diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,5159 @@
+{
+  "best_metric": 0.012959838844835758,
+  "best_model_checkpoint": "/home/paperspace/Data/models/rommel_importgenius_4b8/llm3br256/checkpoint-350",
+  "epoch": 4.97907949790795,
+  "eval_steps": 5,
+  "global_step": 595,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.008368200836820083,
+      "grad_norm": 0.2135079950094223,
+      "learning_rate": 1.6666666666666667e-06,
+      "loss": 0.0776,
+      "step": 1
+    },
+    {
+      "epoch": 0.016736401673640166,
+      "grad_norm": 0.18678432703018188,
+      "learning_rate": 3.3333333333333333e-06,
+      "loss": 0.0663,
+      "step": 2
+    },
+    {
+      "epoch": 0.02510460251046025,
+      "grad_norm": 0.18199113011360168,
+      "learning_rate": 5e-06,
+      "loss": 0.0666,
+      "step": 3
+    },
+    {
+      "epoch": 0.03347280334728033,
+      "grad_norm": 0.13908497989177704,
+      "learning_rate": 6.666666666666667e-06,
+      "loss": 0.0596,
+      "step": 4
+    },
+    {
+      "epoch": 0.04184100418410042,
+      "grad_norm": 0.1845937818288803,
+      "learning_rate": 8.333333333333334e-06,
+      "loss": 0.0672,
+      "step": 5
+    },
+    {
+      "epoch": 0.04184100418410042,
+      "eval_loss": 0.07553808391094208,
+      "eval_runtime": 10.1109,
+      "eval_samples_per_second": 4.945,
+      "eval_steps_per_second": 1.286,
+      "step": 5
+    },
+    {
+      "epoch": 0.0502092050209205,
+      "grad_norm": 0.1750287413597107,
+      "learning_rate": 1e-05,
+      "loss": 0.0618,
+      "step": 6
+    },
+    {
+      "epoch": 0.058577405857740586,
+      "grad_norm": 0.1204298809170723,
+      "learning_rate": 1.1666666666666668e-05,
+      "loss": 0.0529,
+      "step": 7
+    },
+    {
+      "epoch": 0.06694560669456066,
+      "grad_norm": 0.12102019786834717,
+      "learning_rate": 1.3333333333333333e-05,
+      "loss": 0.0519,
+      "step": 8
+    },
+    {
+      "epoch": 0.07531380753138076,
+      "grad_norm": 0.1123175099492073,
+      "learning_rate": 1.5e-05,
+      "loss": 0.0494,
+      "step": 9
+    },
+    {
+      "epoch": 0.08368200836820083,
+      "grad_norm": 0.08419068902730942,
+      "learning_rate": 1.6666666666666667e-05,
+      "loss": 0.0476,
+      "step": 10
+    },
+    {
+      "epoch": 0.08368200836820083,
+      "eval_loss": 0.04517042636871338,
+      "eval_runtime": 8.1834,
+      "eval_samples_per_second": 6.11,
+      "eval_steps_per_second": 1.589,
+      "step": 10
+    },
+    {
+      "epoch": 0.09205020920502092,
+      "grad_norm": 0.0977497473359108,
+      "learning_rate": 1.8333333333333333e-05,
+      "loss": 0.0382,
+      "step": 11
+    },
+    {
+      "epoch": 0.100418410041841,
+      "grad_norm": 0.09294793009757996,
+      "learning_rate": 2e-05,
+      "loss": 0.0374,
+      "step": 12
+    },
+    {
+      "epoch": 0.1087866108786611,
+      "grad_norm": 0.0773313045501709,
+      "learning_rate": 2.1666666666666667e-05,
+      "loss": 0.0365,
+      "step": 13
+    },
+    {
+      "epoch": 0.11715481171548117,
+      "grad_norm": 0.07890919595956802,
+      "learning_rate": 2.3333333333333336e-05,
+      "loss": 0.0323,
+      "step": 14
+    },
+    {
+      "epoch": 0.12552301255230125,
+      "grad_norm": 0.0590483732521534,
+      "learning_rate": 2.5e-05,
+      "loss": 0.0337,
+      "step": 15
+    },
+    {
+      "epoch": 0.12552301255230125,
+      "eval_loss": 0.03555193170905113,
+      "eval_runtime": 8.2187,
+      "eval_samples_per_second": 6.084,
+      "eval_steps_per_second": 1.582,
+      "step": 15
+    },
+    {
+      "epoch": 0.13389121338912133,
+      "grad_norm": 0.05162455514073372,
+      "learning_rate": 2.6666666666666667e-05,
+      "loss": 0.0336,
+      "step": 16
+    },
+    {
+      "epoch": 0.14225941422594143,
+      "grad_norm": 0.04780131205916405,
+      "learning_rate": 2.8333333333333335e-05,
+      "loss": 0.0303,
+      "step": 17
+    },
+    {
+      "epoch": 0.1506276150627615,
+      "grad_norm": 0.05417422577738762,
+      "learning_rate": 3e-05,
+      "loss": 0.0303,
+      "step": 18
+    },
+    {
+      "epoch": 0.1589958158995816,
+      "grad_norm": 0.05216076225042343,
+      "learning_rate": 3.1666666666666666e-05,
+      "loss": 0.0329,
+      "step": 19
+    },
+    {
+      "epoch": 0.16736401673640167,
+      "grad_norm": 0.04367639496922493,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.0333,
+      "step": 20
+    },
+    {
+      "epoch": 0.16736401673640167,
+      "eval_loss": 0.030791474506258965,
+      "eval_runtime": 8.1829,
+      "eval_samples_per_second": 6.11,
+      "eval_steps_per_second": 1.589,
+      "step": 20
+    },
+    {
+      "epoch": 0.17573221757322174,
+      "grad_norm": 0.0510227307677269,
+      "learning_rate": 3.5e-05,
+      "loss": 0.0302,
+      "step": 21
+    },
+    {
+      "epoch": 0.18410041841004185,
+      "grad_norm": 0.03975889831781387,
+      "learning_rate": 3.6666666666666666e-05,
+      "loss": 0.0293,
+      "step": 22
+    },
+    {
+      "epoch": 0.19246861924686193,
+      "grad_norm": 0.033559419214725494,
+      "learning_rate": 3.8333333333333334e-05,
+      "loss": 0.0276,
+      "step": 23
+    },
+    {
+      "epoch": 0.200836820083682,
+      "grad_norm": 0.03443528711795807,
+      "learning_rate": 4e-05,
+      "loss": 0.0255,
+      "step": 24
+    },
+    {
+      "epoch": 0.20920502092050208,
+      "grad_norm": 0.04452834278345108,
+      "learning_rate": 4.166666666666667e-05,
+      "loss": 0.0258,
+      "step": 25
+    },
+    {
+      "epoch": 0.20920502092050208,
+      "eval_loss": 0.027234511449933052,
+      "eval_runtime": 8.2238,
+      "eval_samples_per_second": 6.08,
+      "eval_steps_per_second": 1.581,
+      "step": 25
+    },
+    {
+      "epoch": 0.2175732217573222,
+      "grad_norm": 0.04225548356771469,
+      "learning_rate": 4.3333333333333334e-05,
+      "loss": 0.0265,
+      "step": 26
+    },
+    {
+      "epoch": 0.22594142259414227,
+      "grad_norm": 0.035125982016325,
+      "learning_rate": 4.5e-05,
+      "loss": 0.0248,
+      "step": 27
+    },
+    {
+      "epoch": 0.23430962343096234,
+      "grad_norm": 0.0293523371219635,
+      "learning_rate": 4.666666666666667e-05,
+      "loss": 0.0233,
+      "step": 28
+    },
+    {
+      "epoch": 0.24267782426778242,
+      "grad_norm": 0.046518724411726,
+      "learning_rate": 4.8333333333333334e-05,
+      "loss": 0.025,
+      "step": 29
+    },
+    {
+      "epoch": 0.2510460251046025,
+      "grad_norm": 0.031807683408260345,
+      "learning_rate": 5e-05,
+      "loss": 0.023,
+      "step": 30
+    },
+    {
+      "epoch": 0.2510460251046025,
+      "eval_loss": 0.025461601093411446,
+      "eval_runtime": 8.1735,
+      "eval_samples_per_second": 6.117,
+      "eval_steps_per_second": 1.591,
+      "step": 30
+    },
+    {
+      "epoch": 0.2594142259414226,
+      "grad_norm": 0.03979400172829628,
+      "learning_rate": 5.166666666666667e-05,
+      "loss": 0.0225,
+      "step": 31
+    },
+    {
+      "epoch": 0.26778242677824265,
+      "grad_norm": 0.030477149412035942,
+      "learning_rate": 5.333333333333333e-05,
+      "loss": 0.0212,
+      "step": 32
+    },
+    {
+      "epoch": 0.27615062761506276,
+      "grad_norm": 0.029403740540146828,
+      "learning_rate": 5.500000000000001e-05,
+      "loss": 0.0223,
+      "step": 33
+    },
+    {
+      "epoch": 0.28451882845188287,
+      "grad_norm": 0.03240867331624031,
+      "learning_rate": 5.666666666666667e-05,
+      "loss": 0.024,
+      "step": 34
+    },
+    {
+      "epoch": 0.2928870292887029,
+      "grad_norm": 0.032563403248786926,
+      "learning_rate": 5.833333333333334e-05,
+      "loss": 0.0202,
+      "step": 35
+    },
+    {
+      "epoch": 0.2928870292887029,
+      "eval_loss": 0.023448586463928223,
+      "eval_runtime": 8.2019,
+      "eval_samples_per_second": 6.096,
+      "eval_steps_per_second": 1.585,
+      "step": 35
+    },
+    {
+      "epoch": 0.301255230125523,
+      "grad_norm": 0.029798880219459534,
+      "learning_rate": 6e-05,
+      "loss": 0.0214,
+      "step": 36
+    },
+    {
+      "epoch": 0.30962343096234307,
+      "grad_norm": 0.02481868863105774,
+      "learning_rate": 6.166666666666667e-05,
+      "loss": 0.0175,
+      "step": 37
+    },
+    {
+      "epoch": 0.3179916317991632,
+      "grad_norm": 0.029312577098608017,
+      "learning_rate": 6.333333333333333e-05,
+      "loss": 0.0207,
+      "step": 38
+    },
+    {
+      "epoch": 0.3263598326359833,
+      "grad_norm": 0.035519231110811234,
+      "learning_rate": 6.500000000000001e-05,
+      "loss": 0.019,
+      "step": 39
+    },
+    {
+      "epoch": 0.33472803347280333,
+      "grad_norm": 0.019138505682349205,
+      "learning_rate": 6.666666666666667e-05,
+      "loss": 0.0188,
+      "step": 40
+    },
+    {
+      "epoch": 0.33472803347280333,
+      "eval_loss": 0.02175285294651985,
+      "eval_runtime": 8.1801,
+      "eval_samples_per_second": 6.112,
+      "eval_steps_per_second": 1.589,
+      "step": 40
+    },
+    {
+      "epoch": 0.34309623430962344,
+      "grad_norm": 0.028007732704281807,
+      "learning_rate": 6.833333333333333e-05,
+      "loss": 0.0198,
+      "step": 41
+    },
+    {
+      "epoch": 0.3514644351464435,
+      "grad_norm": 0.02277914620935917,
+      "learning_rate": 7e-05,
+      "loss": 0.0214,
+      "step": 42
+    },
+    {
+      "epoch": 0.3598326359832636,
+      "grad_norm": 0.03379930928349495,
+      "learning_rate": 7.166666666666667e-05,
+      "loss": 0.0228,
+      "step": 43
+    },
+    {
+      "epoch": 0.3682008368200837,
+      "grad_norm": 0.02658325619995594,
+      "learning_rate": 7.333333333333333e-05,
+      "loss": 0.0201,
+      "step": 44
+    },
+    {
+      "epoch": 0.37656903765690375,
+      "grad_norm": 0.0228717178106308,
+      "learning_rate": 7.500000000000001e-05,
+      "loss": 0.0185,
+      "step": 45
+    },
+    {
+      "epoch": 0.37656903765690375,
+      "eval_loss": 0.02076469548046589,
+      "eval_runtime": 8.1924,
+      "eval_samples_per_second": 6.103,
+      "eval_steps_per_second": 1.587,
+      "step": 45
+    },
+    {
+      "epoch": 0.38493723849372385,
+      "grad_norm": 0.022528983652591705,
+      "learning_rate": 7.666666666666667e-05,
+      "loss": 0.0175,
+      "step": 46
+    },
+    {
+      "epoch": 0.39330543933054396,
+      "grad_norm": 0.02600259892642498,
+      "learning_rate": 7.833333333333333e-05,
+      "loss": 0.0199,
+      "step": 47
+    },
+    {
+      "epoch": 0.401673640167364,
+      "grad_norm": 0.022605430334806442,
+      "learning_rate": 8e-05,
+      "loss": 0.0178,
+      "step": 48
+    },
+    {
+      "epoch": 0.4100418410041841,
+      "grad_norm": 0.020774831995368004,
+      "learning_rate": 8.166666666666667e-05,
+      "loss": 0.0168,
+      "step": 49
+    },
+    {
+      "epoch": 0.41841004184100417,
+      "grad_norm": 0.02565855346620083,
+      "learning_rate": 8.333333333333334e-05,
+      "loss": 0.0199,
+      "step": 50
+    },
+    {
+      "epoch": 0.41841004184100417,
+      "eval_loss": 0.020035894587635994,
+      "eval_runtime": 8.1839,
+      "eval_samples_per_second": 6.11,
+      "eval_steps_per_second": 1.588,
+      "step": 50
+    },
+    {
+      "epoch": 0.42677824267782427,
+      "grad_norm": 0.027278423309326172,
+      "learning_rate": 8.5e-05,
+      "loss": 0.0208,
+      "step": 51
+    },
+    {
+      "epoch": 0.4351464435146444,
+      "grad_norm": 0.022894414141774178,
+      "learning_rate": 8.666666666666667e-05,
+      "loss": 0.0166,
+      "step": 52
+    },
+    {
+      "epoch": 0.4435146443514644,
+      "grad_norm": 0.02473956160247326,
+      "learning_rate": 8.833333333333333e-05,
+      "loss": 0.021,
+      "step": 53
+    },
+    {
+      "epoch": 0.45188284518828453,
+      "grad_norm": 0.02041654661297798,
+      "learning_rate": 9e-05,
+      "loss": 0.0189,
+      "step": 54
+    },
+    {
+      "epoch": 0.4602510460251046,
+      "grad_norm": 0.022939356043934822,
+      "learning_rate": 9.166666666666667e-05,
+      "loss": 0.0198,
+      "step": 55
+    },
+    {
+      "epoch": 0.4602510460251046,
+      "eval_loss": 0.019482074305415154,
+      "eval_runtime": 8.206,
+      "eval_samples_per_second": 6.093,
+      "eval_steps_per_second": 1.584,
+      "step": 55
+    },
+    {
+      "epoch": 0.4686192468619247,
+      "grad_norm": 0.018536586314439774,
+      "learning_rate": 9.333333333333334e-05,
+      "loss": 0.0176,
+      "step": 56
+    },
+    {
+      "epoch": 0.4769874476987448,
+      "grad_norm": 0.02208622545003891,
+      "learning_rate": 9.5e-05,
+      "loss": 0.0189,
+      "step": 57
+    },
+    {
+      "epoch": 0.48535564853556484,
+      "grad_norm": 0.021715424954891205,
+      "learning_rate": 9.666666666666667e-05,
+      "loss": 0.0176,
+      "step": 58
+    },
+    {
+      "epoch": 0.49372384937238495,
+      "grad_norm": 0.022242875769734383,
+      "learning_rate": 9.833333333333333e-05,
+      "loss": 0.0208,
+      "step": 59
+    },
+    {
+      "epoch": 0.502092050209205,
+      "grad_norm": 0.022756585851311684,
+      "learning_rate": 0.0001,
+      "loss": 0.0179,
+      "step": 60
+    },
+    {
+      "epoch": 0.502092050209205,
+      "eval_loss": 0.018894275650382042,
+      "eval_runtime": 8.1933,
+      "eval_samples_per_second": 6.103,
+      "eval_steps_per_second": 1.587,
+      "step": 60
+    },
+    {
+      "epoch": 0.5104602510460251,
+      "grad_norm": 0.021328022703528404,
+      "learning_rate": 9.999913795300544e-05,
+      "loss": 0.0183,
+      "step": 61
+    },
+    {
+      "epoch": 0.5188284518828452,
+      "grad_norm": 0.018391713500022888,
+      "learning_rate": 9.999655184174672e-05,
+      "loss": 0.0172,
+      "step": 62
+    },
+    {
+      "epoch": 0.5271966527196653,
+      "grad_norm": 0.019118035212159157,
+      "learning_rate": 9.999224175539785e-05,
+      "loss": 0.0172,
+      "step": 63
+    },
+    {
+      "epoch": 0.5355648535564853,
+      "grad_norm": 0.024857187643647194,
+      "learning_rate": 9.99862078425787e-05,
+      "loss": 0.0192,
+      "step": 64
+    },
+    {
+      "epoch": 0.5439330543933054,
+      "grad_norm": 0.025595078244805336,
+      "learning_rate": 9.997845031134992e-05,
+      "loss": 0.0185,
+      "step": 65
+    },
+    {
+      "epoch": 0.5439330543933054,
+      "eval_loss": 0.018585730344057083,
+      "eval_runtime": 8.172,
+      "eval_samples_per_second": 6.118,
+      "eval_steps_per_second": 1.591,
+      "step": 65
+    },
+    {
+      "epoch": 0.5523012552301255,
+      "grad_norm": 0.02300618588924408,
+      "learning_rate": 9.996896942920578e-05,
+      "loss": 0.0174,
+      "step": 66
+    },
+    {
+      "epoch": 0.5606694560669456,
+      "grad_norm": 0.022543633356690407,
+      "learning_rate": 9.99577655230649e-05,
+      "loss": 0.0155,
+      "step": 67
+    },
+    {
+      "epoch": 0.5690376569037657,
+      "grad_norm": 0.028140394017100334,
+      "learning_rate": 9.994483897925905e-05,
+      "loss": 0.0167,
+      "step": 68
+    },
+    {
+      "epoch": 0.5774058577405857,
+      "grad_norm": 0.027847809717059135,
+      "learning_rate": 9.993019024351974e-05,
+      "loss": 0.0184,
+      "step": 69
+    },
+    {
+      "epoch": 0.5857740585774058,
+      "grad_norm": 0.02762455679476261,
+      "learning_rate": 9.991381982096292e-05,
+      "loss": 0.0174,
+      "step": 70
+    },
+    {
+      "epoch": 0.5857740585774058,
+      "eval_loss": 0.018603580072522163,
+      "eval_runtime": 8.1911,
+      "eval_samples_per_second": 6.104,
+      "eval_steps_per_second": 1.587,
+      "step": 70
+    },
+    {
+      "epoch": 0.5941422594142259,
+      "grad_norm": 0.02379084751009941,
+      "learning_rate": 9.989572827607153e-05,
+      "loss": 0.0151,
+      "step": 71
+    },
+    {
+      "epoch": 0.602510460251046,
+      "grad_norm": 0.028387486934661865,
+      "learning_rate": 9.987591623267606e-05,
+      "loss": 0.0192,
+      "step": 72
+    },
+    {
+      "epoch": 0.6108786610878661,
+      "grad_norm": 0.029841607436537743,
+      "learning_rate": 9.9854384373933e-05,
+      "loss": 0.0166,
+      "step": 73
+    },
+    {
+      "epoch": 0.6192468619246861,
+      "grad_norm": 0.024572154507040977,
+      "learning_rate": 9.983113344230129e-05,
+      "loss": 0.0181,
+      "step": 74
+    },
+    {
+      "epoch": 0.6276150627615062,
+      "grad_norm": 0.024675268679857254,
+      "learning_rate": 9.98061642395168e-05,
+      "loss": 0.0157,
+      "step": 75
+    },
+    {
+      "epoch": 0.6276150627615062,
+      "eval_loss": 0.0182618610560894,
+      "eval_runtime": 8.1895,
+      "eval_samples_per_second": 6.105,
+      "eval_steps_per_second": 1.587,
+      "step": 75
+    },
+    {
+      "epoch": 0.6359832635983264,
+      "grad_norm": 0.019830523058772087,
+      "learning_rate": 9.977947762656455e-05,
+      "loss": 0.0183,
+      "step": 76
+    },
+    {
+      "epoch": 0.6443514644351465,
+      "grad_norm": 0.017791615799069405,
+      "learning_rate": 9.975107452364913e-05,
+      "loss": 0.0174,
+      "step": 77
+    },
+    {
+      "epoch": 0.6527196652719666,
+      "grad_norm": 0.01926375925540924,
+      "learning_rate": 9.972095591016294e-05,
+      "loss": 0.0158,
+      "step": 78
+    },
+    {
+      "epoch": 0.6610878661087866,
+      "grad_norm": 0.022102929651737213,
+      "learning_rate": 9.968912282465236e-05,
+      "loss": 0.0155,
+      "step": 79
+    },
+    {
+      "epoch": 0.6694560669456067,
+      "grad_norm": 0.024682268500328064,
+      "learning_rate": 9.965557636478203e-05,
+      "loss": 0.0175,
+      "step": 80
+    },
+    {
+      "epoch": 0.6694560669456067,
+      "eval_loss": 0.017632750794291496,
+      "eval_runtime": 8.184,
+      "eval_samples_per_second": 6.11,
+      "eval_steps_per_second": 1.588,
+      "step": 80
+    },
+    {
+      "epoch": 0.6778242677824268,
+      "grad_norm": 0.022384189069271088,
+      "learning_rate": 9.962031768729693e-05,
+      "loss": 0.0184,
+      "step": 81
+    },
+    {
+      "epoch": 0.6861924686192469,
+      "grad_norm": 0.024247104302048683,
+      "learning_rate": 9.958334800798256e-05,
+      "loss": 0.0162,
+      "step": 82
+    },
+    {
+      "epoch": 0.694560669456067,
+      "grad_norm": 0.02073194645345211,
+      "learning_rate": 9.954466860162295e-05,
+      "loss": 0.0146,
+      "step": 83
+    },
+    {
+      "epoch": 0.702928870292887,
+      "grad_norm": 0.021671872586011887,
+      "learning_rate": 9.950428080195675e-05,
+      "loss": 0.0183,
+      "step": 84
+    },
+    {
+      "epoch": 0.7112970711297071,
+      "grad_norm": 0.022358493879437447,
+      "learning_rate": 9.94621860016312e-05,
+      "loss": 0.0175,
+      "step": 85
+    },
+    {
+      "epoch": 0.7112970711297071,
+      "eval_loss": 0.01763150468468666,
+      "eval_runtime": 8.1866,
+      "eval_samples_per_second": 6.108,
+      "eval_steps_per_second": 1.588,
+      "step": 85
+    },
+    {
+      "epoch": 0.7196652719665272,
+      "grad_norm": 0.027475032955408096,
+      "learning_rate": 9.941838565215413e-05,
+      "loss": 0.018,
+      "step": 86
+    },
+    {
+      "epoch": 0.7280334728033473,
+      "grad_norm": 0.022491537034511566,
+      "learning_rate": 9.937288126384396e-05,
+      "loss": 0.0189,
+      "step": 87
+    },
+    {
+      "epoch": 0.7364016736401674,
+      "grad_norm": 0.017695261165499687,
+      "learning_rate": 9.932567440577751e-05,
+      "loss": 0.0169,
+      "step": 88
+    },
+    {
+      "epoch": 0.7447698744769874,
+      "grad_norm": 0.021199015900492668,
+      "learning_rate": 9.9276766705736e-05,
+      "loss": 0.016,
+      "step": 89
+    },
+    {
+      "epoch": 0.7531380753138075,
+      "grad_norm": 0.02306659147143364,
+      "learning_rate": 9.922615985014887e-05,
+      "loss": 0.0164,
+      "step": 90
+    },
+    {
+      "epoch": 0.7531380753138075,
+      "eval_loss": 0.017131321132183075,
+      "eval_runtime": 8.1843,
+      "eval_samples_per_second": 6.109,
+      "eval_steps_per_second": 1.588,
+      "step": 90
+    },
+    {
+      "epoch": 0.7615062761506276,
+      "grad_norm": 0.023978663608431816,
+      "learning_rate": 9.91738555840356e-05,
+      "loss": 0.0176,
+      "step": 91
+    },
+    {
+      "epoch": 0.7698744769874477,
+      "grad_norm": 0.020384812727570534,
+      "learning_rate": 9.911985571094564e-05,
+      "loss": 0.0178,
+      "step": 92
+    },
+    {
+      "epoch": 0.7782426778242678,
+      "grad_norm": 0.02201078273355961,
+      "learning_rate": 9.906416209289608e-05,
+      "loss": 0.0161,
+      "step": 93
+    },
+    {
+      "epoch": 0.7866108786610879,
+      "grad_norm": 0.021811528131365776,
+      "learning_rate": 9.900677665030762e-05,
+      "loss": 0.0161,
+      "step": 94
+    },
+    {
+      "epoch": 0.7949790794979079,
+      "grad_norm": 0.0257060918956995,
+      "learning_rate": 9.894770136193814e-05,
+      "loss": 0.0182,
+      "step": 95
+    },
+    {
+      "epoch": 0.7949790794979079,
+      "eval_loss": 0.016829807311296463,
+      "eval_runtime": 8.1842,
+      "eval_samples_per_second": 6.109,
+      "eval_steps_per_second": 1.588,
+      "step": 95
+    },
+    {
+      "epoch": 0.803347280334728,
+      "grad_norm": 0.018415937200188637,
+      "learning_rate": 9.888693826481467e-05,
+      "loss": 0.0154,
+      "step": 96
+    },
+    {
+      "epoch": 0.8117154811715481,
+      "grad_norm": 0.02051800675690174,
+      "learning_rate": 9.882448945416298e-05,
+      "loss": 0.0168,
+      "step": 97
+    },
+    {
+      "epoch": 0.8200836820083682,
+      "grad_norm": 0.01812230981886387,
+      "learning_rate": 9.87603570833355e-05,
+      "loss": 0.0142,
+      "step": 98
+    },
+    {
+      "epoch": 0.8284518828451883,
+      "grad_norm": 0.02094101719558239,
+      "learning_rate": 9.869454336373689e-05,
+      "loss": 0.0163,
+      "step": 99
+    },
+    {
+      "epoch": 0.8368200836820083,
+      "grad_norm": 0.02226194366812706,
+      "learning_rate": 9.862705056474795e-05,
+      "loss": 0.019,
+      "step": 100
+    },
+    {
+      "epoch": 0.8368200836820083,
+      "eval_loss": 0.016707511618733406,
+      "eval_runtime": 8.1876,
+      "eval_samples_per_second": 6.107,
+      "eval_steps_per_second": 1.588,
+      "step": 100
+    },
+    {
+      "epoch": 0.8451882845188284,
+      "grad_norm": 0.018342219293117523,
+      "learning_rate": 9.855788101364722e-05,
+      "loss": 0.0139,
+      "step": 101
+    },
+    {
+      "epoch": 0.8535564853556485,
+      "grad_norm": 0.024607762694358826,
+      "learning_rate": 9.848703709553089e-05,
+      "loss": 0.017,
+      "step": 102
+    },
+    {
+      "epoch": 0.8619246861924686,
+      "grad_norm": 0.021850740537047386,
+      "learning_rate": 9.841452125323041e-05,
+      "loss": 0.0171,
+      "step": 103
+    },
+    {
+      "epoch": 0.8702928870292888,
+      "grad_norm": 0.022089634090662003,
+      "learning_rate": 9.834033598722831e-05,
+      "loss": 0.0151,
+      "step": 104
+    },
+    {
+      "epoch": 0.8786610878661087,
+      "grad_norm": 0.01841079257428646,
+      "learning_rate": 9.826448385557207e-05,
+      "loss": 0.0163,
+      "step": 105
+    },
+    {
+      "epoch": 0.8786610878661087,
+      "eval_loss": 0.01583385095000267,
+      "eval_runtime": 8.2041,
+      "eval_samples_per_second": 6.094,
+      "eval_steps_per_second": 1.585,
+      "step": 105
+    },
+    {
+      "epoch": 0.8870292887029289,
+      "grad_norm": 0.020294038578867912,
+      "learning_rate": 9.818696747378573e-05,
+      "loss": 0.0177,
+      "step": 106
+    },
+    {
+      "epoch": 0.895397489539749,
+      "grad_norm": 0.015671249479055405,
+      "learning_rate": 9.810778951477986e-05,
+      "loss": 0.0144,
+      "step": 107
+    },
+    {
+      "epoch": 0.9037656903765691,
+      "grad_norm": 0.018139947205781937,
+      "learning_rate": 9.802695270875932e-05,
+      "loss": 0.0154,
+      "step": 108
+    },
+    {
+      "epoch": 0.9121338912133892,
+      "grad_norm": 0.018227294087409973,
+      "learning_rate": 9.794445984312915e-05,
+      "loss": 0.0149,
+      "step": 109
+    },
+    {
+      "epoch": 0.9205020920502092,
+      "grad_norm": 0.015631545335054398,
+      "learning_rate": 9.786031376239842e-05,
+      "loss": 0.0145,
+      "step": 110
+    },
+    {
+      "epoch": 0.9205020920502092,
+      "eval_loss": 0.0157511904835701,
+      "eval_runtime": 8.1802,
+      "eval_samples_per_second": 6.112,
+      "eval_steps_per_second": 1.589,
+      "step": 110
+    },
+    {
+      "epoch": 0.9288702928870293,
+      "grad_norm": 0.019501695409417152,
+      "learning_rate": 9.777451736808216e-05,
+      "loss": 0.0138,
+      "step": 111
+    },
+    {
+      "epoch": 0.9372384937238494,
+      "grad_norm": 0.018380407243967056,
+      "learning_rate": 9.768707361860134e-05,
+      "loss": 0.0155,
+      "step": 112
+    },
+    {
+      "epoch": 0.9456066945606695,
+      "grad_norm": 0.016666430979967117,
+      "learning_rate": 9.75979855291808e-05,
+      "loss": 0.0153,
+      "step": 113
+    },
+    {
+      "epoch": 0.9539748953974896,
+      "grad_norm": 0.019959956407546997,
+      "learning_rate": 9.750725617174534e-05,
+      "loss": 0.0162,
+      "step": 114
+    },
+    {
+      "epoch": 0.9623430962343096,
+      "grad_norm": 0.019904401153326035,
+      "learning_rate": 9.741488867481376e-05,
+      "loss": 0.0165,
+      "step": 115
+    },
+    {
+      "epoch": 0.9623430962343096,
+      "eval_loss": 0.015494490042328835,
+      "eval_runtime": 8.1797,
+      "eval_samples_per_second": 6.113,
+      "eval_steps_per_second": 1.589,
+      "step": 115
+    },
+    {
+      "epoch": 0.9707112970711297,
+      "grad_norm": 0.02116929367184639,
+      "learning_rate": 9.7320886223391e-05,
+      "loss": 0.0153,
+      "step": 116
+    },
+    {
+      "epoch": 0.9790794979079498,
+      "grad_norm": 0.019681984558701515,
+      "learning_rate": 9.722525205885825e-05,
+      "loss": 0.0149,
+      "step": 117
+    },
+    {
+      "epoch": 0.9874476987447699,
+      "grad_norm": 0.02032177522778511,
+      "learning_rate": 9.712798947886128e-05,
+      "loss": 0.0148,
+      "step": 118
+    },
+    {
+      "epoch": 0.99581589958159,
+      "grad_norm": 0.021595895290374756,
+      "learning_rate": 9.702910183719671e-05,
+      "loss": 0.0199,
+      "step": 119
+    },
+    {
+      "epoch": 1.00418410041841,
+      "grad_norm": 0.027728823944926262,
+      "learning_rate": 9.692859254369631e-05,
+      "loss": 0.0205,
+      "step": 120
+    },
+    {
+      "epoch": 1.00418410041841,
+      "eval_loss": 0.015181933529675007,
+      "eval_runtime": 8.1872,
+      "eval_samples_per_second": 6.107,
+      "eval_steps_per_second": 1.588,
+      "step": 120
+    },
+    {
+      "epoch": 1.0125523012552302,
+      "grad_norm": 0.017031384631991386,
+      "learning_rate": 9.682646506410943e-05,
+      "loss": 0.0152,
+      "step": 121
+    },
+    {
+      "epoch": 1.0209205020920502,
+      "grad_norm": 0.01707589253783226,
+      "learning_rate": 9.672272291998355e-05,
+      "loss": 0.0127,
+      "step": 122
+    },
+    {
+      "epoch": 1.0292887029288702,
+      "grad_norm": 0.02119339071214199,
+      "learning_rate": 9.661736968854283e-05,
+      "loss": 0.0143,
+      "step": 123
+    },
+    {
+      "epoch": 1.0376569037656904,
+      "grad_norm": 0.01858825981616974,
+      "learning_rate": 9.651040900256471e-05,
+      "loss": 0.0145,
+      "step": 124
+    },
+    {
+      "epoch": 1.0460251046025104,
+      "grad_norm": 0.013188479468226433,
+      "learning_rate": 9.640184455025471e-05,
+      "loss": 0.0105,
+      "step": 125
+    },
+    {
+      "epoch": 1.0460251046025104,
+      "eval_loss": 0.015529554337263107,
+      "eval_runtime": 8.1918,
+      "eval_samples_per_second": 6.104,
+      "eval_steps_per_second": 1.587,
+      "step": 125
+    },
+    {
+      "epoch": 1.0543933054393306,
+      "grad_norm": 0.01512962393462658,
+      "learning_rate": 9.629168007511924e-05,
+      "loss": 0.0121,
+      "step": 126
+    },
+    {
+      "epoch": 1.0627615062761506,
+      "grad_norm": 0.018388476222753525,
+      "learning_rate": 9.617991937583648e-05,
+      "loss": 0.0154,
+      "step": 127
+    },
+    {
+      "epoch": 1.0711297071129706,
+      "grad_norm": 0.018059156835079193,
+      "learning_rate": 9.60665663061254e-05,
+      "loss": 0.0163,
+      "step": 128
+    },
+    {
+      "epoch": 1.0794979079497908,
+      "grad_norm": 0.015875034034252167,
+      "learning_rate": 9.595162477461298e-05,
+      "loss": 0.0126,
+      "step": 129
+    },
+    {
+      "epoch": 1.0878661087866108,
+      "grad_norm": 0.01882511004805565,
+      "learning_rate": 9.583509874469923e-05,
+      "loss": 0.0147,
+      "step": 130
+    },
+    {
+      "epoch": 1.0878661087866108,
+      "eval_loss": 0.015730947256088257,
+      "eval_runtime": 8.1997,
+      "eval_samples_per_second": 6.098,
+      "eval_steps_per_second": 1.585,
+      "step": 130
+    },
+    {
+      "epoch": 1.096234309623431,
+      "grad_norm": 0.01482024509459734,
+      "learning_rate": 9.571699223442074e-05,
+      "loss": 0.0129,
+      "step": 131
+    },
+    {
+      "epoch": 1.104602510460251,
+      "grad_norm": 0.017418276518583298,
+      "learning_rate": 9.559730931631198e-05,
+      "loss": 0.0137,
+      "step": 132
+    },
+    {
+      "epoch": 1.112970711297071,
+      "grad_norm": 0.017272206023335457,
+      "learning_rate": 9.5476054117265e-05,
+      "loss": 0.0119,
+      "step": 133
+    },
+    {
+      "epoch": 1.1213389121338913,
+      "grad_norm": 0.02229718305170536,
+      "learning_rate": 9.535323081838691e-05,
+      "loss": 0.0153,
+      "step": 134
+    },
+    {
+      "epoch": 1.1297071129707112,
+      "grad_norm": 0.021549543365836143,
+      "learning_rate": 9.522884365485598e-05,
+      "loss": 0.0148,
+      "step": 135
+    },
+    {
+      "epoch": 1.1297071129707112,
+      "eval_loss": 0.015986260026693344,
+      "eval_runtime": 8.1904,
+      "eval_samples_per_second": 6.105,
+      "eval_steps_per_second": 1.587,
+      "step": 135
+    },
+    {
+      "epoch": 1.1380753138075315,
+      "grad_norm": 0.01844065822660923,
+      "learning_rate": 9.510289691577544e-05,
+      "loss": 0.0138,
+      "step": 136
+    },
+    {
+      "epoch": 1.1464435146443515,
+      "grad_norm": 0.022018112242221832,
+      "learning_rate": 9.49753949440256e-05,
+      "loss": 0.017,
+      "step": 137
+    },
+    {
+      "epoch": 1.1548117154811715,
+      "grad_norm": 0.015306939370930195,
+      "learning_rate": 9.484634213611411e-05,
+      "loss": 0.0114,
+      "step": 138
+    },
+    {
+      "epoch": 1.1631799163179917,
+      "grad_norm": 0.017746033146977425,
+      "learning_rate": 9.471574294202439e-05,
+      "loss": 0.0132,
+      "step": 139
+    },
+    {
+      "epoch": 1.1715481171548117,
+      "grad_norm": 0.01575053483247757,
+      "learning_rate": 9.458360186506213e-05,
+      "loss": 0.0115,
+      "step": 140
+    },
+    {
+      "epoch": 1.1715481171548117,
+      "eval_loss": 0.015280088409781456,
+      "eval_runtime": 8.1915,
+      "eval_samples_per_second": 6.104,
+      "eval_steps_per_second": 1.587,
+      "step": 140
+    },
+    {
+      "epoch": 1.1799163179916319,
+      "grad_norm": 0.019290335476398468,
+      "learning_rate": 9.444992346170007e-05,
+      "loss": 0.0131,
+      "step": 141
+    },
+    {
+      "epoch": 1.1882845188284519,
+      "grad_norm": 0.0183683130890131,
+      "learning_rate": 9.431471234142087e-05,
+      "loss": 0.0136,
+      "step": 142
+    },
+    {
+      "epoch": 1.196652719665272,
+      "grad_norm": 0.020060362294316292,
+      "learning_rate": 9.41779731665581e-05,
+      "loss": 0.0149,
+      "step": 143
+    },
+    {
+      "epoch": 1.205020920502092,
+      "grad_norm": 0.017927493900060654,
+      "learning_rate": 9.403971065213555e-05,
+      "loss": 0.0128,
+      "step": 144
+    },
+    {
+      "epoch": 1.213389121338912,
+      "grad_norm": 0.022669749334454536,
+      "learning_rate": 9.389992956570462e-05,
+      "loss": 0.0166,
+      "step": 145
+    },
+    {
+      "epoch": 1.213389121338912,
+      "eval_loss": 0.015301547013223171,
+      "eval_runtime": 8.1792,
+      "eval_samples_per_second": 6.113,
+      "eval_steps_per_second": 1.589,
+      "step": 145
+    },
+    {
+      "epoch": 1.2217573221757323,
+      "grad_norm": 0.022278638556599617,
+      "learning_rate": 9.375863472717993e-05,
+      "loss": 0.0171,
+      "step": 146
+    },
+    {
+      "epoch": 1.2301255230125523,
+      "grad_norm": 0.019489524886012077,
+      "learning_rate": 9.361583100867313e-05,
+      "loss": 0.013,
+      "step": 147
+    },
+    {
+      "epoch": 1.2384937238493725,
+      "grad_norm": 0.018918950110673904,
+      "learning_rate": 9.347152333432485e-05,
+      "loss": 0.0139,
+      "step": 148
+    },
+    {
+      "epoch": 1.2468619246861925,
+      "grad_norm": 0.01829618029296398,
+      "learning_rate": 9.332571668013499e-05,
+      "loss": 0.0147,
+      "step": 149
+    },
+    {
+      "epoch": 1.2552301255230125,
+      "grad_norm": 0.018324939534068108,
+      "learning_rate": 9.317841607379107e-05,
+      "loss": 0.015,
+      "step": 150
+    },
+    {
+      "epoch": 1.2552301255230125,
+      "eval_loss": 0.015605509281158447,
+      "eval_runtime": 8.195,
+      "eval_samples_per_second": 6.101,
+      "eval_steps_per_second": 1.586,
+      "step": 150
+    },
+    {
+      "epoch": 1.2635983263598327,
+      "grad_norm": 0.015184380114078522,
+      "learning_rate": 9.302962659449486e-05,
+      "loss": 0.0125,
+      "step": 151
+    },
+    {
+      "epoch": 1.2719665271966527,
+      "grad_norm": 0.02084585465490818,
+      "learning_rate": 9.287935337278733e-05,
+      "loss": 0.0139,
+      "step": 152
+    },
+    {
+      "epoch": 1.280334728033473,
+      "grad_norm": 0.022940685972571373,
+      "learning_rate": 9.272760159037164e-05,
+      "loss": 0.0168,
+      "step": 153
+    },
+    {
+      "epoch": 1.288702928870293,
+      "grad_norm": 0.014851168729364872,
+      "learning_rate": 9.25743764799345e-05,
+      "loss": 0.0107,
+      "step": 154
+    },
+    {
+      "epoch": 1.297071129707113,
+      "grad_norm": 0.01901235058903694,
+      "learning_rate": 9.241968332496575e-05,
+      "loss": 0.0148,
+      "step": 155
+    },
+    {
+      "epoch": 1.297071129707113,
+      "eval_loss": 0.015689246356487274,
+      "eval_runtime": 8.1747,
+      "eval_samples_per_second": 6.116,
+      "eval_steps_per_second": 1.59,
+      "step": 155
+    },
+    {
+      "epoch": 1.3054393305439331,
+      "grad_norm": 0.019857851788401604,
+      "learning_rate": 9.226352745957616e-05,
+      "loss": 0.0121,
+      "step": 156
+    },
+    {
+      "epoch": 1.3138075313807531,
+      "grad_norm": 0.020181361585855484,
+      "learning_rate": 9.210591426831352e-05,
+      "loss": 0.0138,
+      "step": 157
+    },
+    {
+      "epoch": 1.3221757322175733,
+      "grad_norm": 0.0159543976187706,
+      "learning_rate": 9.194684918597694e-05,
+      "loss": 0.0119,
+      "step": 158
+    },
+    {
+      "epoch": 1.3305439330543933,
+      "grad_norm": 0.015190091915428638,
+      "learning_rate": 9.178633769742945e-05,
+      "loss": 0.0122,
+      "step": 159
+    },
+    {
+      "epoch": 1.3389121338912133,
+      "grad_norm": 0.01558399386703968,
+      "learning_rate": 9.162438533740892e-05,
+      "loss": 0.0112,
+      "step": 160
+    },
+    {
+      "epoch": 1.3389121338912133,
+      "eval_loss": 0.01591303013265133,
+      "eval_runtime": 8.203,
+      "eval_samples_per_second": 6.095,
+      "eval_steps_per_second": 1.585,
+      "step": 160
+    },
+    {
+      "epoch": 1.3472803347280335,
+      "grad_norm": 0.020452240481972694,
+      "learning_rate": 9.146099769033715e-05,
+      "loss": 0.0141,
+      "step": 161
+    },
+    {
+      "epoch": 1.3556485355648535,
+      "grad_norm": 0.01579088531434536,
+      "learning_rate": 9.129618039012736e-05,
+      "loss": 0.0123,
+      "step": 162
+    },
+    {
+      "epoch": 1.3640167364016738,
+      "grad_norm": 0.01603122055530548,
+      "learning_rate": 9.112993911998984e-05,
+      "loss": 0.0142,
+      "step": 163
+    },
+    {
+      "epoch": 1.3723849372384938,
+      "grad_norm": 0.018658515065908432,
+      "learning_rate": 9.096227961223613e-05,
+      "loss": 0.0149,
+      "step": 164
+    },
+    {
+      "epoch": 1.3807531380753137,
+      "grad_norm": 0.01548719685524702,
+      "learning_rate": 9.07932076480812e-05,
+      "loss": 0.0128,
+      "step": 165
+    },
+    {
+      "epoch": 1.3807531380753137,
+      "eval_loss": 0.015296611934900284,
+      "eval_runtime": 8.1994,
+      "eval_samples_per_second": 6.098,
+      "eval_steps_per_second": 1.585,
+      "step": 165
+    },
+    {
+      "epoch": 1.389121338912134,
+      "grad_norm": 0.019366098567843437,
+      "learning_rate": 9.06227290574442e-05,
+      "loss": 0.0136,
+      "step": 166
+    },
+    {
+      "epoch": 1.397489539748954,
+      "grad_norm": 0.015139399096369743,
+      "learning_rate": 9.045084971874738e-05,
+      "loss": 0.0111,
+      "step": 167
+    },
+    {
+      "epoch": 1.4058577405857742,
+      "grad_norm": 0.02105526067316532,
+      "learning_rate": 9.027757555871344e-05,
+      "loss": 0.0146,
+      "step": 168
+    },
+    {
+      "epoch": 1.4142259414225942,
+      "grad_norm": 0.020205028355121613,
+      "learning_rate": 9.010291255216117e-05,
+      "loss": 0.0142,
+      "step": 169
+    },
+    {
+      "epoch": 1.4225941422594142,
+      "grad_norm": 0.018409285694360733,
+      "learning_rate": 8.99268667217993e-05,
+      "loss": 0.0125,
+      "step": 170
+    },
+    {
+      "epoch": 1.4225941422594142,
+      "eval_loss": 0.015141828916966915,
+      "eval_runtime": 8.1926,
+      "eval_samples_per_second": 6.103,
+      "eval_steps_per_second": 1.587,
+      "step": 170
+    },
+    {
+      "epoch": 1.4309623430962344,
+      "grad_norm": 0.0186452716588974,
+      "learning_rate": 8.974944413801906e-05,
+      "loss": 0.0158,
+      "step": 171
+    },
+    {
+      "epoch": 1.4393305439330544,
+      "grad_norm": 0.017251404002308846,
+      "learning_rate": 8.957065091868459e-05,
+      "loss": 0.0138,
+      "step": 172
+    },
+    {
+      "epoch": 1.4476987447698746,
+      "grad_norm": 0.015452460385859013,
+      "learning_rate": 8.939049322892223e-05,
+      "loss": 0.0104,
+      "step": 173
+    },
+    {
+      "epoch": 1.4560669456066946,
+      "grad_norm": 0.018598852679133415,
+      "learning_rate": 8.920897728090777e-05,
+      "loss": 0.0142,
+      "step": 174
+    },
+    {
+      "epoch": 1.4644351464435146,
+      "grad_norm": 0.017056932672858238,
+      "learning_rate": 8.902610933365229e-05,
+      "loss": 0.0137,
+      "step": 175
+    },
+    {
+      "epoch": 1.4644351464435146,
+      "eval_loss": 0.015005652792751789,
+      "eval_runtime": 8.1766,
+      "eval_samples_per_second": 6.115,
+      "eval_steps_per_second": 1.59,
+      "step": 175
+    },
+    {
+      "epoch": 1.4728033472803348,
+      "grad_norm": 0.017095167189836502,
+      "learning_rate": 8.88418956927864e-05,
+      "loss": 0.0143,
+      "step": 176
+    },
+    {
+      "epoch": 1.4811715481171548,
+      "grad_norm": 0.016802068799734116,
+      "learning_rate": 8.865634271034268e-05,
+      "loss": 0.0144,
+      "step": 177
+    },
+    {
+      "epoch": 1.489539748953975,
+      "grad_norm": 0.015921777114272118,
+      "learning_rate": 8.846945678453677e-05,
+      "loss": 0.0138,
+      "step": 178
+    },
+    {
+      "epoch": 1.497907949790795,
+      "grad_norm": 0.016825497150421143,
+      "learning_rate": 8.828124435954672e-05,
+      "loss": 0.0147,
+      "step": 179
+    },
+    {
+      "epoch": 1.506276150627615,
+      "grad_norm": 0.01446506567299366,
+      "learning_rate": 8.809171192529073e-05,
+      "loss": 0.0131,
+      "step": 180
+    },
+    {
+      "epoch": 1.506276150627615,
+      "eval_loss": 0.014517219737172127,
+      "eval_runtime": 8.201,
+      "eval_samples_per_second": 6.097,
+      "eval_steps_per_second": 1.585,
+      "step": 180
+    },
+    {
+      "epoch": 1.514644351464435,
+      "grad_norm": 0.016049686819314957,
+      "learning_rate": 8.790086601720339e-05,
+      "loss": 0.0137,
+      "step": 181
+    },
+    {
+      "epoch": 1.5230125523012552,
+      "grad_norm": 0.015877775847911835,
+      "learning_rate": 8.77087132160104e-05,
+      "loss": 0.0126,
+      "step": 182
+    },
+    {
+      "epoch": 1.5313807531380754,
+      "grad_norm": 0.01960177719593048,
+      "learning_rate": 8.751526014750153e-05,
+      "loss": 0.0149,
+      "step": 183
+    },
+    {
+      "epoch": 1.5397489539748954,
+      "grad_norm": 0.0134076913818717,
+      "learning_rate": 8.732051348230223e-05,
+      "loss": 0.0114,
+      "step": 184
+    },
+    {
+      "epoch": 1.5481171548117154,
+      "grad_norm": 0.012596129439771175,
+      "learning_rate": 8.712447993564361e-05,
+      "loss": 0.0105,
+      "step": 185
+    },
+    {
+      "epoch": 1.5481171548117154,
+      "eval_loss": 0.014493227005004883,
+      "eval_runtime": 8.1918,
+      "eval_samples_per_second": 6.104,
+      "eval_steps_per_second": 1.587,
+      "step": 185
+    },
+    {
+      "epoch": 1.5564853556485354,
+      "grad_norm": 0.016738103702664375,
+      "learning_rate": 8.692716626713084e-05,
+      "loss": 0.0133,
+      "step": 186
+    },
+    {
+      "epoch": 1.5648535564853556,
+      "grad_norm": 0.02182360552251339,
+      "learning_rate": 8.672857928051013e-05,
+      "loss": 0.014,
+      "step": 187
+    },
+    {
+      "epoch": 1.5732217573221758,
+      "grad_norm": 0.017175616696476936,
+      "learning_rate": 8.652872582343408e-05,
+      "loss": 0.011,
+      "step": 188
+    },
+    {
+      "epoch": 1.5815899581589958,
+      "grad_norm": 0.01974942907691002,
+      "learning_rate": 8.632761278722556e-05,
+      "loss": 0.0147,
+      "step": 189
+    },
+    {
+      "epoch": 1.5899581589958158,
+      "grad_norm": 0.016432059928774834,
+      "learning_rate": 8.612524710664012e-05,
+      "loss": 0.0126,
+      "step": 190
+    },
+    {
+      "epoch": 1.5899581589958158,
+      "eval_loss": 0.014403114095330238,
+      "eval_runtime": 8.1978,
+      "eval_samples_per_second": 6.099,
+      "eval_steps_per_second": 1.586,
+      "step": 190
+    },
+    {
+      "epoch": 1.5983263598326358,
+      "grad_norm": 0.01811247691512108,
+      "learning_rate": 8.592163575962682e-05,
+      "loss": 0.0128,
+      "step": 191
+    },
+    {
+      "epoch": 1.606694560669456,
+      "grad_norm": 0.017857994884252548,
+      "learning_rate": 8.571678576708766e-05,
+      "loss": 0.0117,
+      "step": 192
+    },
+    {
+      "epoch": 1.6150627615062763,
+      "grad_norm": 0.02201293222606182,
+      "learning_rate": 8.551070419263545e-05,
+      "loss": 0.017,
+      "step": 193
+    },
+    {
+      "epoch": 1.6234309623430963,
+      "grad_norm": 0.018607452511787415,
+      "learning_rate": 8.530339814235027e-05,
+      "loss": 0.0123,
+      "step": 194
+    },
+    {
+      "epoch": 1.6317991631799162,
+      "grad_norm": 0.015770630910992622,
+      "learning_rate": 8.509487476453442e-05,
+      "loss": 0.0119,
+      "step": 195
+    },
+    {
+      "epoch": 1.6317991631799162,
+      "eval_loss": 0.014461501501500607,
+      "eval_runtime": 8.1846,
+      "eval_samples_per_second": 6.109,
+      "eval_steps_per_second": 1.588,
+      "step": 195
+    },
+    {
+      "epoch": 1.6401673640167362,
+      "grad_norm": 0.015161648392677307,
+      "learning_rate": 8.488514124946594e-05,
+      "loss": 0.0137,
+      "step": 196
+    },
+    {
+      "epoch": 1.6485355648535565,
+      "grad_norm": 0.015444417484104633,
+      "learning_rate": 8.46742048291507e-05,
+      "loss": 0.0116,
+      "step": 197
+    },
+    {
+      "epoch": 1.6569037656903767,
+      "grad_norm": 0.018221288919448853,
+      "learning_rate": 8.446207277707296e-05,
+      "loss": 0.0135,
+      "step": 198
+    },
+    {
+      "epoch": 1.6652719665271967,
+      "grad_norm": 0.017886707559227943,
+      "learning_rate": 8.424875240794467e-05,
+      "loss": 0.0117,
+      "step": 199
+    },
+    {
+      "epoch": 1.6736401673640167,
+      "grad_norm": 0.021269524469971657,
+      "learning_rate": 8.403425107745316e-05,
+      "loss": 0.016,
+      "step": 200
+    },
+    {
+      "epoch": 1.6736401673640167,
+      "eval_loss": 0.014672570861876011,
+      "eval_runtime": 8.2192,
+      "eval_samples_per_second": 6.083,
+      "eval_steps_per_second": 1.582,
+      "step": 200
+    },
+    {
+      "epoch": 1.6820083682008367,
+      "grad_norm": 0.014137690886855125,
+      "learning_rate": 8.381857618200746e-05,
+      "loss": 0.0107,
+      "step": 201
+    },
+    {
+      "epoch": 1.6903765690376569,
+      "grad_norm": 0.017183968797326088,
+      "learning_rate": 8.360173515848343e-05,
+      "loss": 0.0127,
+      "step": 202
+    },
+    {
+      "epoch": 1.698744769874477,
+      "grad_norm": 0.018049761652946472,
+      "learning_rate": 8.33837354839672e-05,
+      "loss": 0.0099,
+      "step": 203
+    },
+    {
+      "epoch": 1.707112970711297,
+      "grad_norm": 0.020697997882962227,
+      "learning_rate": 8.316458467549729e-05,
+      "loss": 0.0131,
+      "step": 204
+    },
+    {
+      "epoch": 1.715481171548117,
+      "grad_norm": 0.022819431498646736,
+      "learning_rate": 8.294429028980556e-05,
+      "loss": 0.0143,
+      "step": 205
+    },
+    {
+      "epoch": 1.715481171548117,
+      "eval_loss": 0.014957732520997524,
+      "eval_runtime": 8.2053,
+      "eval_samples_per_second": 6.094,
+      "eval_steps_per_second": 1.584,
+      "step": 205
+    },
+    {
+      "epoch": 1.723849372384937,
+      "grad_norm": 0.020023057237267494,
+      "learning_rate": 8.272285992305653e-05,
+      "loss": 0.014,
+      "step": 206
+    },
+    {
+      "epoch": 1.7322175732217573,
+      "grad_norm": 0.01683293841779232,
+      "learning_rate": 8.250030121058547e-05,
+      "loss": 0.0112,
+      "step": 207
+    },
+    {
+      "epoch": 1.7405857740585775,
+      "grad_norm": 0.01992541179060936,
+      "learning_rate": 8.227662182663518e-05,
+      "loss": 0.0129,
+      "step": 208
+    },
+    {
+      "epoch": 1.7489539748953975,
+      "grad_norm": 0.01862913742661476,
+      "learning_rate": 8.205182948409125e-05,
+      "loss": 0.0131,
+      "step": 209
+    },
+    {
+      "epoch": 1.7573221757322175,
+      "grad_norm": 0.01816224865615368,
+      "learning_rate": 8.182593193421625e-05,
+      "loss": 0.0139,
+      "step": 210
+    },
+    {
+      "epoch": 1.7573221757322175,
+      "eval_loss": 0.014965851791203022,
+      "eval_runtime": 8.2076,
+      "eval_samples_per_second": 6.092,
+      "eval_steps_per_second": 1.584,
+      "step": 210
+    },
+    {
+      "epoch": 1.7656903765690377,
+      "grad_norm": 0.015293323434889317,
+      "learning_rate": 8.159893696638232e-05,
+      "loss": 0.0122,
+      "step": 211
+    },
+    {
+      "epoch": 1.7740585774058577,
+      "grad_norm": 0.013619404286146164,
+      "learning_rate": 8.137085240780262e-05,
+      "loss": 0.0107,
+      "step": 212
+    },
+    {
+      "epoch": 1.782426778242678,
+      "grad_norm": 0.015562871471047401,
+      "learning_rate": 8.114168612326153e-05,
+      "loss": 0.0106,
+      "step": 213
+    },
+    {
+      "epoch": 1.790794979079498,
+      "grad_norm": 0.017086800187826157,
+      "learning_rate": 8.091144601484332e-05,
+      "loss": 0.0131,
+      "step": 214
+    },
+    {
+      "epoch": 1.799163179916318,
+      "grad_norm": 0.020527249202132225,
+      "learning_rate": 8.06801400216597e-05,
+      "loss": 0.0139,
+      "step": 215
+    },
+    {
+      "epoch": 1.799163179916318,
+      "eval_loss": 0.01450226828455925,
+      "eval_runtime": 8.1927,
+      "eval_samples_per_second": 6.103,
+      "eval_steps_per_second": 1.587,
+      "step": 215
+    },
+    {
+      "epoch": 1.8075313807531381,
+      "grad_norm": 0.015211045742034912,
+      "learning_rate": 8.044777611957613e-05,
+      "loss": 0.0114,
+      "step": 216
+    },
+    {
+      "epoch": 1.8158995815899581,
+      "grad_norm": 0.01953182741999626,
+      "learning_rate": 8.021436232093675e-05,
+      "loss": 0.0136,
+      "step": 217
+    },
+    {
+      "epoch": 1.8242677824267783,
+      "grad_norm": 0.018059900030493736,
+      "learning_rate": 7.997990667428811e-05,
+      "loss": 0.013,
+      "step": 218
+    },
+    {
+      "epoch": 1.8326359832635983,
+      "grad_norm": 0.019347647204995155,
+      "learning_rate": 7.974441726410164e-05,
+      "loss": 0.0142,
+      "step": 219
+    },
+    {
+      "epoch": 1.8410041841004183,
+      "grad_norm": 0.016715556383132935,
+      "learning_rate": 7.950790221049484e-05,
+      "loss": 0.0161,
+      "step": 220
+    },
+    {
+      "epoch": 1.8410041841004183,
+      "eval_loss": 0.014255787245929241,
+      "eval_runtime": 8.1832,
+      "eval_samples_per_second": 6.11,
+      "eval_steps_per_second": 1.589,
+      "step": 220
+    },
+    {
+      "epoch": 1.8493723849372385,
+      "grad_norm": 0.015245197340846062,
+      "learning_rate": 7.927036966895138e-05,
+      "loss": 0.0117,
+      "step": 221
+    },
+    {
+      "epoch": 1.8577405857740585,
+      "grad_norm": 0.016146808862686157,
+      "learning_rate": 7.90318278300398e-05,
+      "loss": 0.0132,
+      "step": 222
+    },
+    {
+      "epoch": 1.8661087866108788,
+      "grad_norm": 0.01640206389129162,
+      "learning_rate": 7.879228491913111e-05,
+      "loss": 0.012,
+      "step": 223
+    },
+    {
+      "epoch": 1.8744769874476988,
+      "grad_norm": 0.01924493908882141,
+      "learning_rate": 7.855174919611517e-05,
+      "loss": 0.0135,
+      "step": 224
+    },
+    {
+      "epoch": 1.8828451882845187,
+      "grad_norm": 0.01706533320248127,
+      "learning_rate": 7.831022895511587e-05,
+      "loss": 0.0098,
+      "step": 225
+    },
+    {
+      "epoch": 1.8828451882845187,
+      "eval_loss": 0.013838106766343117,
+      "eval_runtime": 8.1961,
+      "eval_samples_per_second": 6.1,
+      "eval_steps_per_second": 1.586,
+      "step": 225
+    },
+    {
+      "epoch": 1.891213389121339,
+      "grad_norm": 0.01932067796587944,
+      "learning_rate": 7.80677325242051e-05,
+      "loss": 0.0144,
+      "step": 226
+    },
+    {
+      "epoch": 1.899581589958159,
+      "grad_norm": 0.018316643312573433,
+      "learning_rate": 7.782426826511564e-05,
+      "loss": 0.0124,
+      "step": 227
+    },
+    {
+      "epoch": 1.9079497907949792,
+      "grad_norm": 0.016697222366929054,
+      "learning_rate": 7.757984457295284e-05,
+      "loss": 0.0125,
+      "step": 228
+    },
+    {
+      "epoch": 1.9163179916317992,
+      "grad_norm": 0.014732821844518185,
+      "learning_rate": 7.733446987590505e-05,
+      "loss": 0.0115,
+      "step": 229
+    },
+    {
+      "epoch": 1.9246861924686192,
+      "grad_norm": 0.01892097480595112,
+      "learning_rate": 7.708815263495308e-05,
+      "loss": 0.0108,
+      "step": 230
+    },
+    {
+      "epoch": 1.9246861924686192,
+      "eval_loss": 0.014023682102560997,
+      "eval_runtime": 8.1828,
+      "eval_samples_per_second": 6.11,
+      "eval_steps_per_second": 1.589,
+      "step": 230
+    },
+    {
+      "epoch": 1.9330543933054394,
+      "grad_norm": 0.018761448562145233,
+      "learning_rate": 7.684090134357839e-05,
+      "loss": 0.014,
+      "step": 231
+    },
+    {
+      "epoch": 1.9414225941422594,
+      "grad_norm": 0.01908651366829872,
+      "learning_rate": 7.65927245274703e-05,
+      "loss": 0.0102,
+      "step": 232
+    },
+    {
+      "epoch": 1.9497907949790796,
+      "grad_norm": 0.018542446196079254,
+      "learning_rate": 7.634363074423199e-05,
+      "loss": 0.0133,
+      "step": 233
+    },
+    {
+      "epoch": 1.9581589958158996,
+      "grad_norm": 0.018298540264368057,
+      "learning_rate": 7.609362858308531e-05,
+      "loss": 0.0139,
+      "step": 234
+    },
+    {
+      "epoch": 1.9665271966527196,
+      "grad_norm": 0.01675599068403244,
+      "learning_rate": 7.58427266645747e-05,
+      "loss": 0.0117,
+      "step": 235
+    },
+    {
+      "epoch": 1.9665271966527196,
+      "eval_loss": 0.014116828329861164,
+      "eval_runtime": 8.1955,
+      "eval_samples_per_second": 6.101,
+      "eval_steps_per_second": 1.586,
+      "step": 235
+    },
+    {
+      "epoch": 1.9748953974895398,
+      "grad_norm": 0.020569469779729843,
+      "learning_rate": 7.559093364027001e-05,
+      "loss": 0.0141,
+      "step": 236
+    },
+    {
+      "epoch": 1.9832635983263598,
+      "grad_norm": 0.016174696385860443,
+      "learning_rate": 7.5338258192468e-05,
+      "loss": 0.0105,
+      "step": 237
+    },
+    {
+      "epoch": 1.99163179916318,
+      "grad_norm": 0.02071259915828705,
+      "learning_rate": 7.508470903389311e-05,
+      "loss": 0.0138,
+      "step": 238
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 0.022858325392007828,
+      "learning_rate": 7.48302949073969e-05,
+      "loss": 0.0166,
+      "step": 239
+    },
+    {
+      "epoch": 2.00836820083682,
+      "grad_norm": 0.01647881604731083,
+      "learning_rate": 7.457502458565672e-05,
+      "loss": 0.0109,
+      "step": 240
+    },
+    {
+      "epoch": 2.00836820083682,
+      "eval_loss": 0.013845333829522133,
+      "eval_runtime": 8.1876,
+      "eval_samples_per_second": 6.107,
+      "eval_steps_per_second": 1.588,
+      "step": 240
+    },
+    {
+      "epoch": 2.01673640167364,
+      "grad_norm": 0.01600758358836174,
+      "learning_rate": 7.431890687087313e-05,
+      "loss": 0.0102,
+      "step": 241
+    },
+    {
+      "epoch": 2.0251046025104604,
+      "grad_norm": 0.013506948947906494,
+      "learning_rate": 7.406195059446636e-05,
+      "loss": 0.0103,
+      "step": 242
+    },
+    {
+      "epoch": 2.0334728033472804,
+      "grad_norm": 0.019151534885168076,
+      "learning_rate": 7.380416461677187e-05,
+      "loss": 0.0119,
+      "step": 243
+    },
+    {
+      "epoch": 2.0418410041841004,
+      "grad_norm": 0.01597500592470169,
+      "learning_rate": 7.354555782673473e-05,
+      "loss": 0.0107,
+      "step": 244
+    },
+    {
+      "epoch": 2.0502092050209204,
+      "grad_norm": 0.014115000143647194,
+      "learning_rate": 7.328613914160318e-05,
+      "loss": 0.0093,
+      "step": 245
+    },
+    {
+      "epoch": 2.0502092050209204,
+      "eval_loss": 0.014546514488756657,
+      "eval_runtime": 8.1884,
+      "eval_samples_per_second": 6.106,
+      "eval_steps_per_second": 1.588,
+      "step": 245
+    },
+    {
+      "epoch": 2.0585774058577404,
+      "grad_norm": 0.01916162669658661,
+      "learning_rate": 7.302591750662116e-05,
+      "loss": 0.0117,
+      "step": 246
+    },
+    {
+      "epoch": 2.066945606694561,
+      "grad_norm": 0.019530242308974266,
+      "learning_rate": 7.276490189471977e-05,
+      "loss": 0.01,
+      "step": 247
+    },
+    {
+      "epoch": 2.075313807531381,
+      "grad_norm": 0.02205420657992363,
+      "learning_rate": 7.250310130620799e-05,
+      "loss": 0.0111,
+      "step": 248
+    },
+    {
+      "epoch": 2.083682008368201,
+      "grad_norm": 0.025347299873828888,
+      "learning_rate": 7.224052476846222e-05,
+      "loss": 0.0115,
+      "step": 249
+    },
+    {
+      "epoch": 2.092050209205021,
+      "grad_norm": 0.018653597682714462,
+      "learning_rate": 7.197718133561508e-05,
+      "loss": 0.0102,
+      "step": 250
+    },
+    {
+      "epoch": 2.092050209205021,
+      "eval_loss": 0.014336950145661831,
+      "eval_runtime": 8.2008,
+      "eval_samples_per_second": 6.097,
+      "eval_steps_per_second": 1.585,
+      "step": 250
+    },
+    {
+      "epoch": 2.100418410041841,
+      "grad_norm": 0.01541041024029255,
+      "learning_rate": 7.171308008824316e-05,
+      "loss": 0.0105,
+      "step": 251
+    },
+    {
+      "epoch": 2.1087866108786613,
+      "grad_norm": 0.014182745479047298,
+      "learning_rate": 7.144823013305394e-05,
+      "loss": 0.0085,
+      "step": 252
+    },
+    {
+      "epoch": 2.1171548117154813,
+      "grad_norm": 0.022554123774170876,
+      "learning_rate": 7.118264060257169e-05,
+      "loss": 0.0116,
+      "step": 253
+    },
+    {
+      "epoch": 2.1255230125523012,
+      "grad_norm": 0.02267191931605339,
+      "learning_rate": 7.091632065482272e-05,
+      "loss": 0.0115,
+      "step": 254
+    },
+    {
+      "epoch": 2.1338912133891212,
+      "grad_norm": 0.01669103093445301,
+      "learning_rate": 7.064927947301943e-05,
+      "loss": 0.0104,
+      "step": 255
+    },
+    {
+      "epoch": 2.1338912133891212,
+      "eval_loss": 0.014133421704173088,
+      "eval_runtime": 8.1833,
+      "eval_samples_per_second": 6.11,
+      "eval_steps_per_second": 1.589,
+      "step": 255
+    },
+    {
+      "epoch": 2.1422594142259412,
+      "grad_norm": 0.01666867360472679,
+      "learning_rate": 7.038152626524373e-05,
+      "loss": 0.0097,
+      "step": 256
+    },
+    {
+      "epoch": 2.1506276150627617,
+      "grad_norm": 0.0202019102871418,
+      "learning_rate": 7.011307026412958e-05,
+      "loss": 0.0106,
+      "step": 257
+    },
+    {
+      "epoch": 2.1589958158995817,
+      "grad_norm": 0.017287522554397583,
+      "learning_rate": 6.984392072654453e-05,
+      "loss": 0.0105,
+      "step": 258
+    },
+    {
+      "epoch": 2.1673640167364017,
+      "grad_norm": 0.014798184856772423,
+      "learning_rate": 6.957408693327055e-05,
+      "loss": 0.0096,
+      "step": 259
+    },
+    {
+      "epoch": 2.1757322175732217,
+      "grad_norm": 0.016632311046123505,
+      "learning_rate": 6.930357818868409e-05,
+      "loss": 0.0108,
+      "step": 260
+    },
+    {
+      "epoch": 2.1757322175732217,
+      "eval_loss": 0.0147059615701437,
+      "eval_runtime": 8.1958,
+      "eval_samples_per_second": 6.101,
+      "eval_steps_per_second": 1.586,
+      "step": 260
+    },
+    {
+      "epoch": 2.1841004184100417,
+      "grad_norm": 0.016555817797780037,
+      "learning_rate": 6.903240382043514e-05,
+      "loss": 0.011,
+      "step": 261
+    },
+    {
+      "epoch": 2.192468619246862,
+      "grad_norm": 0.015979807823896408,
+      "learning_rate": 6.876057317912569e-05,
+      "loss": 0.0108,
+      "step": 262
+    },
+    {
+      "epoch": 2.200836820083682,
+      "grad_norm": 0.014860883355140686,
+      "learning_rate": 6.848809563798721e-05,
+      "loss": 0.0106,
+      "step": 263
+    },
+    {
+      "epoch": 2.209205020920502,
+      "grad_norm": 0.016682744026184082,
+      "learning_rate": 6.821498059255752e-05,
+      "loss": 0.0101,
+      "step": 264
+    },
+    {
+      "epoch": 2.217573221757322,
+      "grad_norm": 0.017817750573158264,
+      "learning_rate": 6.79412374603568e-05,
+      "loss": 0.0104,
+      "step": 265
+    },
+    {
+      "epoch": 2.217573221757322,
+      "eval_loss": 0.014249371364712715,
+      "eval_runtime": 8.1965,
+      "eval_samples_per_second": 6.1,
+      "eval_steps_per_second": 1.586,
+      "step": 265
+    },
+    {
+      "epoch": 2.225941422594142,
+      "grad_norm": 0.014692910015583038,
+      "learning_rate": 6.766687568056281e-05,
+      "loss": 0.0093,
+      "step": 266
+    },
+    {
+      "epoch": 2.2343096234309625,
+      "grad_norm": 0.01461927779018879,
+      "learning_rate": 6.739190471368548e-05,
+      "loss": 0.0099,
+      "step": 267
+    },
+    {
+      "epoch": 2.2426778242677825,
+      "grad_norm": 0.01550222933292389,
+      "learning_rate": 6.711633404124062e-05,
+      "loss": 0.0105,
+      "step": 268
+    },
+    {
+      "epoch": 2.2510460251046025,
+      "grad_norm": 0.01469459943473339,
+      "learning_rate": 6.684017316542302e-05,
+      "loss": 0.01,
+      "step": 269
+    },
+    {
+      "epoch": 2.2594142259414225,
+      "grad_norm": 0.018585605546832085,
+      "learning_rate": 6.656343160877881e-05,
+      "loss": 0.0103,
+      "step": 270
+    },
+    {
+      "epoch": 2.2594142259414225,
+      "eval_loss": 0.014431845396757126,
+      "eval_runtime": 8.1882,
+      "eval_samples_per_second": 6.106,
+      "eval_steps_per_second": 1.588,
+      "step": 270
+    },
+    {
+      "epoch": 2.2677824267782425,
+      "grad_norm": 0.01749185100197792,
+      "learning_rate": 6.628611891387706e-05,
+      "loss": 0.0098,
+      "step": 271
+    },
+    {
+      "epoch": 2.276150627615063,
+      "grad_norm": 0.019719472154974937,
+      "learning_rate": 6.600824464298082e-05,
+      "loss": 0.0103,
+      "step": 272
+    },
+    {
+      "epoch": 2.284518828451883,
+      "grad_norm": 0.019530480727553368,
+      "learning_rate": 6.572981837771726e-05,
+      "loss": 0.0106,
+      "step": 273
+    },
+    {
+      "epoch": 2.292887029288703,
+      "grad_norm": 0.020081467926502228,
+      "learning_rate": 6.545084971874738e-05,
+      "loss": 0.0105,
+      "step": 274
+    },
+    {
+      "epoch": 2.301255230125523,
+      "grad_norm": 0.016815155744552612,
+      "learning_rate": 6.517134828543496e-05,
+      "loss": 0.0107,
+      "step": 275
+    },
+    {
+      "epoch": 2.301255230125523,
+      "eval_loss": 0.01435365341603756,
+      "eval_runtime": 8.1841,
+      "eval_samples_per_second": 6.109,
+      "eval_steps_per_second": 1.588,
+      "step": 275
+    },
+    {
+      "epoch": 2.309623430962343,
+      "grad_norm": 0.02080889418721199,
+      "learning_rate": 6.489132371551482e-05,
+      "loss": 0.0107,
+      "step": 276
+    },
+    {
+      "epoch": 2.3179916317991633,
+      "grad_norm": 0.019815733656287193,
+      "learning_rate": 6.461078566476054e-05,
+      "loss": 0.0108,
+      "step": 277
+    },
+    {
+      "epoch": 2.3263598326359833,
+      "grad_norm": 0.020922286435961723,
+      "learning_rate": 6.43297438066514e-05,
+      "loss": 0.0115,
+      "step": 278
+    },
+    {
+      "epoch": 2.3347280334728033,
+      "grad_norm": 0.01711447723209858,
+      "learning_rate": 6.404820783203901e-05,
+      "loss": 0.0101,
+      "step": 279
+    },
+    {
+      "epoch": 2.3430962343096233,
+      "grad_norm": 0.016599150374531746,
+      "learning_rate": 6.3766187448813e-05,
+      "loss": 0.0104,
+      "step": 280
+    },
+    {
+      "epoch": 2.3430962343096233,
+      "eval_loss": 0.01405297126621008,
+      "eval_runtime": 8.1827,
+      "eval_samples_per_second": 6.11,
+      "eval_steps_per_second": 1.589,
+      "step": 280
+    },
+    {
+      "epoch": 2.3514644351464433,
+      "grad_norm": 0.01762828603386879,
+      "learning_rate": 6.348369238156631e-05,
+      "loss": 0.0099,
+      "step": 281
+    },
+    {
+      "epoch": 2.3598326359832638,
+      "grad_norm": 0.014734679833054543,
+      "learning_rate": 6.32007323712599e-05,
+      "loss": 0.0094,
+      "step": 282
+    },
+    {
+      "epoch": 2.3682008368200838,
+      "grad_norm": 0.01771124079823494,
+      "learning_rate": 6.29173171748868e-05,
+      "loss": 0.0106,
+      "step": 283
+    },
+    {
+      "epoch": 2.3765690376569037,
+      "grad_norm": 0.01341992523521185,
+      "learning_rate": 6.263345656513576e-05,
+      "loss": 0.0096,
+      "step": 284
+    },
+    {
+      "epoch": 2.3849372384937237,
+      "grad_norm": 0.014316578395664692,
+      "learning_rate": 6.234916033005421e-05,
+      "loss": 0.0092,
+      "step": 285
+    },
+    {
+      "epoch": 2.3849372384937237,
+      "eval_loss": 0.014250432141125202,
+      "eval_runtime": 8.1935,
+      "eval_samples_per_second": 6.102,
+      "eval_steps_per_second": 1.587,
+      "step": 285
+    },
+    {
+      "epoch": 2.393305439330544,
+      "grad_norm": 0.017734510824084282,
+      "learning_rate": 6.206443827271073e-05,
+      "loss": 0.011,
+      "step": 286
+    },
+    {
+      "epoch": 2.401673640167364,
+      "grad_norm": 0.016910862177610397,
+      "learning_rate": 6.17793002108571e-05,
+      "loss": 0.0101,
+      "step": 287
+    },
+    {
+      "epoch": 2.410041841004184,
+      "grad_norm": 0.015546726994216442,
+      "learning_rate": 6.149375597658967e-05,
+      "loss": 0.0092,
+      "step": 288
+    },
+    {
+      "epoch": 2.418410041841004,
+      "grad_norm": 0.01636183261871338,
+      "learning_rate": 6.12078154160104e-05,
+      "loss": 0.0105,
+      "step": 289
+    },
+    {
+      "epoch": 2.426778242677824,
+      "grad_norm": 0.01692620851099491,
+      "learning_rate": 6.092148838888732e-05,
+      "loss": 0.0107,
+      "step": 290
+    },
+    {
+      "epoch": 2.426778242677824,
+      "eval_loss": 0.01402355171740055,
+      "eval_runtime": 8.183,
+      "eval_samples_per_second": 6.11,
+      "eval_steps_per_second": 1.589,
+      "step": 290
+    },
+    {
+      "epoch": 2.435146443514644,
+      "grad_norm": 0.01607348956167698,
+      "learning_rate": 6.0634784768314564e-05,
+      "loss": 0.0107,
+      "step": 291
+    },
+    {
+      "epoch": 2.4435146443514646,
+      "grad_norm": 0.019470971077680588,
+      "learning_rate": 6.0347714440371924e-05,
+      "loss": 0.0095,
+      "step": 292
+    },
+    {
+      "epoch": 2.4518828451882846,
+      "grad_norm": 0.014622722752392292,
+      "learning_rate": 6.0060287303783924e-05,
+      "loss": 0.011,
+      "step": 293
+    },
+    {
+      "epoch": 2.4602510460251046,
+      "grad_norm": 0.02121814712882042,
+      "learning_rate": 5.977251326957852e-05,
+      "loss": 0.0099,
+      "step": 294
+    },
+    {
+      "epoch": 2.4686192468619246,
+      "grad_norm": 0.018752707168459892,
+      "learning_rate": 5.9484402260745386e-05,
+      "loss": 0.0112,
+      "step": 295
+    },
+    {
+      "epoch": 2.4686192468619246,
+      "eval_loss": 0.014348245225846767,
+      "eval_runtime": 8.19,
+      "eval_samples_per_second": 6.105,
+      "eval_steps_per_second": 1.587,
+      "step": 295
+    },
+    {
+      "epoch": 2.476987447698745,
+      "grad_norm": 0.016945000737905502,
+      "learning_rate": 5.919596421189368e-05,
+      "loss": 0.0096,
+      "step": 296
+    },
+    {
+      "epoch": 2.485355648535565,
+      "grad_norm": 0.0150624830275774,
+      "learning_rate": 5.8907209068909554e-05,
+      "loss": 0.0085,
+      "step": 297
+    },
+    {
+      "epoch": 2.493723849372385,
+      "grad_norm": 0.01675923727452755,
+      "learning_rate": 5.8618146788613105e-05,
+      "loss": 0.0101,
+      "step": 298
+    },
+    {
+      "epoch": 2.502092050209205,
+      "grad_norm": 0.01698216423392296,
+      "learning_rate": 5.832878733841513e-05,
+      "loss": 0.0102,
+      "step": 299
+    },
+    {
+      "epoch": 2.510460251046025,
+      "grad_norm": 0.01541865710169077,
+      "learning_rate": 5.803914069597342e-05,
+      "loss": 0.01,
+      "step": 300
+    },
+    {
+      "epoch": 2.510460251046025,
+      "eval_loss": 0.014326265081763268,
+      "eval_runtime": 8.1855,
+      "eval_samples_per_second": 6.108,
+      "eval_steps_per_second": 1.588,
+      "step": 300
+    },
+    {
+      "epoch": 2.518828451882845,
+      "grad_norm": 0.015604431740939617,
+      "learning_rate": 5.7749216848848664e-05,
+      "loss": 0.0092,
+      "step": 301
+    },
+    {
+      "epoch": 2.5271966527196654,
+      "grad_norm": 0.019720997661352158,
+      "learning_rate": 5.7459025794160136e-05,
+      "loss": 0.0109,
+      "step": 302
+    },
+    {
+      "epoch": 2.5355648535564854,
+      "grad_norm": 0.014376216568052769,
+      "learning_rate": 5.716857753824086e-05,
+      "loss": 0.0089,
+      "step": 303
+    },
+    {
+      "epoch": 2.5439330543933054,
+      "grad_norm": 0.015419202856719494,
+      "learning_rate": 5.6877882096292715e-05,
+      "loss": 0.0095,
+      "step": 304
+    },
+    {
+      "epoch": 2.5523012552301254,
+      "grad_norm": 0.018142806366086006,
+      "learning_rate": 5.6586949492040944e-05,
+      "loss": 0.0096,
+      "step": 305
+    },
+    {
+      "epoch": 2.5523012552301254,
+      "eval_loss": 0.01382070779800415,
+      "eval_runtime": 8.2064,
+      "eval_samples_per_second": 6.093,
+      "eval_steps_per_second": 1.584,
+      "step": 305
+    },
+    {
+      "epoch": 2.560669456066946,
+      "grad_norm": 0.016867833212018013,
+      "learning_rate": 5.629578975738865e-05,
+      "loss": 0.0099,
+      "step": 306
+    },
+    {
+      "epoch": 2.569037656903766,
+      "grad_norm": 0.01875397376716137,
+      "learning_rate": 5.6004412932070835e-05,
+      "loss": 0.0104,
+      "step": 307
+    },
+    {
+      "epoch": 2.577405857740586,
+      "grad_norm": 0.018324118107557297,
+      "learning_rate": 5.57128290633081e-05,
+      "loss": 0.0096,
+      "step": 308
+    },
+    {
+      "epoch": 2.585774058577406,
+      "grad_norm": 0.014931274577975273,
+      "learning_rate": 5.542104820546041e-05,
+      "loss": 0.0095,
+      "step": 309
+    },
+    {
+      "epoch": 2.594142259414226,
+      "grad_norm": 0.020742323249578476,
+      "learning_rate": 5.512908041968018e-05,
+      "loss": 0.0096,
+      "step": 310
+    },
+    {
+      "epoch": 2.594142259414226,
+      "eval_loss": 0.013671735301613808,
+      "eval_runtime": 8.1946,
+      "eval_samples_per_second": 6.102,
+      "eval_steps_per_second": 1.586,
+      "step": 310
+    },
+    {
+      "epoch": 2.602510460251046,
+      "grad_norm": 0.016727490350604057,
+      "learning_rate": 5.483693577356552e-05,
+      "loss": 0.01,
+      "step": 311
+    },
+    {
+      "epoch": 2.6108786610878663,
+      "grad_norm": 0.015919320285320282,
+      "learning_rate": 5.454462434081302e-05,
+      "loss": 0.0099,
+      "step": 312
+    },
+    {
+      "epoch": 2.6192468619246863,
+      "grad_norm": 0.01895716041326523,
+      "learning_rate": 5.425215620087033e-05,
+      "loss": 0.0097,
+      "step": 313
+    },
+    {
+      "epoch": 2.6276150627615062,
+      "grad_norm": 0.0199336726218462,
+      "learning_rate": 5.395954143858871e-05,
+      "loss": 0.0108,
+      "step": 314
+    },
+    {
+      "epoch": 2.6359832635983262,
+      "grad_norm": 0.016174420714378357,
+      "learning_rate": 5.36667901438752e-05,
+      "loss": 0.0099,
+      "step": 315
+    },
+    {
+      "epoch": 2.6359832635983262,
+      "eval_loss": 0.013724744319915771,
+      "eval_runtime": 8.2179,
+      "eval_samples_per_second": 6.084,
+      "eval_steps_per_second": 1.582,
+      "step": 315
+    },
+    {
+      "epoch": 2.6443514644351467,
+      "grad_norm": 0.0191593486815691,
+      "learning_rate": 5.337391241134476e-05,
+      "loss": 0.0104,
+      "step": 316
+    },
+    {
+      "epoch": 2.6527196652719667,
+      "grad_norm": 0.017469942569732666,
+      "learning_rate": 5.3080918339972175e-05,
+      "loss": 0.0101,
+      "step": 317
+    },
+    {
+      "epoch": 2.6610878661087867,
+      "grad_norm": 0.024253906682133675,
+      "learning_rate": 5.278781803274375e-05,
+      "loss": 0.0109,
+      "step": 318
+    },
+    {
+      "epoch": 2.6694560669456067,
+      "grad_norm": 0.020112276077270508,
+      "learning_rate": 5.249462159630908e-05,
+      "loss": 0.0105,
+      "step": 319
+    },
+    {
+      "epoch": 2.6778242677824267,
+      "grad_norm": 0.017300646752119064,
+      "learning_rate": 5.220133914063239e-05,
+      "loss": 0.009,
+      "step": 320
+    },
+    {
+      "epoch": 2.6778242677824267,
+      "eval_loss": 0.013836627826094627,
+      "eval_runtime": 8.1921,
+      "eval_samples_per_second": 6.103,
+      "eval_steps_per_second": 1.587,
+      "step": 320
+    },
+    {
+      "epoch": 2.6861924686192467,
+      "grad_norm": 0.01683647558093071,
+      "learning_rate": 5.1907980778644114e-05,
+      "loss": 0.0097,
+      "step": 321
+    },
+    {
+      "epoch": 2.694560669456067,
+      "grad_norm": 0.02241523563861847,
+      "learning_rate": 5.161455662589201e-05,
+      "loss": 0.0112,
+      "step": 322
+    },
+    {
+      "epoch": 2.702928870292887,
+      "grad_norm": 0.017464108765125275,
+      "learning_rate": 5.132107680019241e-05,
+      "loss": 0.0106,
+      "step": 323
+    },
+    {
+      "epoch": 2.711297071129707,
+      "grad_norm": 0.014207069762051105,
+      "learning_rate": 5.102755142128143e-05,
+      "loss": 0.0092,
+      "step": 324
+    },
+    {
+      "epoch": 2.719665271966527,
+      "grad_norm": 0.01582857221364975,
+      "learning_rate": 5.073399061046583e-05,
+      "loss": 0.0097,
+      "step": 325
+    },
+    {
+      "epoch": 2.719665271966527,
+      "eval_loss": 0.01370433159172535,
+      "eval_runtime": 8.1877,
+      "eval_samples_per_second": 6.107,
+      "eval_steps_per_second": 1.588,
+      "step": 325
+    },
+    {
+      "epoch": 2.7280334728033475,
+      "grad_norm": 0.015251656994223595,
+      "learning_rate": 5.044040449027426e-05,
+      "loss": 0.0091,
+      "step": 326
+    },
+    {
+      "epoch": 2.7364016736401675,
+      "grad_norm": 0.01971331425011158,
+      "learning_rate": 5.0146803184108006e-05,
+      "loss": 0.0097,
+      "step": 327
+    },
+    {
+      "epoch": 2.7447698744769875,
+      "grad_norm": 0.017107218503952026,
+      "learning_rate": 4.9853196815892e-05,
+      "loss": 0.0105,
+      "step": 328
+    },
+    {
+      "epoch": 2.7531380753138075,
+      "grad_norm": 0.016256961971521378,
+      "learning_rate": 4.9559595509725744e-05,
+      "loss": 0.0091,
+      "step": 329
+    },
+    {
+      "epoch": 2.7615062761506275,
+      "grad_norm": 0.016966141760349274,
+      "learning_rate": 4.926600938953418e-05,
+      "loss": 0.0097,
+      "step": 330
+    },
+    {
+      "epoch": 2.7615062761506275,
+      "eval_loss": 0.01360410638153553,
+      "eval_runtime": 8.2179,
+      "eval_samples_per_second": 6.084,
+      "eval_steps_per_second": 1.582,
+      "step": 330
+    },
+    {
+      "epoch": 2.7698744769874475,
+      "grad_norm": 0.016130153089761734,
+      "learning_rate": 4.89724485787186e-05,
+      "loss": 0.0094,
+      "step": 331
+    },
+    {
+      "epoch": 2.778242677824268,
+      "grad_norm": 0.016650136560201645,
+      "learning_rate": 4.86789231998076e-05,
+      "loss": 0.0092,
+      "step": 332
+    },
+    {
+      "epoch": 2.786610878661088,
+      "grad_norm": 0.015621655620634556,
+      "learning_rate": 4.8385443374108e-05,
+      "loss": 0.0093,
+      "step": 333
+    },
+    {
+      "epoch": 2.794979079497908,
+      "grad_norm": 0.024522442370653152,
+      "learning_rate": 4.809201922135589e-05,
+      "loss": 0.0116,
+      "step": 334
+    },
+    {
+      "epoch": 2.803347280334728,
+      "grad_norm": 0.016756724566221237,
+      "learning_rate": 4.7798660859367615e-05,
+      "loss": 0.0108,
+      "step": 335
+    },
+    {
+      "epoch": 2.803347280334728,
+      "eval_loss": 0.013578813523054123,
+      "eval_runtime": 8.195,
+      "eval_samples_per_second": 6.101,
+      "eval_steps_per_second": 1.586,
+      "step": 335
+    },
+    {
+      "epoch": 2.8117154811715483,
+      "grad_norm": 0.01872294582426548,
+      "learning_rate": 4.750537840369095e-05,
+      "loss": 0.0097,
+      "step": 336
+    },
+    {
+      "epoch": 2.8200836820083683,
+      "grad_norm": 0.01649293303489685,
+      "learning_rate": 4.721218196725627e-05,
+      "loss": 0.0092,
+      "step": 337
+    },
+    {
+      "epoch": 2.8284518828451883,
+      "grad_norm": 0.017666228115558624,
+      "learning_rate": 4.691908166002784e-05,
+      "loss": 0.0114,
+      "step": 338
+    },
+    {
+      "epoch": 2.8368200836820083,
+      "grad_norm": 0.014805554412305355,
+      "learning_rate": 4.662608758865524e-05,
+      "loss": 0.0102,
+      "step": 339
+    },
+    {
+      "epoch": 2.8451882845188283,
+      "grad_norm": 0.015506643801927567,
+      "learning_rate": 4.633320985612481e-05,
+      "loss": 0.0092,
+      "step": 340
+    },
+    {
+      "epoch": 2.8451882845188283,
+      "eval_loss": 0.01316875871270895,
+      "eval_runtime": 8.1877,
+      "eval_samples_per_second": 6.107,
+      "eval_steps_per_second": 1.588,
+      "step": 340
+    },
+    {
+      "epoch": 2.8535564853556483,
+      "grad_norm": 0.016277095302939415,
+      "learning_rate": 4.60404585614113e-05,
+      "loss": 0.0093,
+      "step": 341
+    },
+    {
+      "epoch": 2.8619246861924688,
+      "grad_norm": 0.015483858995139599,
+      "learning_rate": 4.5747843799129684e-05,
+      "loss": 0.0101,
+      "step": 342
+    },
+    {
+      "epoch": 2.8702928870292888,
+      "grad_norm": 0.014321762137115002,
+      "learning_rate": 4.545537565918699e-05,
+      "loss": 0.0098,
+      "step": 343
+    },
+    {
+      "epoch": 2.8786610878661087,
+      "grad_norm": 0.014206680469214916,
+      "learning_rate": 4.5163064226434475e-05,
+      "loss": 0.009,
+      "step": 344
+    },
+    {
+      "epoch": 2.8870292887029287,
+      "grad_norm": 0.013214725069701672,
+      "learning_rate": 4.487091958031984e-05,
+      "loss": 0.0092,
+      "step": 345
+    },
+    {
+      "epoch": 2.8870292887029287,
+      "eval_loss": 0.013155767694115639,
+      "eval_runtime": 8.1971,
+      "eval_samples_per_second": 6.1,
+      "eval_steps_per_second": 1.586,
+      "step": 345
+    },
+    {
+      "epoch": 2.895397489539749,
+      "grad_norm": 0.01570652797818184,
+      "learning_rate": 4.457895179453961e-05,
+      "loss": 0.0101,
+      "step": 346
+    },
+    {
+      "epoch": 2.903765690376569,
+      "grad_norm": 0.013275294564664364,
+      "learning_rate": 4.42871709366919e-05,
+      "loss": 0.0096,
+      "step": 347
+    },
+    {
+      "epoch": 2.912133891213389,
+      "grad_norm": 0.017738599330186844,
+      "learning_rate": 4.3995587067929177e-05,
+      "loss": 0.0106,
+      "step": 348
+    },
+    {
+      "epoch": 2.920502092050209,
+      "grad_norm": 0.02087695151567459,
+      "learning_rate": 4.3704210242611346e-05,
+      "loss": 0.0105,
+      "step": 349
+    },
+    {
+      "epoch": 2.928870292887029,
+      "grad_norm": 0.01522638276219368,
+      "learning_rate": 4.341305050795907e-05,
+      "loss": 0.0095,
+      "step": 350
+    },
+    {
+      "epoch": 2.928870292887029,
+      "eval_loss": 0.012959838844835758,
+      "eval_runtime": 8.1869,
+      "eval_samples_per_second": 6.107,
+      "eval_steps_per_second": 1.588,
+      "step": 350
+    },
+    {
+      "epoch": 2.937238493723849,
+      "grad_norm": 0.015606192871928215,
+      "learning_rate": 4.31221179037073e-05,
+      "loss": 0.01,
+      "step": 351
+    },
+    {
+      "epoch": 2.9456066945606696,
+      "grad_norm": 0.019583778455853462,
+      "learning_rate": 4.283142246175915e-05,
+      "loss": 0.0096,
+      "step": 352
+    },
+    {
+      "epoch": 2.9539748953974896,
+      "grad_norm": 0.01509869284927845,
+      "learning_rate": 4.2540974205839876e-05,
+      "loss": 0.0094,
+      "step": 353
+    },
+    {
+      "epoch": 2.9623430962343096,
+      "grad_norm": 0.018710950389504433,
+      "learning_rate": 4.2250783151151334e-05,
+      "loss": 0.011,
+      "step": 354
+    },
+    {
+      "epoch": 2.9707112970711296,
+      "grad_norm": 0.015060734003782272,
+      "learning_rate": 4.1960859304026594e-05,
+      "loss": 0.0094,
+      "step": 355
+    },
+    {
+      "epoch": 2.9707112970711296,
+      "eval_loss": 0.012675998732447624,
+      "eval_runtime": 8.1889,
+      "eval_samples_per_second": 6.106,
+      "eval_steps_per_second": 1.588,
+      "step": 355
+    },
+    {
+      "epoch": 2.97907949790795,
+      "grad_norm": 0.02269388735294342,
+      "learning_rate": 4.167121266158488e-05,
+      "loss": 0.0102,
+      "step": 356
+    },
+    {
+      "epoch": 2.98744769874477,
+      "grad_norm": 0.014178493991494179,
+      "learning_rate": 4.1381853211386914e-05,
+      "loss": 0.0085,
+      "step": 357
+    },
+    {
+      "epoch": 2.99581589958159,
+      "grad_norm": 0.017629047855734825,
+      "learning_rate": 4.109279093109045e-05,
+      "loss": 0.0098,
+      "step": 358
+    },
+    {
+      "epoch": 3.00418410041841,
+      "grad_norm": 0.02921808324754238,
+      "learning_rate": 4.080403578810631e-05,
+      "loss": 0.0133,
+      "step": 359
+    },
+    {
+      "epoch": 3.01255230125523,
+      "grad_norm": 0.016827603802084923,
+      "learning_rate": 4.051559773925462e-05,
+      "loss": 0.0088,
+      "step": 360
+    },
+    {
+      "epoch": 3.01255230125523,
+      "eval_loss": 0.012736815959215164,
+      "eval_runtime": 8.1835,
+      "eval_samples_per_second": 6.11,
+      "eval_steps_per_second": 1.589,
+      "step": 360
+    },
+    {
+      "epoch": 3.02092050209205,
+      "grad_norm": 0.015738021582365036,
+      "learning_rate": 4.022748673042149e-05,
+      "loss": 0.0093,
+      "step": 361
+    },
+    {
+      "epoch": 3.0292887029288704,
+      "grad_norm": 0.013376789167523384,
+      "learning_rate": 3.9939712696216094e-05,
+      "loss": 0.007,
+      "step": 362
+    },
+    {
+      "epoch": 3.0376569037656904,
+      "grad_norm": 0.015625160187482834,
+      "learning_rate": 3.965228555962808e-05,
+      "loss": 0.0082,
+      "step": 363
+    },
+    {
+      "epoch": 3.0460251046025104,
+      "grad_norm": 0.01764129102230072,
+      "learning_rate": 3.9365215231685434e-05,
+      "loss": 0.0074,
+      "step": 364
+    },
+    {
+      "epoch": 3.0543933054393304,
+      "grad_norm": 0.015254415571689606,
+      "learning_rate": 3.907851161111269e-05,
+      "loss": 0.0086,
+      "step": 365
+    },
+    {
+      "epoch": 3.0543933054393304,
+      "eval_loss": 0.013066377490758896,
+      "eval_runtime": 8.1923,
+      "eval_samples_per_second": 6.103,
+      "eval_steps_per_second": 1.587,
+      "step": 365
+    },
+    {
+      "epoch": 3.062761506276151,
+      "grad_norm": 0.016880949959158897,
+      "learning_rate": 3.879218458398962e-05,
+      "loss": 0.0095,
+      "step": 366
+    },
+    {
+      "epoch": 3.071129707112971,
+      "grad_norm": 0.012788892723619938,
+      "learning_rate": 3.8506244023410345e-05,
+      "loss": 0.0076,
+      "step": 367
+    },
+    {
+      "epoch": 3.079497907949791,
+      "grad_norm": 0.017358217388391495,
+      "learning_rate": 3.82206997891429e-05,
+      "loss": 0.0073,
+      "step": 368
+    },
+    {
+      "epoch": 3.087866108786611,
+      "grad_norm": 0.017496395856142044,
+      "learning_rate": 3.7935561727289265e-05,
+      "loss": 0.0079,
+      "step": 369
+    },
+    {
+      "epoch": 3.096234309623431,
+      "grad_norm": 0.016393601894378662,
+      "learning_rate": 3.7650839669945804e-05,
+      "loss": 0.0094,
+      "step": 370
+    },
+    {
+      "epoch": 3.096234309623431,
+      "eval_loss": 0.01338866539299488,
+      "eval_runtime": 8.1844,
+      "eval_samples_per_second": 6.109,
+      "eval_steps_per_second": 1.588,
+      "step": 370
+    },
+    {
+      "epoch": 3.104602510460251,
+      "grad_norm": 0.019458502531051636,
+      "learning_rate": 3.736654343486425e-05,
+      "loss": 0.0069,
+      "step": 371
+    },
+    {
+      "epoch": 3.1129707112970713,
+      "grad_norm": 0.01921907253563404,
+      "learning_rate": 3.708268282511321e-05,
+      "loss": 0.0075,
+      "step": 372
+    },
+    {
+      "epoch": 3.1213389121338913,
+      "grad_norm": 0.021092800423502922,
+      "learning_rate": 3.679926762874012e-05,
+      "loss": 0.0086,
+      "step": 373
+    },
+    {
+      "epoch": 3.1297071129707112,
+      "grad_norm": 0.025811336934566498,
+      "learning_rate": 3.65163076184337e-05,
+      "loss": 0.0088,
+      "step": 374
+    },
+    {
+      "epoch": 3.1380753138075312,
+      "grad_norm": 0.017477277666330338,
+      "learning_rate": 3.623381255118702e-05,
+      "loss": 0.0075,
+      "step": 375
+    },
+    {
+      "epoch": 3.1380753138075312,
+      "eval_loss": 0.013682931661605835,
+      "eval_runtime": 8.1808,
+      "eval_samples_per_second": 6.112,
+      "eval_steps_per_second": 1.589,
+      "step": 375
+    },
+    {
+      "epoch": 3.1464435146443517,
+      "grad_norm": 0.017589591443538666,
+      "learning_rate": 3.5951792167961e-05,
+      "loss": 0.0075,
+      "step": 376
+    },
+    {
+      "epoch": 3.1548117154811717,
+      "grad_norm": 0.020391661673784256,
+      "learning_rate": 3.567025619334861e-05,
+      "loss": 0.0091,
+      "step": 377
+    },
+    {
+      "epoch": 3.1631799163179917,
+      "grad_norm": 0.016537083312869072,
+      "learning_rate": 3.538921433523949e-05,
+      "loss": 0.0062,
+      "step": 378
+    },
+    {
+      "epoch": 3.1715481171548117,
+      "grad_norm": 0.01650119759142399,
+      "learning_rate": 3.5108676284485176e-05,
+      "loss": 0.0076,
+      "step": 379
+    },
+    {
+      "epoch": 3.1799163179916317,
+      "grad_norm": 0.01610580086708069,
+      "learning_rate": 3.482865171456505e-05,
+      "loss": 0.0068,
+      "step": 380
+    },
+    {
+      "epoch": 3.1799163179916317,
+      "eval_loss": 0.013625282794237137,
+      "eval_runtime": 8.2399,
+      "eval_samples_per_second": 6.068,
+      "eval_steps_per_second": 1.578,
+      "step": 380
+    },
+    {
+      "epoch": 3.1882845188284517,
+      "grad_norm": 0.0224465299397707,
+      "learning_rate": 3.4549150281252636e-05,
+      "loss": 0.0098,
+      "step": 381
+    },
+    {
+      "epoch": 3.196652719665272,
+      "grad_norm": 0.02336002141237259,
+      "learning_rate": 3.427018162228276e-05,
+      "loss": 0.0079,
+      "step": 382
+    },
+    {
+      "epoch": 3.205020920502092,
+      "grad_norm": 0.01989358477294445,
+      "learning_rate": 3.3991755357019196e-05,
+      "loss": 0.0082,
+      "step": 383
+    },
+    {
+      "epoch": 3.213389121338912,
+      "grad_norm": 0.016892483457922935,
+      "learning_rate": 3.371388108612293e-05,
+      "loss": 0.0068,
+      "step": 384
+    },
+    {
+      "epoch": 3.221757322175732,
+      "grad_norm": 0.018852969631552696,
+      "learning_rate": 3.343656839122121e-05,
+      "loss": 0.0096,
+      "step": 385
+    },
+    {
+      "epoch": 3.221757322175732,
+      "eval_loss": 0.013560273684561253,
+      "eval_runtime": 8.2379,
+      "eval_samples_per_second": 6.07,
+      "eval_steps_per_second": 1.578,
+      "step": 385
+    },
+    {
+      "epoch": 3.2301255230125525,
+      "grad_norm": 0.012373587116599083,
+      "learning_rate": 3.3159826834576994e-05,
+      "loss": 0.0059,
+      "step": 386
+    },
+    {
+      "epoch": 3.2384937238493725,
+      "grad_norm": 0.016591370105743408,
+      "learning_rate": 3.2883665958759394e-05,
+      "loss": 0.0082,
+      "step": 387
+    },
+    {
+      "epoch": 3.2468619246861925,
+      "grad_norm": 0.01938340999186039,
+      "learning_rate": 3.260809528631452e-05,
+      "loss": 0.008,
+      "step": 388
+    },
+    {
+      "epoch": 3.2552301255230125,
+      "grad_norm": 0.015290241688489914,
+      "learning_rate": 3.233312431943719e-05,
+      "loss": 0.0067,
+      "step": 389
+    },
+    {
+      "epoch": 3.2635983263598325,
+      "grad_norm": 0.017185404896736145,
+      "learning_rate": 3.205876253964321e-05,
+      "loss": 0.0088,
+      "step": 390
+    },
+    {
+      "epoch": 3.2635983263598325,
+      "eval_loss": 0.013662769459187984,
+      "eval_runtime": 8.2288,
+      "eval_samples_per_second": 6.076,
+      "eval_steps_per_second": 1.58,
+      "step": 390
+    },
+    {
+      "epoch": 3.2719665271966525,
+      "grad_norm": 0.019198745489120483,
+      "learning_rate": 3.178501940744249e-05,
+      "loss": 0.0066,
+      "step": 391
+    },
+    {
+      "epoch": 3.280334728033473,
+      "grad_norm": 0.01907048560678959,
+      "learning_rate": 3.1511904362012805e-05,
+      "loss": 0.0086,
+      "step": 392
+    },
+    {
+      "epoch": 3.288702928870293,
+      "grad_norm": 0.01648963801562786,
+      "learning_rate": 3.123942682087433e-05,
+      "loss": 0.0066,
+      "step": 393
+    },
+    {
+      "epoch": 3.297071129707113,
+      "grad_norm": 0.015369053930044174,
+      "learning_rate": 3.096759617956486e-05,
+      "loss": 0.0082,
+      "step": 394
+    },
+    {
+      "epoch": 3.305439330543933,
+      "grad_norm": 0.015744755044579506,
+      "learning_rate": 3.069642181131592e-05,
+      "loss": 0.008,
+      "step": 395
+    },
+    {
+      "epoch": 3.305439330543933,
+      "eval_loss": 0.01378766167908907,
+      "eval_runtime": 8.1991,
+      "eval_samples_per_second": 6.098,
+      "eval_steps_per_second": 1.586,
+      "step": 395
+    },
+    {
+      "epoch": 3.3138075313807533,
+      "grad_norm": 0.028769735246896744,
+      "learning_rate": 3.042591306672946e-05,
+      "loss": 0.0089,
+      "step": 396
+    },
+    {
+      "epoch": 3.3221757322175733,
+      "grad_norm": 0.01585283689200878,
+      "learning_rate": 3.0156079273455483e-05,
+      "loss": 0.0071,
+      "step": 397
+    },
+    {
+      "epoch": 3.3305439330543933,
+      "grad_norm": 0.016848471015691757,
+      "learning_rate": 2.9886929735870416e-05,
+      "loss": 0.0079,
+      "step": 398
+    },
+    {
+      "epoch": 3.3389121338912133,
+      "grad_norm": 0.015105370432138443,
+      "learning_rate": 2.9618473734756264e-05,
+      "loss": 0.0079,
+      "step": 399
+    },
+    {
+      "epoch": 3.3472803347280333,
+      "grad_norm": 0.01963132992386818,
+      "learning_rate": 2.935072052698059e-05,
+      "loss": 0.0085,
+      "step": 400
+    },
+    {
+      "epoch": 3.3472803347280333,
+      "eval_loss": 0.013703197240829468,
+      "eval_runtime": 8.1849,
+      "eval_samples_per_second": 6.109,
+      "eval_steps_per_second": 1.588,
+      "step": 400
+    },
+    {
+      "epoch": 3.3556485355648533,
+      "grad_norm": 0.015131447464227676,
+      "learning_rate": 2.908367934517729e-05,
+      "loss": 0.0079,
+      "step": 401
+    },
+    {
+      "epoch": 3.3640167364016738,
+      "grad_norm": 0.01849834993481636,
+      "learning_rate": 2.881735939742831e-05,
+      "loss": 0.0077,
+      "step": 402
+    },
+    {
+      "epoch": 3.3723849372384938,
+      "grad_norm": 0.01487911120057106,
+      "learning_rate": 2.8551769866946077e-05,
+      "loss": 0.0075,
+      "step": 403
+    },
+    {
+      "epoch": 3.3807531380753137,
+      "grad_norm": 0.020853998139500618,
+      "learning_rate": 2.8286919911756836e-05,
+      "loss": 0.0079,
+      "step": 404
+    },
+    {
+      "epoch": 3.3891213389121337,
+      "grad_norm": 0.026485485956072807,
+      "learning_rate": 2.8022818664384944e-05,
+      "loss": 0.0091,
+      "step": 405
+    },
+    {
+      "epoch": 3.3891213389121337,
+      "eval_loss": 0.013571958057582378,
+      "eval_runtime": 8.2051,
+      "eval_samples_per_second": 6.094,
+      "eval_steps_per_second": 1.584,
+      "step": 405
+    },
+    {
+      "epoch": 3.397489539748954,
+      "grad_norm": 0.015749875456094742,
+      "learning_rate": 2.77594752315378e-05,
+      "loss": 0.007,
+      "step": 406
+    },
+    {
+      "epoch": 3.405857740585774,
+      "grad_norm": 0.01729268953204155,
+      "learning_rate": 2.7496898693792028e-05,
+      "loss": 0.0094,
+      "step": 407
+    },
+    {
+      "epoch": 3.414225941422594,
+      "grad_norm": 0.01651543565094471,
+      "learning_rate": 2.7235098105280243e-05,
+      "loss": 0.0079,
+      "step": 408
+    },
+    {
+      "epoch": 3.422594142259414,
+      "grad_norm": 0.026226142421364784,
+      "learning_rate": 2.697408249337886e-05,
+      "loss": 0.0084,
+      "step": 409
+    },
+    {
+      "epoch": 3.430962343096234,
+      "grad_norm": 0.01325182057917118,
+      "learning_rate": 2.671386085839682e-05,
+      "loss": 0.0049,
+      "step": 410
+    },
+    {
+      "epoch": 3.430962343096234,
+      "eval_loss": 0.013367210514843464,
+      "eval_runtime": 8.1901,
+      "eval_samples_per_second": 6.105,
+      "eval_steps_per_second": 1.587,
+      "step": 410
+    },
+    {
+      "epoch": 3.439330543933054,
+      "grad_norm": 0.01933259144425392,
+      "learning_rate": 2.6454442173265282e-05,
+      "loss": 0.0089,
+      "step": 411
+    },
+    {
+      "epoch": 3.4476987447698746,
+      "grad_norm": 0.017439622431993484,
+      "learning_rate": 2.6195835383228127e-05,
+      "loss": 0.0084,
+      "step": 412
+    },
+    {
+      "epoch": 3.4560669456066946,
+      "grad_norm": 0.01493872795253992,
+      "learning_rate": 2.5938049405533627e-05,
+      "loss": 0.0071,
+      "step": 413
+    },
+    {
+      "epoch": 3.4644351464435146,
+      "grad_norm": 0.01432398147881031,
+      "learning_rate": 2.5681093129126864e-05,
+      "loss": 0.0072,
+      "step": 414
+    },
+    {
+      "epoch": 3.4728033472803346,
+      "grad_norm": 0.017259176820516586,
+      "learning_rate": 2.542497541434329e-05,
+      "loss": 0.0072,
+      "step": 415
+    },
+    {
+      "epoch": 3.4728033472803346,
+      "eval_loss": 0.013123424723744392,
+      "eval_runtime": 8.2011,
+      "eval_samples_per_second": 6.097,
+      "eval_steps_per_second": 1.585,
+      "step": 415
+    },
+    {
+      "epoch": 3.481171548117155,
+      "grad_norm": 0.0195622555911541,
+      "learning_rate": 2.5169705092603125e-05,
+      "loss": 0.0082,
+      "step": 416
+    },
+    {
+      "epoch": 3.489539748953975,
+      "grad_norm": 0.020763585343956947,
+      "learning_rate": 2.491529096610692e-05,
+      "loss": 0.0079,
+      "step": 417
+    },
+    {
+      "epoch": 3.497907949790795,
+      "grad_norm": 0.02019021287560463,
+      "learning_rate": 2.4661741807532006e-05,
+      "loss": 0.0088,
+      "step": 418
+    },
+    {
+      "epoch": 3.506276150627615,
+      "grad_norm": 0.017682604491710663,
+      "learning_rate": 2.4409066359729997e-05,
+      "loss": 0.0089,
+      "step": 419
+    },
+    {
+      "epoch": 3.514644351464435,
+      "grad_norm": 0.015951137989759445,
+      "learning_rate": 2.4157273335425297e-05,
+      "loss": 0.0063,
+      "step": 420
+    },
+    {
+      "epoch": 3.514644351464435,
+      "eval_loss": 0.013349597342312336,
+      "eval_runtime": 8.1856,
+      "eval_samples_per_second": 6.108,
+      "eval_steps_per_second": 1.588,
+      "step": 420
+    },
+    {
+      "epoch": 3.523012552301255,
+      "grad_norm": 0.02157527022063732,
+      "learning_rate": 2.390637141691471e-05,
+      "loss": 0.0078,
+      "step": 421
+    },
+    {
+      "epoch": 3.5313807531380754,
+      "grad_norm": 0.015403731726109982,
+      "learning_rate": 2.3656369255768017e-05,
+      "loss": 0.0065,
+      "step": 422
+    },
+    {
+      "epoch": 3.5397489539748954,
+      "grad_norm": 0.016804836690425873,
+      "learning_rate": 2.340727547252969e-05,
+      "loss": 0.0081,
+      "step": 423
+    },
+    {
+      "epoch": 3.5481171548117154,
+      "grad_norm": 0.016539255157113075,
+      "learning_rate": 2.3159098656421614e-05,
+      "loss": 0.0092,
+      "step": 424
+    },
+    {
+      "epoch": 3.5564853556485354,
+      "grad_norm": 0.016467688605189323,
+      "learning_rate": 2.291184736504695e-05,
+      "loss": 0.0076,
+      "step": 425
+    },
+    {
+      "epoch": 3.5564853556485354,
+      "eval_loss": 0.013108824379742146,
+      "eval_runtime": 8.1932,
+      "eval_samples_per_second": 6.103,
+      "eval_steps_per_second": 1.587,
+      "step": 425
+    },
+    {
+      "epoch": 3.564853556485356,
+      "grad_norm": 0.017600081861019135,
+      "learning_rate": 2.2665530124094962e-05,
+      "loss": 0.0084,
+      "step": 426
+    },
+    {
+      "epoch": 3.573221757322176,
+      "grad_norm": 0.02098635584115982,
+      "learning_rate": 2.2420155427047167e-05,
+      "loss": 0.0086,
+      "step": 427
+    },
+    {
+      "epoch": 3.581589958158996,
+      "grad_norm": 0.017134375870227814,
+      "learning_rate": 2.2175731734884365e-05,
+      "loss": 0.0075,
+      "step": 428
+    },
+    {
+      "epoch": 3.589958158995816,
+      "grad_norm": 0.017826305702328682,
+      "learning_rate": 2.1932267475794917e-05,
+      "loss": 0.0071,
+      "step": 429
+    },
+    {
+      "epoch": 3.598326359832636,
+      "grad_norm": 0.016442451626062393,
+      "learning_rate": 2.1689771044884148e-05,
+      "loss": 0.0076,
+      "step": 430
+    },
+    {
+      "epoch": 3.598326359832636,
+      "eval_loss": 0.012915154919028282,
+      "eval_runtime": 8.1885,
+      "eval_samples_per_second": 6.106,
+      "eval_steps_per_second": 1.588,
+      "step": 430
+    },
+    {
+      "epoch": 3.606694560669456,
+      "grad_norm": 0.026688870042562485,
+      "learning_rate": 2.1448250803884834e-05,
+      "loss": 0.0087,
+      "step": 431
+    },
+    {
+      "epoch": 3.6150627615062763,
+      "grad_norm": 0.016501056030392647,
+      "learning_rate": 2.120771508086889e-05,
+      "loss": 0.0066,
+      "step": 432
+    },
+    {
+      "epoch": 3.6234309623430963,
+      "grad_norm": 0.013682221062481403,
+      "learning_rate": 2.0968172169960197e-05,
+      "loss": 0.0067,
+      "step": 433
+    },
+    {
+      "epoch": 3.6317991631799162,
+      "grad_norm": 0.018159842118620872,
+      "learning_rate": 2.0729630331048615e-05,
+      "loss": 0.0088,
+      "step": 434
+    },
+    {
+      "epoch": 3.6401673640167362,
+      "grad_norm": 0.01990005373954773,
+      "learning_rate": 2.0492097789505178e-05,
+      "loss": 0.0074,
+      "step": 435
+    },
+    {
+      "epoch": 3.6401673640167362,
+      "eval_loss": 0.013006238266825676,
+      "eval_runtime": 8.2323,
+      "eval_samples_per_second": 6.074,
+      "eval_steps_per_second": 1.579,
+      "step": 435
+    },
+    {
+      "epoch": 3.6485355648535567,
+      "grad_norm": 0.01765499822795391,
+      "learning_rate": 2.0255582735898383e-05,
+      "loss": 0.0063,
+      "step": 436
+    },
+    {
+      "epoch": 3.6569037656903767,
+      "grad_norm": 0.019951600581407547,
+      "learning_rate": 2.0020093325711896e-05,
+      "loss": 0.0084,
+      "step": 437
+    },
+    {
+      "epoch": 3.6652719665271967,
+      "grad_norm": 0.019513117149472237,
+      "learning_rate": 1.9785637679063258e-05,
+      "loss": 0.0064,
+      "step": 438
+    },
+    {
+      "epoch": 3.6736401673640167,
+      "grad_norm": 0.027216019108891487,
+      "learning_rate": 1.9552223880423882e-05,
+      "loss": 0.0092,
+      "step": 439
+    },
+    {
+      "epoch": 3.6820083682008367,
+      "grad_norm": 0.018059376627206802,
+      "learning_rate": 1.931985997834031e-05,
+      "loss": 0.0074,
+      "step": 440
+    },
+    {
+      "epoch": 3.6820083682008367,
+      "eval_loss": 0.01317664049565792,
+      "eval_runtime": 8.1934,
+      "eval_samples_per_second": 6.102,
+      "eval_steps_per_second": 1.587,
+      "step": 440
+    },
+    {
+      "epoch": 3.6903765690376567,
+      "grad_norm": 0.020217331126332283,
+      "learning_rate": 1.908855398515669e-05,
+      "loss": 0.0079,
+      "step": 441
+    },
+    {
+      "epoch": 3.698744769874477,
+      "grad_norm": 0.019336042925715446,
+      "learning_rate": 1.8858313876738458e-05,
+      "loss": 0.0069,
+      "step": 442
+    },
+    {
+      "epoch": 3.707112970711297,
+      "grad_norm": 0.020630570128560066,
+      "learning_rate": 1.8629147592197372e-05,
+      "loss": 0.0083,
+      "step": 443
+    },
+    {
+      "epoch": 3.715481171548117,
+      "grad_norm": 0.01856216788291931,
+      "learning_rate": 1.840106303361771e-05,
+      "loss": 0.0075,
+      "step": 444
+    },
+    {
+      "epoch": 3.723849372384937,
+      "grad_norm": 0.01747502014040947,
+      "learning_rate": 1.8174068065783765e-05,
+      "loss": 0.0067,
+      "step": 445
+    },
+    {
+      "epoch": 3.723849372384937,
+      "eval_loss": 0.013197257183492184,
+      "eval_runtime": 8.2112,
+      "eval_samples_per_second": 6.089,
+      "eval_steps_per_second": 1.583,
+      "step": 445
+    },
+    {
+      "epoch": 3.7322175732217575,
+      "grad_norm": 0.018087679520249367,
+      "learning_rate": 1.7948170515908757e-05,
+      "loss": 0.0077,
+      "step": 446
+    },
+    {
+      "epoch": 3.7405857740585775,
+      "grad_norm": 0.01658761128783226,
+      "learning_rate": 1.7723378173364836e-05,
+      "loss": 0.0081,
+      "step": 447
+    },
+    {
+      "epoch": 3.7489539748953975,
+      "grad_norm": 0.017800970003008842,
+      "learning_rate": 1.7499698789414536e-05,
+      "loss": 0.0076,
+      "step": 448
+    },
+    {
+      "epoch": 3.7573221757322175,
+      "grad_norm": 0.018941445276141167,
+      "learning_rate": 1.727714007694348e-05,
+      "loss": 0.0092,
+      "step": 449
+    },
+    {
+      "epoch": 3.7656903765690375,
+      "grad_norm": 0.014227015897631645,
+      "learning_rate": 1.705570971019445e-05,
+      "loss": 0.0064,
+      "step": 450
+    },
+    {
+      "epoch": 3.7656903765690375,
+      "eval_loss": 0.013023738749325275,
+      "eval_runtime": 8.1848,
+      "eval_samples_per_second": 6.109,
+      "eval_steps_per_second": 1.588,
+      "step": 450
+    },
+    {
+      "epoch": 3.7740585774058575,
+      "grad_norm": 0.01811639405786991,
+      "learning_rate": 1.683541532450272e-05,
+      "loss": 0.0077,
+      "step": 451
+    },
+    {
+      "epoch": 3.782426778242678,
+      "grad_norm": 0.01967625319957733,
+      "learning_rate": 1.6616264516032807e-05,
+      "loss": 0.0083,
+      "step": 452
+    },
+    {
+      "epoch": 3.790794979079498,
+      "grad_norm": 0.0179898664355278,
+      "learning_rate": 1.6398264841516564e-05,
+      "loss": 0.0076,
+      "step": 453
+    },
+    {
+      "epoch": 3.799163179916318,
+      "grad_norm": 0.015487701632082462,
+      "learning_rate": 1.618142381799256e-05,
+      "loss": 0.0069,
+      "step": 454
+    },
+    {
+      "epoch": 3.8075313807531384,
+      "grad_norm": 0.017041724175214767,
+      "learning_rate": 1.5965748922546876e-05,
+      "loss": 0.0091,
+      "step": 455
+    },
+    {
+      "epoch": 3.8075313807531384,
+      "eval_loss": 0.013014528900384903,
+      "eval_runtime": 8.1813,
+      "eval_samples_per_second": 6.112,
+      "eval_steps_per_second": 1.589,
+      "step": 455
+    },
+    {
+      "epoch": 3.8158995815899583,
+      "grad_norm": 0.01787850260734558,
+      "learning_rate": 1.5751247592055333e-05,
+      "loss": 0.0069,
+      "step": 456
+    },
+    {
+      "epoch": 3.8242677824267783,
+      "grad_norm": 0.01862199790775776,
+      "learning_rate": 1.5537927222927047e-05,
+      "loss": 0.0086,
+      "step": 457
+    },
+    {
+      "epoch": 3.8326359832635983,
+      "grad_norm": 0.016767242923378944,
+      "learning_rate": 1.5325795170849315e-05,
+      "loss": 0.0068,
+      "step": 458
+    },
+    {
+      "epoch": 3.8410041841004183,
+      "grad_norm": 0.01677015796303749,
+      "learning_rate": 1.5114858750534067e-05,
+      "loss": 0.0078,
+      "step": 459
+    },
+    {
+      "epoch": 3.8493723849372383,
+      "grad_norm": 0.017780860885977745,
+      "learning_rate": 1.4905125235465589e-05,
+      "loss": 0.0074,
+      "step": 460
+    },
+    {
+      "epoch": 3.8493723849372383,
+      "eval_loss": 0.013112700544297695,
+      "eval_runtime": 8.184,
+      "eval_samples_per_second": 6.109,
+      "eval_steps_per_second": 1.588,
+      "step": 460
+    },
+    {
+      "epoch": 3.8577405857740583,
+      "grad_norm": 0.02164328470826149,
+      "learning_rate": 1.4696601857649738e-05,
+      "loss": 0.0075,
+      "step": 461
+    },
+    {
+      "epoch": 3.8661087866108788,
+      "grad_norm": 0.01868240348994732,
+      "learning_rate": 1.4489295807364556e-05,
+      "loss": 0.0082,
+      "step": 462
+    },
+    {
+      "epoch": 3.8744769874476988,
+      "grad_norm": 0.018218357115983963,
+      "learning_rate": 1.4283214232912345e-05,
+      "loss": 0.0073,
+      "step": 463
+    },
+    {
+      "epoch": 3.8828451882845187,
+      "grad_norm": 0.02058660052716732,
+      "learning_rate": 1.4078364240373192e-05,
+      "loss": 0.0063,
+      "step": 464
+    },
+    {
+      "epoch": 3.891213389121339,
+      "grad_norm": 0.03692828491330147,
+      "learning_rate": 1.38747528933599e-05,
+      "loss": 0.0076,
+      "step": 465
+    },
+    {
+      "epoch": 3.891213389121339,
+      "eval_loss": 0.01321639958769083,
+      "eval_runtime": 8.1882,
+      "eval_samples_per_second": 6.106,
+      "eval_steps_per_second": 1.588,
+      "step": 465
+    },
+    {
+      "epoch": 3.899581589958159,
+      "grad_norm": 0.015658119693398476,
+      "learning_rate": 1.3672387212774457e-05,
+      "loss": 0.0093,
+      "step": 466
+    },
+    {
+      "epoch": 3.907949790794979,
+      "grad_norm": 0.014811305329203606,
+      "learning_rate": 1.3471274176565935e-05,
+      "loss": 0.0068,
+      "step": 467
+    },
+    {
+      "epoch": 3.916317991631799,
+      "grad_norm": 0.01890689507126808,
+      "learning_rate": 1.327142071948988e-05,
+      "loss": 0.0076,
+      "step": 468
+    },
+    {
+      "epoch": 3.924686192468619,
+      "grad_norm": 0.016337791457772255,
+      "learning_rate": 1.3072833732869167e-05,
+      "loss": 0.0079,
+      "step": 469
+    },
+    {
+      "epoch": 3.933054393305439,
+      "grad_norm": 0.02025439403951168,
+      "learning_rate": 1.28755200643564e-05,
+      "loss": 0.007,
+      "step": 470
+    },
+    {
+      "epoch": 3.933054393305439,
+      "eval_loss": 0.013240823522210121,
+      "eval_runtime": 8.1806,
+      "eval_samples_per_second": 6.112,
+      "eval_steps_per_second": 1.589,
+      "step": 470
+    },
+    {
+      "epoch": 3.941422594142259,
+      "grad_norm": 0.01858791708946228,
+      "learning_rate": 1.267948651769777e-05,
+      "loss": 0.0082,
+      "step": 471
+    },
+    {
+      "epoch": 3.9497907949790796,
+      "grad_norm": 0.015129518695175648,
+      "learning_rate": 1.2484739852498467e-05,
+      "loss": 0.0067,
+      "step": 472
+    },
+    {
+      "epoch": 3.9581589958158996,
+      "grad_norm": 0.01916913501918316,
+      "learning_rate": 1.2291286783989597e-05,
+      "loss": 0.0083,
+      "step": 473
+    },
+    {
+      "epoch": 3.9665271966527196,
+      "grad_norm": 0.020614786073565483,
+      "learning_rate": 1.2099133982796612e-05,
+      "loss": 0.0083,
+      "step": 474
+    },
+    {
+      "epoch": 3.97489539748954,
+      "grad_norm": 0.016557354480028152,
+      "learning_rate": 1.190828807470929e-05,
+      "loss": 0.0082,
+      "step": 475
+    },
+    {
+      "epoch": 3.97489539748954,
+      "eval_loss": 0.013233880512416363,
+      "eval_runtime": 8.1955,
+      "eval_samples_per_second": 6.101,
+      "eval_steps_per_second": 1.586,
+      "step": 475
+    },
+    {
+      "epoch": 3.98326359832636,
+      "grad_norm": 0.01762431487441063,
+      "learning_rate": 1.1718755640453288e-05,
+      "loss": 0.0059,
+      "step": 476
+    },
+    {
+      "epoch": 3.99163179916318,
+      "grad_norm": 0.01950531266629696,
+      "learning_rate": 1.1530543215463235e-05,
+      "loss": 0.0085,
+      "step": 477
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 0.02553708292543888,
+      "learning_rate": 1.1343657289657333e-05,
+      "loss": 0.0115,
+      "step": 478
+    },
+    {
+      "epoch": 4.00836820083682,
+      "grad_norm": 0.011678546667098999,
+      "learning_rate": 1.1158104307213612e-05,
+      "loss": 0.0066,
+      "step": 479
+    },
+    {
+      "epoch": 4.01673640167364,
+      "grad_norm": 0.016237011179327965,
+      "learning_rate": 1.0973890666347702e-05,
+      "loss": 0.0059,
+      "step": 480
+    },
+    {
+      "epoch": 4.01673640167364,
+      "eval_loss": 0.013261526823043823,
+      "eval_runtime": 8.1877,
+      "eval_samples_per_second": 6.107,
+      "eval_steps_per_second": 1.588,
+      "step": 480
+    },
+    {
+      "epoch": 4.02510460251046,
+      "grad_norm": 0.015487432479858398,
+      "learning_rate": 1.0791022719092231e-05,
+      "loss": 0.0064,
+      "step": 481
+    },
+    {
+      "epoch": 4.03347280334728,
+      "grad_norm": 0.015571733005344868,
+      "learning_rate": 1.0609506771077765e-05,
+      "loss": 0.0067,
+      "step": 482
+    },
+    {
+      "epoch": 4.0418410041841,
+      "grad_norm": 0.013560911640524864,
+      "learning_rate": 1.0429349081315404e-05,
+      "loss": 0.0058,
+      "step": 483
+    },
+    {
+      "epoch": 4.050209205020921,
+      "grad_norm": 0.014713162556290627,
+      "learning_rate": 1.025055586198096e-05,
+      "loss": 0.0064,
+      "step": 484
+    },
+    {
+      "epoch": 4.058577405857741,
+      "grad_norm": 0.011582295410335064,
+      "learning_rate": 1.0073133278200703e-05,
+      "loss": 0.0066,
+      "step": 485
+    },
+    {
+      "epoch": 4.058577405857741,
+      "eval_loss": 0.013517594896256924,
+      "eval_runtime": 8.1969,
+      "eval_samples_per_second": 6.1,
+      "eval_steps_per_second": 1.586,
+      "step": 485
+    },
+    {
+      "epoch": 4.066945606694561,
+      "grad_norm": 0.016617678105831146,
+      "learning_rate": 9.897087447838848e-06,
+      "loss": 0.0068,
+      "step": 486
+    },
+    {
+      "epoch": 4.075313807531381,
+      "grad_norm": 0.013949776999652386,
+      "learning_rate": 9.72242444128656e-06,
+      "loss": 0.0068,
+      "step": 487
+    },
+    {
+      "epoch": 4.083682008368201,
+      "grad_norm": 0.0159038994461298,
+      "learning_rate": 9.549150281252633e-06,
+      "loss": 0.0059,
+      "step": 488
+    },
+    {
+      "epoch": 4.092050209205021,
+      "grad_norm": 0.023309912532567978,
+      "learning_rate": 9.377270942555816e-06,
+      "loss": 0.0065,
+      "step": 489
+    },
+    {
+      "epoch": 4.100418410041841,
+      "grad_norm": 0.014613127335906029,
+      "learning_rate": 9.206792351918808e-06,
+      "loss": 0.0063,
+      "step": 490
+    },
+    {
+      "epoch": 4.100418410041841,
+      "eval_loss": 0.013989124447107315,
+      "eval_runtime": 8.2202,
+      "eval_samples_per_second": 6.083,
+      "eval_steps_per_second": 1.581,
+      "step": 490
+    },
+    {
+      "epoch": 4.108786610878661,
+      "grad_norm": 0.03452659025788307,
+      "learning_rate": 9.037720387763877e-06,
+      "loss": 0.0062,
+      "step": 491
+    },
+    {
+      "epoch": 4.117154811715481,
+      "grad_norm": 0.014380170032382011,
+      "learning_rate": 8.870060880010161e-06,
+      "loss": 0.0072,
+      "step": 492
+    },
+    {
+      "epoch": 4.125523012552302,
+      "grad_norm": 0.014284541830420494,
+      "learning_rate": 8.703819609872655e-06,
+      "loss": 0.0065,
+      "step": 493
+    },
+    {
+      "epoch": 4.133891213389122,
+      "grad_norm": 0.028412694111466408,
+      "learning_rate": 8.539002309662863e-06,
+      "loss": 0.0068,
+      "step": 494
+    },
+    {
+      "epoch": 4.142259414225942,
+      "grad_norm": 0.01495958399027586,
+      "learning_rate": 8.375614662591098e-06,
+      "loss": 0.0059,
+      "step": 495
+    },
+    {
+      "epoch": 4.142259414225942,
+      "eval_loss": 0.01436224952340126,
+      "eval_runtime": 8.1984,
+      "eval_samples_per_second": 6.099,
+      "eval_steps_per_second": 1.586,
+      "step": 495
+    },
+    {
+      "epoch": 4.150627615062762,
+      "grad_norm": 0.01771622896194458,
+      "learning_rate": 8.213662302570564e-06,
+      "loss": 0.0072,
+      "step": 496
+    },
+    {
+      "epoch": 4.158995815899582,
+      "grad_norm": 0.02444896660745144,
+      "learning_rate": 8.053150814023075e-06,
+      "loss": 0.006,
+      "step": 497
+    },
+    {
+      "epoch": 4.167364016736402,
+      "grad_norm": 0.016917061060667038,
+      "learning_rate": 7.894085731686484e-06,
+      "loss": 0.0068,
+      "step": 498
+    },
+    {
+      "epoch": 4.175732217573222,
+      "grad_norm": 0.020339468494057655,
+      "learning_rate": 7.736472540423839e-06,
+      "loss": 0.0062,
+      "step": 499
+    },
+    {
+      "epoch": 4.184100418410042,
+      "grad_norm": 0.01869555190205574,
+      "learning_rate": 7.5803166750342545e-06,
+      "loss": 0.0066,
+      "step": 500
+    },
+    {
+      "epoch": 4.184100418410042,
+      "eval_loss": 0.014202076941728592,
+      "eval_runtime": 8.1932,
+      "eval_samples_per_second": 6.103,
+      "eval_steps_per_second": 1.587,
+      "step": 500
+    },
+    {
+      "epoch": 4.192468619246862,
+      "grad_norm": 0.017789356410503387,
+      "learning_rate": 7.425623520065506e-06,
+      "loss": 0.0068,
+      "step": 501
+    },
+    {
+      "epoch": 4.200836820083682,
+      "grad_norm": 0.018326247110962868,
+      "learning_rate": 7.272398409628362e-06,
+      "loss": 0.0058,
+      "step": 502
+    },
+    {
+      "epoch": 4.209205020920502,
+      "grad_norm": 0.01593657210469246,
+      "learning_rate": 7.120646627212668e-06,
+      "loss": 0.0065,
+      "step": 503
+    },
+    {
+      "epoch": 4.2175732217573225,
+      "grad_norm": 0.01636459119617939,
+      "learning_rate": 6.970373405505149e-06,
+      "loss": 0.0059,
+      "step": 504
+    },
+    {
+      "epoch": 4.2259414225941425,
+      "grad_norm": 0.014171932823956013,
+      "learning_rate": 6.8215839262089465e-06,
+      "loss": 0.0055,
+      "step": 505
+    },
+    {
+      "epoch": 4.2259414225941425,
+      "eval_loss": 0.01415976881980896,
+      "eval_runtime": 8.2304,
+      "eval_samples_per_second": 6.075,
+      "eval_steps_per_second": 1.58,
+      "step": 505
+    },
+    {
+      "epoch": 4.2343096234309625,
+      "grad_norm": 0.013969158753752708,
+      "learning_rate": 6.674283319865015e-06,
+      "loss": 0.0065,
+      "step": 506
+    },
+    {
+      "epoch": 4.2426778242677825,
+      "grad_norm": 0.020747659727931023,
+      "learning_rate": 6.5284766656751486e-06,
+      "loss": 0.0067,
+      "step": 507
+    },
+    {
+      "epoch": 4.2510460251046025,
+      "grad_norm": 0.015698375180363655,
+      "learning_rate": 6.384168991326872e-06,
+      "loss": 0.0058,
+      "step": 508
+    },
+    {
+      "epoch": 4.2594142259414225,
+      "grad_norm": 0.018806830048561096,
+      "learning_rate": 6.241365272820065e-06,
+      "loss": 0.0065,
+      "step": 509
+    },
+    {
+      "epoch": 4.2677824267782425,
+      "grad_norm": 0.01827147603034973,
+      "learning_rate": 6.100070434295379e-06,
+      "loss": 0.0067,
+      "step": 510
+    },
+    {
+      "epoch": 4.2677824267782425,
+      "eval_loss": 0.014154573902487755,
+      "eval_runtime": 8.2041,
+      "eval_samples_per_second": 6.095,
+      "eval_steps_per_second": 1.585,
+      "step": 510
+    },
+    {
+      "epoch": 4.2761506276150625,
+      "grad_norm": 0.016821326687932014,
+      "learning_rate": 5.96028934786445e-06,
+      "loss": 0.0057,
+      "step": 511
+    },
+    {
+      "epoch": 4.2845188284518825,
+      "grad_norm": 0.01477895863354206,
+      "learning_rate": 5.822026833441901e-06,
+      "loss": 0.0061,
+      "step": 512
+    },
+    {
+      "epoch": 4.292887029288703,
+      "grad_norm": 0.02095729485154152,
+      "learning_rate": 5.685287658579124e-06,
+      "loss": 0.0067,
+      "step": 513
+    },
+    {
+      "epoch": 4.301255230125523,
+      "grad_norm": 0.014905648306012154,
+      "learning_rate": 5.550076538299931e-06,
+      "loss": 0.0064,
+      "step": 514
+    },
+    {
+      "epoch": 4.309623430962343,
+      "grad_norm": 0.01747983880341053,
+      "learning_rate": 5.416398134937878e-06,
+      "loss": 0.0065,
+      "step": 515
+    },
+    {
+      "epoch": 4.309623430962343,
+      "eval_loss": 0.014257834292948246,
+      "eval_runtime": 8.1812,
+      "eval_samples_per_second": 6.112,
+      "eval_steps_per_second": 1.589,
+      "step": 515
+    },
+    {
+      "epoch": 4.317991631799163,
+      "grad_norm": 0.01598619669675827,
+      "learning_rate": 5.284257057975628e-06,
+      "loss": 0.0063,
+      "step": 516
+    },
+    {
+      "epoch": 4.326359832635983,
+      "grad_norm": 0.016550792381167412,
+      "learning_rate": 5.153657863885902e-06,
+      "loss": 0.0056,
+      "step": 517
+    },
+    {
+      "epoch": 4.334728033472803,
+      "grad_norm": 0.02542988397181034,
+      "learning_rate": 5.024605055974408e-06,
+      "loss": 0.0065,
+      "step": 518
+    },
+    {
+      "epoch": 4.343096234309623,
+      "grad_norm": 0.01975710317492485,
+      "learning_rate": 4.8971030842245635e-06,
+      "loss": 0.0066,
+      "step": 519
+    },
+    {
+      "epoch": 4.351464435146443,
+      "grad_norm": 0.01818425953388214,
+      "learning_rate": 4.771156345144018e-06,
+      "loss": 0.0062,
+      "step": 520
+    },
+    {
+      "epoch": 4.351464435146443,
+      "eval_loss": 0.014165216125547886,
+      "eval_runtime": 8.1986,
+      "eval_samples_per_second": 6.099,
+      "eval_steps_per_second": 1.586,
+      "step": 520
+    },
+    {
+      "epoch": 4.359832635983263,
+      "grad_norm": 0.01841098628938198,
+      "learning_rate": 4.646769181613098e-06,
+      "loss": 0.0063,
+      "step": 521
+    },
+    {
+      "epoch": 4.368200836820083,
+      "grad_norm": 0.01820644736289978,
+      "learning_rate": 4.52394588273502e-06,
+      "loss": 0.0065,
+      "step": 522
+    },
+    {
+      "epoch": 4.376569037656903,
+      "grad_norm": 0.01679069548845291,
+      "learning_rate": 4.40269068368801e-06,
+      "loss": 0.0065,
+      "step": 523
+    },
+    {
+      "epoch": 4.384937238493724,
+      "grad_norm": 0.017241882160305977,
+      "learning_rate": 4.283007765579267e-06,
+      "loss": 0.0062,
+      "step": 524
+    },
+    {
+      "epoch": 4.393305439330544,
+      "grad_norm": 0.02121458761394024,
+      "learning_rate": 4.164901255300779e-06,
+      "loss": 0.0065,
+      "step": 525
+    },
+    {
+      "epoch": 4.393305439330544,
+      "eval_loss": 0.01406473945826292,
+      "eval_runtime": 8.2004,
+      "eval_samples_per_second": 6.097,
+      "eval_steps_per_second": 1.585,
+      "step": 525
+    },
+    {
+      "epoch": 4.401673640167364,
+      "grad_norm": 0.01692904345691204,
+      "learning_rate": 4.048375225387036e-06,
+      "loss": 0.0057,
+      "step": 526
+    },
+    {
+      "epoch": 4.410041841004184,
+      "grad_norm": 0.017979208379983902,
+      "learning_rate": 3.9334336938746e-06,
+      "loss": 0.0059,
+      "step": 527
+    },
+    {
+      "epoch": 4.418410041841004,
+      "grad_norm": 0.015546801500022411,
+      "learning_rate": 3.82008062416353e-06,
+      "loss": 0.0064,
+      "step": 528
+    },
+    {
+      "epoch": 4.426778242677824,
+      "grad_norm": 0.014850549399852753,
+      "learning_rate": 3.7083199248807656e-06,
+      "loss": 0.0058,
+      "step": 529
+    },
+    {
+      "epoch": 4.435146443514644,
+      "grad_norm": 0.02002605050802231,
+      "learning_rate": 3.5981554497452884e-06,
+      "loss": 0.007,
+      "step": 530
+    },
+    {
+      "epoch": 4.435146443514644,
+      "eval_loss": 0.013939818367362022,
+      "eval_runtime": 8.1895,
+      "eval_samples_per_second": 6.105,
+      "eval_steps_per_second": 1.587,
+      "step": 530
+    },
+    {
+      "epoch": 4.443514644351464,
+      "grad_norm": 0.015825215727090836,
+      "learning_rate": 3.4895909974352935e-06,
+      "loss": 0.0066,
+      "step": 531
+    },
+    {
+      "epoch": 4.451882845188284,
+      "grad_norm": 0.017673788592219353,
+      "learning_rate": 3.38263031145718e-06,
+      "loss": 0.007,
+      "step": 532
+    },
+    {
+      "epoch": 4.460251046025105,
+      "grad_norm": 0.014746863394975662,
+      "learning_rate": 3.27727708001645e-06,
+      "loss": 0.0055,
+      "step": 533
+    },
+    {
+      "epoch": 4.468619246861925,
+      "grad_norm": 0.017045453190803528,
+      "learning_rate": 3.173534935890582e-06,
+      "loss": 0.0058,
+      "step": 534
+    },
+    {
+      "epoch": 4.476987447698745,
+      "grad_norm": 0.0173475444316864,
+      "learning_rate": 3.0714074563037044e-06,
+      "loss": 0.0058,
+      "step": 535
+    },
+    {
+      "epoch": 4.476987447698745,
+      "eval_loss": 0.013857071287930012,
+      "eval_runtime": 8.227,
+      "eval_samples_per_second": 6.078,
+      "eval_steps_per_second": 1.58,
+      "step": 535
+    },
+    {
+      "epoch": 4.485355648535565,
+      "grad_norm": 0.014112849719822407,
+      "learning_rate": 2.970898162803287e-06,
+      "loss": 0.0062,
+      "step": 536
+    },
+    {
+      "epoch": 4.493723849372385,
+      "grad_norm": 0.018640313297510147,
+      "learning_rate": 2.8720105211387183e-06,
+      "loss": 0.006,
+      "step": 537
+    },
+    {
+      "epoch": 4.502092050209205,
+      "grad_norm": 0.018944760784506798,
+      "learning_rate": 2.774747941141764e-06,
+      "loss": 0.0065,
+      "step": 538
+    },
+    {
+      "epoch": 4.510460251046025,
+      "grad_norm": 0.017405930906534195,
+      "learning_rate": 2.6791137766090157e-06,
+      "loss": 0.0068,
+      "step": 539
+    },
+    {
+      "epoch": 4.518828451882845,
+      "grad_norm": 0.019485093653202057,
+      "learning_rate": 2.5851113251862345e-06,
+      "loss": 0.0056,
+      "step": 540
+    },
+    {
+      "epoch": 4.518828451882845,
+      "eval_loss": 0.01391106192022562,
+      "eval_runtime": 8.1885,
+      "eval_samples_per_second": 6.106,
+      "eval_steps_per_second": 1.588,
+      "step": 540
+    },
+    {
+      "epoch": 4.527196652719665,
+      "grad_norm": 0.01877741701900959,
+      "learning_rate": 2.4927438282546567e-06,
+      "loss": 0.0068,
+      "step": 541
+    },
+    {
+      "epoch": 4.535564853556485,
+      "grad_norm": 0.0198155976831913,
+      "learning_rate": 2.4020144708192026e-06,
+      "loss": 0.006,
+      "step": 542
+    },
+    {
+      "epoch": 4.543933054393305,
+      "grad_norm": 0.01476176455616951,
+      "learning_rate": 2.3129263813986724e-06,
+      "loss": 0.0062,
+      "step": 543
+    },
+    {
+      "epoch": 4.552301255230126,
+      "grad_norm": 0.01914817839860916,
+      "learning_rate": 2.225482631917847e-06,
+      "loss": 0.0067,
+      "step": 544
+    },
+    {
+      "epoch": 4.560669456066946,
+      "grad_norm": 0.0210384763777256,
+      "learning_rate": 2.1396862376015903e-06,
+      "loss": 0.0062,
+      "step": 545
+    },
+    {
+      "epoch": 4.560669456066946,
+      "eval_loss": 0.013888537883758545,
+      "eval_runtime": 8.1978,
+      "eval_samples_per_second": 6.099,
+      "eval_steps_per_second": 1.586,
+      "step": 545
+    },
+    {
+      "epoch": 4.569037656903766,
+      "grad_norm": 0.021067198365926743,
+      "learning_rate": 2.0555401568708554e-06,
+      "loss": 0.0065,
+      "step": 546
+    },
+    {
+      "epoch": 4.577405857740586,
+      "grad_norm": 0.017656171694397926,
+      "learning_rate": 1.9730472912406816e-06,
+      "loss": 0.0071,
+      "step": 547
+    },
+    {
+      "epoch": 4.585774058577406,
+      "grad_norm": 0.017184708267450333,
+      "learning_rate": 1.8922104852201528e-06,
+      "loss": 0.0064,
+      "step": 548
+    },
+    {
+      "epoch": 4.594142259414226,
+      "grad_norm": 0.02077127993106842,
+      "learning_rate": 1.8130325262142755e-06,
+      "loss": 0.0066,
+      "step": 549
+    },
+    {
+      "epoch": 4.602510460251046,
+      "grad_norm": 0.013661603443324566,
+      "learning_rate": 1.7355161444279344e-06,
+      "loss": 0.0061,
+      "step": 550
+    },
+    {
+      "epoch": 4.602510460251046,
+      "eval_loss": 0.013935999944806099,
+      "eval_runtime": 8.2068,
+      "eval_samples_per_second": 6.093,
+      "eval_steps_per_second": 1.584,
+      "step": 550
+    },
+    {
+      "epoch": 4.610878661087866,
+      "grad_norm": 0.033053044229745865,
+      "learning_rate": 1.6596640127716778e-06,
+      "loss": 0.0074,
+      "step": 551
+    },
+    {
+      "epoch": 4.619246861924686,
+      "grad_norm": 0.019544832408428192,
+      "learning_rate": 1.5854787467695975e-06,
+      "loss": 0.0066,
+      "step": 552
+    },
+    {
+      "epoch": 4.627615062761507,
+      "grad_norm": 0.017127815634012222,
+      "learning_rate": 1.512962904469112e-06,
+      "loss": 0.0057,
+      "step": 553
+    },
+    {
+      "epoch": 4.635983263598327,
+      "grad_norm": 0.024544665589928627,
+      "learning_rate": 1.442118986352775e-06,
+      "loss": 0.0064,
+      "step": 554
+    },
+    {
+      "epoch": 4.644351464435147,
+      "grad_norm": 0.016828225925564766,
+      "learning_rate": 1.3729494352520578e-06,
+      "loss": 0.0061,
+      "step": 555
+    },
+    {
+      "epoch": 4.644351464435147,
+      "eval_loss": 0.013867921195924282,
+      "eval_runtime": 8.1969,
+      "eval_samples_per_second": 6.1,
+      "eval_steps_per_second": 1.586,
+      "step": 555
+    },
+    {
+      "epoch": 4.652719665271967,
+      "grad_norm": 0.016187148168683052,
+      "learning_rate": 1.3054566362631082e-06,
+      "loss": 0.0068,
+      "step": 556
+    },
+    {
+      "epoch": 4.661087866108787,
+      "grad_norm": 0.018801085650920868,
+      "learning_rate": 1.2396429166645073e-06,
+      "loss": 0.0064,
+      "step": 557
+    },
+    {
+      "epoch": 4.669456066945607,
+      "grad_norm": 0.01601874642074108,
+      "learning_rate": 1.175510545837022e-06,
+      "loss": 0.0062,
+      "step": 558
+    },
+    {
+      "epoch": 4.677824267782427,
+      "grad_norm": 0.02034141682088375,
+      "learning_rate": 1.1130617351853446e-06,
+      "loss": 0.0057,
+      "step": 559
+    },
+    {
+      "epoch": 4.686192468619247,
+      "grad_norm": 0.017557624727487564,
+      "learning_rate": 1.0522986380618605e-06,
+      "loss": 0.0068,
+      "step": 560
+    },
+    {
+      "epoch": 4.686192468619247,
+      "eval_loss": 0.01384110189974308,
+      "eval_runtime": 8.1896,
+      "eval_samples_per_second": 6.105,
+      "eval_steps_per_second": 1.587,
+      "step": 560
+    },
+    {
+      "epoch": 4.694560669456067,
+      "grad_norm": 0.015541495755314827,
+      "learning_rate": 9.932233496923903e-07,
+      "loss": 0.0063,
+      "step": 561
+    },
+    {
+      "epoch": 4.702928870292887,
+      "grad_norm": 0.01691366545855999,
+      "learning_rate": 9.358379071039147e-07,
+      "loss": 0.0069,
+      "step": 562
+    },
+    {
+      "epoch": 4.711297071129707,
+      "grad_norm": 0.017451630905270576,
+      "learning_rate": 8.801442890543743e-07,
+      "loss": 0.0057,
+      "step": 563
+    },
+    {
+      "epoch": 4.7196652719665275,
+      "grad_norm": 0.013958767987787724,
+      "learning_rate": 8.261444159644083e-07,
+      "loss": 0.0059,
+      "step": 564
+    },
+    {
+      "epoch": 4.7280334728033475,
+      "grad_norm": 0.015628065913915634,
+      "learning_rate": 7.738401498511405e-07,
+      "loss": 0.0069,
+      "step": 565
+    },
+    {
+      "epoch": 4.7280334728033475,
+      "eval_loss": 0.013867964036762714,
+      "eval_runtime": 8.201,
+      "eval_samples_per_second": 6.097,
+      "eval_steps_per_second": 1.585,
+      "step": 565
+    },
+    {
+      "epoch": 4.7364016736401675,
+      "grad_norm": 0.017734922468662262,
+      "learning_rate": 7.232332942639974e-07,
+      "loss": 0.0066,
+      "step": 566
+    },
+    {
+      "epoch": 4.7447698744769875,
+      "grad_norm": 0.01975751295685768,
+      "learning_rate": 6.743255942224913e-07,
+      "loss": 0.0063,
+      "step": 567
+    },
+    {
+      "epoch": 4.7531380753138075,
+      "grad_norm": 0.01823136769235134,
+      "learning_rate": 6.27118736156046e-07,
+      "loss": 0.0063,
+      "step": 568
+    },
+    {
+      "epoch": 4.7615062761506275,
+      "grad_norm": 0.01851455122232437,
+      "learning_rate": 5.816143478458714e-07,
+      "loss": 0.0059,
+      "step": 569
+    },
+    {
+      "epoch": 4.7698744769874475,
+      "grad_norm": 0.019744612276554108,
+      "learning_rate": 5.378139983688135e-07,
+      "loss": 0.0063,
+      "step": 570
+    },
+    {
+      "epoch": 4.7698744769874475,
+      "eval_loss": 0.013921505771577358,
+      "eval_runtime": 8.2248,
+      "eval_samples_per_second": 6.079,
+      "eval_steps_per_second": 1.581,
+      "step": 570
+    },
+    {
+      "epoch": 4.7782426778242675,
+      "grad_norm": 0.017613310366868973,
+      "learning_rate": 4.957191980432541e-07,
+      "loss": 0.0061,
+      "step": 571
+    },
+    {
+      "epoch": 4.786610878661088,
+      "grad_norm": 0.01842423528432846,
+      "learning_rate": 4.5533139837704574e-07,
+      "loss": 0.0065,
+      "step": 572
+    },
+    {
+      "epoch": 4.794979079497908,
+      "grad_norm": 0.015000440180301666,
+      "learning_rate": 4.1665199201744167e-07,
+      "loss": 0.0065,
+      "step": 573
+    },
+    {
+      "epoch": 4.803347280334728,
+      "grad_norm": 0.01420409232378006,
+      "learning_rate": 3.7968231270307244e-07,
+      "loss": 0.006,
+      "step": 574
+    },
+    {
+      "epoch": 4.811715481171548,
+      "grad_norm": 0.014016808941960335,
+      "learning_rate": 3.4442363521798304e-07,
+      "loss": 0.0065,
+      "step": 575
+    },
+    {
+      "epoch": 4.811715481171548,
+      "eval_loss": 0.01387822162359953,
+      "eval_runtime": 8.1941,
+      "eval_samples_per_second": 6.102,
+      "eval_steps_per_second": 1.587,
+      "step": 575
+    },
+    {
+      "epoch": 4.820083682008368,
+      "grad_norm": 0.025155773386359215,
+      "learning_rate": 3.1087717534764584e-07,
+      "loss": 0.0065,
+      "step": 576
+    },
+    {
+      "epoch": 4.828451882845188,
+      "grad_norm": 0.017688661813735962,
+      "learning_rate": 2.7904408983706607e-07,
+      "loss": 0.0062,
+      "step": 577
+    },
+    {
+      "epoch": 4.836820083682008,
+      "grad_norm": 0.014572778716683388,
+      "learning_rate": 2.489254763508642e-07,
+      "loss": 0.0063,
+      "step": 578
+    },
+    {
+      "epoch": 4.845188284518828,
+      "grad_norm": 0.015580548904836178,
+      "learning_rate": 2.2052237343545e-07,
+      "loss": 0.0058,
+      "step": 579
+    },
+    {
+      "epoch": 4.853556485355648,
+      "grad_norm": 0.01487270649522543,
+      "learning_rate": 1.938357604832075e-07,
+      "loss": 0.0064,
+      "step": 580
+    },
+    {
+      "epoch": 4.853556485355648,
+      "eval_loss": 0.013862421736121178,
+      "eval_runtime": 8.192,
+      "eval_samples_per_second": 6.104,
+      "eval_steps_per_second": 1.587,
+      "step": 580
+    },
+    {
+      "epoch": 4.861924686192468,
+      "grad_norm": 0.020253285765647888,
+      "learning_rate": 1.6886655769871029e-07,
+      "loss": 0.0058,
+      "step": 581
+    },
+    {
+      "epoch": 4.870292887029288,
+      "grad_norm": 0.01656149886548519,
+      "learning_rate": 1.4561562606701385e-07,
+      "loss": 0.0062,
+      "step": 582
+    },
+    {
+      "epoch": 4.878661087866108,
+      "grad_norm": 0.016599709168076515,
+      "learning_rate": 1.240837673239459e-07,
+      "loss": 0.0065,
+      "step": 583
+    },
+    {
+      "epoch": 4.887029288702929,
+      "grad_norm": 0.01712622120976448,
+      "learning_rate": 1.0427172392847295e-07,
+      "loss": 0.0062,
+      "step": 584
+    },
+    {
+      "epoch": 4.895397489539749,
+      "grad_norm": 0.01667969673871994,
+      "learning_rate": 8.618017903708197e-08,
+      "loss": 0.0062,
+      "step": 585
+    },
+    {
+      "epoch": 4.895397489539749,
+      "eval_loss": 0.013880819082260132,
+      "eval_runtime": 8.2034,
+      "eval_samples_per_second": 6.095,
+      "eval_steps_per_second": 1.585,
+      "step": 585
+    },
+    {
+      "epoch": 4.903765690376569,
+      "grad_norm": 0.019158054143190384,
+      "learning_rate": 6.980975648026578e-08,
+      "loss": 0.0061,
+      "step": 586
+    },
+    {
+      "epoch": 4.912133891213389,
+      "grad_norm": 0.013462690636515617,
+      "learning_rate": 5.516102074095697e-08,
+      "loss": 0.0062,
+      "step": 587
+    },
+    {
+      "epoch": 4.920502092050209,
+      "grad_norm": 0.014567787759006023,
+      "learning_rate": 4.2234476935099074e-08,
+      "loss": 0.0062,
+      "step": 588
+    },
+    {
+      "epoch": 4.928870292887029,
+      "grad_norm": 0.014899435453116894,
+      "learning_rate": 3.103057079422711e-08,
+      "loss": 0.0057,
+      "step": 589
+    },
+    {
+      "epoch": 4.937238493723849,
+      "grad_norm": 0.013964378274977207,
+      "learning_rate": 2.154968865007989e-08,
+      "loss": 0.0065,
+      "step": 590
+    },
+    {
+      "epoch": 4.937238493723849,
+      "eval_loss": 0.013855883851647377,
+      "eval_runtime": 8.2021,
+      "eval_samples_per_second": 6.096,
+      "eval_steps_per_second": 1.585,
+      "step": 590
+    },
+    {
+      "epoch": 4.945606694560669,
+      "grad_norm": 0.013118981383740902,
+      "learning_rate": 1.3792157421299579e-08,
+      "loss": 0.006,
+      "step": 591
+    },
+    {
+      "epoch": 4.95397489539749,
+      "grad_norm": 0.01663290336728096,
+      "learning_rate": 7.758244602151799e-09,
+      "loss": 0.0057,
+      "step": 592
+    },
+    {
+      "epoch": 4.96234309623431,
+      "grad_norm": 0.01918199472129345,
+      "learning_rate": 3.448158253277489e-09,
+      "loss": 0.0059,
+      "step": 593
+    },
+    {
+      "epoch": 4.97071129707113,
+      "grad_norm": 0.016706887632608414,
+      "learning_rate": 8.620469945708198e-10,
+      "loss": 0.0062,
+      "step": 594
+    },
+    {
+      "epoch": 4.97907949790795,
+      "grad_norm": 0.016998644918203354,
+      "learning_rate": 0.0,
+      "loss": 0.0055,
+      "step": 595
+    },
+    {
+      "epoch": 4.97907949790795,
+      "eval_loss": 0.01388928946107626,
+      "eval_runtime": 8.1966,
+      "eval_samples_per_second": 6.1,
+      "eval_steps_per_second": 1.586,
+      "step": 595
+    },
+    {
+      "epoch": 4.97907949790795,
+      "step": 595,
+      "total_flos": 1.199536935920124e+18,
+      "train_loss": 0.012196063881880846,
+      "train_runtime": 9595.7495,
+      "train_samples_per_second": 1.991,
+      "train_steps_per_second": 0.062
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 595,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 5,
+  "save_steps": 50,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.199536935920124e+18,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}