| { | |
| "best_metric": 0.7395176291465759, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-150", | |
| "epoch": 0.09832841691248771, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0006555227794165847, | |
| "grad_norm": 0.2373015284538269, | |
| "learning_rate": 5e-06, | |
| "loss": 0.8465, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0006555227794165847, | |
| "eval_loss": 1.0142732858657837, | |
| "eval_runtime": 197.6224, | |
| "eval_samples_per_second": 13.005, | |
| "eval_steps_per_second": 6.502, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0013110455588331695, | |
| "grad_norm": 0.28709691762924194, | |
| "learning_rate": 1e-05, | |
| "loss": 0.9051, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.0019665683382497543, | |
| "grad_norm": 0.3491666615009308, | |
| "learning_rate": 1.5e-05, | |
| "loss": 0.9242, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.002622091117666339, | |
| "grad_norm": 0.3462134897708893, | |
| "learning_rate": 2e-05, | |
| "loss": 0.918, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.0032776138970829235, | |
| "grad_norm": 0.28024375438690186, | |
| "learning_rate": 2.5e-05, | |
| "loss": 0.8966, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.003933136676499509, | |
| "grad_norm": 0.29582786560058594, | |
| "learning_rate": 3e-05, | |
| "loss": 0.9393, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.004588659455916093, | |
| "grad_norm": 0.308539479970932, | |
| "learning_rate": 3.5e-05, | |
| "loss": 0.9266, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.005244182235332678, | |
| "grad_norm": 0.2761336863040924, | |
| "learning_rate": 4e-05, | |
| "loss": 0.9699, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.0058997050147492625, | |
| "grad_norm": 0.2822662591934204, | |
| "learning_rate": 4.5e-05, | |
| "loss": 0.9126, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.006555227794165847, | |
| "grad_norm": 0.2669486999511719, | |
| "learning_rate": 5e-05, | |
| "loss": 0.942, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.007210750573582432, | |
| "grad_norm": 0.2491784691810608, | |
| "learning_rate": 5.500000000000001e-05, | |
| "loss": 0.9112, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.007866273352999017, | |
| "grad_norm": 0.22748543322086334, | |
| "learning_rate": 6e-05, | |
| "loss": 0.8695, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.008521796132415601, | |
| "grad_norm": 0.2544311285018921, | |
| "learning_rate": 6.500000000000001e-05, | |
| "loss": 0.8578, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.009177318911832186, | |
| "grad_norm": 0.24937686324119568, | |
| "learning_rate": 7e-05, | |
| "loss": 0.8732, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.00983284169124877, | |
| "grad_norm": 0.25596895813941956, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 0.8058, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.010488364470665356, | |
| "grad_norm": 0.2677898406982422, | |
| "learning_rate": 8e-05, | |
| "loss": 0.8617, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.011143887250081941, | |
| "grad_norm": 0.27905362844467163, | |
| "learning_rate": 8.5e-05, | |
| "loss": 0.8626, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.011799410029498525, | |
| "grad_norm": 0.25651392340660095, | |
| "learning_rate": 9e-05, | |
| "loss": 0.8348, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.01245493280891511, | |
| "grad_norm": 0.2798261046409607, | |
| "learning_rate": 9.5e-05, | |
| "loss": 0.8304, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.013110455588331694, | |
| "grad_norm": 0.26732340455055237, | |
| "learning_rate": 0.0001, | |
| "loss": 0.8372, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.01376597836774828, | |
| "grad_norm": 0.25484147667884827, | |
| "learning_rate": 9.999238475781957e-05, | |
| "loss": 0.8088, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.014421501147164863, | |
| "grad_norm": 0.2536450922489166, | |
| "learning_rate": 9.99695413509548e-05, | |
| "loss": 0.8054, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.015077023926581449, | |
| "grad_norm": 0.2574830949306488, | |
| "learning_rate": 9.99314767377287e-05, | |
| "loss": 0.854, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.015732546705998034, | |
| "grad_norm": 0.2322714626789093, | |
| "learning_rate": 9.987820251299122e-05, | |
| "loss": 0.813, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.01638806948541462, | |
| "grad_norm": 0.2515113055706024, | |
| "learning_rate": 9.980973490458728e-05, | |
| "loss": 0.8065, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.017043592264831202, | |
| "grad_norm": 0.24610289931297302, | |
| "learning_rate": 9.972609476841367e-05, | |
| "loss": 0.7927, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.017699115044247787, | |
| "grad_norm": 0.24373719096183777, | |
| "learning_rate": 9.962730758206611e-05, | |
| "loss": 0.8351, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.018354637823664373, | |
| "grad_norm": 0.24467414617538452, | |
| "learning_rate": 9.951340343707852e-05, | |
| "loss": 0.813, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.01901016060308096, | |
| "grad_norm": 0.23405858874320984, | |
| "learning_rate": 9.938441702975689e-05, | |
| "loss": 0.8165, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.01966568338249754, | |
| "grad_norm": 0.24256618320941925, | |
| "learning_rate": 9.924038765061042e-05, | |
| "loss": 0.7617, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.020321206161914126, | |
| "grad_norm": 0.2607885003089905, | |
| "learning_rate": 9.908135917238321e-05, | |
| "loss": 0.7873, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.02097672894133071, | |
| "grad_norm": 0.2551879286766052, | |
| "learning_rate": 9.890738003669029e-05, | |
| "loss": 0.7735, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.021632251720747297, | |
| "grad_norm": 0.2627292275428772, | |
| "learning_rate": 9.871850323926177e-05, | |
| "loss": 0.8321, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.022287774500163882, | |
| "grad_norm": 0.29124778509140015, | |
| "learning_rate": 9.851478631379982e-05, | |
| "loss": 0.7953, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.022943297279580464, | |
| "grad_norm": 0.26475605368614197, | |
| "learning_rate": 9.829629131445342e-05, | |
| "loss": 0.7745, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.02359882005899705, | |
| "grad_norm": 0.2607128322124481, | |
| "learning_rate": 9.806308479691595e-05, | |
| "loss": 0.7786, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.024254342838413635, | |
| "grad_norm": 0.2592867314815521, | |
| "learning_rate": 9.781523779815179e-05, | |
| "loss": 0.7604, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.02490986561783022, | |
| "grad_norm": 0.25272342562675476, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 0.7898, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.025565388397246803, | |
| "grad_norm": 0.2672499716281891, | |
| "learning_rate": 9.727592877996585e-05, | |
| "loss": 0.7601, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.02622091117666339, | |
| "grad_norm": 0.29855430126190186, | |
| "learning_rate": 9.698463103929542e-05, | |
| "loss": 0.8142, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.026876433956079974, | |
| "grad_norm": 0.3550460934638977, | |
| "learning_rate": 9.667902132486009e-05, | |
| "loss": 0.7327, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.02753195673549656, | |
| "grad_norm": 0.29364699125289917, | |
| "learning_rate": 9.635919272833938e-05, | |
| "loss": 0.7738, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.028187479514913145, | |
| "grad_norm": 0.32683664560317993, | |
| "learning_rate": 9.602524267262203e-05, | |
| "loss": 0.7557, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.028843002294329727, | |
| "grad_norm": 0.31279873847961426, | |
| "learning_rate": 9.567727288213005e-05, | |
| "loss": 0.7635, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.029498525073746312, | |
| "grad_norm": 0.30517908930778503, | |
| "learning_rate": 9.53153893518325e-05, | |
| "loss": 0.7931, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.030154047853162898, | |
| "grad_norm": 0.3420467972755432, | |
| "learning_rate": 9.493970231495835e-05, | |
| "loss": 0.8058, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.030809570632579483, | |
| "grad_norm": 0.3307579755783081, | |
| "learning_rate": 9.45503262094184e-05, | |
| "loss": 0.7355, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.03146509341199607, | |
| "grad_norm": 0.6385186910629272, | |
| "learning_rate": 9.414737964294636e-05, | |
| "loss": 0.7978, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.032120616191412654, | |
| "grad_norm": 0.6117073893547058, | |
| "learning_rate": 9.373098535696979e-05, | |
| "loss": 0.7463, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.03277613897082924, | |
| "grad_norm": 0.9974555373191833, | |
| "learning_rate": 9.330127018922194e-05, | |
| "loss": 0.7761, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.03277613897082924, | |
| "eval_loss": 0.8445320129394531, | |
| "eval_runtime": 199.0087, | |
| "eval_samples_per_second": 12.914, | |
| "eval_steps_per_second": 6.457, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.03343166175024582, | |
| "grad_norm": 0.45800045132637024, | |
| "learning_rate": 9.285836503510562e-05, | |
| "loss": 0.8983, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.034087184529662404, | |
| "grad_norm": 0.394972026348114, | |
| "learning_rate": 9.24024048078213e-05, | |
| "loss": 0.8675, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.03474270730907899, | |
| "grad_norm": 0.31172430515289307, | |
| "learning_rate": 9.193352839727121e-05, | |
| "loss": 0.8291, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.035398230088495575, | |
| "grad_norm": 0.22538970410823822, | |
| "learning_rate": 9.145187862775209e-05, | |
| "loss": 0.7883, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.03605375286791216, | |
| "grad_norm": 0.20795033872127533, | |
| "learning_rate": 9.09576022144496e-05, | |
| "loss": 0.752, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.036709275647328746, | |
| "grad_norm": 0.2245747447013855, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.7782, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.03736479842674533, | |
| "grad_norm": 0.19541659951210022, | |
| "learning_rate": 8.993177550236464e-05, | |
| "loss": 0.7945, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.03802032120616192, | |
| "grad_norm": 0.1979525238275528, | |
| "learning_rate": 8.940053768033609e-05, | |
| "loss": 0.8191, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.0386758439855785, | |
| "grad_norm": 0.21741431951522827, | |
| "learning_rate": 8.885729807284856e-05, | |
| "loss": 0.8025, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.03933136676499508, | |
| "grad_norm": 0.2003290057182312, | |
| "learning_rate": 8.83022221559489e-05, | |
| "loss": 0.7927, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.039986889544411666, | |
| "grad_norm": 0.21580374240875244, | |
| "learning_rate": 8.773547901113862e-05, | |
| "loss": 0.8017, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.04064241232382825, | |
| "grad_norm": 0.20684310793876648, | |
| "learning_rate": 8.715724127386972e-05, | |
| "loss": 0.8022, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.04129793510324484, | |
| "grad_norm": 0.20254190266132355, | |
| "learning_rate": 8.656768508095853e-05, | |
| "loss": 0.838, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.04195345788266142, | |
| "grad_norm": 0.2019943743944168, | |
| "learning_rate": 8.596699001693255e-05, | |
| "loss": 0.7911, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.04260898066207801, | |
| "grad_norm": 0.19427140057086945, | |
| "learning_rate": 8.535533905932738e-05, | |
| "loss": 0.7759, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.043264503441494594, | |
| "grad_norm": 0.19595062732696533, | |
| "learning_rate": 8.473291852294987e-05, | |
| "loss": 0.7871, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.04392002622091118, | |
| "grad_norm": 0.2017510086297989, | |
| "learning_rate": 8.409991800312493e-05, | |
| "loss": 0.8125, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.044575549000327765, | |
| "grad_norm": 0.22166839241981506, | |
| "learning_rate": 8.345653031794292e-05, | |
| "loss": 0.8228, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.04523107177974434, | |
| "grad_norm": 0.19774554669857025, | |
| "learning_rate": 8.280295144952536e-05, | |
| "loss": 0.7538, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.04588659455916093, | |
| "grad_norm": 0.20463129878044128, | |
| "learning_rate": 8.213938048432697e-05, | |
| "loss": 0.7343, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.046542117338577514, | |
| "grad_norm": 0.22063031792640686, | |
| "learning_rate": 8.146601955249188e-05, | |
| "loss": 0.7974, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.0471976401179941, | |
| "grad_norm": 0.20627279579639435, | |
| "learning_rate": 8.07830737662829e-05, | |
| "loss": 0.7727, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.047853162897410685, | |
| "grad_norm": 0.2119884043931961, | |
| "learning_rate": 8.009075115760243e-05, | |
| "loss": 0.7318, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.04850868567682727, | |
| "grad_norm": 0.24390168488025665, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 0.7685, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.049164208456243856, | |
| "grad_norm": 0.21942038834095, | |
| "learning_rate": 7.86788218175523e-05, | |
| "loss": 0.7906, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.04981973123566044, | |
| "grad_norm": 0.20912085473537445, | |
| "learning_rate": 7.795964517353735e-05, | |
| "loss": 0.7866, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.05047525401507703, | |
| "grad_norm": 0.21996894478797913, | |
| "learning_rate": 7.723195175075136e-05, | |
| "loss": 0.7765, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.051130776794493606, | |
| "grad_norm": 0.2343721091747284, | |
| "learning_rate": 7.649596321166024e-05, | |
| "loss": 0.772, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.05178629957391019, | |
| "grad_norm": 0.23541605472564697, | |
| "learning_rate": 7.575190374550272e-05, | |
| "loss": 0.7334, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.05244182235332678, | |
| "grad_norm": 0.24592329561710358, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 0.7377, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.05309734513274336, | |
| "grad_norm": 0.227330282330513, | |
| "learning_rate": 7.424048101231686e-05, | |
| "loss": 0.7109, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.05375286791215995, | |
| "grad_norm": 0.22559882700443268, | |
| "learning_rate": 7.347357813929454e-05, | |
| "loss": 0.7267, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.05440839069157653, | |
| "grad_norm": 0.2438517063856125, | |
| "learning_rate": 7.269952498697734e-05, | |
| "loss": 0.7595, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.05506391347099312, | |
| "grad_norm": 0.24995003640651703, | |
| "learning_rate": 7.191855733945387e-05, | |
| "loss": 0.7756, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.055719436250409704, | |
| "grad_norm": 0.2487722635269165, | |
| "learning_rate": 7.113091308703498e-05, | |
| "loss": 0.7314, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.05637495902982629, | |
| "grad_norm": 0.2623230516910553, | |
| "learning_rate": 7.033683215379002e-05, | |
| "loss": 0.7438, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.05703048180924287, | |
| "grad_norm": 0.28028956055641174, | |
| "learning_rate": 6.953655642446368e-05, | |
| "loss": 0.7467, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.057686004588659454, | |
| "grad_norm": 0.28336915373802185, | |
| "learning_rate": 6.873032967079561e-05, | |
| "loss": 0.7406, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.05834152736807604, | |
| "grad_norm": 0.3217703700065613, | |
| "learning_rate": 6.7918397477265e-05, | |
| "loss": 0.7428, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.058997050147492625, | |
| "grad_norm": 0.2729189693927765, | |
| "learning_rate": 6.710100716628344e-05, | |
| "loss": 0.7873, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.05965257292690921, | |
| "grad_norm": 0.2695431709289551, | |
| "learning_rate": 6.627840772285784e-05, | |
| "loss": 0.7392, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.060308095706325796, | |
| "grad_norm": 0.2915373742580414, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 0.7131, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.06096361848574238, | |
| "grad_norm": 0.783824622631073, | |
| "learning_rate": 6.461858523613684e-05, | |
| "loss": 0.7253, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.061619141265158967, | |
| "grad_norm": 0.2767072916030884, | |
| "learning_rate": 6.378186779084995e-05, | |
| "loss": 0.7848, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.06227466404457555, | |
| "grad_norm": 0.29190996289253235, | |
| "learning_rate": 6.294095225512603e-05, | |
| "loss": 0.7908, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.06293018682399214, | |
| "grad_norm": 0.30580437183380127, | |
| "learning_rate": 6.209609477998338e-05, | |
| "loss": 0.7437, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.06358570960340872, | |
| "grad_norm": 0.3551975190639496, | |
| "learning_rate": 6.124755271719325e-05, | |
| "loss": 0.7034, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.06424123238282531, | |
| "grad_norm": 0.3854832351207733, | |
| "learning_rate": 6.0395584540887963e-05, | |
| "loss": 0.7428, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.06489675516224189, | |
| "grad_norm": 0.4369230270385742, | |
| "learning_rate": 5.9540449768827246e-05, | |
| "loss": 0.7368, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.06555227794165848, | |
| "grad_norm": 0.7531588673591614, | |
| "learning_rate": 5.868240888334653e-05, | |
| "loss": 0.7087, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.06555227794165848, | |
| "eval_loss": 0.772973358631134, | |
| "eval_runtime": 198.9609, | |
| "eval_samples_per_second": 12.917, | |
| "eval_steps_per_second": 6.459, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.06620780072107506, | |
| "grad_norm": 0.29382604360580444, | |
| "learning_rate": 5.782172325201155e-05, | |
| "loss": 0.8081, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.06686332350049164, | |
| "grad_norm": 0.25580349564552307, | |
| "learning_rate": 5.695865504800327e-05, | |
| "loss": 0.7691, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.06751884627990823, | |
| "grad_norm": 0.26038458943367004, | |
| "learning_rate": 5.6093467170257374e-05, | |
| "loss": 0.7871, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.06817436905932481, | |
| "grad_norm": 0.21882936358451843, | |
| "learning_rate": 5.522642316338268e-05, | |
| "loss": 0.7262, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.0688298918387414, | |
| "grad_norm": 0.1874031126499176, | |
| "learning_rate": 5.435778713738292e-05, | |
| "loss": 0.7742, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.06948541461815798, | |
| "grad_norm": 0.19271279871463776, | |
| "learning_rate": 5.348782368720626e-05, | |
| "loss": 0.7742, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.07014093739757457, | |
| "grad_norm": 0.1871943324804306, | |
| "learning_rate": 5.26167978121472e-05, | |
| "loss": 0.7608, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.07079646017699115, | |
| "grad_norm": 0.18977497518062592, | |
| "learning_rate": 5.174497483512506e-05, | |
| "loss": 0.7577, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.07145198295640774, | |
| "grad_norm": 0.20853860676288605, | |
| "learning_rate": 5.0872620321864185e-05, | |
| "loss": 0.7623, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.07210750573582432, | |
| "grad_norm": 0.1942238062620163, | |
| "learning_rate": 5e-05, | |
| "loss": 0.7691, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.0727630285152409, | |
| "grad_norm": 0.19266293942928314, | |
| "learning_rate": 4.912737967813583e-05, | |
| "loss": 0.7152, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.07341855129465749, | |
| "grad_norm": 0.2014968991279602, | |
| "learning_rate": 4.825502516487497e-05, | |
| "loss": 0.7467, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.07407407407407407, | |
| "grad_norm": 0.20719727873802185, | |
| "learning_rate": 4.738320218785281e-05, | |
| "loss": 0.7312, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.07472959685349066, | |
| "grad_norm": 0.2144216001033783, | |
| "learning_rate": 4.6512176312793736e-05, | |
| "loss": 0.7853, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.07538511963290724, | |
| "grad_norm": 0.21068166196346283, | |
| "learning_rate": 4.564221286261709e-05, | |
| "loss": 0.7404, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.07604064241232383, | |
| "grad_norm": 0.20504605770111084, | |
| "learning_rate": 4.477357683661734e-05, | |
| "loss": 0.7685, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.07669616519174041, | |
| "grad_norm": 0.21056416630744934, | |
| "learning_rate": 4.390653282974264e-05, | |
| "loss": 0.758, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.077351687971157, | |
| "grad_norm": 0.20693372189998627, | |
| "learning_rate": 4.3041344951996746e-05, | |
| "loss": 0.7626, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.07800721075057358, | |
| "grad_norm": 0.20952555537223816, | |
| "learning_rate": 4.2178276747988446e-05, | |
| "loss": 0.7542, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.07866273352999016, | |
| "grad_norm": 0.2200106978416443, | |
| "learning_rate": 4.131759111665349e-05, | |
| "loss": 0.7802, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.07931825630940675, | |
| "grad_norm": 0.21421992778778076, | |
| "learning_rate": 4.045955023117276e-05, | |
| "loss": 0.7863, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.07997377908882333, | |
| "grad_norm": 0.21937255561351776, | |
| "learning_rate": 3.960441545911204e-05, | |
| "loss": 0.778, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.08062930186823992, | |
| "grad_norm": 0.21657036244869232, | |
| "learning_rate": 3.875244728280676e-05, | |
| "loss": 0.7492, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.0812848246476565, | |
| "grad_norm": 0.22081716358661652, | |
| "learning_rate": 3.790390522001662e-05, | |
| "loss": 0.7652, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.0819403474270731, | |
| "grad_norm": 0.22119349241256714, | |
| "learning_rate": 3.705904774487396e-05, | |
| "loss": 0.7341, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.08259587020648967, | |
| "grad_norm": 0.22376029193401337, | |
| "learning_rate": 3.6218132209150045e-05, | |
| "loss": 0.7702, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.08325139298590627, | |
| "grad_norm": 0.22649244964122772, | |
| "learning_rate": 3.5381414763863166e-05, | |
| "loss": 0.722, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.08390691576532285, | |
| "grad_norm": 0.23790930211544037, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 0.7523, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.08456243854473942, | |
| "grad_norm": 0.24429242312908173, | |
| "learning_rate": 3.372159227714218e-05, | |
| "loss": 0.7745, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.08521796132415602, | |
| "grad_norm": 0.222866490483284, | |
| "learning_rate": 3.289899283371657e-05, | |
| "loss": 0.7027, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.0858734841035726, | |
| "grad_norm": 0.2363133728504181, | |
| "learning_rate": 3.2081602522734986e-05, | |
| "loss": 0.7988, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.08652900688298919, | |
| "grad_norm": 0.24787373840808868, | |
| "learning_rate": 3.12696703292044e-05, | |
| "loss": 0.7659, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.08718452966240577, | |
| "grad_norm": 0.24052570760250092, | |
| "learning_rate": 3.046344357553632e-05, | |
| "loss": 0.7198, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.08784005244182236, | |
| "grad_norm": 0.29197588562965393, | |
| "learning_rate": 2.9663167846209998e-05, | |
| "loss": 0.7033, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.08849557522123894, | |
| "grad_norm": 0.24380062520503998, | |
| "learning_rate": 2.886908691296504e-05, | |
| "loss": 0.7448, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.08915109800065553, | |
| "grad_norm": 0.24123196303844452, | |
| "learning_rate": 2.8081442660546125e-05, | |
| "loss": 0.7246, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.08980662078007211, | |
| "grad_norm": 0.24851925671100616, | |
| "learning_rate": 2.7300475013022663e-05, | |
| "loss": 0.7595, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.09046214355948869, | |
| "grad_norm": 0.24264392256736755, | |
| "learning_rate": 2.6526421860705473e-05, | |
| "loss": 0.7071, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.09111766633890528, | |
| "grad_norm": 0.2609306871891022, | |
| "learning_rate": 2.575951898768315e-05, | |
| "loss": 0.7124, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.09177318911832186, | |
| "grad_norm": 0.2576271593570709, | |
| "learning_rate": 2.500000000000001e-05, | |
| "loss": 0.7195, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.09242871189773845, | |
| "grad_norm": 0.27716317772865295, | |
| "learning_rate": 2.4248096254497288e-05, | |
| "loss": 0.7594, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.09308423467715503, | |
| "grad_norm": 0.2704521119594574, | |
| "learning_rate": 2.350403678833976e-05, | |
| "loss": 0.7154, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.09373975745657162, | |
| "grad_norm": 0.2818368375301361, | |
| "learning_rate": 2.2768048249248648e-05, | |
| "loss": 0.7471, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.0943952802359882, | |
| "grad_norm": 0.3111891746520996, | |
| "learning_rate": 2.2040354826462668e-05, | |
| "loss": 0.7962, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.09505080301540479, | |
| "grad_norm": 0.3065095543861389, | |
| "learning_rate": 2.132117818244771e-05, | |
| "loss": 0.7228, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.09570632579482137, | |
| "grad_norm": 0.273107647895813, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 0.7028, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.09636184857423795, | |
| "grad_norm": 0.30496013164520264, | |
| "learning_rate": 1.9909248842397584e-05, | |
| "loss": 0.7111, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.09701737135365454, | |
| "grad_norm": 0.3292425572872162, | |
| "learning_rate": 1.9216926233717085e-05, | |
| "loss": 0.7902, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.09767289413307112, | |
| "grad_norm": 0.3946918547153473, | |
| "learning_rate": 1.8533980447508137e-05, | |
| "loss": 0.7038, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.09832841691248771, | |
| "grad_norm": 1.1070607900619507, | |
| "learning_rate": 1.7860619515673033e-05, | |
| "loss": 0.7469, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.09832841691248771, | |
| "eval_loss": 0.7395176291465759, | |
| "eval_runtime": 199.0516, | |
| "eval_samples_per_second": 12.911, | |
| "eval_steps_per_second": 6.456, | |
| "step": 150 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 200, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 50, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 5, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 2.0024155320405197e+17, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |