040978936810431296, + "loss": 1.5072, + "theoretical_loss": 3.853724156343744, + "tokens_seen": 622460928 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040977933801404214, + "loss": 1.4474, + "theoretical_loss": 3.853682405051722, + "tokens_seen": 622526464 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 236377, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4054521322250366, + "objective/train/theoretical_loss": 3.8536406593853503, + "objective/train/tokens_used": 622592000, + "theoretical_loss": 3.8536406593853503, + "tokens_seen": 622592000 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004097693079237713, + "loss": 1.4734, + "theoretical_loss": 3.8536406593853503, + "tokens_seen": 622592000 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040975927783350055, + "loss": 1.4249, + "theoretical_loss": 3.8535989193432787, + "tokens_seen": 622657536 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004097492477432297, + "loss": 1.4373, + "theoretical_loss": 3.8535571849241577, + "tokens_seen": 622723072 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004097392176529589, + "loss": 1.5025, + "theoretical_loss": 3.8535154561266385, + "tokens_seen": 622788608 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040972918756268804, + "loss": 1.4618, + "theoretical_loss": 3.8534737329493725, + "tokens_seen": 622854144 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004097191574724173, + "loss": 1.3824, + "theoretical_loss": 3.853432015391011, + "tokens_seen": 622919680 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040970912738214646, + "loss": 1.5337, + "theoretical_loss": 3.8533903034502064, + "tokens_seen": 622985216 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040969909729187564, + "loss": 1.5038, + "theoretical_loss": 3.8533485971256125, + "tokens_seen": 623050752 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004096890672016048, + "loss": 1.4449, + "theoretical_loss": 3.8533068964158805, + "tokens_seen": 623116288 + }, + { + "epoch": 0.19, + "learning_rate": 0.000409679037111334, + "loss": 1.5643, + "theoretical_loss": 3.8532652013196653, + "tokens_seen": 623181824 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004096690070210632, + "loss": 1.529, + "theoretical_loss": 3.853223511835621, + "tokens_seen": 623247360 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004096589769307924, + "loss": 1.3799, + "theoretical_loss": 3.853181827962402, + "tokens_seen": 623312896 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040964894684052155, + "loss": 1.4884, + "theoretical_loss": 3.853140149698664, + "tokens_seen": 623378432 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004096389167502508, + "loss": 1.4428, + "theoretical_loss": 3.8530984770430607, + "tokens_seen": 623443968 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040962888665997996, + "loss": 1.3913, + "theoretical_loss": 3.85305680999425, + "tokens_seen": 623509504 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040961885656970914, + "loss": 1.4302, + "theoretical_loss": 3.8530151485508872, + "tokens_seen": 623575040 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004096088264794383, + "loss": 1.4474, + "theoretical_loss": 3.85297349271163, + "tokens_seen": 623640576 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004095987963891675, + "loss": 1.4998, + "theoretical_loss": 3.852931842475135, + "tokens_seen": 623706112 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004095887662988967, + "loss": 1.4123, + "theoretical_loss": 3.8528901978400603, + "tokens_seen": 623771648 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004095787362086259, + "loss": 1.4588, + "theoretical_loss": 3.852848558805065, + "tokens_seen": 623837184 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040956870611835505, + "loss": 1.5625, + "theoretical_loss": 3.8528069253688066, + "tokens_seen": 623902720 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004095586760280843, + "loss": 1.5329, + "theoretical_loss": 3.852765297529945, + "tokens_seen": 623968256 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004095486459378134, + "loss": 1.442, + "theoretical_loss": 3.8527236752871397, + "tokens_seen": 624033792 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040953861584754265, + "loss": 1.5234, + "theoretical_loss": 3.8526820586390507, + "tokens_seen": 624099328 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040952858575727183, + "loss": 1.4263, + "theoretical_loss": 3.8526404475843385, + "tokens_seen": 624164864 + }, + { + "epoch": 0.19, + "learning_rate": 0.000409518555667001, + "loss": 1.4774, + "theoretical_loss": 3.8525988421216653, + "tokens_seen": 624230400 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004095085255767302, + "loss": 1.4295, + "theoretical_loss": 3.852557242249691, + "tokens_seen": 624295936 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040949849548645937, + "loss": 1.4351, + "theoretical_loss": 3.8525156479670786, + "tokens_seen": 624361472 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040948846539618855, + "loss": 1.4466, + "theoretical_loss": 3.8524740592724904, + "tokens_seen": 624427008 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004094784353059178, + "loss": 1.5741, + "theoretical_loss": 3.8524324761645894, + "tokens_seen": 624492544 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004094684052156469, + "loss": 1.4403, + "theoretical_loss": 3.8523908986420383, + "tokens_seen": 624558080 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040945837512537615, + "loss": 1.4955, + "theoretical_loss": 3.8523493267035014, + "tokens_seen": 624623616 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040944834503510533, + "loss": 1.474, + "theoretical_loss": 3.852307760347643, + "tokens_seen": 624689152 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004094383149448345, + "loss": 1.4157, + "theoretical_loss": 3.852266199573128, + "tokens_seen": 624754688 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004094282848545637, + "loss": 1.4945, + "theoretical_loss": 3.852224644378621, + "tokens_seen": 624820224 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004094182547642929, + "loss": 1.5238, + "theoretical_loss": 3.852183094762789, + "tokens_seen": 624885760 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040940822467402206, + "loss": 1.4318, + "theoretical_loss": 3.852141550724297, + "tokens_seen": 624951296 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004093981945837513, + "loss": 1.4943, + "theoretical_loss": 3.852100012261811, + "tokens_seen": 625016832 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004093881644934804, + "loss": 1.4695, + "theoretical_loss": 3.8520584793739996, + "tokens_seen": 625082368 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040937813440320965, + "loss": 1.387, + "theoretical_loss": 3.852016952059529, + "tokens_seen": 625147904 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004093681043129388, + "loss": 1.4652, + "theoretical_loss": 3.8519754303170677, + "tokens_seen": 625213440 + }, + { + "epoch": 0.19, + "learning_rate": 0.000409358074222668, + "loss": 1.6497, + "theoretical_loss": 3.8519339141452846, + "tokens_seen": 625278976 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004093480441323972, + "loss": 1.4476, + "theoretical_loss": 3.8518924035428475, + "tokens_seen": 625344512 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004093380140421264, + "loss": 1.3538, + "theoretical_loss": 3.8518508985084265, + "tokens_seen": 625410048 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040932798395185556, + "loss": 1.4262, + "theoretical_loss": 3.8518093990406914, + "tokens_seen": 625475584 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004093179538615848, + "loss": 1.4269, + "theoretical_loss": 3.8517679051383116, + "tokens_seen": 625541120 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004093079237713139, + "loss": 1.3536, + "theoretical_loss": 3.851726416799959, + "tokens_seen": 625606656 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040929789368104316, + "loss": 1.4275, + "theoretical_loss": 3.8516849340243042, + "tokens_seen": 625672192 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004092878635907723, + "loss": 1.4147, + "theoretical_loss": 3.8516434568100184, + "tokens_seen": 625737728 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004092778335005015, + "loss": 1.4848, + "theoretical_loss": 3.8516019851557743, + "tokens_seen": 625803264 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 237523, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5203742980957031, + "objective/train/theoretical_loss": 3.8515605190602447, + "objective/train/tokens_used": 625868800, + "theoretical_loss": 3.8515605190602447, + "tokens_seen": 625868800 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004092678034102307, + "loss": 1.5103, + "theoretical_loss": 3.8515605190602447, + "tokens_seen": 625868800 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004092577733199599, + "loss": 1.3983, + "theoretical_loss": 3.8515190585221015, + "tokens_seen": 625934336 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040924774322968906, + "loss": 1.4788, + "theoretical_loss": 3.851477603540019, + "tokens_seen": 625999872 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040923771313941824, + "loss": 1.5458, + "theoretical_loss": 3.8514361541126703, + "tokens_seen": 626065408 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004092276830491474, + "loss": 1.3937, + "theoretical_loss": 3.85139471023873, + "tokens_seen": 626130944 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040921765295887666, + "loss": 1.3796, + "theoretical_loss": 3.8513532719168744, + "tokens_seen": 626196480 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004092076228686058, + "loss": 1.4851, + "theoretical_loss": 3.851311839145777, + "tokens_seen": 626262016 + }, + { + "epoch": 0.19, + "learning_rate": 0.000409197592778335, + "loss": 1.5446, + "theoretical_loss": 3.851270411924114, + "tokens_seen": 626327552 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040918756268806415, + "loss": 1.4458, + "theoretical_loss": 3.8512289902505605, + "tokens_seen": 626393088 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004091775325977934, + "loss": 1.4688, + "theoretical_loss": 3.8511875741237955, + "tokens_seen": 626458624 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040916750250752257, + "loss": 1.4487, + "theoretical_loss": 3.851146163542495, + "tokens_seen": 626524160 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040915747241725175, + "loss": 1.3749, + "theoretical_loss": 3.8511047585053353, + "tokens_seen": 626589696 + }, + { + "epoch": 0.19, + "learning_rate": 0.000409147442326981, + "loss": 1.4522, + "theoretical_loss": 3.851063359010996, + "tokens_seen": 626655232 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040913741223671016, + "loss": 1.6046, + "theoretical_loss": 3.851021965058155, + "tokens_seen": 626720768 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040912738214643934, + "loss": 1.5427, + "theoretical_loss": 3.850980576645491, + "tokens_seen": 626786304 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004091173520561685, + "loss": 1.4857, + "theoretical_loss": 3.8509391937716835, + "tokens_seen": 626851840 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004091073219658977, + "loss": 1.3436, + "theoretical_loss": 3.8508978164354124, + "tokens_seen": 626917376 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004090972918756269, + "loss": 1.5356, + "theoretical_loss": 3.8508564446353573, + "tokens_seen": 626982912 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004090872617853561, + "loss": 1.4934, + "theoretical_loss": 3.8508150783702, + "tokens_seen": 627048448 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040907723169508525, + "loss": 1.3576, + "theoretical_loss": 3.850773717638621, + "tokens_seen": 627113984 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004090672016048145, + "loss": 1.4205, + "theoretical_loss": 3.850732362439302, + "tokens_seen": 627179520 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004090571715145436, + "loss": 1.4048, + "theoretical_loss": 3.850691012770925, + "tokens_seen": 627245056 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040904714142427285, + "loss": 1.4951, + "theoretical_loss": 3.8506496686321725, + "tokens_seen": 627310592 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040903711133400203, + "loss": 1.4289, + "theoretical_loss": 3.8506083300217275, + "tokens_seen": 627376128 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004090270812437312, + "loss": 1.5701, + "theoretical_loss": 3.850566996938274, + "tokens_seen": 627441664 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004090170511534604, + "loss": 1.4997, + "theoretical_loss": 3.8505256693804943, + "tokens_seen": 627507200 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040900702106318957, + "loss": 1.4789, + "theoretical_loss": 3.8504843473470745, + "tokens_seen": 627572736 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040899699097291875, + "loss": 1.4495, + "theoretical_loss": 3.850443030836698, + "tokens_seen": 627638272 + }, + { + "epoch": 0.19, + "learning_rate": 0.000408986960882648, + "loss": 1.3469, + "theoretical_loss": 3.850401719848051, + "tokens_seen": 627703808 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004089769307923771, + "loss": 1.4405, + "theoretical_loss": 3.8503604143798187, + "tokens_seen": 627769344 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040896690070210635, + "loss": 1.5274, + "theoretical_loss": 3.850319114430687, + "tokens_seen": 627834880 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040895687061183553, + "loss": 1.4463, + "theoretical_loss": 3.8502778199993433, + "tokens_seen": 627900416 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004089468405215647, + "loss": 1.5044, + "theoretical_loss": 3.850236531084473, + "tokens_seen": 627965952 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004089368104312939, + "loss": 1.4932, + "theoretical_loss": 3.8501952476847654, + "tokens_seen": 628031488 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004089267803410231, + "loss": 1.4684, + "theoretical_loss": 3.8501539697989076, + "tokens_seen": 628097024 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040891675025075226, + "loss": 1.3825, + "theoretical_loss": 3.8501126974255877, + "tokens_seen": 628162560 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004089067201604815, + "loss": 1.4347, + "theoretical_loss": 3.850071430563495, + "tokens_seen": 628228096 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004088966900702106, + "loss": 1.553, + "theoretical_loss": 3.850030169211319, + "tokens_seen": 628293632 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040888665997993985, + "loss": 1.4727, + "theoretical_loss": 3.8499889133677483, + "tokens_seen": 628359168 + }, + { + "epoch": 0.19, + "learning_rate": 0.000408876629889669, + "loss": 1.4826, + "theoretical_loss": 3.8499476630314735, + "tokens_seen": 628424704 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004088665997993982, + "loss": 1.5094, + "theoretical_loss": 3.849906418201186, + "tokens_seen": 628490240 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004088565697091274, + "loss": 1.5303, + "theoretical_loss": 3.849865178875576, + "tokens_seen": 628555776 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004088465396188566, + "loss": 1.4609, + "theoretical_loss": 3.8498239450533354, + "tokens_seen": 628621312 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040883650952858576, + "loss": 1.485, + "theoretical_loss": 3.8497827167331558, + "tokens_seen": 628686848 + }, + { + "epoch": 0.19, + "learning_rate": 0.000408826479438315, + "loss": 1.4937, + "theoretical_loss": 3.84974149391373, + "tokens_seen": 628752384 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004088164493480441, + "loss": 1.506, + "theoretical_loss": 3.849700276593751, + "tokens_seen": 628817920 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040880641925777336, + "loss": 1.4237, + "theoretical_loss": 3.8496590647719113, + "tokens_seen": 628883456 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004087963891675025, + "loss": 1.4461, + "theoretical_loss": 3.849617858446905, + "tokens_seen": 628948992 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004087863590772317, + "loss": 1.4665, + "theoretical_loss": 3.8495766576174266, + "tokens_seen": 629014528 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004087763289869609, + "loss": 1.4107, + "theoretical_loss": 3.8495354622821703, + "tokens_seen": 629080064 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 239008, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.59373140335083, + "objective/train/theoretical_loss": 3.8494942724398316, + "objective/train/tokens_used": 629145600, + "theoretical_loss": 3.8494942724398316, + "tokens_seen": 629145600 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004087662988966901, + "loss": 1.5474, + "theoretical_loss": 3.8494942724398316, + "tokens_seen": 629145600 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040875626880641926, + "loss": 1.4975, + "theoretical_loss": 3.8494530880891054, + "tokens_seen": 629211136 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040874623871614844, + "loss": 1.5348, + "theoretical_loss": 3.8494119092286887, + "tokens_seen": 629276672 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004087362086258776, + "loss": 1.4181, + "theoretical_loss": 3.849370735857277, + "tokens_seen": 629342208 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040872617853560686, + "loss": 1.4787, + "theoretical_loss": 3.849329567973567, + "tokens_seen": 629407744 + }, + { + "epoch": 0.19, + "learning_rate": 0.000408716148445336, + "loss": 1.5644, + "theoretical_loss": 3.8492884055762566, + "tokens_seen": 629473280 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004087061183550652, + "loss": 1.46, + "theoretical_loss": 3.849247248664043, + "tokens_seen": 629538816 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040869608826479435, + "loss": 1.5113, + "theoretical_loss": 3.8492060972356255, + "tokens_seen": 629604352 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004086860581745236, + "loss": 1.5009, + "theoretical_loss": 3.8491649512897017, + "tokens_seen": 629669888 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040867602808425277, + "loss": 1.429, + "theoretical_loss": 3.849123810824971, + "tokens_seen": 629735424 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040866599799398195, + "loss": 1.5112, + "theoretical_loss": 3.8490826758401324, + "tokens_seen": 629800960 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040865596790371113, + "loss": 1.4333, + "theoretical_loss": 3.849041546333887, + "tokens_seen": 629866496 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040864593781344036, + "loss": 1.5208, + "theoretical_loss": 3.8490004223049343, + "tokens_seen": 629932032 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004086359077231695, + "loss": 1.4901, + "theoretical_loss": 3.8489593037519754, + "tokens_seen": 629997568 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004086258776328987, + "loss": 1.5068, + "theoretical_loss": 3.8489181906737118, + "tokens_seen": 630063104 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040861584754262785, + "loss": 1.4122, + "theoretical_loss": 3.848877083068845, + "tokens_seen": 630128640 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004086058174523571, + "loss": 1.4875, + "theoretical_loss": 3.8488359809360766, + "tokens_seen": 630194176 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040859578736208627, + "loss": 1.452, + "theoretical_loss": 3.8487948842741107, + "tokens_seen": 630259712 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040858575727181545, + "loss": 1.5443, + "theoretical_loss": 3.8487537930816487, + "tokens_seen": 630325248 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040857572718154463, + "loss": 1.414, + "theoretical_loss": 3.8487127073573957, + "tokens_seen": 630390784 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004085656970912738, + "loss": 1.4372, + "theoretical_loss": 3.8486716271000545, + "tokens_seen": 630456320 + }, + { + "epoch": 0.19, + "learning_rate": 0.000408555667001003, + "loss": 1.5161, + "theoretical_loss": 3.8486305523083297, + "tokens_seen": 630521856 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040854563691073223, + "loss": 1.4782, + "theoretical_loss": 3.848589482980927, + "tokens_seen": 630587392 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040853560682046136, + "loss": 1.5419, + "theoretical_loss": 3.8485484191165504, + "tokens_seen": 630652928 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004085255767301906, + "loss": 1.4543, + "theoretical_loss": 3.8485073607139064, + "tokens_seen": 630718464 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004085155466399197, + "loss": 1.5181, + "theoretical_loss": 3.8484663077717007, + "tokens_seen": 630784000 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040850551654964895, + "loss": 1.4539, + "theoretical_loss": 3.84842526028864, + "tokens_seen": 630849536 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040849548645937814, + "loss": 1.4592, + "theoretical_loss": 3.8483842182634325, + "tokens_seen": 630915072 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004084854563691073, + "loss": 1.5185, + "theoretical_loss": 3.848343181694784, + "tokens_seen": 630980608 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004084754262788365, + "loss": 1.5619, + "theoretical_loss": 3.8483021505814032, + "tokens_seen": 631046144 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040846539618856573, + "loss": 1.5387, + "theoretical_loss": 3.8482611249219985, + "tokens_seen": 631111680 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040845536609829486, + "loss": 1.5029, + "theoretical_loss": 3.8482201047152786, + "tokens_seen": 631177216 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004084453360080241, + "loss": 1.553, + "theoretical_loss": 3.8481790899599524, + "tokens_seen": 631242752 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004084353059177532, + "loss": 1.3732, + "theoretical_loss": 3.84813808065473, + "tokens_seen": 631308288 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040842527582748246, + "loss": 1.5194, + "theoretical_loss": 3.8480970767983216, + "tokens_seen": 631373824 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040841524573721164, + "loss": 1.5241, + "theoretical_loss": 3.8480560783894373, + "tokens_seen": 631439360 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004084052156469408, + "loss": 1.4714, + "theoretical_loss": 3.8480150854267885, + "tokens_seen": 631504896 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040839518555667005, + "loss": 1.5257, + "theoretical_loss": 3.847974097909087, + "tokens_seen": 631570432 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004083851554663992, + "loss": 1.3438, + "theoretical_loss": 3.8479331158350436, + "tokens_seen": 631635968 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004083751253761284, + "loss": 1.3682, + "theoretical_loss": 3.8478921392033714, + "tokens_seen": 631701504 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004083650952858576, + "loss": 1.5677, + "theoretical_loss": 3.8478511680127827, + "tokens_seen": 631767040 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004083550651955868, + "loss": 1.423, + "theoretical_loss": 3.847810202261991, + "tokens_seen": 631832576 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040834503510531596, + "loss": 1.4907, + "theoretical_loss": 3.8477692419497105, + "tokens_seen": 631898112 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004083350050150452, + "loss": 1.46, + "theoretical_loss": 3.847728287074654, + "tokens_seen": 631963648 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004083249749247743, + "loss": 1.4684, + "theoretical_loss": 3.847687337635537, + "tokens_seen": 632029184 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040831494483450356, + "loss": 1.4954, + "theoretical_loss": 3.8476463936310736, + "tokens_seen": 632094720 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004083049147442327, + "loss": 1.5006, + "theoretical_loss": 3.8476054550599796, + "tokens_seen": 632160256 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004082948846539619, + "loss": 1.5042, + "theoretical_loss": 3.8475645219209715, + "tokens_seen": 632225792 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004082848545636911, + "loss": 1.5858, + "theoretical_loss": 3.847523594212765, + "tokens_seen": 632291328 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004082748244734203, + "loss": 1.5238, + "theoretical_loss": 3.8474826719340762, + "tokens_seen": 632356864 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 240233, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.55375337600708, + "objective/train/theoretical_loss": 3.847441755083623, + "objective/train/tokens_used": 632422400, + "theoretical_loss": 3.847441755083623, + "tokens_seen": 632422400 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040826479438314946, + "loss": 1.4838, + "theoretical_loss": 3.847441755083623, + "tokens_seen": 632422400 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040825476429287864, + "loss": 1.4225, + "theoretical_loss": 3.8474008436601226, + "tokens_seen": 632487936 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004082447342026078, + "loss": 1.5938, + "theoretical_loss": 3.8473599376622927, + "tokens_seen": 632553472 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040823470411233706, + "loss": 1.4514, + "theoretical_loss": 3.8473190370888526, + "tokens_seen": 632619008 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004082246740220662, + "loss": 1.406, + "theoretical_loss": 3.8472781419385207, + "tokens_seen": 632684544 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004082146439317954, + "loss": 1.5145, + "theoretical_loss": 3.8472372522100162, + "tokens_seen": 632750080 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040820461384152455, + "loss": 1.4881, + "theoretical_loss": 3.8471963679020593, + "tokens_seen": 632815616 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004081945837512538, + "loss": 1.4837, + "theoretical_loss": 3.8471554890133692, + "tokens_seen": 632881152 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040818455366098297, + "loss": 1.5274, + "theoretical_loss": 3.847114615542667, + "tokens_seen": 632946688 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040817452357071215, + "loss": 1.5706, + "theoretical_loss": 3.8470737474886745, + "tokens_seen": 633012224 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040816449348044133, + "loss": 1.5385, + "theoretical_loss": 3.847032884850112, + "tokens_seen": 633077760 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040815446339017056, + "loss": 1.4834, + "theoretical_loss": 3.846992027625702, + "tokens_seen": 633143296 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004081444332998997, + "loss": 1.5643, + "theoretical_loss": 3.8469511758141666, + "tokens_seen": 633208832 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004081344032096289, + "loss": 1.5013, + "theoretical_loss": 3.8469103294142286, + "tokens_seen": 633274368 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040812437311935805, + "loss": 1.4697, + "theoretical_loss": 3.8468694884246117, + "tokens_seen": 633339904 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004081143430290873, + "loss": 1.4758, + "theoretical_loss": 3.8468286528440396, + "tokens_seen": 633405440 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040810431293881647, + "loss": 1.4622, + "theoretical_loss": 3.846787822671235, + "tokens_seen": 633470976 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040809428284854565, + "loss": 1.3689, + "theoretical_loss": 3.8467469979049245, + "tokens_seen": 633536512 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040808425275827483, + "loss": 1.3322, + "theoretical_loss": 3.8467061785438315, + "tokens_seen": 633602048 + }, + { + "epoch": 0.19, + "learning_rate": 0.000408074222668004, + "loss": 1.5531, + "theoretical_loss": 3.8466653645866815, + "tokens_seen": 633667584 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004080641925777332, + "loss": 1.4611, + "theoretical_loss": 3.8466245560322005, + "tokens_seen": 633733120 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040805416248746243, + "loss": 1.5641, + "theoretical_loss": 3.8465837528791154, + "tokens_seen": 633798656 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040804413239719156, + "loss": 1.5417, + "theoretical_loss": 3.846542955126152, + "tokens_seen": 633864192 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004080341023069208, + "loss": 1.4123, + "theoretical_loss": 3.8465021627720377, + "tokens_seen": 633929728 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004080240722166499, + "loss": 1.3928, + "theoretical_loss": 3.8464613758155, + "tokens_seen": 633995264 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040801404212637915, + "loss": 1.4234, + "theoretical_loss": 3.8464205942552665, + "tokens_seen": 634060800 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040800401203610834, + "loss": 1.4487, + "theoretical_loss": 3.8463798180900666, + "tokens_seen": 634126336 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004079939819458375, + "loss": 1.4627, + "theoretical_loss": 3.8463390473186285, + "tokens_seen": 634191872 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004079839518555667, + "loss": 1.5134, + "theoretical_loss": 3.8462982819396814, + "tokens_seen": 634257408 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040797392176529593, + "loss": 1.5516, + "theoretical_loss": 3.846257521951955, + "tokens_seen": 634322944 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040796389167502506, + "loss": 1.4213, + "theoretical_loss": 3.84621676735418, + "tokens_seen": 634388480 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004079538615847543, + "loss": 1.5221, + "theoretical_loss": 3.8461760181450857, + "tokens_seen": 634454016 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004079438314944834, + "loss": 1.3679, + "theoretical_loss": 3.8461352743234043, + "tokens_seen": 634519552 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040793380140421266, + "loss": 1.4603, + "theoretical_loss": 3.8460945358878664, + "tokens_seen": 634585088 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040792377131394184, + "loss": 1.4604, + "theoretical_loss": 3.846053802837204, + "tokens_seen": 634650624 + }, + { + "epoch": 0.19, + "learning_rate": 0.000407913741223671, + "loss": 1.4958, + "theoretical_loss": 3.8460130751701502, + "tokens_seen": 634716160 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004079037111334002, + "loss": 1.4071, + "theoretical_loss": 3.845972352885437, + "tokens_seen": 634781696 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004078936810431294, + "loss": 1.5076, + "theoretical_loss": 3.845931635981797, + "tokens_seen": 634847232 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040788365095285856, + "loss": 1.4689, + "theoretical_loss": 3.845890924457965, + "tokens_seen": 634912768 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004078736208625878, + "loss": 1.4663, + "theoretical_loss": 3.845850218312674, + "tokens_seen": 634978304 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004078635907723169, + "loss": 1.4238, + "theoretical_loss": 3.8458095175446587, + "tokens_seen": 635043840 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040785356068204616, + "loss": 1.4262, + "theoretical_loss": 3.8457688221526545, + "tokens_seen": 635109376 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004078435305917753, + "loss": 1.5009, + "theoretical_loss": 3.8457281321353953, + "tokens_seen": 635174912 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004078335005015045, + "loss": 1.4975, + "theoretical_loss": 3.8456874474916183, + "tokens_seen": 635240448 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004078234704112337, + "loss": 1.4486, + "theoretical_loss": 3.8456467682200586, + "tokens_seen": 635305984 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004078134403209629, + "loss": 1.4478, + "theoretical_loss": 3.8456060943194537, + "tokens_seen": 635371520 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040780341023069207, + "loss": 1.5208, + "theoretical_loss": 3.845565425788539, + "tokens_seen": 635437056 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004077933801404213, + "loss": 1.5336, + "theoretical_loss": 3.8455247626260536, + "tokens_seen": 635502592 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040778335005015043, + "loss": 1.4195, + "theoretical_loss": 3.845484104830734, + "tokens_seen": 635568128 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040777331995987966, + "loss": 1.6264, + "theoretical_loss": 3.84544345240132, + "tokens_seen": 635633664 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 241399, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3744878768920898, + "objective/train/theoretical_loss": 3.8454028053365494, + "objective/train/tokens_used": 635699200, + "theoretical_loss": 3.8454028053365494, + "tokens_seen": 635699200 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004077632898696088, + "loss": 1.5181, + "theoretical_loss": 3.8454028053365494, + "tokens_seen": 635699200 + }, + { + "epoch": 0.19, + "learning_rate": 0.000407753259779338, + "loss": 1.552, + "theoretical_loss": 3.8453621636351607, + "tokens_seen": 635764736 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004077432296890672, + "loss": 1.4303, + "theoretical_loss": 3.8453215272958943, + "tokens_seen": 635830272 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004077331995987964, + "loss": 1.5279, + "theoretical_loss": 3.84528089631749, + "tokens_seen": 635895808 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040772316950852557, + "loss": 1.4305, + "theoretical_loss": 3.845240270698688, + "tokens_seen": 635961344 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040771313941825475, + "loss": 1.4737, + "theoretical_loss": 3.845199650438229, + "tokens_seen": 636026880 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040770310932798393, + "loss": 1.6246, + "theoretical_loss": 3.845159035534855, + "tokens_seen": 636092416 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040769307923771317, + "loss": 1.4365, + "theoretical_loss": 3.845118425987307, + "tokens_seen": 636157952 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004076830491474423, + "loss": 1.5043, + "theoretical_loss": 3.845077821794327, + "tokens_seen": 636223488 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040767301905717153, + "loss": 1.4489, + "theoretical_loss": 3.845037222954658, + "tokens_seen": 636289024 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004076629889669007, + "loss": 1.462, + "theoretical_loss": 3.8449966294670426, + "tokens_seen": 636354560 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004076529588766299, + "loss": 1.4371, + "theoretical_loss": 3.8449560413302244, + "tokens_seen": 636420096 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004076429287863591, + "loss": 1.3754, + "theoretical_loss": 3.844915458542947, + "tokens_seen": 636485632 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040763289869608825, + "loss": 1.5655, + "theoretical_loss": 3.844874881103955, + "tokens_seen": 636551168 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004076228686058175, + "loss": 1.5179, + "theoretical_loss": 3.844834309011992, + "tokens_seen": 636616704 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040761283851554667, + "loss": 1.4941, + "theoretical_loss": 3.8447937422658045, + "tokens_seen": 636682240 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040760280842527585, + "loss": 1.3824, + "theoretical_loss": 3.8447531808641373, + "tokens_seen": 636747776 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040759277833500503, + "loss": 1.6265, + "theoretical_loss": 3.844712624805736, + "tokens_seen": 636813312 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004075827482447342, + "loss": 1.4057, + "theoretical_loss": 3.844672074089347, + "tokens_seen": 636878848 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004075727181544634, + "loss": 1.4601, + "theoretical_loss": 3.844631528713718, + "tokens_seen": 636944384 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040756268806419263, + "loss": 1.4174, + "theoretical_loss": 3.844590988677595, + "tokens_seen": 637009920 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040755265797392176, + "loss": 1.496, + "theoretical_loss": 3.844550453979726, + "tokens_seen": 637075456 + }, + { + "epoch": 0.19, + "learning_rate": 0.000407542627883651, + "loss": 1.5457, + "theoretical_loss": 3.8445099246188597, + "tokens_seen": 637140992 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004075325977933801, + "loss": 1.4698, + "theoretical_loss": 3.8444694005937436, + "tokens_seen": 637206528 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040752256770310935, + "loss": 1.5454, + "theoretical_loss": 3.844428881903127, + "tokens_seen": 637272064 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040751253761283854, + "loss": 1.5112, + "theoretical_loss": 3.844388368545759, + "tokens_seen": 637337600 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004075025075225677, + "loss": 1.5157, + "theoretical_loss": 3.84434786052039, + "tokens_seen": 637403136 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004074924774322969, + "loss": 1.5016, + "theoretical_loss": 3.8443073578257687, + "tokens_seen": 637468672 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040748244734202613, + "loss": 1.5854, + "theoretical_loss": 3.8442668604606474, + "tokens_seen": 637534208 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040747241725175526, + "loss": 1.4955, + "theoretical_loss": 3.844226368423776, + "tokens_seen": 637599744 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004074623871614845, + "loss": 1.5026, + "theoretical_loss": 3.844185881713906, + "tokens_seen": 637665280 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004074523570712136, + "loss": 1.5405, + "theoretical_loss": 3.844145400329789, + "tokens_seen": 637730816 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040744232698094286, + "loss": 1.4783, + "theoretical_loss": 3.8441049242701784, + "tokens_seen": 637796352 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040743229689067204, + "loss": 1.5723, + "theoretical_loss": 3.8440644535338255, + "tokens_seen": 637861888 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004074222668004012, + "loss": 1.4987, + "theoretical_loss": 3.844023988119484, + "tokens_seen": 637927424 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004074122367101304, + "loss": 1.4809, + "theoretical_loss": 3.8439835280259076, + "tokens_seen": 637992960 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004074022066198596, + "loss": 1.4853, + "theoretical_loss": 3.84394307325185, + "tokens_seen": 638058496 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040739217652958876, + "loss": 1.534, + "theoretical_loss": 3.8439026237960654, + "tokens_seen": 638124032 + }, + { + "epoch": 0.19, + "learning_rate": 0.000407382146439318, + "loss": 1.4812, + "theoretical_loss": 3.8438621796573087, + "tokens_seen": 638189568 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004073721163490471, + "loss": 1.5317, + "theoretical_loss": 3.843821740834335, + "tokens_seen": 638255104 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040736208625877636, + "loss": 1.4243, + "theoretical_loss": 3.8437813073258997, + "tokens_seen": 638320640 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004073520561685055, + "loss": 1.5213, + "theoretical_loss": 3.8437408791307597, + "tokens_seen": 638386176 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004073420260782347, + "loss": 1.5128, + "theoretical_loss": 3.843700456247671, + "tokens_seen": 638451712 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004073319959879639, + "loss": 1.4526, + "theoretical_loss": 3.84366003867539, + "tokens_seen": 638517248 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004073219658976931, + "loss": 1.4432, + "theoretical_loss": 3.8436196264126736, + "tokens_seen": 638582784 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040731193580742227, + "loss": 1.4571, + "theoretical_loss": 3.843579219458281, + "tokens_seen": 638648320 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004073019057171515, + "loss": 1.4845, + "theoretical_loss": 3.843538817810969, + "tokens_seen": 638713856 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040729187562688063, + "loss": 1.4259, + "theoretical_loss": 3.843498421469497, + "tokens_seen": 638779392 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040728184553660986, + "loss": 1.412, + "theoretical_loss": 3.843458030432623, + "tokens_seen": 638844928 + }, + { + "epoch": 0.19, + "learning_rate": 0.000407271815446339, + "loss": 1.5095, + "theoretical_loss": 3.8434176446991075, + "tokens_seen": 638910464 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 242533, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4696460962295532, + "objective/train/theoretical_loss": 3.8433772642677093, + "objective/train/tokens_used": 638976000, + "theoretical_loss": 3.8433772642677093, + "tokens_seen": 638976000 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004072617853560682, + "loss": 1.4438, + "theoretical_loss": 3.8433772642677093, + "tokens_seen": 638976000 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004072517552657974, + "loss": 1.5749, + "theoretical_loss": 3.8433368891371895, + "tokens_seen": 639041536 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004072417251755266, + "loss": 1.5434, + "theoretical_loss": 3.843296519306308, + "tokens_seen": 639107072 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040723169508525577, + "loss": 1.4582, + "theoretical_loss": 3.843256154773826, + "tokens_seen": 639172608 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040722166499498495, + "loss": 1.4411, + "theoretical_loss": 3.843215795538505, + "tokens_seen": 639238144 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040721163490471413, + "loss": 1.4801, + "theoretical_loss": 3.843175441599107, + "tokens_seen": 639303680 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040720160481444337, + "loss": 1.5339, + "theoretical_loss": 3.8431350929543946, + "tokens_seen": 639369216 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004071915747241725, + "loss": 1.4482, + "theoretical_loss": 3.8430947496031296, + "tokens_seen": 639434752 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040718154463390173, + "loss": 1.3983, + "theoretical_loss": 3.8430544115440752, + "tokens_seen": 639500288 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004071715145436309, + "loss": 1.3787, + "theoretical_loss": 3.8430140787759957, + "tokens_seen": 639565824 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004071614844533601, + "loss": 1.3852, + "theoretical_loss": 3.842973751297655, + "tokens_seen": 639631360 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040715145436308927, + "loss": 1.4919, + "theoretical_loss": 3.8429334291078168, + "tokens_seen": 639696896 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040714142427281845, + "loss": 1.4033, + "theoretical_loss": 3.842893112205246, + "tokens_seen": 639762432 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040713139418254763, + "loss": 1.4148, + "theoretical_loss": 3.842852800588709, + "tokens_seen": 639827968 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040712136409227687, + "loss": 1.4544, + "theoretical_loss": 3.8428124942569695, + "tokens_seen": 639893504 + }, + { + "epoch": 0.19, + "learning_rate": 0.000407111334002006, + "loss": 1.4255, + "theoretical_loss": 3.842772193208795, + "tokens_seen": 639959040 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040710130391173523, + "loss": 1.4125, + "theoretical_loss": 3.8427318974429507, + "tokens_seen": 640024576 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040709127382146436, + "loss": 1.5097, + "theoretical_loss": 3.8426916069582044, + "tokens_seen": 640090112 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004070812437311936, + "loss": 1.4204, + "theoretical_loss": 3.842651321753323, + "tokens_seen": 640155648 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004070712136409228, + "loss": 1.6027, + "theoretical_loss": 3.8426110418270745, + "tokens_seen": 640221184 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040706118355065196, + "loss": 1.4345, + "theoretical_loss": 3.842570767178227, + "tokens_seen": 640286720 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040705115346038114, + "loss": 1.4244, + "theoretical_loss": 3.8425304978055483, + "tokens_seen": 640352256 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004070411233701103, + "loss": 1.6019, + "theoretical_loss": 3.8424902337078084, + "tokens_seen": 640417792 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004070310932798395, + "loss": 1.3977, + "theoretical_loss": 3.8424499748837753, + "tokens_seen": 640483328 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040702106318956874, + "loss": 1.4847, + "theoretical_loss": 3.8424097213322197, + "tokens_seen": 640548864 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040701103309929786, + "loss": 1.3936, + "theoretical_loss": 3.8423694730519116, + "tokens_seen": 640614400 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004070010030090271, + "loss": 1.4475, + "theoretical_loss": 3.8423292300416216, + "tokens_seen": 640679936 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004069909729187563, + "loss": 1.4144, + "theoretical_loss": 3.8422889923001207, + "tokens_seen": 640745472 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040698094282848546, + "loss": 1.5375, + "theoretical_loss": 3.84224875982618, + "tokens_seen": 640811008 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040697091273821464, + "loss": 1.3975, + "theoretical_loss": 3.8422085326185718, + "tokens_seen": 640876544 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004069608826479438, + "loss": 1.4939, + "theoretical_loss": 3.842168310676067, + "tokens_seen": 640942080 + }, + { + "epoch": 0.19, + "learning_rate": 0.000406950852557673, + "loss": 1.5089, + "theoretical_loss": 3.8421280939974407, + "tokens_seen": 641007616 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040694082246740224, + "loss": 1.495, + "theoretical_loss": 3.8420878825814637, + "tokens_seen": 641073152 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040693079237713137, + "loss": 1.4601, + "theoretical_loss": 3.84204767642691, + "tokens_seen": 641138688 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004069207622868606, + "loss": 1.5297, + "theoretical_loss": 3.8420074755325544, + "tokens_seen": 641204224 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040691073219658973, + "loss": 1.4352, + "theoretical_loss": 3.8419672798971707, + "tokens_seen": 641269760 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040690070210631896, + "loss": 1.4493, + "theoretical_loss": 3.8419270895195323, + "tokens_seen": 641335296 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004068906720160482, + "loss": 1.4524, + "theoretical_loss": 3.8418869043984163, + "tokens_seen": 641400832 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004068806419257773, + "loss": 1.4702, + "theoretical_loss": 3.841846724532597, + "tokens_seen": 641466368 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040687061183550656, + "loss": 1.4857, + "theoretical_loss": 3.8418065499208502, + "tokens_seen": 641531904 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004068605817452357, + "loss": 1.5205, + "theoretical_loss": 3.8417663805619533, + "tokens_seen": 641597440 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004068505516549649, + "loss": 1.4163, + "theoretical_loss": 3.841726216454682, + "tokens_seen": 641662976 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004068405215646941, + "loss": 1.4316, + "theoretical_loss": 3.8416860575978142, + "tokens_seen": 641728512 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004068304914744233, + "loss": 1.515, + "theoretical_loss": 3.841645903990127, + "tokens_seen": 641794048 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040682046138415247, + "loss": 1.3826, + "theoretical_loss": 3.841605755630398, + "tokens_seen": 641859584 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004068104312938817, + "loss": 1.4388, + "theoretical_loss": 3.8415656125174067, + "tokens_seen": 641925120 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040680040120361083, + "loss": 1.5548, + "theoretical_loss": 3.8415254746499308, + "tokens_seen": 641990656 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040679037111334006, + "loss": 1.4562, + "theoretical_loss": 3.8414853420267496, + "tokens_seen": 642056192 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004067803410230692, + "loss": 1.5796, + "theoretical_loss": 3.841445214646644, + "tokens_seen": 642121728 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004067703109327984, + "loss": 1.5255, + "theoretical_loss": 3.8414050925083925, + "tokens_seen": 642187264 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 243591, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4463163614273071, + "objective/train/theoretical_loss": 3.8413649756107757, + "objective/train/tokens_used": 642252800, + "theoretical_loss": 3.8413649756107757, + "tokens_seen": 642252800 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004067602808425276, + "loss": 1.4828, + "theoretical_loss": 3.8413649756107757, + "tokens_seen": 642252800 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004067502507522568, + "loss": 1.4681, + "theoretical_loss": 3.8413248639525754, + "tokens_seen": 642318336 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040674022066198597, + "loss": 1.3534, + "theoretical_loss": 3.8412847575325717, + "tokens_seen": 642383872 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040673019057171515, + "loss": 1.5101, + "theoretical_loss": 3.8412446563495477, + "tokens_seen": 642449408 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040672016048144433, + "loss": 1.3954, + "theoretical_loss": 3.8412045604022835, + "tokens_seen": 642514944 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040671013039117357, + "loss": 1.4459, + "theoretical_loss": 3.8411644696895633, + "tokens_seen": 642580480 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004067001003009027, + "loss": 1.5561, + "theoretical_loss": 3.8411243842101688, + "tokens_seen": 642646016 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040669007021063193, + "loss": 1.5331, + "theoretical_loss": 3.841084303962884, + "tokens_seen": 642711552 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004066800401203611, + "loss": 1.5217, + "theoretical_loss": 3.841044228946492, + "tokens_seen": 642777088 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004066700100300903, + "loss": 1.4643, + "theoretical_loss": 3.8410041591597777, + "tokens_seen": 642842624 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040665997993981947, + "loss": 1.3853, + "theoretical_loss": 3.840964094601525, + "tokens_seen": 642908160 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040664994984954865, + "loss": 1.4052, + "theoretical_loss": 3.8409240352705187, + "tokens_seen": 642973696 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040663991975927783, + "loss": 1.4418, + "theoretical_loss": 3.840883981165544, + "tokens_seen": 643039232 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040662988966900707, + "loss": 1.5754, + "theoretical_loss": 3.8408439322853876, + "tokens_seen": 643104768 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004066198595787362, + "loss": 1.3814, + "theoretical_loss": 3.8408038886288347, + "tokens_seen": 643170304 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040660982948846543, + "loss": 1.4879, + "theoretical_loss": 3.840763850194672, + "tokens_seen": 643235840 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040659979939819456, + "loss": 1.3695, + "theoretical_loss": 3.8407238169816864, + "tokens_seen": 643301376 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004065897693079238, + "loss": 1.465, + "theoretical_loss": 3.840683788988666, + "tokens_seen": 643366912 + }, + { + "epoch": 0.19, + "learning_rate": 0.000406579739217653, + "loss": 1.4517, + "theoretical_loss": 3.840643766214397, + "tokens_seen": 643432448 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040656970912738216, + "loss": 1.4153, + "theoretical_loss": 3.840603748657669, + "tokens_seen": 643497984 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040655967903711134, + "loss": 1.4654, + "theoretical_loss": 3.8405637363172693, + "tokens_seen": 643563520 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004065496489468405, + "loss": 1.4906, + "theoretical_loss": 3.8405237291919883, + "tokens_seen": 643629056 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004065396188565697, + "loss": 1.4414, + "theoretical_loss": 3.8404837272806143, + "tokens_seen": 643694592 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040652958876629894, + "loss": 1.4034, + "theoretical_loss": 3.840443730581937, + "tokens_seen": 643760128 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040651955867602806, + "loss": 1.5925, + "theoretical_loss": 3.8404037390947474, + "tokens_seen": 643825664 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004065095285857573, + "loss": 1.3593, + "theoretical_loss": 3.8403637528178356, + "tokens_seen": 643891200 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004064994984954865, + "loss": 1.3928, + "theoretical_loss": 3.8403237717499925, + "tokens_seen": 643956736 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040648946840521566, + "loss": 1.5258, + "theoretical_loss": 3.8402837958900093, + "tokens_seen": 644022272 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040647943831494484, + "loss": 1.5661, + "theoretical_loss": 3.840243825236678, + "tokens_seen": 644087808 + }, + { + "epoch": 0.2, + "learning_rate": 0.000406469408224674, + "loss": 1.4625, + "theoretical_loss": 3.840203859788791, + "tokens_seen": 644153344 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004064593781344032, + "loss": 1.4396, + "theoretical_loss": 3.840163899545141, + "tokens_seen": 644218880 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040644934804413244, + "loss": 1.4006, + "theoretical_loss": 3.8401239445045205, + "tokens_seen": 644284416 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040643931795386157, + "loss": 1.4113, + "theoretical_loss": 3.840083994665723, + "tokens_seen": 644349952 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004064292878635908, + "loss": 1.3794, + "theoretical_loss": 3.840044050027543, + "tokens_seen": 644415488 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040641925777331993, + "loss": 1.4938, + "theoretical_loss": 3.8400041105887732, + "tokens_seen": 644481024 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040640922768304916, + "loss": 1.4107, + "theoretical_loss": 3.8399641763482095, + "tokens_seen": 644546560 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040639919759277834, + "loss": 1.4803, + "theoretical_loss": 3.8399242473046464, + "tokens_seen": 644612096 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004063891675025075, + "loss": 1.4238, + "theoretical_loss": 3.8398843234568796, + "tokens_seen": 644677632 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004063791374122367, + "loss": 1.5052, + "theoretical_loss": 3.8398444048037046, + "tokens_seen": 644743168 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004063691073219659, + "loss": 1.548, + "theoretical_loss": 3.8398044913439175, + "tokens_seen": 644808704 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040635907723169507, + "loss": 1.4933, + "theoretical_loss": 3.839764583076316, + "tokens_seen": 644874240 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004063490471414243, + "loss": 1.4742, + "theoretical_loss": 3.839724679999695, + "tokens_seen": 644939776 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040633901705115343, + "loss": 1.5027, + "theoretical_loss": 3.8396847821128537, + "tokens_seen": 645005312 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040632898696088267, + "loss": 1.5246, + "theoretical_loss": 3.83964488941459, + "tokens_seen": 645070848 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040631895687061185, + "loss": 1.3878, + "theoretical_loss": 3.8396050019037005, + "tokens_seen": 645136384 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040630892678034103, + "loss": 1.4297, + "theoretical_loss": 3.8395651195789853, + "tokens_seen": 645201920 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004062988966900702, + "loss": 1.4457, + "theoretical_loss": 3.839525242439243, + "tokens_seen": 645267456 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004062888665997994, + "loss": 1.3897, + "theoretical_loss": 3.8394853704832723, + "tokens_seen": 645332992 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040627883650952857, + "loss": 1.4264, + "theoretical_loss": 3.8394455037098743, + "tokens_seen": 645398528 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004062688064192578, + "loss": 1.4613, + "theoretical_loss": 3.8394056421178484, + "tokens_seen": 645464064 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 244772, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3522366285324097, + "objective/train/theoretical_loss": 3.8393657857059953, + "objective/train/tokens_used": 645529600, + "theoretical_loss": 3.8393657857059953, + "tokens_seen": 645529600 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040625877632898693, + "loss": 1.407, + "theoretical_loss": 3.8393657857059953, + "tokens_seen": 645529600 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040624874623871617, + "loss": 1.4675, + "theoretical_loss": 3.839325934473116, + "tokens_seen": 645595136 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004062387161484453, + "loss": 1.4728, + "theoretical_loss": 3.839286088418012, + "tokens_seen": 645660672 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040622868605817453, + "loss": 1.4079, + "theoretical_loss": 3.839246247539485, + "tokens_seen": 645726208 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004062186559679037, + "loss": 1.4519, + "theoretical_loss": 3.8392064118363374, + "tokens_seen": 645791744 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004062086258776329, + "loss": 1.417, + "theoretical_loss": 3.839166581307372, + "tokens_seen": 645857280 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004061985957873621, + "loss": 1.4399, + "theoretical_loss": 3.8391267559513915, + "tokens_seen": 645922816 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040618856569709126, + "loss": 1.4991, + "theoretical_loss": 3.839086935767199, + "tokens_seen": 645988352 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040617853560682044, + "loss": 1.4537, + "theoretical_loss": 3.839047120753599, + "tokens_seen": 646053888 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040616850551654967, + "loss": 1.3691, + "theoretical_loss": 3.8390073109093947, + "tokens_seen": 646119424 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004061584754262788, + "loss": 1.4589, + "theoretical_loss": 3.8389675062333923, + "tokens_seen": 646184960 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040614844533600804, + "loss": 1.518, + "theoretical_loss": 3.838927706724395, + "tokens_seen": 646250496 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040613841524573727, + "loss": 1.5547, + "theoretical_loss": 3.8388879123812094, + "tokens_seen": 646316032 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004061283851554664, + "loss": 1.523, + "theoretical_loss": 3.8388481232026406, + "tokens_seen": 646381568 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040611835506519563, + "loss": 1.4274, + "theoretical_loss": 3.838808339187495, + "tokens_seen": 646447104 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040610832497492476, + "loss": 1.478, + "theoretical_loss": 3.8387685603345805, + "tokens_seen": 646512640 + }, + { + "epoch": 0.2, + "learning_rate": 0.000406098294884654, + "loss": 1.4865, + "theoretical_loss": 3.838728786642702, + "tokens_seen": 646578176 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004060882647943832, + "loss": 1.4388, + "theoretical_loss": 3.838689018110667, + "tokens_seen": 646643712 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040607823470411236, + "loss": 1.449, + "theoretical_loss": 3.838649254737285, + "tokens_seen": 646709248 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040606820461384154, + "loss": 1.4147, + "theoretical_loss": 3.838609496521363, + "tokens_seen": 646774784 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004060581745235707, + "loss": 1.3566, + "theoretical_loss": 3.8385697434617096, + "tokens_seen": 646840320 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004060481444332999, + "loss": 1.3783, + "theoretical_loss": 3.838529995557134, + "tokens_seen": 646905856 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040603811434302914, + "loss": 1.4342, + "theoretical_loss": 3.8384902528064457, + "tokens_seen": 646971392 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040602808425275826, + "loss": 1.5477, + "theoretical_loss": 3.8384505152084545, + "tokens_seen": 647036928 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004060180541624875, + "loss": 1.5328, + "theoretical_loss": 3.8384107827619696, + "tokens_seen": 647102464 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004060080240722167, + "loss": 1.5211, + "theoretical_loss": 3.8383710554658026, + "tokens_seen": 647168000 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040599799398194586, + "loss": 1.4276, + "theoretical_loss": 3.838331333318764, + "tokens_seen": 647233536 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040598796389167504, + "loss": 1.5599, + "theoretical_loss": 3.8382916163196654, + "tokens_seen": 647299072 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004059779338014042, + "loss": 1.5079, + "theoretical_loss": 3.8382519044673176, + "tokens_seen": 647364608 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004059679037111334, + "loss": 1.3873, + "theoretical_loss": 3.8382121977605346, + "tokens_seen": 647430144 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040595787362086264, + "loss": 1.5602, + "theoretical_loss": 3.8381724961981267, + "tokens_seen": 647495680 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040594784353059177, + "loss": 1.539, + "theoretical_loss": 3.8381327997789088, + "tokens_seen": 647561216 + }, + { + "epoch": 0.2, + "learning_rate": 0.000405937813440321, + "loss": 1.4421, + "theoretical_loss": 3.8380931085016927, + "tokens_seen": 647626752 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040592778335005013, + "loss": 1.3795, + "theoretical_loss": 3.8380534223652933, + "tokens_seen": 647692288 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040591775325977936, + "loss": 1.462, + "theoretical_loss": 3.838013741368523, + "tokens_seen": 647757824 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040590772316950854, + "loss": 1.4608, + "theoretical_loss": 3.837974065510198, + "tokens_seen": 647823360 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004058976930792377, + "loss": 1.4416, + "theoretical_loss": 3.837934394789133, + "tokens_seen": 647888896 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004058876629889669, + "loss": 1.512, + "theoretical_loss": 3.8378947292041428, + "tokens_seen": 647954432 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004058776328986961, + "loss": 1.5783, + "theoretical_loss": 3.8378550687540427, + "tokens_seen": 648019968 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040586760280842527, + "loss": 1.3745, + "theoretical_loss": 3.8378154134376494, + "tokens_seen": 648085504 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004058575727181545, + "loss": 1.5266, + "theoretical_loss": 3.837775763253779, + "tokens_seen": 648151040 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040584754262788363, + "loss": 1.4565, + "theoretical_loss": 3.8377361182012484, + "tokens_seen": 648216576 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040583751253761287, + "loss": 1.4056, + "theoretical_loss": 3.8376964782788754, + "tokens_seen": 648282112 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040582748244734205, + "loss": 1.4881, + "theoretical_loss": 3.8376568434854765, + "tokens_seen": 648347648 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040581745235707123, + "loss": 1.3632, + "theoretical_loss": 3.837617213819871, + "tokens_seen": 648413184 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004058074222668004, + "loss": 1.4531, + "theoretical_loss": 3.837577589280876, + "tokens_seen": 648478720 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004057973921765296, + "loss": 1.3627, + "theoretical_loss": 3.8375379698673115, + "tokens_seen": 648544256 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040578736208625877, + "loss": 1.4922, + "theoretical_loss": 3.837498355577996, + "tokens_seen": 648609792 + }, + { + "epoch": 0.2, + "learning_rate": 0.000405777331995988, + "loss": 1.4402, + "theoretical_loss": 3.837458746411749, + "tokens_seen": 648675328 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040576730190571713, + "loss": 1.5378, + "theoretical_loss": 3.8374191423673913, + "tokens_seen": 648740864 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 245985, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5602788925170898, + "objective/train/theoretical_loss": 3.8373795434437423, + "objective/train/tokens_used": 648806400, + "theoretical_loss": 3.8373795434437423, + "tokens_seen": 648806400 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040575727181544637, + "loss": 1.557, + "theoretical_loss": 3.8373795434437423, + "tokens_seen": 648806400 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004057472417251755, + "loss": 1.4374, + "theoretical_loss": 3.8373399496396234, + "tokens_seen": 648871936 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040573721163490473, + "loss": 1.3517, + "theoretical_loss": 3.8373003609538556, + "tokens_seen": 648937472 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004057271815446339, + "loss": 1.5152, + "theoretical_loss": 3.8372607773852603, + "tokens_seen": 649003008 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004057171514543631, + "loss": 1.534, + "theoretical_loss": 3.8372211989326592, + "tokens_seen": 649068544 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004057071213640923, + "loss": 1.4714, + "theoretical_loss": 3.8371816255948756, + "tokens_seen": 649134080 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040569709127382146, + "loss": 1.5109, + "theoretical_loss": 3.837142057370731, + "tokens_seen": 649199616 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040568706118355064, + "loss": 1.3694, + "theoretical_loss": 3.837102494259049, + "tokens_seen": 649265152 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004056770310932799, + "loss": 1.4837, + "theoretical_loss": 3.8370629362586532, + "tokens_seen": 649330688 + }, + { + "epoch": 0.2, + "learning_rate": 0.000405667001003009, + "loss": 1.5187, + "theoretical_loss": 3.8370233833683676, + "tokens_seen": 649396224 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040565697091273824, + "loss": 1.5502, + "theoretical_loss": 3.8369838355870165, + "tokens_seen": 649461760 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004056469408224674, + "loss": 1.4914, + "theoretical_loss": 3.836944292913424, + "tokens_seen": 649527296 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004056369107321966, + "loss": 1.4613, + "theoretical_loss": 3.8369047553464153, + "tokens_seen": 649592832 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004056268806419258, + "loss": 1.368, + "theoretical_loss": 3.836865222884817, + "tokens_seen": 649658368 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040561685055165496, + "loss": 1.5396, + "theoretical_loss": 3.8368256955274527, + "tokens_seen": 649723904 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040560682046138414, + "loss": 1.4447, + "theoretical_loss": 3.8367861732731505, + "tokens_seen": 649789440 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004055967903711134, + "loss": 1.4762, + "theoretical_loss": 3.836746656120737, + "tokens_seen": 649854976 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004055867602808425, + "loss": 1.4607, + "theoretical_loss": 3.8367071440690372, + "tokens_seen": 649920512 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040557673019057174, + "loss": 1.459, + "theoretical_loss": 3.836667637116881, + "tokens_seen": 649986048 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040556670010030087, + "loss": 1.5614, + "theoretical_loss": 3.8366281352630947, + "tokens_seen": 650051584 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004055566700100301, + "loss": 1.4339, + "theoretical_loss": 3.8365886385065067, + "tokens_seen": 650117120 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004055466399197593, + "loss": 1.4294, + "theoretical_loss": 3.836549146845946, + "tokens_seen": 650182656 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040553660982948846, + "loss": 1.4597, + "theoretical_loss": 3.8365096602802407, + "tokens_seen": 650248192 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040552657973921764, + "loss": 1.5398, + "theoretical_loss": 3.83647017880822, + "tokens_seen": 650313728 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004055165496489469, + "loss": 1.3957, + "theoretical_loss": 3.836430702428715, + "tokens_seen": 650379264 + }, + { + "epoch": 0.2, + "learning_rate": 0.000405506519558676, + "loss": 1.3709, + "theoretical_loss": 3.8363912311405546, + "tokens_seen": 650444800 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040549648946840524, + "loss": 1.4905, + "theoretical_loss": 3.8363517649425702, + "tokens_seen": 650510336 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040548645937813437, + "loss": 1.4484, + "theoretical_loss": 3.8363123038335916, + "tokens_seen": 650575872 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004054764292878636, + "loss": 1.5118, + "theoretical_loss": 3.8362728478124506, + "tokens_seen": 650641408 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004054663991975928, + "loss": 1.5089, + "theoretical_loss": 3.8362333968779794, + "tokens_seen": 650706944 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040545636910732197, + "loss": 1.4443, + "theoretical_loss": 3.836193951029008, + "tokens_seen": 650772480 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040544633901705115, + "loss": 1.4787, + "theoretical_loss": 3.8361545102643717, + "tokens_seen": 650838016 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040543630892678033, + "loss": 1.5317, + "theoretical_loss": 3.836115074582901, + "tokens_seen": 650903552 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004054262788365095, + "loss": 1.4528, + "theoretical_loss": 3.8360756439834303, + "tokens_seen": 650969088 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040541624874623874, + "loss": 1.5078, + "theoretical_loss": 3.836036218464792, + "tokens_seen": 651034624 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040540621865596787, + "loss": 1.5976, + "theoretical_loss": 3.8359967980258216, + "tokens_seen": 651100160 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004053961885656971, + "loss": 1.5748, + "theoretical_loss": 3.8359573826653524, + "tokens_seen": 651165696 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004053861584754263, + "loss": 1.5247, + "theoretical_loss": 3.8359179723822194, + "tokens_seen": 651231232 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040537612838515547, + "loss": 1.4541, + "theoretical_loss": 3.835878567175257, + "tokens_seen": 651296768 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004053660982948847, + "loss": 1.5007, + "theoretical_loss": 3.8358391670433023, + "tokens_seen": 651362304 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040535606820461383, + "loss": 1.4954, + "theoretical_loss": 3.8357997719851893, + "tokens_seen": 651427840 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040534603811434307, + "loss": 1.4337, + "theoretical_loss": 3.835760381999756, + "tokens_seen": 651493376 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040533600802407225, + "loss": 1.5275, + "theoretical_loss": 3.835720997085838, + "tokens_seen": 651558912 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040532597793380143, + "loss": 1.372, + "theoretical_loss": 3.835681617242272, + "tokens_seen": 651624448 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004053159478435306, + "loss": 1.4968, + "theoretical_loss": 3.8356422424678964, + "tokens_seen": 651689984 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004053059177532598, + "loss": 1.4706, + "theoretical_loss": 3.835602872761549, + "tokens_seen": 651755520 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040529588766298897, + "loss": 1.4052, + "theoretical_loss": 3.8355635081220667, + "tokens_seen": 651821056 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004052858575727182, + "loss": 1.4712, + "theoretical_loss": 3.8355241485482887, + "tokens_seen": 651886592 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040527582748244733, + "loss": 1.4215, + "theoretical_loss": 3.835484794039055, + "tokens_seen": 651952128 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040526579739217657, + "loss": 1.4548, + "theoretical_loss": 3.835445444593203, + "tokens_seen": 652017664 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 247088, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4889452457427979, + "objective/train/theoretical_loss": 3.8354061002095743, + "objective/train/tokens_used": 652083200, + "theoretical_loss": 3.8354061002095743, + "tokens_seen": 652083200 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004052557673019057, + "loss": 1.5313, + "theoretical_loss": 3.8354061002095743, + "tokens_seen": 652083200 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040524573721163493, + "loss": 1.4389, + "theoretical_loss": 3.835366760887008, + "tokens_seen": 652148736 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004052357071213641, + "loss": 1.3893, + "theoretical_loss": 3.8353274266243442, + "tokens_seen": 652214272 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004052256770310933, + "loss": 1.5599, + "theoretical_loss": 3.835288097420425, + "tokens_seen": 652279808 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004052156469408225, + "loss": 1.4855, + "theoretical_loss": 3.83524877327409, + "tokens_seen": 652345344 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040520561685055166, + "loss": 1.3383, + "theoretical_loss": 3.8352094541841817, + "tokens_seen": 652410880 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040519558676028084, + "loss": 1.5861, + "theoretical_loss": 3.8351701401495424, + "tokens_seen": 652476416 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004051855566700101, + "loss": 1.4352, + "theoretical_loss": 3.8351308311690144, + "tokens_seen": 652541952 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004051755265797392, + "loss": 1.4676, + "theoretical_loss": 3.83509152724144, + "tokens_seen": 652607488 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040516549648946844, + "loss": 1.4799, + "theoretical_loss": 3.8350522283656625, + "tokens_seen": 652673024 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004051554663991976, + "loss": 1.4345, + "theoretical_loss": 3.8350129345405257, + "tokens_seen": 652738560 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004051454363089268, + "loss": 1.4804, + "theoretical_loss": 3.8349736457648733, + "tokens_seen": 652804096 + }, + { + "epoch": 0.2, + "learning_rate": 0.000405135406218656, + "loss": 1.4886, + "theoretical_loss": 3.83493436203755, + "tokens_seen": 652869632 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040512537612838516, + "loss": 1.5299, + "theoretical_loss": 3.834895083357399, + "tokens_seen": 652935168 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040511534603811434, + "loss": 1.4493, + "theoretical_loss": 3.8348558097232672, + "tokens_seen": 653000704 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004051053159478436, + "loss": 1.4935, + "theoretical_loss": 3.834816541133999, + "tokens_seen": 653066240 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004050952858575727, + "loss": 1.4505, + "theoretical_loss": 3.83477727758844, + "tokens_seen": 653131776 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040508525576730194, + "loss": 1.4873, + "theoretical_loss": 3.8347380190854374, + "tokens_seen": 653197312 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040507522567703107, + "loss": 1.4459, + "theoretical_loss": 3.8346987656238367, + "tokens_seen": 653262848 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004050651955867603, + "loss": 1.5089, + "theoretical_loss": 3.8346595172024855, + "tokens_seen": 653328384 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004050551654964895, + "loss": 1.4894, + "theoretical_loss": 3.834620273820231, + "tokens_seen": 653393920 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040504513540621866, + "loss": 1.5621, + "theoretical_loss": 3.834581035475921, + "tokens_seen": 653459456 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040503510531594784, + "loss": 1.4915, + "theoretical_loss": 3.8345418021684035, + "tokens_seen": 653524992 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004050250752256771, + "loss": 1.4635, + "theoretical_loss": 3.8345025738965264, + "tokens_seen": 653590528 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004050150451354062, + "loss": 1.4342, + "theoretical_loss": 3.8344633506591395, + "tokens_seen": 653656064 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040500501504513544, + "loss": 1.4109, + "theoretical_loss": 3.8344241324550907, + "tokens_seen": 653721600 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040499498495486457, + "loss": 1.4311, + "theoretical_loss": 3.8343849192832313, + "tokens_seen": 653787136 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004049849548645938, + "loss": 1.3613, + "theoretical_loss": 3.83434571114241, + "tokens_seen": 653852672 + }, + { + "epoch": 0.2, + "learning_rate": 0.000404974924774323, + "loss": 1.5702, + "theoretical_loss": 3.834306508031478, + "tokens_seen": 653918208 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040496489468405217, + "loss": 1.3813, + "theoretical_loss": 3.8342673099492854, + "tokens_seen": 653983744 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040495486459378135, + "loss": 1.5285, + "theoretical_loss": 3.8342281168946832, + "tokens_seen": 654049280 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040494483450351053, + "loss": 1.469, + "theoretical_loss": 3.8341889288665234, + "tokens_seen": 654114816 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004049348044132397, + "loss": 1.4299, + "theoretical_loss": 3.8341497458636575, + "tokens_seen": 654180352 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040492477432296894, + "loss": 1.4142, + "theoretical_loss": 3.834110567884938, + "tokens_seen": 654245888 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040491474423269807, + "loss": 1.4652, + "theoretical_loss": 3.8340713949292176, + "tokens_seen": 654311424 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004049047141424273, + "loss": 1.5894, + "theoretical_loss": 3.834032226995349, + "tokens_seen": 654376960 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040489468405215643, + "loss": 1.5356, + "theoretical_loss": 3.8339930640821853, + "tokens_seen": 654442496 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040488465396188567, + "loss": 1.4903, + "theoretical_loss": 3.8339539061885812, + "tokens_seen": 654508032 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040487462387161485, + "loss": 1.4376, + "theoretical_loss": 3.83391475331339, + "tokens_seen": 654573568 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040486459378134403, + "loss": 1.3568, + "theoretical_loss": 3.833875605455466, + "tokens_seen": 654639104 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004048545636910732, + "loss": 1.4774, + "theoretical_loss": 3.8338364626136654, + "tokens_seen": 654704640 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040484453360080245, + "loss": 1.5485, + "theoretical_loss": 3.833797324786842, + "tokens_seen": 654770176 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004048345035105316, + "loss": 1.3811, + "theoretical_loss": 3.8337581919738524, + "tokens_seen": 654835712 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004048244734202608, + "loss": 1.5637, + "theoretical_loss": 3.8337190641735512, + "tokens_seen": 654901248 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040481444332998994, + "loss": 1.5884, + "theoretical_loss": 3.833679941384797, + "tokens_seen": 654966784 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040480441323971917, + "loss": 1.4552, + "theoretical_loss": 3.833640823606445, + "tokens_seen": 655032320 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040479438314944835, + "loss": 1.5219, + "theoretical_loss": 3.8336017108373523, + "tokens_seen": 655097856 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040478435305917753, + "loss": 1.4452, + "theoretical_loss": 3.833562603076377, + "tokens_seen": 655163392 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004047743229689067, + "loss": 1.5145, + "theoretical_loss": 3.833523500322377, + "tokens_seen": 655228928 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004047642928786359, + "loss": 1.4623, + "theoretical_loss": 3.83348440257421, + "tokens_seen": 655294464 + }, + { + "debugging/length": 498.1666666666667, + "debugging/num_segments": 12, + "debugging/score": 0.0034189909721288694, + "debugging/score_std": 0.0030829856842896496, + "epoch": 0.2, + "objective/train/docs_used": 248293, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5483520030975342, + "objective/train/theoretical_loss": 3.833445309830735, + "objective/train/tokens_used": 655360000, + "theoretical_loss": 3.833445309830735, + "tokens_seen": 655360000 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004047542627883651, + "loss": 1.4895, + "theoretical_loss": 3.833445309830735, + "tokens_seen": 655360000 + } + ], + "max_steps": 50354, + "num_train_epochs": 9223372036854775807, + "total_flos": 3.3445380096e+17, + "trial_name": null, + "trial_params": null +}