diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,6570 @@
+{
+  "best_global_step": 400,
+  "best_metric": 0.9147540983606557,
+  "best_model_checkpoint": "./modelka_bebro/checkpoint-400",
+  "epoch": 20.0,
+  "eval_steps": 20,
+  "global_step": 860,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.023255813953488372,
+      "grad_norm": 0.4809555113315582,
+      "learning_rate": 0.0,
+      "loss": 0.6913,
+      "step": 1
+    },
+    {
+      "epoch": 0.046511627906976744,
+      "grad_norm": 0.4111059904098511,
+      "learning_rate": 5.813953488372093e-07,
+      "loss": 0.6908,
+      "step": 2
+    },
+    {
+      "epoch": 0.06976744186046512,
+      "grad_norm": 0.4677865207195282,
+      "learning_rate": 1.1627906976744186e-06,
+      "loss": 0.6945,
+      "step": 3
+    },
+    {
+      "epoch": 0.09302325581395349,
+      "grad_norm": 0.42459365725517273,
+      "learning_rate": 1.744186046511628e-06,
+      "loss": 0.6974,
+      "step": 4
+    },
+    {
+      "epoch": 0.11627906976744186,
+      "grad_norm": 0.5455716252326965,
+      "learning_rate": 2.325581395348837e-06,
+      "loss": 0.6876,
+      "step": 5
+    },
+    {
+      "epoch": 0.13953488372093023,
+      "grad_norm": 0.3835294842720032,
+      "learning_rate": 2.9069767441860468e-06,
+      "loss": 0.6882,
+      "step": 6
+    },
+    {
+      "epoch": 0.16279069767441862,
+      "grad_norm": 0.6538524627685547,
+      "learning_rate": 3.488372093023256e-06,
+      "loss": 0.6888,
+      "step": 7
+    },
+    {
+      "epoch": 0.18604651162790697,
+      "grad_norm": 0.40157362818717957,
+      "learning_rate": 4.0697674418604655e-06,
+      "loss": 0.6855,
+      "step": 8
+    },
+    {
+      "epoch": 0.20930232558139536,
+      "grad_norm": 0.43001478910446167,
+      "learning_rate": 4.651162790697674e-06,
+      "loss": 0.6918,
+      "step": 9
+    },
+    {
+      "epoch": 0.23255813953488372,
+      "grad_norm": 0.5613086819648743,
+      "learning_rate": 5.232558139534884e-06,
+      "loss": 0.6977,
+      "step": 10
+    },
+    {
+      "epoch": 0.2558139534883721,
+      "grad_norm": 0.3894985020160675,
+      "learning_rate": 5.8139534883720935e-06,
+      "loss": 0.6847,
+      "step": 11
+    },
+    {
+      "epoch": 0.27906976744186046,
+      "grad_norm": 0.4243035912513733,
+      "learning_rate": 6.395348837209303e-06,
+      "loss": 0.6934,
+      "step": 12
+    },
+    {
+      "epoch": 0.3023255813953488,
+      "grad_norm": 0.5996053218841553,
+      "learning_rate": 6.976744186046512e-06,
+      "loss": 0.6966,
+      "step": 13
+    },
+    {
+      "epoch": 0.32558139534883723,
+      "grad_norm": 0.5445254445075989,
+      "learning_rate": 7.558139534883721e-06,
+      "loss": 0.695,
+      "step": 14
+    },
+    {
+      "epoch": 0.3488372093023256,
+      "grad_norm": 0.433915913105011,
+      "learning_rate": 8.139534883720931e-06,
+      "loss": 0.6855,
+      "step": 15
+    },
+    {
+      "epoch": 0.37209302325581395,
+      "grad_norm": 0.43263381719589233,
+      "learning_rate": 8.72093023255814e-06,
+      "loss": 0.6901,
+      "step": 16
+    },
+    {
+      "epoch": 0.3953488372093023,
+      "grad_norm": 0.5314894318580627,
+      "learning_rate": 9.302325581395349e-06,
+      "loss": 0.6966,
+      "step": 17
+    },
+    {
+      "epoch": 0.4186046511627907,
+      "grad_norm": 0.3562123477458954,
+      "learning_rate": 9.883720930232558e-06,
+      "loss": 0.6877,
+      "step": 18
+    },
+    {
+      "epoch": 0.4418604651162791,
+      "grad_norm": 0.4751898944377899,
+      "learning_rate": 1.0465116279069768e-05,
+      "loss": 0.6834,
+      "step": 19
+    },
+    {
+      "epoch": 0.46511627906976744,
+      "grad_norm": 0.4273492395877838,
+      "learning_rate": 1.1046511627906977e-05,
+      "loss": 0.6885,
+      "step": 20
+    },
+    {
+      "epoch": 0.46511627906976744,
+      "eval_accuracy": 0.5901639344262295,
+      "eval_f1": 0.5314093976229397,
+      "eval_loss": 0.68656325340271,
+      "eval_precision": 0.6364057433997221,
+      "eval_recall": 0.5761178599121219,
+      "eval_runtime": 0.0751,
+      "eval_samples_per_second": 4062.686,
+      "eval_steps_per_second": 66.601,
+      "step": 20
+    },
+    {
+      "epoch": 0.4883720930232558,
+      "grad_norm": 0.5205795764923096,
+      "learning_rate": 1.1627906976744187e-05,
+      "loss": 0.6908,
+      "step": 21
+    },
+    {
+      "epoch": 0.5116279069767442,
+      "grad_norm": 0.4466778635978699,
+      "learning_rate": 1.2209302325581395e-05,
+      "loss": 0.6861,
+      "step": 22
+    },
+    {
+      "epoch": 0.5348837209302325,
+      "grad_norm": 0.44298067688941956,
+      "learning_rate": 1.2790697674418606e-05,
+      "loss": 0.6832,
+      "step": 23
+    },
+    {
+      "epoch": 0.5581395348837209,
+      "grad_norm": 0.39592233300209045,
+      "learning_rate": 1.3372093023255814e-05,
+      "loss": 0.6877,
+      "step": 24
+    },
+    {
+      "epoch": 0.5813953488372093,
+      "grad_norm": 0.4810555875301361,
+      "learning_rate": 1.3953488372093024e-05,
+      "loss": 0.68,
+      "step": 25
+    },
+    {
+      "epoch": 0.6046511627906976,
+      "grad_norm": 0.47530311346054077,
+      "learning_rate": 1.4534883720930233e-05,
+      "loss": 0.6837,
+      "step": 26
+    },
+    {
+      "epoch": 0.627906976744186,
+      "grad_norm": 0.5110817551612854,
+      "learning_rate": 1.5116279069767441e-05,
+      "loss": 0.6908,
+      "step": 27
+    },
+    {
+      "epoch": 0.6511627906976745,
+      "grad_norm": 0.44798141717910767,
+      "learning_rate": 1.569767441860465e-05,
+      "loss": 0.6756,
+      "step": 28
+    },
+    {
+      "epoch": 0.6744186046511628,
+      "grad_norm": 0.4494743347167969,
+      "learning_rate": 1.6279069767441862e-05,
+      "loss": 0.6813,
+      "step": 29
+    },
+    {
+      "epoch": 0.6976744186046512,
+      "grad_norm": 0.4449148178100586,
+      "learning_rate": 1.686046511627907e-05,
+      "loss": 0.6786,
+      "step": 30
+    },
+    {
+      "epoch": 0.7209302325581395,
+      "grad_norm": 0.49257639050483704,
+      "learning_rate": 1.744186046511628e-05,
+      "loss": 0.6821,
+      "step": 31
+    },
+    {
+      "epoch": 0.7441860465116279,
+      "grad_norm": 0.49307772517204285,
+      "learning_rate": 1.802325581395349e-05,
+      "loss": 0.6856,
+      "step": 32
+    },
+    {
+      "epoch": 0.7674418604651163,
+      "grad_norm": 0.4002581536769867,
+      "learning_rate": 1.8604651162790697e-05,
+      "loss": 0.6804,
+      "step": 33
+    },
+    {
+      "epoch": 0.7906976744186046,
+      "grad_norm": 0.44210293889045715,
+      "learning_rate": 1.918604651162791e-05,
+      "loss": 0.6739,
+      "step": 34
+    },
+    {
+      "epoch": 0.813953488372093,
+      "grad_norm": 0.3651827871799469,
+      "learning_rate": 1.9767441860465116e-05,
+      "loss": 0.6839,
+      "step": 35
+    },
+    {
+      "epoch": 0.8372093023255814,
+      "grad_norm": 0.5120950937271118,
+      "learning_rate": 2.0348837209302328e-05,
+      "loss": 0.6746,
+      "step": 36
+    },
+    {
+      "epoch": 0.8604651162790697,
+      "grad_norm": 0.4111502766609192,
+      "learning_rate": 2.0930232558139536e-05,
+      "loss": 0.6805,
+      "step": 37
+    },
+    {
+      "epoch": 0.8837209302325582,
+      "grad_norm": 0.44530022144317627,
+      "learning_rate": 2.1511627906976744e-05,
+      "loss": 0.6833,
+      "step": 38
+    },
+    {
+      "epoch": 0.9069767441860465,
+      "grad_norm": 0.4006333649158478,
+      "learning_rate": 2.2093023255813955e-05,
+      "loss": 0.6769,
+      "step": 39
+    },
+    {
+      "epoch": 0.9302325581395349,
+      "grad_norm": 0.4231868386268616,
+      "learning_rate": 2.2674418604651163e-05,
+      "loss": 0.6721,
+      "step": 40
+    },
+    {
+      "epoch": 0.9302325581395349,
+      "eval_accuracy": 0.659016393442623,
+      "eval_f1": 0.649208174820844,
+      "eval_loss": 0.6683934926986694,
+      "eval_precision": 0.6947368421052631,
+      "eval_recall": 0.6673559059188421,
+      "eval_runtime": 0.0782,
+      "eval_samples_per_second": 3899.573,
+      "eval_steps_per_second": 63.927,
+      "step": 40
+    },
+    {
+      "epoch": 0.9534883720930233,
+      "grad_norm": 0.4570966064929962,
+      "learning_rate": 2.3255813953488374e-05,
+      "loss": 0.679,
+      "step": 41
+    },
+    {
+      "epoch": 0.9767441860465116,
+      "grad_norm": 0.6141464710235596,
+      "learning_rate": 2.3837209302325582e-05,
+      "loss": 0.6768,
+      "step": 42
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.6506465673446655,
+      "learning_rate": 2.441860465116279e-05,
+      "loss": 0.6573,
+      "step": 43
+    },
+    {
+      "epoch": 1.0232558139534884,
+      "grad_norm": 1.0706639289855957,
+      "learning_rate": 2.5e-05,
+      "loss": 0.6865,
+      "step": 44
+    },
+    {
+      "epoch": 1.0465116279069768,
+      "grad_norm": 0.4623822867870331,
+      "learning_rate": 2.5581395348837212e-05,
+      "loss": 0.6538,
+      "step": 45
+    },
+    {
+      "epoch": 1.069767441860465,
+      "grad_norm": 0.4986269176006317,
+      "learning_rate": 2.616279069767442e-05,
+      "loss": 0.6585,
+      "step": 46
+    },
+    {
+      "epoch": 1.0930232558139534,
+      "grad_norm": 0.48053082823753357,
+      "learning_rate": 2.674418604651163e-05,
+      "loss": 0.6506,
+      "step": 47
+    },
+    {
+      "epoch": 1.1162790697674418,
+      "grad_norm": 0.5987160801887512,
+      "learning_rate": 2.7325581395348836e-05,
+      "loss": 0.6691,
+      "step": 48
+    },
+    {
+      "epoch": 1.1395348837209303,
+      "grad_norm": 0.685001015663147,
+      "learning_rate": 2.7906976744186048e-05,
+      "loss": 0.6512,
+      "step": 49
+    },
+    {
+      "epoch": 1.1627906976744187,
+      "grad_norm": 0.5840403437614441,
+      "learning_rate": 2.848837209302326e-05,
+      "loss": 0.6759,
+      "step": 50
+    },
+    {
+      "epoch": 1.1860465116279069,
+      "grad_norm": 0.6035524606704712,
+      "learning_rate": 2.9069767441860467e-05,
+      "loss": 0.6502,
+      "step": 51
+    },
+    {
+      "epoch": 1.2093023255813953,
+      "grad_norm": 0.6236196160316467,
+      "learning_rate": 2.9651162790697678e-05,
+      "loss": 0.6575,
+      "step": 52
+    },
+    {
+      "epoch": 1.2325581395348837,
+      "grad_norm": 0.5840880274772644,
+      "learning_rate": 3.0232558139534883e-05,
+      "loss": 0.6438,
+      "step": 53
+    },
+    {
+      "epoch": 1.255813953488372,
+      "grad_norm": 0.7563631534576416,
+      "learning_rate": 3.081395348837209e-05,
+      "loss": 0.645,
+      "step": 54
+    },
+    {
+      "epoch": 1.2790697674418605,
+      "grad_norm": 0.8720300793647766,
+      "learning_rate": 3.13953488372093e-05,
+      "loss": 0.6551,
+      "step": 55
+    },
+    {
+      "epoch": 1.302325581395349,
+      "grad_norm": 0.5763499140739441,
+      "learning_rate": 3.197674418604651e-05,
+      "loss": 0.6304,
+      "step": 56
+    },
+    {
+      "epoch": 1.3255813953488373,
+      "grad_norm": 0.5487861037254333,
+      "learning_rate": 3.2558139534883724e-05,
+      "loss": 0.6373,
+      "step": 57
+    },
+    {
+      "epoch": 1.3488372093023255,
+      "grad_norm": 0.5714741945266724,
+      "learning_rate": 3.313953488372093e-05,
+      "loss": 0.638,
+      "step": 58
+    },
+    {
+      "epoch": 1.372093023255814,
+      "grad_norm": 0.7604700326919556,
+      "learning_rate": 3.372093023255814e-05,
+      "loss": 0.6537,
+      "step": 59
+    },
+    {
+      "epoch": 1.3953488372093024,
+      "grad_norm": 0.8128901720046997,
+      "learning_rate": 3.430232558139535e-05,
+      "loss": 0.6573,
+      "step": 60
+    },
+    {
+      "epoch": 1.3953488372093024,
+      "eval_accuracy": 0.760655737704918,
+      "eval_f1": 0.7589037239168805,
+      "eval_loss": 0.6135660409927368,
+      "eval_precision": 0.7778153153153153,
+      "eval_recall": 0.7656801929869905,
+      "eval_runtime": 0.0732,
+      "eval_samples_per_second": 4169.193,
+      "eval_steps_per_second": 68.347,
+      "step": 60
+    },
+    {
+      "epoch": 1.4186046511627908,
+      "grad_norm": 0.6136202216148376,
+      "learning_rate": 3.488372093023256e-05,
+      "loss": 0.6257,
+      "step": 61
+    },
+    {
+      "epoch": 1.441860465116279,
+      "grad_norm": 0.6842197179794312,
+      "learning_rate": 3.5465116279069774e-05,
+      "loss": 0.6065,
+      "step": 62
+    },
+    {
+      "epoch": 1.4651162790697674,
+      "grad_norm": 0.6682329177856445,
+      "learning_rate": 3.604651162790698e-05,
+      "loss": 0.6114,
+      "step": 63
+    },
+    {
+      "epoch": 1.4883720930232558,
+      "grad_norm": 0.7913668155670166,
+      "learning_rate": 3.662790697674418e-05,
+      "loss": 0.636,
+      "step": 64
+    },
+    {
+      "epoch": 1.5116279069767442,
+      "grad_norm": 0.7657461762428284,
+      "learning_rate": 3.7209302325581394e-05,
+      "loss": 0.6067,
+      "step": 65
+    },
+    {
+      "epoch": 1.5348837209302326,
+      "grad_norm": 0.9346985816955566,
+      "learning_rate": 3.7790697674418606e-05,
+      "loss": 0.5902,
+      "step": 66
+    },
+    {
+      "epoch": 1.558139534883721,
+      "grad_norm": 0.7872809767723083,
+      "learning_rate": 3.837209302325582e-05,
+      "loss": 0.5731,
+      "step": 67
+    },
+    {
+      "epoch": 1.5813953488372094,
+      "grad_norm": 0.8864495754241943,
+      "learning_rate": 3.895348837209303e-05,
+      "loss": 0.581,
+      "step": 68
+    },
+    {
+      "epoch": 1.6046511627906976,
+      "grad_norm": 0.8121451139450073,
+      "learning_rate": 3.953488372093023e-05,
+      "loss": 0.5734,
+      "step": 69
+    },
+    {
+      "epoch": 1.627906976744186,
+      "grad_norm": 1.0192168951034546,
+      "learning_rate": 4.0116279069767444e-05,
+      "loss": 0.5645,
+      "step": 70
+    },
+    {
+      "epoch": 1.6511627906976745,
+      "grad_norm": 0.9971368312835693,
+      "learning_rate": 4.0697674418604655e-05,
+      "loss": 0.5948,
+      "step": 71
+    },
+    {
+      "epoch": 1.6744186046511627,
+      "grad_norm": 0.9530318975448608,
+      "learning_rate": 4.127906976744187e-05,
+      "loss": 0.5129,
+      "step": 72
+    },
+    {
+      "epoch": 1.697674418604651,
+      "grad_norm": 0.980096161365509,
+      "learning_rate": 4.186046511627907e-05,
+      "loss": 0.5767,
+      "step": 73
+    },
+    {
+      "epoch": 1.7209302325581395,
+      "grad_norm": 0.991642951965332,
+      "learning_rate": 4.2441860465116276e-05,
+      "loss": 0.5403,
+      "step": 74
+    },
+    {
+      "epoch": 1.744186046511628,
+      "grad_norm": 0.8262429237365723,
+      "learning_rate": 4.302325581395349e-05,
+      "loss": 0.5467,
+      "step": 75
+    },
+    {
+      "epoch": 1.7674418604651163,
+      "grad_norm": 1.1981412172317505,
+      "learning_rate": 4.36046511627907e-05,
+      "loss": 0.5081,
+      "step": 76
+    },
+    {
+      "epoch": 1.7906976744186047,
+      "grad_norm": 0.9658230543136597,
+      "learning_rate": 4.418604651162791e-05,
+      "loss": 0.5647,
+      "step": 77
+    },
+    {
+      "epoch": 1.8139534883720931,
+      "grad_norm": 1.1805201768875122,
+      "learning_rate": 4.476744186046512e-05,
+      "loss": 0.4536,
+      "step": 78
+    },
+    {
+      "epoch": 1.8372093023255816,
+      "grad_norm": 1.101920485496521,
+      "learning_rate": 4.5348837209302326e-05,
+      "loss": 0.5361,
+      "step": 79
+    },
+    {
+      "epoch": 1.8604651162790697,
+      "grad_norm": 1.0350321531295776,
+      "learning_rate": 4.593023255813954e-05,
+      "loss": 0.5207,
+      "step": 80
+    },
+    {
+      "epoch": 1.8604651162790697,
+      "eval_accuracy": 0.8032786885245902,
+      "eval_f1": 0.8030224748127099,
+      "eval_loss": 0.4702507555484772,
+      "eval_precision": 0.8029135823721811,
+      "eval_recall": 0.8032006547772896,
+      "eval_runtime": 0.0743,
+      "eval_samples_per_second": 4102.555,
+      "eval_steps_per_second": 67.255,
+      "step": 80
+    },
+    {
+      "epoch": 1.8837209302325582,
+      "grad_norm": 1.0638865232467651,
+      "learning_rate": 4.651162790697675e-05,
+      "loss": 0.5351,
+      "step": 81
+    },
+    {
+      "epoch": 1.9069767441860463,
+      "grad_norm": 1.7043132781982422,
+      "learning_rate": 4.709302325581396e-05,
+      "loss": 0.4186,
+      "step": 82
+    },
+    {
+      "epoch": 1.9302325581395348,
+      "grad_norm": 1.0048333406448364,
+      "learning_rate": 4.7674418604651164e-05,
+      "loss": 0.5134,
+      "step": 83
+    },
+    {
+      "epoch": 1.9534883720930232,
+      "grad_norm": 1.5068817138671875,
+      "learning_rate": 4.8255813953488375e-05,
+      "loss": 0.4723,
+      "step": 84
+    },
+    {
+      "epoch": 1.9767441860465116,
+      "grad_norm": 1.1574243307113647,
+      "learning_rate": 4.883720930232558e-05,
+      "loss": 0.4904,
+      "step": 85
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 1.2925702333450317,
+      "learning_rate": 4.941860465116279e-05,
+      "loss": 0.5023,
+      "step": 86
+    },
+    {
+      "epoch": 2.0232558139534884,
+      "grad_norm": 1.679551124572754,
+      "learning_rate": 5e-05,
+      "loss": 0.4305,
+      "step": 87
+    },
+    {
+      "epoch": 2.046511627906977,
+      "grad_norm": 1.1237919330596924,
+      "learning_rate": 5.0581395348837214e-05,
+      "loss": 0.475,
+      "step": 88
+    },
+    {
+      "epoch": 2.0697674418604652,
+      "grad_norm": 1.6137198209762573,
+      "learning_rate": 5.1162790697674425e-05,
+      "loss": 0.4115,
+      "step": 89
+    },
+    {
+      "epoch": 2.0930232558139537,
+      "grad_norm": 1.814559817314148,
+      "learning_rate": 5.1744186046511636e-05,
+      "loss": 0.4894,
+      "step": 90
+    },
+    {
+      "epoch": 2.116279069767442,
+      "grad_norm": 2.1933960914611816,
+      "learning_rate": 5.232558139534884e-05,
+      "loss": 0.3742,
+      "step": 91
+    },
+    {
+      "epoch": 2.13953488372093,
+      "grad_norm": 1.3443443775177002,
+      "learning_rate": 5.290697674418605e-05,
+      "loss": 0.396,
+      "step": 92
+    },
+    {
+      "epoch": 2.1627906976744184,
+      "grad_norm": 1.5833871364593506,
+      "learning_rate": 5.348837209302326e-05,
+      "loss": 0.3518,
+      "step": 93
+    },
+    {
+      "epoch": 2.186046511627907,
+      "grad_norm": 1.5842630863189697,
+      "learning_rate": 5.406976744186046e-05,
+      "loss": 0.4174,
+      "step": 94
+    },
+    {
+      "epoch": 2.2093023255813953,
+      "grad_norm": 1.2995549440383911,
+      "learning_rate": 5.465116279069767e-05,
+      "loss": 0.3873,
+      "step": 95
+    },
+    {
+      "epoch": 2.2325581395348837,
+      "grad_norm": 2.144974946975708,
+      "learning_rate": 5.5232558139534884e-05,
+      "loss": 0.4805,
+      "step": 96
+    },
+    {
+      "epoch": 2.255813953488372,
+      "grad_norm": 1.4312434196472168,
+      "learning_rate": 5.5813953488372095e-05,
+      "loss": 0.4575,
+      "step": 97
+    },
+    {
+      "epoch": 2.2790697674418605,
+      "grad_norm": 2.25107741355896,
+      "learning_rate": 5.6395348837209306e-05,
+      "loss": 0.4139,
+      "step": 98
+    },
+    {
+      "epoch": 2.302325581395349,
+      "grad_norm": 1.6673591136932373,
+      "learning_rate": 5.697674418604652e-05,
+      "loss": 0.4151,
+      "step": 99
+    },
+    {
+      "epoch": 2.3255813953488373,
+      "grad_norm": 1.3179768323898315,
+      "learning_rate": 5.755813953488373e-05,
+      "loss": 0.2997,
+      "step": 100
+    },
+    {
+      "epoch": 2.3255813953488373,
+      "eval_accuracy": 0.839344262295082,
+      "eval_f1": 0.839344262295082,
+      "eval_loss": 0.3802424669265747,
+      "eval_precision": 0.8408718876540019,
+      "eval_recall": 0.8408718876540019,
+      "eval_runtime": 0.0744,
+      "eval_samples_per_second": 4101.555,
+      "eval_steps_per_second": 67.239,
+      "step": 100
+    },
+    {
+      "epoch": 2.3488372093023258,
+      "grad_norm": 2.245741844177246,
+      "learning_rate": 5.8139534883720933e-05,
+      "loss": 0.3442,
+      "step": 101
+    },
+    {
+      "epoch": 2.3720930232558137,
+      "grad_norm": 2.9141011238098145,
+      "learning_rate": 5.8720930232558145e-05,
+      "loss": 0.3489,
+      "step": 102
+    },
+    {
+      "epoch": 2.395348837209302,
+      "grad_norm": 1.829503059387207,
+      "learning_rate": 5.9302325581395356e-05,
+      "loss": 0.3544,
+      "step": 103
+    },
+    {
+      "epoch": 2.4186046511627906,
+      "grad_norm": 2.233344793319702,
+      "learning_rate": 5.9883720930232554e-05,
+      "loss": 0.3008,
+      "step": 104
+    },
+    {
+      "epoch": 2.441860465116279,
+      "grad_norm": 2.4595718383789062,
+      "learning_rate": 6.0465116279069765e-05,
+      "loss": 0.3853,
+      "step": 105
+    },
+    {
+      "epoch": 2.4651162790697674,
+      "grad_norm": 1.539181113243103,
+      "learning_rate": 6.104651162790698e-05,
+      "loss": 0.3165,
+      "step": 106
+    },
+    {
+      "epoch": 2.488372093023256,
+      "grad_norm": 1.7545955181121826,
+      "learning_rate": 6.162790697674418e-05,
+      "loss": 0.3907,
+      "step": 107
+    },
+    {
+      "epoch": 2.511627906976744,
+      "grad_norm": 3.285968065261841,
+      "learning_rate": 6.22093023255814e-05,
+      "loss": 0.4645,
+      "step": 108
+    },
+    {
+      "epoch": 2.5348837209302326,
+      "grad_norm": 4.625302314758301,
+      "learning_rate": 6.27906976744186e-05,
+      "loss": 0.2975,
+      "step": 109
+    },
+    {
+      "epoch": 2.558139534883721,
+      "grad_norm": 1.9037487506866455,
+      "learning_rate": 6.337209302325582e-05,
+      "loss": 0.3629,
+      "step": 110
+    },
+    {
+      "epoch": 2.5813953488372094,
+      "grad_norm": 2.768249988555908,
+      "learning_rate": 6.395348837209303e-05,
+      "loss": 0.385,
+      "step": 111
+    },
+    {
+      "epoch": 2.604651162790698,
+      "grad_norm": 1.8411540985107422,
+      "learning_rate": 6.453488372093024e-05,
+      "loss": 0.2562,
+      "step": 112
+    },
+    {
+      "epoch": 2.6279069767441863,
+      "grad_norm": 2.3537325859069824,
+      "learning_rate": 6.511627906976745e-05,
+      "loss": 0.2486,
+      "step": 113
+    },
+    {
+      "epoch": 2.6511627906976747,
+      "grad_norm": 4.455109596252441,
+      "learning_rate": 6.569767441860465e-05,
+      "loss": 0.2424,
+      "step": 114
+    },
+    {
+      "epoch": 2.6744186046511627,
+      "grad_norm": 3.032252550125122,
+      "learning_rate": 6.627906976744186e-05,
+      "loss": 0.3554,
+      "step": 115
+    },
+    {
+      "epoch": 2.697674418604651,
+      "grad_norm": 2.142354965209961,
+      "learning_rate": 6.686046511627908e-05,
+      "loss": 0.3701,
+      "step": 116
+    },
+    {
+      "epoch": 2.7209302325581395,
+      "grad_norm": 3.0141141414642334,
+      "learning_rate": 6.744186046511628e-05,
+      "loss": 0.4602,
+      "step": 117
+    },
+    {
+      "epoch": 2.744186046511628,
+      "grad_norm": 2.129042625427246,
+      "learning_rate": 6.802325581395348e-05,
+      "loss": 0.4295,
+      "step": 118
+    },
+    {
+      "epoch": 2.7674418604651163,
+      "grad_norm": 4.583343029022217,
+      "learning_rate": 6.86046511627907e-05,
+      "loss": 0.3599,
+      "step": 119
+    },
+    {
+      "epoch": 2.7906976744186047,
+      "grad_norm": 2.799339532852173,
+      "learning_rate": 6.918604651162791e-05,
+      "loss": 0.4383,
+      "step": 120
+    },
+    {
+      "epoch": 2.7906976744186047,
+      "eval_accuracy": 0.8426229508196721,
+      "eval_f1": 0.8426077234024254,
+      "eval_loss": 0.30933016538619995,
+      "eval_precision": 0.8432928927895371,
+      "eval_recall": 0.8437365382958559,
+      "eval_runtime": 0.0741,
+      "eval_samples_per_second": 4116.681,
+      "eval_steps_per_second": 67.487,
+      "step": 120
+    },
+    {
+      "epoch": 2.813953488372093,
+      "grad_norm": 3.329214096069336,
+      "learning_rate": 6.976744186046513e-05,
+      "loss": 0.2955,
+      "step": 121
+    },
+    {
+      "epoch": 2.8372093023255816,
+      "grad_norm": 1.7641863822937012,
+      "learning_rate": 7.034883720930233e-05,
+      "loss": 0.372,
+      "step": 122
+    },
+    {
+      "epoch": 2.8604651162790695,
+      "grad_norm": 2.8948066234588623,
+      "learning_rate": 7.093023255813955e-05,
+      "loss": 0.3738,
+      "step": 123
+    },
+    {
+      "epoch": 2.883720930232558,
+      "grad_norm": 3.2698655128479004,
+      "learning_rate": 7.151162790697675e-05,
+      "loss": 0.4183,
+      "step": 124
+    },
+    {
+      "epoch": 2.9069767441860463,
+      "grad_norm": 2.3984827995300293,
+      "learning_rate": 7.209302325581396e-05,
+      "loss": 0.3729,
+      "step": 125
+    },
+    {
+      "epoch": 2.9302325581395348,
+      "grad_norm": 2.4366798400878906,
+      "learning_rate": 7.267441860465116e-05,
+      "loss": 0.291,
+      "step": 126
+    },
+    {
+      "epoch": 2.953488372093023,
+      "grad_norm": 1.6613179445266724,
+      "learning_rate": 7.325581395348837e-05,
+      "loss": 0.2673,
+      "step": 127
+    },
+    {
+      "epoch": 2.9767441860465116,
+      "grad_norm": 3.9009366035461426,
+      "learning_rate": 7.383720930232558e-05,
+      "loss": 0.3912,
+      "step": 128
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 6.0683417320251465,
+      "learning_rate": 7.441860465116279e-05,
+      "loss": 0.3879,
+      "step": 129
+    },
+    {
+      "epoch": 3.0232558139534884,
+      "grad_norm": 2.161290407180786,
+      "learning_rate": 7.500000000000001e-05,
+      "loss": 0.2878,
+      "step": 130
+    },
+    {
+      "epoch": 3.046511627906977,
+      "grad_norm": 5.449894905090332,
+      "learning_rate": 7.558139534883721e-05,
+      "loss": 0.3326,
+      "step": 131
+    },
+    {
+      "epoch": 3.0697674418604652,
+      "grad_norm": 2.6764752864837646,
+      "learning_rate": 7.616279069767443e-05,
+      "loss": 0.2813,
+      "step": 132
+    },
+    {
+      "epoch": 3.0930232558139537,
+      "grad_norm": 1.1198005676269531,
+      "learning_rate": 7.674418604651163e-05,
+      "loss": 0.2651,
+      "step": 133
+    },
+    {
+      "epoch": 3.116279069767442,
+      "grad_norm": 1.7498178482055664,
+      "learning_rate": 7.732558139534884e-05,
+      "loss": 0.2171,
+      "step": 134
+    },
+    {
+      "epoch": 3.13953488372093,
+      "grad_norm": 4.505956649780273,
+      "learning_rate": 7.790697674418606e-05,
+      "loss": 0.2529,
+      "step": 135
+    },
+    {
+      "epoch": 3.1627906976744184,
+      "grad_norm": 2.509193181991577,
+      "learning_rate": 7.848837209302326e-05,
+      "loss": 0.2102,
+      "step": 136
+    },
+    {
+      "epoch": 3.186046511627907,
+      "grad_norm": 1.7064465284347534,
+      "learning_rate": 7.906976744186047e-05,
+      "loss": 0.2097,
+      "step": 137
+    },
+    {
+      "epoch": 3.2093023255813953,
+      "grad_norm": 1.7659178972244263,
+      "learning_rate": 7.965116279069767e-05,
+      "loss": 0.1806,
+      "step": 138
+    },
+    {
+      "epoch": 3.2325581395348837,
+      "grad_norm": 3.508608818054199,
+      "learning_rate": 8.023255813953489e-05,
+      "loss": 0.1419,
+      "step": 139
+    },
+    {
+      "epoch": 3.255813953488372,
+      "grad_norm": 2.2445175647735596,
+      "learning_rate": 8.081395348837209e-05,
+      "loss": 0.4201,
+      "step": 140
+    },
+    {
+      "epoch": 3.255813953488372,
+      "eval_accuracy": 0.8622950819672132,
+      "eval_f1": 0.8622225090345896,
+      "eval_loss": 0.31382298469543457,
+      "eval_precision": 0.8622291021671826,
+      "eval_recall": 0.8628844662703541,
+      "eval_runtime": 0.0729,
+      "eval_samples_per_second": 4184.713,
+      "eval_steps_per_second": 68.602,
+      "step": 140
+    },
+    {
+      "epoch": 3.2790697674418605,
+      "grad_norm": 2.019263744354248,
+      "learning_rate": 8.139534883720931e-05,
+      "loss": 0.2023,
+      "step": 141
+    },
+    {
+      "epoch": 3.302325581395349,
+      "grad_norm": 1.6362420320510864,
+      "learning_rate": 8.197674418604652e-05,
+      "loss": 0.1729,
+      "step": 142
+    },
+    {
+      "epoch": 3.3255813953488373,
+      "grad_norm": 2.1134684085845947,
+      "learning_rate": 8.255813953488373e-05,
+      "loss": 0.2688,
+      "step": 143
+    },
+    {
+      "epoch": 3.3488372093023258,
+      "grad_norm": 3.522068738937378,
+      "learning_rate": 8.313953488372094e-05,
+      "loss": 0.199,
+      "step": 144
+    },
+    {
+      "epoch": 3.3720930232558137,
+      "grad_norm": 4.994085788726807,
+      "learning_rate": 8.372093023255814e-05,
+      "loss": 0.2641,
+      "step": 145
+    },
+    {
+      "epoch": 3.395348837209302,
+      "grad_norm": 2.1998469829559326,
+      "learning_rate": 8.430232558139536e-05,
+      "loss": 0.2183,
+      "step": 146
+    },
+    {
+      "epoch": 3.4186046511627906,
+      "grad_norm": 1.9816820621490479,
+      "learning_rate": 8.488372093023255e-05,
+      "loss": 0.2952,
+      "step": 147
+    },
+    {
+      "epoch": 3.441860465116279,
+      "grad_norm": 2.8008265495300293,
+      "learning_rate": 8.546511627906977e-05,
+      "loss": 0.2562,
+      "step": 148
+    },
+    {
+      "epoch": 3.4651162790697674,
+      "grad_norm": 2.082418918609619,
+      "learning_rate": 8.604651162790697e-05,
+      "loss": 0.2626,
+      "step": 149
+    },
+    {
+      "epoch": 3.488372093023256,
+      "grad_norm": 3.3041298389434814,
+      "learning_rate": 8.662790697674419e-05,
+      "loss": 0.2373,
+      "step": 150
+    },
+    {
+      "epoch": 3.511627906976744,
+      "grad_norm": 2.62347674369812,
+      "learning_rate": 8.72093023255814e-05,
+      "loss": 0.2749,
+      "step": 151
+    },
+    {
+      "epoch": 3.5348837209302326,
+      "grad_norm": 3.0671257972717285,
+      "learning_rate": 8.779069767441861e-05,
+      "loss": 0.3538,
+      "step": 152
+    },
+    {
+      "epoch": 3.558139534883721,
+      "grad_norm": 2.737697124481201,
+      "learning_rate": 8.837209302325582e-05,
+      "loss": 0.3537,
+      "step": 153
+    },
+    {
+      "epoch": 3.5813953488372094,
+      "grad_norm": 6.468879699707031,
+      "learning_rate": 8.895348837209302e-05,
+      "loss": 0.5041,
+      "step": 154
+    },
+    {
+      "epoch": 3.604651162790698,
+      "grad_norm": 3.8154025077819824,
+      "learning_rate": 8.953488372093024e-05,
+      "loss": 0.3206,
+      "step": 155
+    },
+    {
+      "epoch": 3.6279069767441863,
+      "grad_norm": 1.7028948068618774,
+      "learning_rate": 9.011627906976745e-05,
+      "loss": 0.3093,
+      "step": 156
+    },
+    {
+      "epoch": 3.6511627906976747,
+      "grad_norm": 2.867501735687256,
+      "learning_rate": 9.069767441860465e-05,
+      "loss": 0.2656,
+      "step": 157
+    },
+    {
+      "epoch": 3.6744186046511627,
+      "grad_norm": 4.052389621734619,
+      "learning_rate": 9.127906976744186e-05,
+      "loss": 0.4386,
+      "step": 158
+    },
+    {
+      "epoch": 3.697674418604651,
+      "grad_norm": 1.909553050994873,
+      "learning_rate": 9.186046511627907e-05,
+      "loss": 0.2907,
+      "step": 159
+    },
+    {
+      "epoch": 3.7209302325581395,
+      "grad_norm": 1.4470328092575073,
+      "learning_rate": 9.244186046511628e-05,
+      "loss": 0.1697,
+      "step": 160
+    },
+    {
+      "epoch": 3.7209302325581395,
+      "eval_accuracy": 0.8721311475409836,
+      "eval_f1": 0.8721256490469893,
+      "eval_loss": 0.26616957783699036,
+      "eval_precision": 0.8730848683077983,
+      "eval_recall": 0.87343844231929,
+      "eval_runtime": 0.0715,
+      "eval_samples_per_second": 4265.076,
+      "eval_steps_per_second": 69.919,
+      "step": 160
+    },
+    {
+      "epoch": 3.744186046511628,
+      "grad_norm": 1.7241488695144653,
+      "learning_rate": 9.30232558139535e-05,
+      "loss": 0.248,
+      "step": 161
+    },
+    {
+      "epoch": 3.7674418604651163,
+      "grad_norm": 1.370429277420044,
+      "learning_rate": 9.36046511627907e-05,
+      "loss": 0.26,
+      "step": 162
+    },
+    {
+      "epoch": 3.7906976744186047,
+      "grad_norm": 2.8360958099365234,
+      "learning_rate": 9.418604651162792e-05,
+      "loss": 0.3161,
+      "step": 163
+    },
+    {
+      "epoch": 3.813953488372093,
+      "grad_norm": 2.840177297592163,
+      "learning_rate": 9.476744186046512e-05,
+      "loss": 0.3643,
+      "step": 164
+    },
+    {
+      "epoch": 3.8372093023255816,
+      "grad_norm": 1.9618759155273438,
+      "learning_rate": 9.534883720930233e-05,
+      "loss": 0.3486,
+      "step": 165
+    },
+    {
+      "epoch": 3.8604651162790695,
+      "grad_norm": 3.201247215270996,
+      "learning_rate": 9.593023255813955e-05,
+      "loss": 0.293,
+      "step": 166
+    },
+    {
+      "epoch": 3.883720930232558,
+      "grad_norm": 3.3960511684417725,
+      "learning_rate": 9.651162790697675e-05,
+      "loss": 0.2797,
+      "step": 167
+    },
+    {
+      "epoch": 3.9069767441860463,
+      "grad_norm": 2.5416486263275146,
+      "learning_rate": 9.709302325581396e-05,
+      "loss": 0.2781,
+      "step": 168
+    },
+    {
+      "epoch": 3.9302325581395348,
+      "grad_norm": 4.113210678100586,
+      "learning_rate": 9.767441860465116e-05,
+      "loss": 0.2734,
+      "step": 169
+    },
+    {
+      "epoch": 3.953488372093023,
+      "grad_norm": 3.773771286010742,
+      "learning_rate": 9.825581395348838e-05,
+      "loss": 0.2775,
+      "step": 170
+    },
+    {
+      "epoch": 3.9767441860465116,
+      "grad_norm": 1.3982651233673096,
+      "learning_rate": 9.883720930232558e-05,
+      "loss": 0.1802,
+      "step": 171
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 1.2582266330718994,
+      "learning_rate": 9.94186046511628e-05,
+      "loss": 0.1667,
+      "step": 172
+    },
+    {
+      "epoch": 4.023255813953488,
+      "grad_norm": 6.577229976654053,
+      "learning_rate": 0.0001,
+      "loss": 0.4147,
+      "step": 173
+    },
+    {
+      "epoch": 4.046511627906977,
+      "grad_norm": 4.923567295074463,
+      "learning_rate": 9.98546511627907e-05,
+      "loss": 0.1983,
+      "step": 174
+    },
+    {
+      "epoch": 4.069767441860465,
+      "grad_norm": 1.9299342632293701,
+      "learning_rate": 9.97093023255814e-05,
+      "loss": 0.2169,
+      "step": 175
+    },
+    {
+      "epoch": 4.093023255813954,
+      "grad_norm": 3.2299985885620117,
+      "learning_rate": 9.95639534883721e-05,
+      "loss": 0.1698,
+      "step": 176
+    },
+    {
+      "epoch": 4.116279069767442,
+      "grad_norm": 2.034163236618042,
+      "learning_rate": 9.94186046511628e-05,
+      "loss": 0.191,
+      "step": 177
+    },
+    {
+      "epoch": 4.1395348837209305,
+      "grad_norm": 4.542312145233154,
+      "learning_rate": 9.927325581395349e-05,
+      "loss": 0.2465,
+      "step": 178
+    },
+    {
+      "epoch": 4.162790697674419,
+      "grad_norm": 1.8868296146392822,
+      "learning_rate": 9.912790697674418e-05,
+      "loss": 0.2437,
+      "step": 179
+    },
+    {
+      "epoch": 4.186046511627907,
+      "grad_norm": 4.527658939361572,
+      "learning_rate": 9.898255813953488e-05,
+      "loss": 0.2774,
+      "step": 180
+    },
+    {
+      "epoch": 4.186046511627907,
+      "eval_accuracy": 0.8852459016393442,
+      "eval_f1": 0.8839686525146466,
+      "eval_loss": 0.3168491721153259,
+      "eval_precision": 0.8926612403786605,
+      "eval_recall": 0.8823770138709399,
+      "eval_runtime": 0.0728,
+      "eval_samples_per_second": 4190.498,
+      "eval_steps_per_second": 68.697,
+      "step": 180
+    },
+    {
+      "epoch": 4.209302325581396,
+      "grad_norm": 6.06718635559082,
+      "learning_rate": 9.883720930232558e-05,
+      "loss": 0.267,
+      "step": 181
+    },
+    {
+      "epoch": 4.232558139534884,
+      "grad_norm": 8.161118507385254,
+      "learning_rate": 9.869186046511628e-05,
+      "loss": 0.459,
+      "step": 182
+    },
+    {
+      "epoch": 4.2558139534883725,
+      "grad_norm": 3.9242475032806396,
+      "learning_rate": 9.854651162790698e-05,
+      "loss": 0.2698,
+      "step": 183
+    },
+    {
+      "epoch": 4.27906976744186,
+      "grad_norm": 6.657914161682129,
+      "learning_rate": 9.840116279069768e-05,
+      "loss": 0.2114,
+      "step": 184
+    },
+    {
+      "epoch": 4.3023255813953485,
+      "grad_norm": 2.23410964012146,
+      "learning_rate": 9.825581395348838e-05,
+      "loss": 0.1337,
+      "step": 185
+    },
+    {
+      "epoch": 4.325581395348837,
+      "grad_norm": 1.6781232357025146,
+      "learning_rate": 9.811046511627908e-05,
+      "loss": 0.1269,
+      "step": 186
+    },
+    {
+      "epoch": 4.348837209302325,
+      "grad_norm": 1.3690869808197021,
+      "learning_rate": 9.796511627906976e-05,
+      "loss": 0.1783,
+      "step": 187
+    },
+    {
+      "epoch": 4.372093023255814,
+      "grad_norm": 2.0360560417175293,
+      "learning_rate": 9.781976744186046e-05,
+      "loss": 0.281,
+      "step": 188
+    },
+    {
+      "epoch": 4.395348837209302,
+      "grad_norm": 3.783336877822876,
+      "learning_rate": 9.767441860465116e-05,
+      "loss": 0.3072,
+      "step": 189
+    },
+    {
+      "epoch": 4.4186046511627906,
+      "grad_norm": 3.0387606620788574,
+      "learning_rate": 9.752906976744186e-05,
+      "loss": 0.2682,
+      "step": 190
+    },
+    {
+      "epoch": 4.441860465116279,
+      "grad_norm": 1.6739147901535034,
+      "learning_rate": 9.738372093023256e-05,
+      "loss": 0.1597,
+      "step": 191
+    },
+    {
+      "epoch": 4.465116279069767,
+      "grad_norm": 4.560539245605469,
+      "learning_rate": 9.723837209302326e-05,
+      "loss": 0.2328,
+      "step": 192
+    },
+    {
+      "epoch": 4.488372093023256,
+      "grad_norm": 2.564955711364746,
+      "learning_rate": 9.709302325581396e-05,
+      "loss": 0.2731,
+      "step": 193
+    },
+    {
+      "epoch": 4.511627906976744,
+      "grad_norm": 6.892247676849365,
+      "learning_rate": 9.694767441860465e-05,
+      "loss": 0.2358,
+      "step": 194
+    },
+    {
+      "epoch": 4.534883720930233,
+      "grad_norm": 2.4123950004577637,
+      "learning_rate": 9.680232558139535e-05,
+      "loss": 0.2267,
+      "step": 195
+    },
+    {
+      "epoch": 4.558139534883721,
+      "grad_norm": 2.6707470417022705,
+      "learning_rate": 9.665697674418605e-05,
+      "loss": 0.1415,
+      "step": 196
+    },
+    {
+      "epoch": 4.5813953488372094,
+      "grad_norm": 1.8982809782028198,
+      "learning_rate": 9.651162790697675e-05,
+      "loss": 0.2119,
+      "step": 197
+    },
+    {
+      "epoch": 4.604651162790698,
+      "grad_norm": 3.8330225944519043,
+      "learning_rate": 9.636627906976745e-05,
+      "loss": 0.152,
+      "step": 198
+    },
+    {
+      "epoch": 4.627906976744186,
+      "grad_norm": 2.504080057144165,
+      "learning_rate": 9.622093023255815e-05,
+      "loss": 0.1623,
+      "step": 199
+    },
+    {
+      "epoch": 4.651162790697675,
+      "grad_norm": 1.6243199110031128,
+      "learning_rate": 9.607558139534885e-05,
+      "loss": 0.1537,
+      "step": 200
+    },
+    {
+      "epoch": 4.651162790697675,
+      "eval_accuracy": 0.8950819672131147,
+      "eval_f1": 0.8942532720811303,
+      "eval_loss": 0.2364664524793625,
+      "eval_precision": 0.8994351024697845,
+      "eval_recall": 0.892930989919876,
+      "eval_runtime": 0.0733,
+      "eval_samples_per_second": 4158.906,
+      "eval_steps_per_second": 68.179,
+      "step": 200
+    },
+    {
+      "epoch": 4.674418604651163,
+      "grad_norm": 3.362497091293335,
+      "learning_rate": 9.593023255813955e-05,
+      "loss": 0.2097,
+      "step": 201
+    },
+    {
+      "epoch": 4.6976744186046515,
+      "grad_norm": 1.36155366897583,
+      "learning_rate": 9.578488372093024e-05,
+      "loss": 0.134,
+      "step": 202
+    },
+    {
+      "epoch": 4.720930232558139,
+      "grad_norm": 2.335468053817749,
+      "learning_rate": 9.563953488372094e-05,
+      "loss": 0.2128,
+      "step": 203
+    },
+    {
+      "epoch": 4.7441860465116275,
+      "grad_norm": 2.431711196899414,
+      "learning_rate": 9.549418604651163e-05,
+      "loss": 0.2257,
+      "step": 204
+    },
+    {
+      "epoch": 4.767441860465116,
+      "grad_norm": 2.425615072250366,
+      "learning_rate": 9.534883720930233e-05,
+      "loss": 0.1691,
+      "step": 205
+    },
+    {
+      "epoch": 4.790697674418604,
+      "grad_norm": 3.3857626914978027,
+      "learning_rate": 9.520348837209303e-05,
+      "loss": 0.2282,
+      "step": 206
+    },
+    {
+      "epoch": 4.813953488372093,
+      "grad_norm": 2.252249240875244,
+      "learning_rate": 9.505813953488373e-05,
+      "loss": 0.2333,
+      "step": 207
+    },
+    {
+      "epoch": 4.837209302325581,
+      "grad_norm": 4.851521015167236,
+      "learning_rate": 9.491279069767442e-05,
+      "loss": 0.2621,
+      "step": 208
+    },
+    {
+      "epoch": 4.8604651162790695,
+      "grad_norm": 1.249811053276062,
+      "learning_rate": 9.476744186046512e-05,
+      "loss": 0.075,
+      "step": 209
+    },
+    {
+      "epoch": 4.883720930232558,
+      "grad_norm": 1.6059314012527466,
+      "learning_rate": 9.462209302325582e-05,
+      "loss": 0.1878,
+      "step": 210
+    },
+    {
+      "epoch": 4.906976744186046,
+      "grad_norm": 7.576606750488281,
+      "learning_rate": 9.447674418604652e-05,
+      "loss": 0.2258,
+      "step": 211
+    },
+    {
+      "epoch": 4.930232558139535,
+      "grad_norm": 2.4628143310546875,
+      "learning_rate": 9.433139534883722e-05,
+      "loss": 0.1463,
+      "step": 212
+    },
+    {
+      "epoch": 4.953488372093023,
+      "grad_norm": 4.781926155090332,
+      "learning_rate": 9.418604651162792e-05,
+      "loss": 0.2608,
+      "step": 213
+    },
+    {
+      "epoch": 4.976744186046512,
+      "grad_norm": 3.2314810752868652,
+      "learning_rate": 9.40406976744186e-05,
+      "loss": 0.2344,
+      "step": 214
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 3.6826560497283936,
+      "learning_rate": 9.38953488372093e-05,
+      "loss": 0.17,
+      "step": 215
+    },
+    {
+      "epoch": 5.023255813953488,
+      "grad_norm": 1.804560661315918,
+      "learning_rate": 9.375e-05,
+      "loss": 0.1241,
+      "step": 216
+    },
+    {
+      "epoch": 5.046511627906977,
+      "grad_norm": 3.6199300289154053,
+      "learning_rate": 9.36046511627907e-05,
+      "loss": 0.1335,
+      "step": 217
+    },
+    {
+      "epoch": 5.069767441860465,
+      "grad_norm": 2.37376070022583,
+      "learning_rate": 9.34593023255814e-05,
+      "loss": 0.1883,
+      "step": 218
+    },
+    {
+      "epoch": 5.093023255813954,
+      "grad_norm": 4.643337726593018,
+      "learning_rate": 9.33139534883721e-05,
+      "loss": 0.187,
+      "step": 219
+    },
+    {
+      "epoch": 5.116279069767442,
+      "grad_norm": 6.206487655639648,
+      "learning_rate": 9.31686046511628e-05,
+      "loss": 0.166,
+      "step": 220
+    },
+    {
+      "epoch": 5.116279069767442,
+      "eval_accuracy": 0.8819672131147541,
+      "eval_f1": 0.8815073815073815,
+      "eval_loss": 0.27603965997695923,
+      "eval_precision": 0.8827922077922078,
+      "eval_recall": 0.8809123804600671,
+      "eval_runtime": 0.0725,
+      "eval_samples_per_second": 4205.887,
+      "eval_steps_per_second": 68.949,
+      "step": 220
+    },
+    {
+      "epoch": 5.1395348837209305,
+      "grad_norm": 2.4301323890686035,
+      "learning_rate": 9.30232558139535e-05,
+      "loss": 0.0946,
+      "step": 221
+    },
+    {
+      "epoch": 5.162790697674419,
+      "grad_norm": 4.870151519775391,
+      "learning_rate": 9.28779069767442e-05,
+      "loss": 0.2674,
+      "step": 222
+    },
+    {
+      "epoch": 5.186046511627907,
+      "grad_norm": 5.952517032623291,
+      "learning_rate": 9.273255813953488e-05,
+      "loss": 0.2605,
+      "step": 223
+    },
+    {
+      "epoch": 5.209302325581396,
+      "grad_norm": 1.8462685346603394,
+      "learning_rate": 9.258720930232558e-05,
+      "loss": 0.1902,
+      "step": 224
+    },
+    {
+      "epoch": 5.232558139534884,
+      "grad_norm": 2.608715295791626,
+      "learning_rate": 9.244186046511628e-05,
+      "loss": 0.1392,
+      "step": 225
+    },
+    {
+      "epoch": 5.2558139534883725,
+      "grad_norm": 4.296781539916992,
+      "learning_rate": 9.229651162790698e-05,
+      "loss": 0.2599,
+      "step": 226
+    },
+    {
+      "epoch": 5.27906976744186,
+      "grad_norm": 6.4339165687561035,
+      "learning_rate": 9.215116279069768e-05,
+      "loss": 0.162,
+      "step": 227
+    },
+    {
+      "epoch": 5.3023255813953485,
+      "grad_norm": 4.432989597320557,
+      "learning_rate": 9.200581395348837e-05,
+      "loss": 0.1832,
+      "step": 228
+    },
+    {
+      "epoch": 5.325581395348837,
+      "grad_norm": 1.9655786752700806,
+      "learning_rate": 9.186046511627907e-05,
+      "loss": 0.1192,
+      "step": 229
+    },
+    {
+      "epoch": 5.348837209302325,
+      "grad_norm": 4.3318891525268555,
+      "learning_rate": 9.171511627906977e-05,
+      "loss": 0.244,
+      "step": 230
+    },
+    {
+      "epoch": 5.372093023255814,
+      "grad_norm": 2.993924856185913,
+      "learning_rate": 9.156976744186047e-05,
+      "loss": 0.1828,
+      "step": 231
+    },
+    {
+      "epoch": 5.395348837209302,
+      "grad_norm": 2.020063638687134,
+      "learning_rate": 9.142441860465116e-05,
+      "loss": 0.1642,
+      "step": 232
+    },
+    {
+      "epoch": 5.4186046511627906,
+      "grad_norm": 2.384181499481201,
+      "learning_rate": 9.127906976744186e-05,
+      "loss": 0.1752,
+      "step": 233
+    },
+    {
+      "epoch": 5.441860465116279,
+      "grad_norm": 1.9194653034210205,
+      "learning_rate": 9.113372093023255e-05,
+      "loss": 0.0996,
+      "step": 234
+    },
+    {
+      "epoch": 5.465116279069767,
+      "grad_norm": 3.680722713470459,
+      "learning_rate": 9.098837209302325e-05,
+      "loss": 0.1925,
+      "step": 235
+    },
+    {
+      "epoch": 5.488372093023256,
+      "grad_norm": 3.332195520401001,
+      "learning_rate": 9.084302325581395e-05,
+      "loss": 0.13,
+      "step": 236
+    },
+    {
+      "epoch": 5.511627906976744,
+      "grad_norm": 2.6394619941711426,
+      "learning_rate": 9.069767441860465e-05,
+      "loss": 0.193,
+      "step": 237
+    },
+    {
+      "epoch": 5.534883720930233,
+      "grad_norm": 4.5959038734436035,
+      "learning_rate": 9.055232558139536e-05,
+      "loss": 0.2303,
+      "step": 238
+    },
+    {
+      "epoch": 5.558139534883721,
+      "grad_norm": 6.544390678405762,
+      "learning_rate": 9.040697674418606e-05,
+      "loss": 0.2432,
+      "step": 239
+    },
+    {
+      "epoch": 5.5813953488372094,
+      "grad_norm": 1.6620129346847534,
+      "learning_rate": 9.026162790697675e-05,
+      "loss": 0.1378,
+      "step": 240
+    },
+    {
+      "epoch": 5.5813953488372094,
+      "eval_accuracy": 0.8819672131147541,
+      "eval_f1": 0.8816810344827587,
+      "eval_loss": 0.21341156959533691,
+      "eval_precision": 0.8819660110420979,
+      "eval_recall": 0.8814723873524597,
+      "eval_runtime": 0.091,
+      "eval_samples_per_second": 3353.025,
+      "eval_steps_per_second": 54.968,
+      "step": 240
+    },
+    {
+      "epoch": 5.604651162790698,
+      "grad_norm": 1.251617431640625,
+      "learning_rate": 9.011627906976745e-05,
+      "loss": 0.1108,
+      "step": 241
+    },
+    {
+      "epoch": 5.627906976744186,
+      "grad_norm": 2.331637144088745,
+      "learning_rate": 8.997093023255815e-05,
+      "loss": 0.1591,
+      "step": 242
+    },
+    {
+      "epoch": 5.651162790697675,
+      "grad_norm": 2.126330852508545,
+      "learning_rate": 8.982558139534884e-05,
+      "loss": 0.1115,
+      "step": 243
+    },
+    {
+      "epoch": 5.674418604651163,
+      "grad_norm": 2.0312886238098145,
+      "learning_rate": 8.968023255813954e-05,
+      "loss": 0.1009,
+      "step": 244
+    },
+    {
+      "epoch": 5.6976744186046515,
+      "grad_norm": 3.0067551136016846,
+      "learning_rate": 8.953488372093024e-05,
+      "loss": 0.1851,
+      "step": 245
+    },
+    {
+      "epoch": 5.720930232558139,
+      "grad_norm": 3.1301369667053223,
+      "learning_rate": 8.938953488372094e-05,
+      "loss": 0.107,
+      "step": 246
+    },
+    {
+      "epoch": 5.7441860465116275,
+      "grad_norm": 6.043979167938232,
+      "learning_rate": 8.924418604651164e-05,
+      "loss": 0.273,
+      "step": 247
+    },
+    {
+      "epoch": 5.767441860465116,
+      "grad_norm": 3.7104833126068115,
+      "learning_rate": 8.909883720930234e-05,
+      "loss": 0.1299,
+      "step": 248
+    },
+    {
+      "epoch": 5.790697674418604,
+      "grad_norm": 2.3856964111328125,
+      "learning_rate": 8.895348837209302e-05,
+      "loss": 0.1252,
+      "step": 249
+    },
+    {
+      "epoch": 5.813953488372093,
+      "grad_norm": 5.702451705932617,
+      "learning_rate": 8.880813953488372e-05,
+      "loss": 0.1939,
+      "step": 250
+    },
+    {
+      "epoch": 5.837209302325581,
+      "grad_norm": 3.5763566493988037,
+      "learning_rate": 8.866279069767442e-05,
+      "loss": 0.2426,
+      "step": 251
+    },
+    {
+      "epoch": 5.8604651162790695,
+      "grad_norm": 4.04267692565918,
+      "learning_rate": 8.851744186046512e-05,
+      "loss": 0.1966,
+      "step": 252
+    },
+    {
+      "epoch": 5.883720930232558,
+      "grad_norm": 2.391784191131592,
+      "learning_rate": 8.837209302325582e-05,
+      "loss": 0.2284,
+      "step": 253
+    },
+    {
+      "epoch": 5.906976744186046,
+      "grad_norm": 4.925626754760742,
+      "learning_rate": 8.822674418604652e-05,
+      "loss": 0.1559,
+      "step": 254
+    },
+    {
+      "epoch": 5.930232558139535,
+      "grad_norm": 5.659231185913086,
+      "learning_rate": 8.808139534883722e-05,
+      "loss": 0.1527,
+      "step": 255
+    },
+    {
+      "epoch": 5.953488372093023,
+      "grad_norm": 1.3519715070724487,
+      "learning_rate": 8.793604651162792e-05,
+      "loss": 0.097,
+      "step": 256
+    },
+    {
+      "epoch": 5.976744186046512,
+      "grad_norm": 3.517763137817383,
+      "learning_rate": 8.779069767441861e-05,
+      "loss": 0.2737,
+      "step": 257
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 3.0338218212127686,
+      "learning_rate": 8.76453488372093e-05,
+      "loss": 0.1154,
+      "step": 258
+    },
+    {
+      "epoch": 6.023255813953488,
+      "grad_norm": 2.716304063796997,
+      "learning_rate": 8.75e-05,
+      "loss": 0.2143,
+      "step": 259
+    },
+    {
+      "epoch": 6.046511627906977,
+      "grad_norm": 1.8077433109283447,
+      "learning_rate": 8.73546511627907e-05,
+      "loss": 0.1411,
+      "step": 260
+    },
+    {
+      "epoch": 6.046511627906977,
+      "eval_accuracy": 0.8950819672131147,
+      "eval_f1": 0.8945822172297591,
+      "eval_loss": 0.23519852757453918,
+      "eval_precision": 0.8966414996094767,
+      "eval_recall": 0.8937710002584647,
+      "eval_runtime": 0.0746,
+      "eval_samples_per_second": 4089.283,
+      "eval_steps_per_second": 67.037,
+      "step": 260
+    },
+    {
+      "epoch": 6.069767441860465,
+      "grad_norm": 4.399131774902344,
+      "learning_rate": 8.72093023255814e-05,
+      "loss": 0.1678,
+      "step": 261
+    },
+    {
+      "epoch": 6.093023255813954,
+      "grad_norm": 1.8267767429351807,
+      "learning_rate": 8.70639534883721e-05,
+      "loss": 0.0858,
+      "step": 262
+    },
+    {
+      "epoch": 6.116279069767442,
+      "grad_norm": 4.386148929595947,
+      "learning_rate": 8.69186046511628e-05,
+      "loss": 0.1666,
+      "step": 263
+    },
+    {
+      "epoch": 6.1395348837209305,
+      "grad_norm": 4.314615249633789,
+      "learning_rate": 8.67732558139535e-05,
+      "loss": 0.1782,
+      "step": 264
+    },
+    {
+      "epoch": 6.162790697674419,
+      "grad_norm": 1.9564558267593384,
+      "learning_rate": 8.662790697674419e-05,
+      "loss": 0.203,
+      "step": 265
+    },
+    {
+      "epoch": 6.186046511627907,
+      "grad_norm": 3.557976484298706,
+      "learning_rate": 8.648255813953489e-05,
+      "loss": 0.1347,
+      "step": 266
+    },
+    {
+      "epoch": 6.209302325581396,
+      "grad_norm": 2.464057207107544,
+      "learning_rate": 8.633720930232559e-05,
+      "loss": 0.0906,
+      "step": 267
+    },
+    {
+      "epoch": 6.232558139534884,
+      "grad_norm": 2.893970251083374,
+      "learning_rate": 8.619186046511628e-05,
+      "loss": 0.0893,
+      "step": 268
+    },
+    {
+      "epoch": 6.2558139534883725,
+      "grad_norm": 4.821974277496338,
+      "learning_rate": 8.604651162790697e-05,
+      "loss": 0.1008,
+      "step": 269
+    },
+    {
+      "epoch": 6.27906976744186,
+      "grad_norm": 2.767670154571533,
+      "learning_rate": 8.590116279069767e-05,
+      "loss": 0.1072,
+      "step": 270
+    },
+    {
+      "epoch": 6.3023255813953485,
+      "grad_norm": 2.337888240814209,
+      "learning_rate": 8.575581395348837e-05,
+      "loss": 0.1691,
+      "step": 271
+    },
+    {
+      "epoch": 6.325581395348837,
+      "grad_norm": 2.4074652194976807,
+      "learning_rate": 8.561046511627907e-05,
+      "loss": 0.1671,
+      "step": 272
+    },
+    {
+      "epoch": 6.348837209302325,
+      "grad_norm": 6.366968631744385,
+      "learning_rate": 8.546511627906977e-05,
+      "loss": 0.1741,
+      "step": 273
+    },
+    {
+      "epoch": 6.372093023255814,
+      "grad_norm": 2.4231204986572266,
+      "learning_rate": 8.531976744186047e-05,
+      "loss": 0.1563,
+      "step": 274
+    },
+    {
+      "epoch": 6.395348837209302,
+      "grad_norm": 3.424386739730835,
+      "learning_rate": 8.517441860465117e-05,
+      "loss": 0.2211,
+      "step": 275
+    },
+    {
+      "epoch": 6.4186046511627906,
+      "grad_norm": 6.0630717277526855,
+      "learning_rate": 8.502906976744187e-05,
+      "loss": 0.3506,
+      "step": 276
+    },
+    {
+      "epoch": 6.441860465116279,
+      "grad_norm": 2.5241973400115967,
+      "learning_rate": 8.488372093023255e-05,
+      "loss": 0.1239,
+      "step": 277
+    },
+    {
+      "epoch": 6.465116279069767,
+      "grad_norm": 1.1857047080993652,
+      "learning_rate": 8.473837209302325e-05,
+      "loss": 0.1218,
+      "step": 278
+    },
+    {
+      "epoch": 6.488372093023256,
+      "grad_norm": 2.826883316040039,
+      "learning_rate": 8.459302325581395e-05,
+      "loss": 0.1296,
+      "step": 279
+    },
+    {
+      "epoch": 6.511627906976744,
+      "grad_norm": 3.720126152038574,
+      "learning_rate": 8.444767441860465e-05,
+      "loss": 0.1594,
+      "step": 280
+    },
+    {
+      "epoch": 6.511627906976744,
+      "eval_accuracy": 0.8819672131147541,
+      "eval_f1": 0.8815073815073815,
+      "eval_loss": 0.25715914368629456,
+      "eval_precision": 0.8827922077922078,
+      "eval_recall": 0.8809123804600671,
+      "eval_runtime": 0.0733,
+      "eval_samples_per_second": 4160.082,
+      "eval_steps_per_second": 68.198,
+      "step": 280
+    },
+    {
+      "epoch": 6.534883720930233,
+      "grad_norm": 2.7126853466033936,
+      "learning_rate": 8.430232558139536e-05,
+      "loss": 0.1234,
+      "step": 281
+    },
+    {
+      "epoch": 6.558139534883721,
+      "grad_norm": 2.641782522201538,
+      "learning_rate": 8.415697674418606e-05,
+      "loss": 0.1515,
+      "step": 282
+    },
+    {
+      "epoch": 6.5813953488372094,
+      "grad_norm": 5.715346813201904,
+      "learning_rate": 8.401162790697676e-05,
+      "loss": 0.1463,
+      "step": 283
+    },
+    {
+      "epoch": 6.604651162790698,
+      "grad_norm": 4.03624153137207,
+      "learning_rate": 8.386627906976746e-05,
+      "loss": 0.1425,
+      "step": 284
+    },
+    {
+      "epoch": 6.627906976744186,
+      "grad_norm": 0.8614564538002014,
+      "learning_rate": 8.372093023255814e-05,
+      "loss": 0.1115,
+      "step": 285
+    },
+    {
+      "epoch": 6.651162790697675,
+      "grad_norm": 3.0801596641540527,
+      "learning_rate": 8.357558139534884e-05,
+      "loss": 0.1954,
+      "step": 286
+    },
+    {
+      "epoch": 6.674418604651163,
+      "grad_norm": 2.3671586513519287,
+      "learning_rate": 8.343023255813954e-05,
+      "loss": 0.1223,
+      "step": 287
+    },
+    {
+      "epoch": 6.6976744186046515,
+      "grad_norm": 4.468645095825195,
+      "learning_rate": 8.328488372093024e-05,
+      "loss": 0.1422,
+      "step": 288
+    },
+    {
+      "epoch": 6.720930232558139,
+      "grad_norm": 2.540923833847046,
+      "learning_rate": 8.313953488372094e-05,
+      "loss": 0.127,
+      "step": 289
+    },
+    {
+      "epoch": 6.7441860465116275,
+      "grad_norm": 1.3170732259750366,
+      "learning_rate": 8.299418604651164e-05,
+      "loss": 0.0479,
+      "step": 290
+    },
+    {
+      "epoch": 6.767441860465116,
+      "grad_norm": 3.5731353759765625,
+      "learning_rate": 8.284883720930234e-05,
+      "loss": 0.1118,
+      "step": 291
+    },
+    {
+      "epoch": 6.790697674418604,
+      "grad_norm": 6.928558826446533,
+      "learning_rate": 8.270348837209303e-05,
+      "loss": 0.2011,
+      "step": 292
+    },
+    {
+      "epoch": 6.813953488372093,
+      "grad_norm": 4.524282932281494,
+      "learning_rate": 8.255813953488373e-05,
+      "loss": 0.1252,
+      "step": 293
+    },
+    {
+      "epoch": 6.837209302325581,
+      "grad_norm": 7.167698383331299,
+      "learning_rate": 8.241279069767442e-05,
+      "loss": 0.153,
+      "step": 294
+    },
+    {
+      "epoch": 6.8604651162790695,
+      "grad_norm": 4.981342792510986,
+      "learning_rate": 8.226744186046512e-05,
+      "loss": 0.2423,
+      "step": 295
+    },
+    {
+      "epoch": 6.883720930232558,
+      "grad_norm": 2.139774799346924,
+      "learning_rate": 8.212209302325582e-05,
+      "loss": 0.1548,
+      "step": 296
+    },
+    {
+      "epoch": 6.906976744186046,
+      "grad_norm": 2.5026113986968994,
+      "learning_rate": 8.197674418604652e-05,
+      "loss": 0.0935,
+      "step": 297
+    },
+    {
+      "epoch": 6.930232558139535,
+      "grad_norm": 1.633834719657898,
+      "learning_rate": 8.183139534883721e-05,
+      "loss": 0.1224,
+      "step": 298
+    },
+    {
+      "epoch": 6.953488372093023,
+      "grad_norm": 4.0517168045043945,
+      "learning_rate": 8.168604651162791e-05,
+      "loss": 0.106,
+      "step": 299
+    },
+    {
+      "epoch": 6.976744186046512,
+      "grad_norm": 4.622942924499512,
+      "learning_rate": 8.154069767441861e-05,
+      "loss": 0.1923,
+      "step": 300
+    },
+    {
+      "epoch": 6.976744186046512,
+      "eval_accuracy": 0.8918032786885246,
+      "eval_f1": 0.8912373974778747,
+      "eval_loss": 0.276712030172348,
+      "eval_precision": 0.8937043795620438,
+      "eval_recall": 0.8903463427242182,
+      "eval_runtime": 0.0737,
+      "eval_samples_per_second": 4140.129,
+      "eval_steps_per_second": 67.871,
+      "step": 300
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 3.300420045852661,
+      "learning_rate": 8.139534883720931e-05,
+      "loss": 0.1218,
+      "step": 301
+    },
+    {
+      "epoch": 7.023255813953488,
+      "grad_norm": 2.9694247245788574,
+      "learning_rate": 8.125000000000001e-05,
+      "loss": 0.1067,
+      "step": 302
+    },
+    {
+      "epoch": 7.046511627906977,
+      "grad_norm": 3.805917978286743,
+      "learning_rate": 8.11046511627907e-05,
+      "loss": 0.1165,
+      "step": 303
+    },
+    {
+      "epoch": 7.069767441860465,
+      "grad_norm": 2.0280096530914307,
+      "learning_rate": 8.09593023255814e-05,
+      "loss": 0.119,
+      "step": 304
+    },
+    {
+      "epoch": 7.093023255813954,
+      "grad_norm": 5.049257278442383,
+      "learning_rate": 8.081395348837209e-05,
+      "loss": 0.1838,
+      "step": 305
+    },
+    {
+      "epoch": 7.116279069767442,
+      "grad_norm": 6.135448455810547,
+      "learning_rate": 8.066860465116279e-05,
+      "loss": 0.2193,
+      "step": 306
+    },
+    {
+      "epoch": 7.1395348837209305,
+      "grad_norm": 2.0232150554656982,
+      "learning_rate": 8.052325581395349e-05,
+      "loss": 0.0829,
+      "step": 307
+    },
+    {
+      "epoch": 7.162790697674419,
+      "grad_norm": 3.0096096992492676,
+      "learning_rate": 8.037790697674419e-05,
+      "loss": 0.1394,
+      "step": 308
+    },
+    {
+      "epoch": 7.186046511627907,
+      "grad_norm": 0.9171739220619202,
+      "learning_rate": 8.023255813953489e-05,
+      "loss": 0.1058,
+      "step": 309
+    },
+    {
+      "epoch": 7.209302325581396,
+      "grad_norm": 3.929137706756592,
+      "learning_rate": 8.008720930232559e-05,
+      "loss": 0.2189,
+      "step": 310
+    },
+    {
+      "epoch": 7.232558139534884,
+      "grad_norm": 1.1936514377593994,
+      "learning_rate": 7.994186046511629e-05,
+      "loss": 0.0981,
+      "step": 311
+    },
+    {
+      "epoch": 7.2558139534883725,
+      "grad_norm": 2.542099714279175,
+      "learning_rate": 7.979651162790697e-05,
+      "loss": 0.1052,
+      "step": 312
+    },
+    {
+      "epoch": 7.27906976744186,
+      "grad_norm": 2.42411732673645,
+      "learning_rate": 7.965116279069767e-05,
+      "loss": 0.0834,
+      "step": 313
+    },
+    {
+      "epoch": 7.3023255813953485,
+      "grad_norm": 1.07964289188385,
+      "learning_rate": 7.950581395348837e-05,
+      "loss": 0.0684,
+      "step": 314
+    },
+    {
+      "epoch": 7.325581395348837,
+      "grad_norm": 0.9596773386001587,
+      "learning_rate": 7.936046511627907e-05,
+      "loss": 0.0791,
+      "step": 315
+    },
+    {
+      "epoch": 7.348837209302325,
+      "grad_norm": 3.508575201034546,
+      "learning_rate": 7.921511627906977e-05,
+      "loss": 0.145,
+      "step": 316
+    },
+    {
+      "epoch": 7.372093023255814,
+      "grad_norm": 3.8409645557403564,
+      "learning_rate": 7.906976744186047e-05,
+      "loss": 0.1057,
+      "step": 317
+    },
+    {
+      "epoch": 7.395348837209302,
+      "grad_norm": 5.247814655303955,
+      "learning_rate": 7.892441860465116e-05,
+      "loss": 0.0731,
+      "step": 318
+    },
+    {
+      "epoch": 7.4186046511627906,
+      "grad_norm": 2.5410549640655518,
+      "learning_rate": 7.877906976744186e-05,
+      "loss": 0.2125,
+      "step": 319
+    },
+    {
+      "epoch": 7.441860465116279,
+      "grad_norm": 3.2063095569610596,
+      "learning_rate": 7.863372093023256e-05,
+      "loss": 0.0896,
+      "step": 320
+    },
+    {
+      "epoch": 7.441860465116279,
+      "eval_accuracy": 0.9114754098360656,
+      "eval_f1": 0.9114411381991419,
+      "eval_loss": 0.22706902027130127,
+      "eval_precision": 0.9115497076023391,
+      "eval_recall": 0.9122943051606789,
+      "eval_runtime": 0.0744,
+      "eval_samples_per_second": 4097.456,
+      "eval_steps_per_second": 67.171,
+      "step": 320
+    },
+    {
+      "epoch": 7.465116279069767,
+      "grad_norm": 1.1547659635543823,
+      "learning_rate": 7.848837209302326e-05,
+      "loss": 0.0425,
+      "step": 321
+    },
+    {
+      "epoch": 7.488372093023256,
+      "grad_norm": 10.542337417602539,
+      "learning_rate": 7.834302325581395e-05,
+      "loss": 0.4056,
+      "step": 322
+    },
+    {
+      "epoch": 7.511627906976744,
+      "grad_norm": 6.774454593658447,
+      "learning_rate": 7.819767441860465e-05,
+      "loss": 0.106,
+      "step": 323
+    },
+    {
+      "epoch": 7.534883720930233,
+      "grad_norm": 3.680190086364746,
+      "learning_rate": 7.805232558139536e-05,
+      "loss": 0.1208,
+      "step": 324
+    },
+    {
+      "epoch": 7.558139534883721,
+      "grad_norm": 3.1805880069732666,
+      "learning_rate": 7.790697674418606e-05,
+      "loss": 0.0968,
+      "step": 325
+    },
+    {
+      "epoch": 7.5813953488372094,
+      "grad_norm": 3.887376070022583,
+      "learning_rate": 7.776162790697676e-05,
+      "loss": 0.3548,
+      "step": 326
+    },
+    {
+      "epoch": 7.604651162790698,
+      "grad_norm": 2.4842185974121094,
+      "learning_rate": 7.761627906976745e-05,
+      "loss": 0.0782,
+      "step": 327
+    },
+    {
+      "epoch": 7.627906976744186,
+      "grad_norm": 4.861859321594238,
+      "learning_rate": 7.747093023255815e-05,
+      "loss": 0.1529,
+      "step": 328
+    },
+    {
+      "epoch": 7.651162790697675,
+      "grad_norm": 4.307983875274658,
+      "learning_rate": 7.732558139534884e-05,
+      "loss": 0.1938,
+      "step": 329
+    },
+    {
+      "epoch": 7.674418604651163,
+      "grad_norm": 6.8119587898254395,
+      "learning_rate": 7.718023255813954e-05,
+      "loss": 0.1118,
+      "step": 330
+    },
+    {
+      "epoch": 7.6976744186046515,
+      "grad_norm": 1.8230775594711304,
+      "learning_rate": 7.703488372093024e-05,
+      "loss": 0.0551,
+      "step": 331
+    },
+    {
+      "epoch": 7.720930232558139,
+      "grad_norm": 1.6934735774993896,
+      "learning_rate": 7.688953488372094e-05,
+      "loss": 0.0629,
+      "step": 332
+    },
+    {
+      "epoch": 7.7441860465116275,
+      "grad_norm": 2.537760019302368,
+      "learning_rate": 7.674418604651163e-05,
+      "loss": 0.1084,
+      "step": 333
+    },
+    {
+      "epoch": 7.767441860465116,
+      "grad_norm": 1.6460330486297607,
+      "learning_rate": 7.659883720930233e-05,
+      "loss": 0.0927,
+      "step": 334
+    },
+    {
+      "epoch": 7.790697674418604,
+      "grad_norm": 1.8621909618377686,
+      "learning_rate": 7.645348837209303e-05,
+      "loss": 0.107,
+      "step": 335
+    },
+    {
+      "epoch": 7.813953488372093,
+      "grad_norm": 7.628468036651611,
+      "learning_rate": 7.630813953488373e-05,
+      "loss": 0.1038,
+      "step": 336
+    },
+    {
+      "epoch": 7.837209302325581,
+      "grad_norm": 2.3361170291900635,
+      "learning_rate": 7.616279069767443e-05,
+      "loss": 0.1017,
+      "step": 337
+    },
+    {
+      "epoch": 7.8604651162790695,
+      "grad_norm": 1.8995561599731445,
+      "learning_rate": 7.601744186046513e-05,
+      "loss": 0.0414,
+      "step": 338
+    },
+    {
+      "epoch": 7.883720930232558,
+      "grad_norm": 5.01551628112793,
+      "learning_rate": 7.587209302325581e-05,
+      "loss": 0.1229,
+      "step": 339
+    },
+    {
+      "epoch": 7.906976744186046,
+      "grad_norm": 4.123791694641113,
+      "learning_rate": 7.572674418604651e-05,
+      "loss": 0.099,
+      "step": 340
+    },
+    {
+      "epoch": 7.906976744186046,
+      "eval_accuracy": 0.9016393442622951,
+      "eval_f1": 0.9016129032258065,
+      "eval_loss": 0.22697412967681885,
+      "eval_precision": 0.9018878472520857,
+      "eval_recall": 0.9025803394503318,
+      "eval_runtime": 0.0787,
+      "eval_samples_per_second": 3874.417,
+      "eval_steps_per_second": 63.515,
+      "step": 340
+    },
+    {
+      "epoch": 7.930232558139535,
+      "grad_norm": 1.9877846240997314,
+      "learning_rate": 7.558139534883721e-05,
+      "loss": 0.0625,
+      "step": 341
+    },
+    {
+      "epoch": 7.953488372093023,
+      "grad_norm": 1.6333461999893188,
+      "learning_rate": 7.543604651162791e-05,
+      "loss": 0.051,
+      "step": 342
+    },
+    {
+      "epoch": 7.976744186046512,
+      "grad_norm": 8.968517303466797,
+      "learning_rate": 7.529069767441861e-05,
+      "loss": 0.0781,
+      "step": 343
+    },
+    {
+      "epoch": 8.0,
+      "grad_norm": 2.306673765182495,
+      "learning_rate": 7.514534883720931e-05,
+      "loss": 0.0684,
+      "step": 344
+    },
+    {
+      "epoch": 8.023255813953488,
+      "grad_norm": 5.471800804138184,
+      "learning_rate": 7.500000000000001e-05,
+      "loss": 0.1,
+      "step": 345
+    },
+    {
+      "epoch": 8.046511627906977,
+      "grad_norm": 6.212913513183594,
+      "learning_rate": 7.48546511627907e-05,
+      "loss": 0.0966,
+      "step": 346
+    },
+    {
+      "epoch": 8.069767441860465,
+      "grad_norm": 7.509064197540283,
+      "learning_rate": 7.47093023255814e-05,
+      "loss": 0.2218,
+      "step": 347
+    },
+    {
+      "epoch": 8.093023255813954,
+      "grad_norm": 5.755818843841553,
+      "learning_rate": 7.456395348837209e-05,
+      "loss": 0.2037,
+      "step": 348
+    },
+    {
+      "epoch": 8.116279069767442,
+      "grad_norm": 2.3190879821777344,
+      "learning_rate": 7.441860465116279e-05,
+      "loss": 0.1259,
+      "step": 349
+    },
+    {
+      "epoch": 8.13953488372093,
+      "grad_norm": 0.9036604166030884,
+      "learning_rate": 7.427325581395349e-05,
+      "loss": 0.0662,
+      "step": 350
+    },
+    {
+      "epoch": 8.162790697674419,
+      "grad_norm": 2.2601730823516846,
+      "learning_rate": 7.412790697674419e-05,
+      "loss": 0.0833,
+      "step": 351
+    },
+    {
+      "epoch": 8.186046511627907,
+      "grad_norm": 2.334266185760498,
+      "learning_rate": 7.398255813953489e-05,
+      "loss": 0.0482,
+      "step": 352
+    },
+    {
+      "epoch": 8.209302325581396,
+      "grad_norm": 5.374472141265869,
+      "learning_rate": 7.383720930232558e-05,
+      "loss": 0.1771,
+      "step": 353
+    },
+    {
+      "epoch": 8.232558139534884,
+      "grad_norm": 5.360229969024658,
+      "learning_rate": 7.369186046511628e-05,
+      "loss": 0.2253,
+      "step": 354
+    },
+    {
+      "epoch": 8.255813953488373,
+      "grad_norm": 2.7183334827423096,
+      "learning_rate": 7.354651162790698e-05,
+      "loss": 0.13,
+      "step": 355
+    },
+    {
+      "epoch": 8.279069767441861,
+      "grad_norm": 1.2531534433364868,
+      "learning_rate": 7.340116279069768e-05,
+      "loss": 0.0898,
+      "step": 356
+    },
+    {
+      "epoch": 8.30232558139535,
+      "grad_norm": 5.129657745361328,
+      "learning_rate": 7.325581395348837e-05,
+      "loss": 0.0884,
+      "step": 357
+    },
+    {
+      "epoch": 8.325581395348838,
+      "grad_norm": 0.5411895513534546,
+      "learning_rate": 7.311046511627907e-05,
+      "loss": 0.0227,
+      "step": 358
+    },
+    {
+      "epoch": 8.348837209302326,
+      "grad_norm": 3.1815500259399414,
+      "learning_rate": 7.296511627906976e-05,
+      "loss": 0.1153,
+      "step": 359
+    },
+    {
+      "epoch": 8.372093023255815,
+      "grad_norm": 3.956552743911743,
+      "learning_rate": 7.281976744186046e-05,
+      "loss": 0.1048,
+      "step": 360
+    },
+    {
+      "epoch": 8.372093023255815,
+      "eval_accuracy": 0.8918032786885246,
+      "eval_f1": 0.8917846660000646,
+      "eval_loss": 0.2492106705904007,
+      "eval_precision": 0.8922580645161291,
+      "eval_recall": 0.8928663737399845,
+      "eval_runtime": 0.0761,
+      "eval_samples_per_second": 4010.404,
+      "eval_steps_per_second": 65.744,
+      "step": 360
+    },
+    {
+      "epoch": 8.395348837209303,
+      "grad_norm": 3.400632619857788,
+      "learning_rate": 7.267441860465116e-05,
+      "loss": 0.1008,
+      "step": 361
+    },
+    {
+      "epoch": 8.418604651162791,
+      "grad_norm": 4.204819202423096,
+      "learning_rate": 7.252906976744186e-05,
+      "loss": 0.0525,
+      "step": 362
+    },
+    {
+      "epoch": 8.44186046511628,
+      "grad_norm": 4.265756130218506,
+      "learning_rate": 7.238372093023256e-05,
+      "loss": 0.1789,
+      "step": 363
+    },
+    {
+      "epoch": 8.465116279069768,
+      "grad_norm": 3.7337512969970703,
+      "learning_rate": 7.223837209302326e-05,
+      "loss": 0.2187,
+      "step": 364
+    },
+    {
+      "epoch": 8.488372093023255,
+      "grad_norm": 7.506768226623535,
+      "learning_rate": 7.209302325581396e-05,
+      "loss": 0.1643,
+      "step": 365
+    },
+    {
+      "epoch": 8.511627906976745,
+      "grad_norm": 2.040639877319336,
+      "learning_rate": 7.194767441860464e-05,
+      "loss": 0.0645,
+      "step": 366
+    },
+    {
+      "epoch": 8.534883720930232,
+      "grad_norm": 3.809933662414551,
+      "learning_rate": 7.180232558139535e-05,
+      "loss": 0.1413,
+      "step": 367
+    },
+    {
+      "epoch": 8.55813953488372,
+      "grad_norm": 4.585285186767578,
+      "learning_rate": 7.165697674418605e-05,
+      "loss": 0.073,
+      "step": 368
+    },
+    {
+      "epoch": 8.581395348837209,
+      "grad_norm": 1.4243372678756714,
+      "learning_rate": 7.151162790697675e-05,
+      "loss": 0.0358,
+      "step": 369
+    },
+    {
+      "epoch": 8.604651162790697,
+      "grad_norm": 7.1119513511657715,
+      "learning_rate": 7.136627906976745e-05,
+      "loss": 0.1204,
+      "step": 370
+    },
+    {
+      "epoch": 8.627906976744185,
+      "grad_norm": 4.922991752624512,
+      "learning_rate": 7.122093023255815e-05,
+      "loss": 0.1183,
+      "step": 371
+    },
+    {
+      "epoch": 8.651162790697674,
+      "grad_norm": 2.3582427501678467,
+      "learning_rate": 7.107558139534885e-05,
+      "loss": 0.1057,
+      "step": 372
+    },
+    {
+      "epoch": 8.674418604651162,
+      "grad_norm": 1.6656956672668457,
+      "learning_rate": 7.093023255813955e-05,
+      "loss": 0.0434,
+      "step": 373
+    },
+    {
+      "epoch": 8.69767441860465,
+      "grad_norm": 3.274289608001709,
+      "learning_rate": 7.078488372093023e-05,
+      "loss": 0.1541,
+      "step": 374
+    },
+    {
+      "epoch": 8.720930232558139,
+      "grad_norm": 4.472874641418457,
+      "learning_rate": 7.063953488372093e-05,
+      "loss": 0.1322,
+      "step": 375
+    },
+    {
+      "epoch": 8.744186046511627,
+      "grad_norm": 2.455209493637085,
+      "learning_rate": 7.049418604651163e-05,
+      "loss": 0.1081,
+      "step": 376
+    },
+    {
+      "epoch": 8.767441860465116,
+      "grad_norm": 2.7206978797912598,
+      "learning_rate": 7.034883720930233e-05,
+      "loss": 0.1198,
+      "step": 377
+    },
+    {
+      "epoch": 8.790697674418604,
+      "grad_norm": 6.333116054534912,
+      "learning_rate": 7.020348837209303e-05,
+      "loss": 0.1965,
+      "step": 378
+    },
+    {
+      "epoch": 8.813953488372093,
+      "grad_norm": 1.0274301767349243,
+      "learning_rate": 7.005813953488373e-05,
+      "loss": 0.0388,
+      "step": 379
+    },
+    {
+      "epoch": 8.837209302325581,
+      "grad_norm": 4.29472541809082,
+      "learning_rate": 6.991279069767443e-05,
+      "loss": 0.0699,
+      "step": 380
+    },
+    {
+      "epoch": 8.837209302325581,
+      "eval_accuracy": 0.898360655737705,
+      "eval_f1": 0.8978290703580035,
+      "eval_loss": 0.26147955656051636,
+      "eval_precision": 0.9003302050747306,
+      "eval_recall": 0.8969156543465151,
+      "eval_runtime": 0.0748,
+      "eval_samples_per_second": 4079.802,
+      "eval_steps_per_second": 66.882,
+      "step": 380
+    },
+    {
+      "epoch": 8.86046511627907,
+      "grad_norm": 3.4688355922698975,
+      "learning_rate": 6.976744186046513e-05,
+      "loss": 0.1465,
+      "step": 381
+    },
+    {
+      "epoch": 8.883720930232558,
+      "grad_norm": 4.693603992462158,
+      "learning_rate": 6.962209302325582e-05,
+      "loss": 0.0675,
+      "step": 382
+    },
+    {
+      "epoch": 8.906976744186046,
+      "grad_norm": 1.98982572555542,
+      "learning_rate": 6.947674418604651e-05,
+      "loss": 0.1484,
+      "step": 383
+    },
+    {
+      "epoch": 8.930232558139535,
+      "grad_norm": 1.312723994255066,
+      "learning_rate": 6.933139534883721e-05,
+      "loss": 0.0458,
+      "step": 384
+    },
+    {
+      "epoch": 8.953488372093023,
+      "grad_norm": 2.106917142868042,
+      "learning_rate": 6.918604651162791e-05,
+      "loss": 0.0417,
+      "step": 385
+    },
+    {
+      "epoch": 8.976744186046512,
+      "grad_norm": 4.182035446166992,
+      "learning_rate": 6.90406976744186e-05,
+      "loss": 0.0459,
+      "step": 386
+    },
+    {
+      "epoch": 9.0,
+      "grad_norm": 2.1285786628723145,
+      "learning_rate": 6.88953488372093e-05,
+      "loss": 0.0743,
+      "step": 387
+    },
+    {
+      "epoch": 9.023255813953488,
+      "grad_norm": 4.470812797546387,
+      "learning_rate": 6.875e-05,
+      "loss": 0.0973,
+      "step": 388
+    },
+    {
+      "epoch": 9.046511627906977,
+      "grad_norm": 1.6776498556137085,
+      "learning_rate": 6.86046511627907e-05,
+      "loss": 0.1894,
+      "step": 389
+    },
+    {
+      "epoch": 9.069767441860465,
+      "grad_norm": 3.9984216690063477,
+      "learning_rate": 6.84593023255814e-05,
+      "loss": 0.1704,
+      "step": 390
+    },
+    {
+      "epoch": 9.093023255813954,
+      "grad_norm": 2.3983538150787354,
+      "learning_rate": 6.83139534883721e-05,
+      "loss": 0.0736,
+      "step": 391
+    },
+    {
+      "epoch": 9.116279069767442,
+      "grad_norm": 2.989044427871704,
+      "learning_rate": 6.81686046511628e-05,
+      "loss": 0.0892,
+      "step": 392
+    },
+    {
+      "epoch": 9.13953488372093,
+      "grad_norm": 3.356992483139038,
+      "learning_rate": 6.802325581395348e-05,
+      "loss": 0.0999,
+      "step": 393
+    },
+    {
+      "epoch": 9.162790697674419,
+      "grad_norm": 2.2563319206237793,
+      "learning_rate": 6.787790697674418e-05,
+      "loss": 0.0452,
+      "step": 394
+    },
+    {
+      "epoch": 9.186046511627907,
+      "grad_norm": 5.215000152587891,
+      "learning_rate": 6.773255813953488e-05,
+      "loss": 0.0793,
+      "step": 395
+    },
+    {
+      "epoch": 9.209302325581396,
+      "grad_norm": 5.427519798278809,
+      "learning_rate": 6.758720930232558e-05,
+      "loss": 0.0755,
+      "step": 396
+    },
+    {
+      "epoch": 9.232558139534884,
+      "grad_norm": 7.729520797729492,
+      "learning_rate": 6.744186046511628e-05,
+      "loss": 0.2358,
+      "step": 397
+    },
+    {
+      "epoch": 9.255813953488373,
+      "grad_norm": 5.97319221496582,
+      "learning_rate": 6.729651162790698e-05,
+      "loss": 0.1829,
+      "step": 398
+    },
+    {
+      "epoch": 9.279069767441861,
+      "grad_norm": 4.662299633026123,
+      "learning_rate": 6.715116279069768e-05,
+      "loss": 0.0549,
+      "step": 399
+    },
+    {
+      "epoch": 9.30232558139535,
+      "grad_norm": 5.684605598449707,
+      "learning_rate": 6.700581395348838e-05,
+      "loss": 0.1466,
+      "step": 400
+    },
+    {
+      "epoch": 9.30232558139535,
+      "eval_accuracy": 0.9147540983606557,
+      "eval_f1": 0.9145474137931033,
+      "eval_loss": 0.22707484662532806,
+      "eval_precision": 0.9148550724637681,
+      "eval_recall": 0.9143189454639442,
+      "eval_runtime": 0.0754,
+      "eval_samples_per_second": 4044.013,
+      "eval_steps_per_second": 66.295,
+      "step": 400
+    },
+    {
+      "epoch": 9.325581395348838,
+      "grad_norm": 8.901346206665039,
+      "learning_rate": 6.686046511627908e-05,
+      "loss": 0.1351,
+      "step": 401
+    },
+    {
+      "epoch": 9.348837209302326,
+      "grad_norm": 3.3197011947631836,
+      "learning_rate": 6.671511627906976e-05,
+      "loss": 0.0962,
+      "step": 402
+    },
+    {
+      "epoch": 9.372093023255815,
+      "grad_norm": 5.270369529724121,
+      "learning_rate": 6.656976744186046e-05,
+      "loss": 0.0862,
+      "step": 403
+    },
+    {
+      "epoch": 9.395348837209303,
+      "grad_norm": 1.4588712453842163,
+      "learning_rate": 6.642441860465116e-05,
+      "loss": 0.0286,
+      "step": 404
+    },
+    {
+      "epoch": 9.418604651162791,
+      "grad_norm": 7.928572177886963,
+      "learning_rate": 6.627906976744186e-05,
+      "loss": 0.1137,
+      "step": 405
+    },
+    {
+      "epoch": 9.44186046511628,
+      "grad_norm": 4.6285223960876465,
+      "learning_rate": 6.613372093023256e-05,
+      "loss": 0.0907,
+      "step": 406
+    },
+    {
+      "epoch": 9.465116279069768,
+      "grad_norm": 1.1047800779342651,
+      "learning_rate": 6.598837209302326e-05,
+      "loss": 0.0279,
+      "step": 407
+    },
+    {
+      "epoch": 9.488372093023255,
+      "grad_norm": 2.0423073768615723,
+      "learning_rate": 6.584302325581395e-05,
+      "loss": 0.0528,
+      "step": 408
+    },
+    {
+      "epoch": 9.511627906976745,
+      "grad_norm": 3.1448893547058105,
+      "learning_rate": 6.569767441860465e-05,
+      "loss": 0.0557,
+      "step": 409
+    },
+    {
+      "epoch": 9.534883720930232,
+      "grad_norm": 3.5887062549591064,
+      "learning_rate": 6.555232558139535e-05,
+      "loss": 0.0858,
+      "step": 410
+    },
+    {
+      "epoch": 9.55813953488372,
+      "grad_norm": 4.293898105621338,
+      "learning_rate": 6.540697674418605e-05,
+      "loss": 0.0547,
+      "step": 411
+    },
+    {
+      "epoch": 9.581395348837209,
+      "grad_norm": 3.3504300117492676,
+      "learning_rate": 6.526162790697675e-05,
+      "loss": 0.1361,
+      "step": 412
+    },
+    {
+      "epoch": 9.604651162790697,
+      "grad_norm": 0.5247292518615723,
+      "learning_rate": 6.511627906976745e-05,
+      "loss": 0.0168,
+      "step": 413
+    },
+    {
+      "epoch": 9.627906976744185,
+      "grad_norm": 0.9081286191940308,
+      "learning_rate": 6.497093023255815e-05,
+      "loss": 0.0337,
+      "step": 414
+    },
+    {
+      "epoch": 9.651162790697674,
+      "grad_norm": 5.076108932495117,
+      "learning_rate": 6.482558139534885e-05,
+      "loss": 0.0972,
+      "step": 415
+    },
+    {
+      "epoch": 9.674418604651162,
+      "grad_norm": 2.2447941303253174,
+      "learning_rate": 6.468023255813955e-05,
+      "loss": 0.0468,
+      "step": 416
+    },
+    {
+      "epoch": 9.69767441860465,
+      "grad_norm": 6.019301891326904,
+      "learning_rate": 6.453488372093024e-05,
+      "loss": 0.0967,
+      "step": 417
+    },
+    {
+      "epoch": 9.720930232558139,
+      "grad_norm": 1.4316606521606445,
+      "learning_rate": 6.438953488372094e-05,
+      "loss": 0.0666,
+      "step": 418
+    },
+    {
+      "epoch": 9.744186046511627,
+      "grad_norm": 1.9364207983016968,
+      "learning_rate": 6.424418604651163e-05,
+      "loss": 0.1047,
+      "step": 419
+    },
+    {
+      "epoch": 9.767441860465116,
+      "grad_norm": 4.00551700592041,
+      "learning_rate": 6.409883720930233e-05,
+      "loss": 0.0663,
+      "step": 420
+    },
+    {
+      "epoch": 9.767441860465116,
+      "eval_accuracy": 0.9114754098360656,
+      "eval_f1": 0.9106105834464043,
+      "eval_loss": 0.26964467763900757,
+      "eval_precision": 0.9181201550387597,
+      "eval_recall": 0.9089342638063238,
+      "eval_runtime": 0.0726,
+      "eval_samples_per_second": 4201.095,
+      "eval_steps_per_second": 68.87,
+      "step": 420
+    },
+    {
+      "epoch": 9.790697674418604,
+      "grad_norm": 3.918790817260742,
+      "learning_rate": 6.395348837209303e-05,
+      "loss": 0.0773,
+      "step": 421
+    },
+    {
+      "epoch": 9.813953488372093,
+      "grad_norm": 4.5966901779174805,
+      "learning_rate": 6.380813953488373e-05,
+      "loss": 0.1099,
+      "step": 422
+    },
+    {
+      "epoch": 9.837209302325581,
+      "grad_norm": 3.332273244857788,
+      "learning_rate": 6.366279069767442e-05,
+      "loss": 0.0697,
+      "step": 423
+    },
+    {
+      "epoch": 9.86046511627907,
+      "grad_norm": 3.300297737121582,
+      "learning_rate": 6.351744186046512e-05,
+      "loss": 0.1235,
+      "step": 424
+    },
+    {
+      "epoch": 9.883720930232558,
+      "grad_norm": 1.9253029823303223,
+      "learning_rate": 6.337209302325582e-05,
+      "loss": 0.0276,
+      "step": 425
+    },
+    {
+      "epoch": 9.906976744186046,
+      "grad_norm": 2.729339599609375,
+      "learning_rate": 6.322674418604652e-05,
+      "loss": 0.2399,
+      "step": 426
+    },
+    {
+      "epoch": 9.930232558139535,
+      "grad_norm": 2.6790566444396973,
+      "learning_rate": 6.308139534883722e-05,
+      "loss": 0.0897,
+      "step": 427
+    },
+    {
+      "epoch": 9.953488372093023,
+      "grad_norm": 2.726901054382324,
+      "learning_rate": 6.29360465116279e-05,
+      "loss": 0.0768,
+      "step": 428
+    },
+    {
+      "epoch": 9.976744186046512,
+      "grad_norm": 2.4175329208374023,
+      "learning_rate": 6.27906976744186e-05,
+      "loss": 0.0943,
+      "step": 429
+    },
+    {
+      "epoch": 10.0,
+      "grad_norm": 4.533559799194336,
+      "learning_rate": 6.26453488372093e-05,
+      "loss": 0.127,
+      "step": 430
+    },
+    {
+      "epoch": 10.023255813953488,
+      "grad_norm": 3.8240880966186523,
+      "learning_rate": 6.25e-05,
+      "loss": 0.083,
+      "step": 431
+    },
+    {
+      "epoch": 10.046511627906977,
+      "grad_norm": 2.4785892963409424,
+      "learning_rate": 6.23546511627907e-05,
+      "loss": 0.0292,
+      "step": 432
+    },
+    {
+      "epoch": 10.069767441860465,
+      "grad_norm": 1.059528112411499,
+      "learning_rate": 6.22093023255814e-05,
+      "loss": 0.0619,
+      "step": 433
+    },
+    {
+      "epoch": 10.093023255813954,
+      "grad_norm": 2.8850722312927246,
+      "learning_rate": 6.20639534883721e-05,
+      "loss": 0.1071,
+      "step": 434
+    },
+    {
+      "epoch": 10.116279069767442,
+      "grad_norm": 1.9426606893539429,
+      "learning_rate": 6.19186046511628e-05,
+      "loss": 0.0988,
+      "step": 435
+    },
+    {
+      "epoch": 10.13953488372093,
+      "grad_norm": 2.1318485736846924,
+      "learning_rate": 6.17732558139535e-05,
+      "loss": 0.1346,
+      "step": 436
+    },
+    {
+      "epoch": 10.162790697674419,
+      "grad_norm": 2.793272018432617,
+      "learning_rate": 6.162790697674418e-05,
+      "loss": 0.0296,
+      "step": 437
+    },
+    {
+      "epoch": 10.186046511627907,
+      "grad_norm": 1.0038514137268066,
+      "learning_rate": 6.148255813953488e-05,
+      "loss": 0.0576,
+      "step": 438
+    },
+    {
+      "epoch": 10.209302325581396,
+      "grad_norm": 2.623504400253296,
+      "learning_rate": 6.133720930232558e-05,
+      "loss": 0.0607,
+      "step": 439
+    },
+    {
+      "epoch": 10.232558139534884,
+      "grad_norm": 1.1035139560699463,
+      "learning_rate": 6.119186046511628e-05,
+      "loss": 0.0473,
+      "step": 440
+    },
+    {
+      "epoch": 10.232558139534884,
+      "eval_accuracy": 0.9049180327868852,
+      "eval_f1": 0.9045074224021592,
+      "eval_loss": 0.2681082785129547,
+      "eval_precision": 0.9062147872063795,
+      "eval_recall": 0.9037649694150082,
+      "eval_runtime": 0.0732,
+      "eval_samples_per_second": 4164.7,
+      "eval_steps_per_second": 68.274,
+      "step": 440
+    },
+    {
+      "epoch": 10.255813953488373,
+      "grad_norm": 4.833049774169922,
+      "learning_rate": 6.104651162790698e-05,
+      "loss": 0.0978,
+      "step": 441
+    },
+    {
+      "epoch": 10.279069767441861,
+      "grad_norm": 0.8670517802238464,
+      "learning_rate": 6.0901162790697675e-05,
+      "loss": 0.0284,
+      "step": 442
+    },
+    {
+      "epoch": 10.30232558139535,
+      "grad_norm": 4.661880016326904,
+      "learning_rate": 6.0755813953488374e-05,
+      "loss": 0.1335,
+      "step": 443
+    },
+    {
+      "epoch": 10.325581395348838,
+      "grad_norm": 4.113884449005127,
+      "learning_rate": 6.0610465116279066e-05,
+      "loss": 0.0624,
+      "step": 444
+    },
+    {
+      "epoch": 10.348837209302326,
+      "grad_norm": 4.014869689941406,
+      "learning_rate": 6.0465116279069765e-05,
+      "loss": 0.075,
+      "step": 445
+    },
+    {
+      "epoch": 10.372093023255815,
+      "grad_norm": 0.5572329759597778,
+      "learning_rate": 6.0319767441860464e-05,
+      "loss": 0.0172,
+      "step": 446
+    },
+    {
+      "epoch": 10.395348837209303,
+      "grad_norm": 2.2173776626586914,
+      "learning_rate": 6.017441860465116e-05,
+      "loss": 0.0753,
+      "step": 447
+    },
+    {
+      "epoch": 10.418604651162791,
+      "grad_norm": 4.567299842834473,
+      "learning_rate": 6.002906976744186e-05,
+      "loss": 0.1678,
+      "step": 448
+    },
+    {
+      "epoch": 10.44186046511628,
+      "grad_norm": 2.608336925506592,
+      "learning_rate": 5.9883720930232554e-05,
+      "loss": 0.1871,
+      "step": 449
+    },
+    {
+      "epoch": 10.465116279069768,
+      "grad_norm": 2.773268699645996,
+      "learning_rate": 5.973837209302325e-05,
+      "loss": 0.1034,
+      "step": 450
+    },
+    {
+      "epoch": 10.488372093023255,
+      "grad_norm": 4.1804304122924805,
+      "learning_rate": 5.959302325581395e-05,
+      "loss": 0.1966,
+      "step": 451
+    },
+    {
+      "epoch": 10.511627906976745,
+      "grad_norm": 2.119896650314331,
+      "learning_rate": 5.944767441860465e-05,
+      "loss": 0.0443,
+      "step": 452
+    },
+    {
+      "epoch": 10.534883720930232,
+      "grad_norm": 2.4072678089141846,
+      "learning_rate": 5.9302325581395356e-05,
+      "loss": 0.0509,
+      "step": 453
+    },
+    {
+      "epoch": 10.55813953488372,
+      "grad_norm": 2.580864906311035,
+      "learning_rate": 5.9156976744186055e-05,
+      "loss": 0.1404,
+      "step": 454
+    },
+    {
+      "epoch": 10.581395348837209,
+      "grad_norm": 1.998783826828003,
+      "learning_rate": 5.9011627906976754e-05,
+      "loss": 0.0546,
+      "step": 455
+    },
+    {
+      "epoch": 10.604651162790697,
+      "grad_norm": 1.9264705181121826,
+      "learning_rate": 5.886627906976745e-05,
+      "loss": 0.0753,
+      "step": 456
+    },
+    {
+      "epoch": 10.627906976744185,
+      "grad_norm": 1.8207919597625732,
+      "learning_rate": 5.8720930232558145e-05,
+      "loss": 0.0783,
+      "step": 457
+    },
+    {
+      "epoch": 10.651162790697674,
+      "grad_norm": 3.059663772583008,
+      "learning_rate": 5.8575581395348844e-05,
+      "loss": 0.0829,
+      "step": 458
+    },
+    {
+      "epoch": 10.674418604651162,
+      "grad_norm": 2.6100552082061768,
+      "learning_rate": 5.843023255813954e-05,
+      "loss": 0.1253,
+      "step": 459
+    },
+    {
+      "epoch": 10.69767441860465,
+      "grad_norm": 2.3080873489379883,
+      "learning_rate": 5.828488372093024e-05,
+      "loss": 0.0337,
+      "step": 460
+    },
+    {
+      "epoch": 10.69767441860465,
+      "eval_accuracy": 0.8918032786885246,
+      "eval_f1": 0.8916355336398186,
+      "eval_loss": 0.2402210682630539,
+      "eval_precision": 0.8915439593558943,
+      "eval_recall": 0.8917463599551995,
+      "eval_runtime": 0.0746,
+      "eval_samples_per_second": 4086.997,
+      "eval_steps_per_second": 67.0,
+      "step": 460
+    },
+    {
+      "epoch": 10.720930232558139,
+      "grad_norm": 1.366176724433899,
+      "learning_rate": 5.8139534883720933e-05,
+      "loss": 0.0355,
+      "step": 461
+    },
+    {
+      "epoch": 10.744186046511627,
+      "grad_norm": 3.6105799674987793,
+      "learning_rate": 5.799418604651163e-05,
+      "loss": 0.1579,
+      "step": 462
+    },
+    {
+      "epoch": 10.767441860465116,
+      "grad_norm": 2.1533732414245605,
+      "learning_rate": 5.784883720930233e-05,
+      "loss": 0.0605,
+      "step": 463
+    },
+    {
+      "epoch": 10.790697674418604,
+      "grad_norm": 2.887242078781128,
+      "learning_rate": 5.770348837209303e-05,
+      "loss": 0.0787,
+      "step": 464
+    },
+    {
+      "epoch": 10.813953488372093,
+      "grad_norm": 4.048674583435059,
+      "learning_rate": 5.755813953488373e-05,
+      "loss": 0.0646,
+      "step": 465
+    },
+    {
+      "epoch": 10.837209302325581,
+      "grad_norm": 4.225156307220459,
+      "learning_rate": 5.741279069767442e-05,
+      "loss": 0.0449,
+      "step": 466
+    },
+    {
+      "epoch": 10.86046511627907,
+      "grad_norm": 5.0084123611450195,
+      "learning_rate": 5.726744186046512e-05,
+      "loss": 0.0798,
+      "step": 467
+    },
+    {
+      "epoch": 10.883720930232558,
+      "grad_norm": 2.2451436519622803,
+      "learning_rate": 5.712209302325582e-05,
+      "loss": 0.1,
+      "step": 468
+    },
+    {
+      "epoch": 10.906976744186046,
+      "grad_norm": 2.683537006378174,
+      "learning_rate": 5.697674418604652e-05,
+      "loss": 0.0875,
+      "step": 469
+    },
+    {
+      "epoch": 10.930232558139535,
+      "grad_norm": 2.50972580909729,
+      "learning_rate": 5.683139534883721e-05,
+      "loss": 0.0313,
+      "step": 470
+    },
+    {
+      "epoch": 10.953488372093023,
+      "grad_norm": 8.650345802307129,
+      "learning_rate": 5.668604651162791e-05,
+      "loss": 0.0685,
+      "step": 471
+    },
+    {
+      "epoch": 10.976744186046512,
+      "grad_norm": 2.9768123626708984,
+      "learning_rate": 5.654069767441861e-05,
+      "loss": 0.031,
+      "step": 472
+    },
+    {
+      "epoch": 11.0,
+      "grad_norm": 2.2081665992736816,
+      "learning_rate": 5.6395348837209306e-05,
+      "loss": 0.0335,
+      "step": 473
+    },
+    {
+      "epoch": 11.023255813953488,
+      "grad_norm": 4.979495048522949,
+      "learning_rate": 5.6250000000000005e-05,
+      "loss": 0.0331,
+      "step": 474
+    },
+    {
+      "epoch": 11.046511627906977,
+      "grad_norm": 2.000579833984375,
+      "learning_rate": 5.61046511627907e-05,
+      "loss": 0.0472,
+      "step": 475
+    },
+    {
+      "epoch": 11.069767441860465,
+      "grad_norm": 2.3896875381469727,
+      "learning_rate": 5.5959302325581396e-05,
+      "loss": 0.0745,
+      "step": 476
+    },
+    {
+      "epoch": 11.093023255813954,
+      "grad_norm": 2.0320870876312256,
+      "learning_rate": 5.5813953488372095e-05,
+      "loss": 0.0634,
+      "step": 477
+    },
+    {
+      "epoch": 11.116279069767442,
+      "grad_norm": 2.134622812271118,
+      "learning_rate": 5.5668604651162794e-05,
+      "loss": 0.0387,
+      "step": 478
+    },
+    {
+      "epoch": 11.13953488372093,
+      "grad_norm": 1.0899347066879272,
+      "learning_rate": 5.552325581395349e-05,
+      "loss": 0.0455,
+      "step": 479
+    },
+    {
+      "epoch": 11.162790697674419,
+      "grad_norm": 0.905604898929596,
+      "learning_rate": 5.5377906976744185e-05,
+      "loss": 0.0156,
+      "step": 480
+    },
+    {
+      "epoch": 11.162790697674419,
+      "eval_accuracy": 0.9016393442622951,
+      "eval_f1": 0.9008624425760596,
+      "eval_loss": 0.3055068254470825,
+      "eval_precision": 0.906113154668068,
+      "eval_recall": 0.8995003015421728,
+      "eval_runtime": 0.0798,
+      "eval_samples_per_second": 3820.2,
+      "eval_steps_per_second": 62.626,
+      "step": 480
+    },
+    {
+      "epoch": 11.186046511627907,
+      "grad_norm": 1.7722190618515015,
+      "learning_rate": 5.5232558139534884e-05,
+      "loss": 0.0672,
+      "step": 481
+    },
+    {
+      "epoch": 11.209302325581396,
+      "grad_norm": 3.3152947425842285,
+      "learning_rate": 5.508720930232558e-05,
+      "loss": 0.1093,
+      "step": 482
+    },
+    {
+      "epoch": 11.232558139534884,
+      "grad_norm": 3.6544511318206787,
+      "learning_rate": 5.494186046511628e-05,
+      "loss": 0.0894,
+      "step": 483
+    },
+    {
+      "epoch": 11.255813953488373,
+      "grad_norm": 4.428676128387451,
+      "learning_rate": 5.4796511627906974e-05,
+      "loss": 0.0495,
+      "step": 484
+    },
+    {
+      "epoch": 11.279069767441861,
+      "grad_norm": 3.5235519409179688,
+      "learning_rate": 5.465116279069767e-05,
+      "loss": 0.0313,
+      "step": 485
+    },
+    {
+      "epoch": 11.30232558139535,
+      "grad_norm": 1.8089061975479126,
+      "learning_rate": 5.450581395348837e-05,
+      "loss": 0.0294,
+      "step": 486
+    },
+    {
+      "epoch": 11.325581395348838,
+      "grad_norm": 9.173127174377441,
+      "learning_rate": 5.436046511627907e-05,
+      "loss": 0.0559,
+      "step": 487
+    },
+    {
+      "epoch": 11.348837209302326,
+      "grad_norm": 10.084165573120117,
+      "learning_rate": 5.421511627906977e-05,
+      "loss": 0.0759,
+      "step": 488
+    },
+    {
+      "epoch": 11.372093023255815,
+      "grad_norm": 0.5530898571014404,
+      "learning_rate": 5.406976744186046e-05,
+      "loss": 0.0202,
+      "step": 489
+    },
+    {
+      "epoch": 11.395348837209303,
+      "grad_norm": 4.215567588806152,
+      "learning_rate": 5.392441860465116e-05,
+      "loss": 0.1209,
+      "step": 490
+    },
+    {
+      "epoch": 11.418604651162791,
+      "grad_norm": 0.7583315372467041,
+      "learning_rate": 5.377906976744186e-05,
+      "loss": 0.0054,
+      "step": 491
+    },
+    {
+      "epoch": 11.44186046511628,
+      "grad_norm": 0.2435784786939621,
+      "learning_rate": 5.363372093023256e-05,
+      "loss": 0.0064,
+      "step": 492
+    },
+    {
+      "epoch": 11.465116279069768,
+      "grad_norm": 2.534301519393921,
+      "learning_rate": 5.348837209302326e-05,
+      "loss": 0.0733,
+      "step": 493
+    },
+    {
+      "epoch": 11.488372093023255,
+      "grad_norm": 3.94118332862854,
+      "learning_rate": 5.334302325581395e-05,
+      "loss": 0.0812,
+      "step": 494
+    },
+    {
+      "epoch": 11.511627906976745,
+      "grad_norm": 3.799800157546997,
+      "learning_rate": 5.319767441860465e-05,
+      "loss": 0.1377,
+      "step": 495
+    },
+    {
+      "epoch": 11.534883720930232,
+      "grad_norm": 0.16743521392345428,
+      "learning_rate": 5.305232558139536e-05,
+      "loss": 0.0054,
+      "step": 496
+    },
+    {
+      "epoch": 11.55813953488372,
+      "grad_norm": 1.3862305879592896,
+      "learning_rate": 5.290697674418605e-05,
+      "loss": 0.0364,
+      "step": 497
+    },
+    {
+      "epoch": 11.581395348837209,
+      "grad_norm": 3.8730082511901855,
+      "learning_rate": 5.276162790697675e-05,
+      "loss": 0.0827,
+      "step": 498
+    },
+    {
+      "epoch": 11.604651162790697,
+      "grad_norm": 5.6961188316345215,
+      "learning_rate": 5.261627906976745e-05,
+      "loss": 0.0936,
+      "step": 499
+    },
+    {
+      "epoch": 11.627906976744185,
+      "grad_norm": 0.9591822028160095,
+      "learning_rate": 5.247093023255815e-05,
+      "loss": 0.0146,
+      "step": 500
+    },
+    {
+      "epoch": 11.627906976744185,
+      "eval_accuracy": 0.9016393442622951,
+      "eval_f1": 0.900974025974026,
+      "eval_loss": 0.32393914461135864,
+      "eval_precision": 0.9050144016758314,
+      "eval_recall": 0.899780304988369,
+      "eval_runtime": 0.075,
+      "eval_samples_per_second": 4066.651,
+      "eval_steps_per_second": 66.666,
+      "step": 500
+    },
+    {
+      "epoch": 11.651162790697674,
+      "grad_norm": 5.92799186706543,
+      "learning_rate": 5.232558139534884e-05,
+      "loss": 0.0904,
+      "step": 501
+    },
+    {
+      "epoch": 11.674418604651162,
+      "grad_norm": 9.12846851348877,
+      "learning_rate": 5.218023255813954e-05,
+      "loss": 0.322,
+      "step": 502
+    },
+    {
+      "epoch": 11.69767441860465,
+      "grad_norm": 2.95119571685791,
+      "learning_rate": 5.203488372093024e-05,
+      "loss": 0.0175,
+      "step": 503
+    },
+    {
+      "epoch": 11.720930232558139,
+      "grad_norm": 3.9121649265289307,
+      "learning_rate": 5.188953488372094e-05,
+      "loss": 0.0733,
+      "step": 504
+    },
+    {
+      "epoch": 11.744186046511627,
+      "grad_norm": 2.7404439449310303,
+      "learning_rate": 5.1744186046511636e-05,
+      "loss": 0.0178,
+      "step": 505
+    },
+    {
+      "epoch": 11.767441860465116,
+      "grad_norm": 4.207398891448975,
+      "learning_rate": 5.159883720930233e-05,
+      "loss": 0.1179,
+      "step": 506
+    },
+    {
+      "epoch": 11.790697674418604,
+      "grad_norm": 5.63287878036499,
+      "learning_rate": 5.145348837209303e-05,
+      "loss": 0.0728,
+      "step": 507
+    },
+    {
+      "epoch": 11.813953488372093,
+      "grad_norm": 4.2650556564331055,
+      "learning_rate": 5.1308139534883726e-05,
+      "loss": 0.0409,
+      "step": 508
+    },
+    {
+      "epoch": 11.837209302325581,
+      "grad_norm": 4.405393123626709,
+      "learning_rate": 5.1162790697674425e-05,
+      "loss": 0.1272,
+      "step": 509
+    },
+    {
+      "epoch": 11.86046511627907,
+      "grad_norm": 3.2369868755340576,
+      "learning_rate": 5.1017441860465124e-05,
+      "loss": 0.1354,
+      "step": 510
+    },
+    {
+      "epoch": 11.883720930232558,
+      "grad_norm": 2.492950439453125,
+      "learning_rate": 5.0872093023255816e-05,
+      "loss": 0.116,
+      "step": 511
+    },
+    {
+      "epoch": 11.906976744186046,
+      "grad_norm": 5.160562992095947,
+      "learning_rate": 5.0726744186046515e-05,
+      "loss": 0.0813,
+      "step": 512
+    },
+    {
+      "epoch": 11.930232558139535,
+      "grad_norm": 3.234938383102417,
+      "learning_rate": 5.0581395348837214e-05,
+      "loss": 0.041,
+      "step": 513
+    },
+    {
+      "epoch": 11.953488372093023,
+      "grad_norm": 2.3901174068450928,
+      "learning_rate": 5.043604651162791e-05,
+      "loss": 0.0285,
+      "step": 514
+    },
+    {
+      "epoch": 11.976744186046512,
+      "grad_norm": 1.9587090015411377,
+      "learning_rate": 5.0290697674418605e-05,
+      "loss": 0.0191,
+      "step": 515
+    },
+    {
+      "epoch": 12.0,
+      "grad_norm": 1.5882123708724976,
+      "learning_rate": 5.0145348837209304e-05,
+      "loss": 0.0292,
+      "step": 516
+    },
+    {
+      "epoch": 12.023255813953488,
+      "grad_norm": 1.0756316184997559,
+      "learning_rate": 5e-05,
+      "loss": 0.0348,
+      "step": 517
+    },
+    {
+      "epoch": 12.046511627906977,
+      "grad_norm": 7.145483016967773,
+      "learning_rate": 4.98546511627907e-05,
+      "loss": 0.1204,
+      "step": 518
+    },
+    {
+      "epoch": 12.069767441860465,
+      "grad_norm": 10.459267616271973,
+      "learning_rate": 4.97093023255814e-05,
+      "loss": 0.0682,
+      "step": 519
+    },
+    {
+      "epoch": 12.093023255813954,
+      "grad_norm": 2.6111533641815186,
+      "learning_rate": 4.956395348837209e-05,
+      "loss": 0.1088,
+      "step": 520
+    },
+    {
+      "epoch": 12.093023255813954,
+      "eval_accuracy": 0.9081967213114754,
+      "eval_f1": 0.907471613070989,
+      "eval_loss": 0.29710283875465393,
+      "eval_precision": 0.9127912068663514,
+      "eval_recall": 0.9060696131644697,
+      "eval_runtime": 0.0726,
+      "eval_samples_per_second": 4200.336,
+      "eval_steps_per_second": 68.858,
+      "step": 520
+    },
+    {
+      "epoch": 12.116279069767442,
+      "grad_norm": 1.2339119911193848,
+      "learning_rate": 4.941860465116279e-05,
+      "loss": 0.0222,
+      "step": 521
+    },
+    {
+      "epoch": 12.13953488372093,
+      "grad_norm": 1.6148200035095215,
+      "learning_rate": 4.927325581395349e-05,
+      "loss": 0.0968,
+      "step": 522
+    },
+    {
+      "epoch": 12.162790697674419,
+      "grad_norm": 2.2320408821105957,
+      "learning_rate": 4.912790697674419e-05,
+      "loss": 0.0742,
+      "step": 523
+    },
+    {
+      "epoch": 12.186046511627907,
+      "grad_norm": 1.0352402925491333,
+      "learning_rate": 4.898255813953488e-05,
+      "loss": 0.0287,
+      "step": 524
+    },
+    {
+      "epoch": 12.209302325581396,
+      "grad_norm": 3.2948336601257324,
+      "learning_rate": 4.883720930232558e-05,
+      "loss": 0.0945,
+      "step": 525
+    },
+    {
+      "epoch": 12.232558139534884,
+      "grad_norm": 0.48531392216682434,
+      "learning_rate": 4.869186046511628e-05,
+      "loss": 0.0119,
+      "step": 526
+    },
+    {
+      "epoch": 12.255813953488373,
+      "grad_norm": 3.030161142349243,
+      "learning_rate": 4.854651162790698e-05,
+      "loss": 0.0713,
+      "step": 527
+    },
+    {
+      "epoch": 12.279069767441861,
+      "grad_norm": 1.6946715116500854,
+      "learning_rate": 4.8401162790697676e-05,
+      "loss": 0.0322,
+      "step": 528
+    },
+    {
+      "epoch": 12.30232558139535,
+      "grad_norm": 4.674280643463135,
+      "learning_rate": 4.8255813953488375e-05,
+      "loss": 0.048,
+      "step": 529
+    },
+    {
+      "epoch": 12.325581395348838,
+      "grad_norm": 1.2208576202392578,
+      "learning_rate": 4.8110465116279074e-05,
+      "loss": 0.009,
+      "step": 530
+    },
+    {
+      "epoch": 12.348837209302326,
+      "grad_norm": 7.181972026824951,
+      "learning_rate": 4.796511627906977e-05,
+      "loss": 0.0275,
+      "step": 531
+    },
+    {
+      "epoch": 12.372093023255815,
+      "grad_norm": 1.5175693035125732,
+      "learning_rate": 4.781976744186047e-05,
+      "loss": 0.0246,
+      "step": 532
+    },
+    {
+      "epoch": 12.395348837209303,
+      "grad_norm": 5.098257064819336,
+      "learning_rate": 4.7674418604651164e-05,
+      "loss": 0.0758,
+      "step": 533
+    },
+    {
+      "epoch": 12.418604651162791,
+      "grad_norm": 2.7552084922790527,
+      "learning_rate": 4.752906976744186e-05,
+      "loss": 0.0558,
+      "step": 534
+    },
+    {
+      "epoch": 12.44186046511628,
+      "grad_norm": 2.456735134124756,
+      "learning_rate": 4.738372093023256e-05,
+      "loss": 0.0484,
+      "step": 535
+    },
+    {
+      "epoch": 12.465116279069768,
+      "grad_norm": 1.6237318515777588,
+      "learning_rate": 4.723837209302326e-05,
+      "loss": 0.08,
+      "step": 536
+    },
+    {
+      "epoch": 12.488372093023255,
+      "grad_norm": 3.8842220306396484,
+      "learning_rate": 4.709302325581396e-05,
+      "loss": 0.0858,
+      "step": 537
+    },
+    {
+      "epoch": 12.511627906976745,
+      "grad_norm": 3.0373013019561768,
+      "learning_rate": 4.694767441860465e-05,
+      "loss": 0.0256,
+      "step": 538
+    },
+    {
+      "epoch": 12.534883720930232,
+      "grad_norm": 5.356001853942871,
+      "learning_rate": 4.680232558139535e-05,
+      "loss": 0.1057,
+      "step": 539
+    },
+    {
+      "epoch": 12.55813953488372,
+      "grad_norm": 0.7480242252349854,
+      "learning_rate": 4.665697674418605e-05,
+      "loss": 0.0102,
+      "step": 540
+    },
+    {
+      "epoch": 12.55813953488372,
+      "eval_accuracy": 0.898360655737705,
+      "eval_f1": 0.8983213068212369,
+      "eval_loss": 0.27956801652908325,
+      "eval_precision": 0.8984348125214998,
+      "eval_recall": 0.8991556819160851,
+      "eval_runtime": 0.0733,
+      "eval_samples_per_second": 4158.554,
+      "eval_steps_per_second": 68.173,
+      "step": 540
+    },
+    {
+      "epoch": 12.581395348837209,
+      "grad_norm": 8.634737014770508,
+      "learning_rate": 4.651162790697675e-05,
+      "loss": 0.1041,
+      "step": 541
+    },
+    {
+      "epoch": 12.604651162790697,
+      "grad_norm": 2.154197931289673,
+      "learning_rate": 4.636627906976744e-05,
+      "loss": 0.0883,
+      "step": 542
+    },
+    {
+      "epoch": 12.627906976744185,
+      "grad_norm": 5.6526265144348145,
+      "learning_rate": 4.622093023255814e-05,
+      "loss": 0.0634,
+      "step": 543
+    },
+    {
+      "epoch": 12.651162790697674,
+      "grad_norm": 0.4163094162940979,
+      "learning_rate": 4.607558139534884e-05,
+      "loss": 0.0102,
+      "step": 544
+    },
+    {
+      "epoch": 12.674418604651162,
+      "grad_norm": 2.6310293674468994,
+      "learning_rate": 4.593023255813954e-05,
+      "loss": 0.0613,
+      "step": 545
+    },
+    {
+      "epoch": 12.69767441860465,
+      "grad_norm": 0.8978974223136902,
+      "learning_rate": 4.5784883720930236e-05,
+      "loss": 0.0118,
+      "step": 546
+    },
+    {
+      "epoch": 12.720930232558139,
+      "grad_norm": 3.0897459983825684,
+      "learning_rate": 4.563953488372093e-05,
+      "loss": 0.0592,
+      "step": 547
+    },
+    {
+      "epoch": 12.744186046511627,
+      "grad_norm": 1.5553648471832275,
+      "learning_rate": 4.549418604651163e-05,
+      "loss": 0.0357,
+      "step": 548
+    },
+    {
+      "epoch": 12.767441860465116,
+      "grad_norm": 1.9957389831542969,
+      "learning_rate": 4.5348837209302326e-05,
+      "loss": 0.0415,
+      "step": 549
+    },
+    {
+      "epoch": 12.790697674418604,
+      "grad_norm": 0.22972577810287476,
+      "learning_rate": 4.520348837209303e-05,
+      "loss": 0.0078,
+      "step": 550
+    },
+    {
+      "epoch": 12.813953488372093,
+      "grad_norm": 2.78206205368042,
+      "learning_rate": 4.505813953488372e-05,
+      "loss": 0.026,
+      "step": 551
+    },
+    {
+      "epoch": 12.837209302325581,
+      "grad_norm": 3.2643375396728516,
+      "learning_rate": 4.491279069767442e-05,
+      "loss": 0.0386,
+      "step": 552
+    },
+    {
+      "epoch": 12.86046511627907,
+      "grad_norm": 9.664105415344238,
+      "learning_rate": 4.476744186046512e-05,
+      "loss": 0.1431,
+      "step": 553
+    },
+    {
+      "epoch": 12.883720930232558,
+      "grad_norm": 3.2015540599823,
+      "learning_rate": 4.462209302325582e-05,
+      "loss": 0.0605,
+      "step": 554
+    },
+    {
+      "epoch": 12.906976744186046,
+      "grad_norm": 4.17910099029541,
+      "learning_rate": 4.447674418604651e-05,
+      "loss": 0.0797,
+      "step": 555
+    },
+    {
+      "epoch": 12.930232558139535,
+      "grad_norm": 3.999985933303833,
+      "learning_rate": 4.433139534883721e-05,
+      "loss": 0.117,
+      "step": 556
+    },
+    {
+      "epoch": 12.953488372093023,
+      "grad_norm": 1.1732484102249146,
+      "learning_rate": 4.418604651162791e-05,
+      "loss": 0.0244,
+      "step": 557
+    },
+    {
+      "epoch": 12.976744186046512,
+      "grad_norm": 4.184539794921875,
+      "learning_rate": 4.404069767441861e-05,
+      "loss": 0.0517,
+      "step": 558
+    },
+    {
+      "epoch": 13.0,
+      "grad_norm": 2.3533170223236084,
+      "learning_rate": 4.389534883720931e-05,
+      "loss": 0.0505,
+      "step": 559
+    },
+    {
+      "epoch": 13.023255813953488,
+      "grad_norm": 2.3821420669555664,
+      "learning_rate": 4.375e-05,
+      "loss": 0.0253,
+      "step": 560
+    },
+    {
+      "epoch": 13.023255813953488,
+      "eval_accuracy": 0.8885245901639345,
+      "eval_f1": 0.8879936058066189,
+      "eval_loss": 0.3079434633255005,
+      "eval_precision": 0.8900242992276317,
+      "eval_recall": 0.8872016886361678,
+      "eval_runtime": 0.0743,
+      "eval_samples_per_second": 4102.318,
+      "eval_steps_per_second": 67.251,
+      "step": 560
+    },
+    {
+      "epoch": 13.046511627906977,
+      "grad_norm": 3.6151607036590576,
+      "learning_rate": 4.36046511627907e-05,
+      "loss": 0.08,
+      "step": 561
+    },
+    {
+      "epoch": 13.069767441860465,
+      "grad_norm": 1.870180368423462,
+      "learning_rate": 4.34593023255814e-05,
+      "loss": 0.0475,
+      "step": 562
+    },
+    {
+      "epoch": 13.093023255813954,
+      "grad_norm": 3.9536831378936768,
+      "learning_rate": 4.3313953488372096e-05,
+      "loss": 0.0206,
+      "step": 563
+    },
+    {
+      "epoch": 13.116279069767442,
+      "grad_norm": 6.368700981140137,
+      "learning_rate": 4.3168604651162795e-05,
+      "loss": 0.0485,
+      "step": 564
+    },
+    {
+      "epoch": 13.13953488372093,
+      "grad_norm": 1.1507411003112793,
+      "learning_rate": 4.302325581395349e-05,
+      "loss": 0.0133,
+      "step": 565
+    },
+    {
+      "epoch": 13.162790697674419,
+      "grad_norm": 1.034654140472412,
+      "learning_rate": 4.2877906976744186e-05,
+      "loss": 0.0437,
+      "step": 566
+    },
+    {
+      "epoch": 13.186046511627907,
+      "grad_norm": 0.2083161622285843,
+      "learning_rate": 4.2732558139534885e-05,
+      "loss": 0.0054,
+      "step": 567
+    },
+    {
+      "epoch": 13.209302325581396,
+      "grad_norm": 1.1866053342819214,
+      "learning_rate": 4.2587209302325584e-05,
+      "loss": 0.0134,
+      "step": 568
+    },
+    {
+      "epoch": 13.232558139534884,
+      "grad_norm": 0.6963038444519043,
+      "learning_rate": 4.2441860465116276e-05,
+      "loss": 0.0083,
+      "step": 569
+    },
+    {
+      "epoch": 13.255813953488373,
+      "grad_norm": 3.132704973220825,
+      "learning_rate": 4.2296511627906975e-05,
+      "loss": 0.0567,
+      "step": 570
+    },
+    {
+      "epoch": 13.279069767441861,
+      "grad_norm": 1.62773859500885,
+      "learning_rate": 4.215116279069768e-05,
+      "loss": 0.0803,
+      "step": 571
+    },
+    {
+      "epoch": 13.30232558139535,
+      "grad_norm": 0.7541981935501099,
+      "learning_rate": 4.200581395348838e-05,
+      "loss": 0.0114,
+      "step": 572
+    },
+    {
+      "epoch": 13.325581395348838,
+      "grad_norm": 5.177672386169434,
+      "learning_rate": 4.186046511627907e-05,
+      "loss": 0.0302,
+      "step": 573
+    },
+    {
+      "epoch": 13.348837209302326,
+      "grad_norm": 3.653960704803467,
+      "learning_rate": 4.171511627906977e-05,
+      "loss": 0.0648,
+      "step": 574
+    },
+    {
+      "epoch": 13.372093023255815,
+      "grad_norm": 0.6348634362220764,
+      "learning_rate": 4.156976744186047e-05,
+      "loss": 0.0077,
+      "step": 575
+    },
+    {
+      "epoch": 13.395348837209303,
+      "grad_norm": 4.980737686157227,
+      "learning_rate": 4.142441860465117e-05,
+      "loss": 0.0857,
+      "step": 576
+    },
+    {
+      "epoch": 13.418604651162791,
+      "grad_norm": 4.5483551025390625,
+      "learning_rate": 4.127906976744187e-05,
+      "loss": 0.0456,
+      "step": 577
+    },
+    {
+      "epoch": 13.44186046511628,
+      "grad_norm": 4.054844379425049,
+      "learning_rate": 4.113372093023256e-05,
+      "loss": 0.0837,
+      "step": 578
+    },
+    {
+      "epoch": 13.465116279069768,
+      "grad_norm": 3.3397719860076904,
+      "learning_rate": 4.098837209302326e-05,
+      "loss": 0.0742,
+      "step": 579
+    },
+    {
+      "epoch": 13.488372093023255,
+      "grad_norm": 6.787147045135498,
+      "learning_rate": 4.0843023255813957e-05,
+      "loss": 0.0868,
+      "step": 580
+    },
+    {
+      "epoch": 13.488372093023255,
+      "eval_accuracy": 0.8918032786885246,
+      "eval_f1": 0.8914251194701244,
+      "eval_loss": 0.34987983107566833,
+      "eval_precision": 0.8924277806607853,
+      "eval_recall": 0.8909063496166107,
+      "eval_runtime": 0.0747,
+      "eval_samples_per_second": 4085.209,
+      "eval_steps_per_second": 66.971,
+      "step": 580
+    },
+    {
+      "epoch": 13.511627906976745,
+      "grad_norm": 1.7752641439437866,
+      "learning_rate": 4.0697674418604655e-05,
+      "loss": 0.0476,
+      "step": 581
+    },
+    {
+      "epoch": 13.534883720930232,
+      "grad_norm": 2.80531907081604,
+      "learning_rate": 4.055232558139535e-05,
+      "loss": 0.0202,
+      "step": 582
+    },
+    {
+      "epoch": 13.55813953488372,
+      "grad_norm": 0.5729110836982727,
+      "learning_rate": 4.0406976744186046e-05,
+      "loss": 0.0078,
+      "step": 583
+    },
+    {
+      "epoch": 13.581395348837209,
+      "grad_norm": 5.221317768096924,
+      "learning_rate": 4.0261627906976745e-05,
+      "loss": 0.092,
+      "step": 584
+    },
+    {
+      "epoch": 13.604651162790697,
+      "grad_norm": 3.757556915283203,
+      "learning_rate": 4.0116279069767444e-05,
+      "loss": 0.024,
+      "step": 585
+    },
+    {
+      "epoch": 13.627906976744185,
+      "grad_norm": 6.441154479980469,
+      "learning_rate": 3.997093023255814e-05,
+      "loss": 0.1422,
+      "step": 586
+    },
+    {
+      "epoch": 13.651162790697674,
+      "grad_norm": 2.314418077468872,
+      "learning_rate": 3.9825581395348835e-05,
+      "loss": 0.0815,
+      "step": 587
+    },
+    {
+      "epoch": 13.674418604651162,
+      "grad_norm": 2.129340887069702,
+      "learning_rate": 3.9680232558139534e-05,
+      "loss": 0.0419,
+      "step": 588
+    },
+    {
+      "epoch": 13.69767441860465,
+      "grad_norm": 1.477169394493103,
+      "learning_rate": 3.953488372093023e-05,
+      "loss": 0.0379,
+      "step": 589
+    },
+    {
+      "epoch": 13.720930232558139,
+      "grad_norm": 2.567399740219116,
+      "learning_rate": 3.938953488372093e-05,
+      "loss": 0.1169,
+      "step": 590
+    },
+    {
+      "epoch": 13.744186046511627,
+      "grad_norm": 1.2707808017730713,
+      "learning_rate": 3.924418604651163e-05,
+      "loss": 0.0225,
+      "step": 591
+    },
+    {
+      "epoch": 13.767441860465116,
+      "grad_norm": 0.3642787039279938,
+      "learning_rate": 3.909883720930232e-05,
+      "loss": 0.0053,
+      "step": 592
+    },
+    {
+      "epoch": 13.790697674418604,
+      "grad_norm": 2.2786664962768555,
+      "learning_rate": 3.895348837209303e-05,
+      "loss": 0.043,
+      "step": 593
+    },
+    {
+      "epoch": 13.813953488372093,
+      "grad_norm": 6.119050979614258,
+      "learning_rate": 3.880813953488373e-05,
+      "loss": 0.1129,
+      "step": 594
+    },
+    {
+      "epoch": 13.837209302325581,
+      "grad_norm": 3.127246141433716,
+      "learning_rate": 3.866279069767442e-05,
+      "loss": 0.1094,
+      "step": 595
+    },
+    {
+      "epoch": 13.86046511627907,
+      "grad_norm": 4.50978946685791,
+      "learning_rate": 3.851744186046512e-05,
+      "loss": 0.0714,
+      "step": 596
+    },
+    {
+      "epoch": 13.883720930232558,
+      "grad_norm": 2.3526670932769775,
+      "learning_rate": 3.837209302325582e-05,
+      "loss": 0.1126,
+      "step": 597
+    },
+    {
+      "epoch": 13.906976744186046,
+      "grad_norm": 4.066991806030273,
+      "learning_rate": 3.8226744186046516e-05,
+      "loss": 0.0466,
+      "step": 598
+    },
+    {
+      "epoch": 13.930232558139535,
+      "grad_norm": 1.5120619535446167,
+      "learning_rate": 3.8081395348837215e-05,
+      "loss": 0.0609,
+      "step": 599
+    },
+    {
+      "epoch": 13.953488372093023,
+      "grad_norm": 3.2349016666412354,
+      "learning_rate": 3.793604651162791e-05,
+      "loss": 0.0399,
+      "step": 600
+    },
+    {
+      "epoch": 13.953488372093023,
+      "eval_accuracy": 0.8885245901639345,
+      "eval_f1": 0.8880903047569713,
+      "eval_loss": 0.36928993463516235,
+      "eval_precision": 0.8893939393939394,
+      "eval_recall": 0.887481692082364,
+      "eval_runtime": 0.0753,
+      "eval_samples_per_second": 4048.646,
+      "eval_steps_per_second": 66.371,
+      "step": 600
+    },
+    {
+      "epoch": 13.976744186046512,
+      "grad_norm": 2.7635951042175293,
+      "learning_rate": 3.7790697674418606e-05,
+      "loss": 0.1149,
+      "step": 601
+    },
+    {
+      "epoch": 14.0,
+      "grad_norm": 0.716441810131073,
+      "learning_rate": 3.7645348837209305e-05,
+      "loss": 0.0095,
+      "step": 602
+    },
+    {
+      "epoch": 14.023255813953488,
+      "grad_norm": 1.099993348121643,
+      "learning_rate": 3.7500000000000003e-05,
+      "loss": 0.0179,
+      "step": 603
+    },
+    {
+      "epoch": 14.046511627906977,
+      "grad_norm": 0.9968247413635254,
+      "learning_rate": 3.73546511627907e-05,
+      "loss": 0.0295,
+      "step": 604
+    },
+    {
+      "epoch": 14.069767441860465,
+      "grad_norm": 0.9443352222442627,
+      "learning_rate": 3.7209302325581394e-05,
+      "loss": 0.0121,
+      "step": 605
+    },
+    {
+      "epoch": 14.093023255813954,
+      "grad_norm": 0.24217644333839417,
+      "learning_rate": 3.706395348837209e-05,
+      "loss": 0.0066,
+      "step": 606
+    },
+    {
+      "epoch": 14.116279069767442,
+      "grad_norm": 0.32041826844215393,
+      "learning_rate": 3.691860465116279e-05,
+      "loss": 0.0071,
+      "step": 607
+    },
+    {
+      "epoch": 14.13953488372093,
+      "grad_norm": 4.715015411376953,
+      "learning_rate": 3.677325581395349e-05,
+      "loss": 0.0775,
+      "step": 608
+    },
+    {
+      "epoch": 14.162790697674419,
+      "grad_norm": 1.9597333669662476,
+      "learning_rate": 3.662790697674418e-05,
+      "loss": 0.0232,
+      "step": 609
+    },
+    {
+      "epoch": 14.186046511627907,
+      "grad_norm": 1.3680325746536255,
+      "learning_rate": 3.648255813953488e-05,
+      "loss": 0.0194,
+      "step": 610
+    },
+    {
+      "epoch": 14.209302325581396,
+      "grad_norm": 7.5977396965026855,
+      "learning_rate": 3.633720930232558e-05,
+      "loss": 0.1274,
+      "step": 611
+    },
+    {
+      "epoch": 14.232558139534884,
+      "grad_norm": 7.637692451477051,
+      "learning_rate": 3.619186046511628e-05,
+      "loss": 0.1341,
+      "step": 612
+    },
+    {
+      "epoch": 14.255813953488373,
+      "grad_norm": 10.332743644714355,
+      "learning_rate": 3.604651162790698e-05,
+      "loss": 0.1039,
+      "step": 613
+    },
+    {
+      "epoch": 14.279069767441861,
+      "grad_norm": 1.6394683122634888,
+      "learning_rate": 3.590116279069768e-05,
+      "loss": 0.0126,
+      "step": 614
+    },
+    {
+      "epoch": 14.30232558139535,
+      "grad_norm": 1.69204580783844,
+      "learning_rate": 3.5755813953488376e-05,
+      "loss": 0.0327,
+      "step": 615
+    },
+    {
+      "epoch": 14.325581395348838,
+      "grad_norm": 3.1536524295806885,
+      "learning_rate": 3.5610465116279075e-05,
+      "loss": 0.0347,
+      "step": 616
+    },
+    {
+      "epoch": 14.348837209302326,
+      "grad_norm": 0.6318650841712952,
+      "learning_rate": 3.5465116279069774e-05,
+      "loss": 0.0424,
+      "step": 617
+    },
+    {
+      "epoch": 14.372093023255815,
+      "grad_norm": 5.714948654174805,
+      "learning_rate": 3.5319767441860466e-05,
+      "loss": 0.0339,
+      "step": 618
+    },
+    {
+      "epoch": 14.395348837209303,
+      "grad_norm": 4.141314506530762,
+      "learning_rate": 3.5174418604651165e-05,
+      "loss": 0.045,
+      "step": 619
+    },
+    {
+      "epoch": 14.418604651162791,
+      "grad_norm": 1.1429768800735474,
+      "learning_rate": 3.5029069767441864e-05,
+      "loss": 0.0291,
+      "step": 620
+    },
+    {
+      "epoch": 14.418604651162791,
+      "eval_accuracy": 0.8950819672131147,
+      "eval_f1": 0.8944819235426397,
+      "eval_loss": 0.3720164895057678,
+      "eval_precision": 0.8974286460146188,
+      "eval_recall": 0.8934909968122684,
+      "eval_runtime": 0.0902,
+      "eval_samples_per_second": 3381.6,
+      "eval_steps_per_second": 55.436,
+      "step": 620
+    },
+    {
+      "epoch": 14.44186046511628,
+      "grad_norm": 11.266057014465332,
+      "learning_rate": 3.488372093023256e-05,
+      "loss": 0.0526,
+      "step": 621
+    },
+    {
+      "epoch": 14.465116279069768,
+      "grad_norm": 0.8715966939926147,
+      "learning_rate": 3.4738372093023255e-05,
+      "loss": 0.0256,
+      "step": 622
+    },
+    {
+      "epoch": 14.488372093023255,
+      "grad_norm": 3.1204757690429688,
+      "learning_rate": 3.4593023255813954e-05,
+      "loss": 0.0256,
+      "step": 623
+    },
+    {
+      "epoch": 14.511627906976745,
+      "grad_norm": 2.438506841659546,
+      "learning_rate": 3.444767441860465e-05,
+      "loss": 0.0321,
+      "step": 624
+    },
+    {
+      "epoch": 14.534883720930232,
+      "grad_norm": 3.6766304969787598,
+      "learning_rate": 3.430232558139535e-05,
+      "loss": 0.1027,
+      "step": 625
+    },
+    {
+      "epoch": 14.55813953488372,
+      "grad_norm": 6.79380464553833,
+      "learning_rate": 3.415697674418605e-05,
+      "loss": 0.0597,
+      "step": 626
+    },
+    {
+      "epoch": 14.581395348837209,
+      "grad_norm": 1.6378151178359985,
+      "learning_rate": 3.401162790697674e-05,
+      "loss": 0.0238,
+      "step": 627
+    },
+    {
+      "epoch": 14.604651162790697,
+      "grad_norm": 0.5451012253761292,
+      "learning_rate": 3.386627906976744e-05,
+      "loss": 0.0103,
+      "step": 628
+    },
+    {
+      "epoch": 14.627906976744185,
+      "grad_norm": 2.0606513023376465,
+      "learning_rate": 3.372093023255814e-05,
+      "loss": 0.0223,
+      "step": 629
+    },
+    {
+      "epoch": 14.651162790697674,
+      "grad_norm": 0.4602338373661041,
+      "learning_rate": 3.357558139534884e-05,
+      "loss": 0.0057,
+      "step": 630
+    },
+    {
+      "epoch": 14.674418604651162,
+      "grad_norm": 4.629768371582031,
+      "learning_rate": 3.343023255813954e-05,
+      "loss": 0.1183,
+      "step": 631
+    },
+    {
+      "epoch": 14.69767441860465,
+      "grad_norm": 6.286869525909424,
+      "learning_rate": 3.328488372093023e-05,
+      "loss": 0.0308,
+      "step": 632
+    },
+    {
+      "epoch": 14.720930232558139,
+      "grad_norm": 1.118071436882019,
+      "learning_rate": 3.313953488372093e-05,
+      "loss": 0.0692,
+      "step": 633
+    },
+    {
+      "epoch": 14.744186046511627,
+      "grad_norm": 2.621478319168091,
+      "learning_rate": 3.299418604651163e-05,
+      "loss": 0.0206,
+      "step": 634
+    },
+    {
+      "epoch": 14.767441860465116,
+      "grad_norm": 2.408926486968994,
+      "learning_rate": 3.284883720930233e-05,
+      "loss": 0.0868,
+      "step": 635
+    },
+    {
+      "epoch": 14.790697674418604,
+      "grad_norm": 4.2226104736328125,
+      "learning_rate": 3.2703488372093026e-05,
+      "loss": 0.0165,
+      "step": 636
+    },
+    {
+      "epoch": 14.813953488372093,
+      "grad_norm": 8.624298095703125,
+      "learning_rate": 3.2558139534883724e-05,
+      "loss": 0.1733,
+      "step": 637
+    },
+    {
+      "epoch": 14.837209302325581,
+      "grad_norm": 5.992215633392334,
+      "learning_rate": 3.241279069767442e-05,
+      "loss": 0.0286,
+      "step": 638
+    },
+    {
+      "epoch": 14.86046511627907,
+      "grad_norm": 0.8743571043014526,
+      "learning_rate": 3.226744186046512e-05,
+      "loss": 0.0113,
+      "step": 639
+    },
+    {
+      "epoch": 14.883720930232558,
+      "grad_norm": 0.9226441979408264,
+      "learning_rate": 3.2122093023255814e-05,
+      "loss": 0.0069,
+      "step": 640
+    },
+    {
+      "epoch": 14.883720930232558,
+      "eval_accuracy": 0.8918032786885246,
+      "eval_f1": 0.8915047052355851,
+      "eval_loss": 0.3578951358795166,
+      "eval_precision": 0.8919968919968919,
+      "eval_recall": 0.8911863530628069,
+      "eval_runtime": 0.0756,
+      "eval_samples_per_second": 4035.886,
+      "eval_steps_per_second": 66.162,
+      "step": 640
+    },
+    {
+      "epoch": 14.906976744186046,
+      "grad_norm": 4.145755767822266,
+      "learning_rate": 3.197674418604651e-05,
+      "loss": 0.0439,
+      "step": 641
+    },
+    {
+      "epoch": 14.930232558139535,
+      "grad_norm": 1.2993311882019043,
+      "learning_rate": 3.183139534883721e-05,
+      "loss": 0.0446,
+      "step": 642
+    },
+    {
+      "epoch": 14.953488372093023,
+      "grad_norm": 0.2827296555042267,
+      "learning_rate": 3.168604651162791e-05,
+      "loss": 0.0052,
+      "step": 643
+    },
+    {
+      "epoch": 14.976744186046512,
+      "grad_norm": 2.9335291385650635,
+      "learning_rate": 3.154069767441861e-05,
+      "loss": 0.0644,
+      "step": 644
+    },
+    {
+      "epoch": 15.0,
+      "grad_norm": 7.6974968910217285,
+      "learning_rate": 3.13953488372093e-05,
+      "loss": 0.0456,
+      "step": 645
+    },
+    {
+      "epoch": 15.023255813953488,
+      "grad_norm": 2.137190103530884,
+      "learning_rate": 3.125e-05,
+      "loss": 0.0147,
+      "step": 646
+    },
+    {
+      "epoch": 15.046511627906977,
+      "grad_norm": 0.834999680519104,
+      "learning_rate": 3.11046511627907e-05,
+      "loss": 0.0181,
+      "step": 647
+    },
+    {
+      "epoch": 15.069767441860465,
+      "grad_norm": 0.8868411183357239,
+      "learning_rate": 3.09593023255814e-05,
+      "loss": 0.0467,
+      "step": 648
+    },
+    {
+      "epoch": 15.093023255813954,
+      "grad_norm": 3.1377875804901123,
+      "learning_rate": 3.081395348837209e-05,
+      "loss": 0.0281,
+      "step": 649
+    },
+    {
+      "epoch": 15.116279069767442,
+      "grad_norm": 3.3987367153167725,
+      "learning_rate": 3.066860465116279e-05,
+      "loss": 0.1133,
+      "step": 650
+    },
+    {
+      "epoch": 15.13953488372093,
+      "grad_norm": 0.19906505942344666,
+      "learning_rate": 3.052325581395349e-05,
+      "loss": 0.0032,
+      "step": 651
+    },
+    {
+      "epoch": 15.162790697674419,
+      "grad_norm": 4.011625289916992,
+      "learning_rate": 3.0377906976744187e-05,
+      "loss": 0.0792,
+      "step": 652
+    },
+    {
+      "epoch": 15.186046511627907,
+      "grad_norm": 9.421031951904297,
+      "learning_rate": 3.0232558139534883e-05,
+      "loss": 0.0612,
+      "step": 653
+    },
+    {
+      "epoch": 15.209302325581396,
+      "grad_norm": 3.5618064403533936,
+      "learning_rate": 3.008720930232558e-05,
+      "loss": 0.0159,
+      "step": 654
+    },
+    {
+      "epoch": 15.232558139534884,
+      "grad_norm": 7.569765090942383,
+      "learning_rate": 2.9941860465116277e-05,
+      "loss": 0.1039,
+      "step": 655
+    },
+    {
+      "epoch": 15.255813953488373,
+      "grad_norm": 6.453846454620361,
+      "learning_rate": 2.9796511627906976e-05,
+      "loss": 0.0283,
+      "step": 656
+    },
+    {
+      "epoch": 15.279069767441861,
+      "grad_norm": 0.1914948970079422,
+      "learning_rate": 2.9651162790697678e-05,
+      "loss": 0.0033,
+      "step": 657
+    },
+    {
+      "epoch": 15.30232558139535,
+      "grad_norm": 2.6470947265625,
+      "learning_rate": 2.9505813953488377e-05,
+      "loss": 0.0505,
+      "step": 658
+    },
+    {
+      "epoch": 15.325581395348838,
+      "grad_norm": 5.789623737335205,
+      "learning_rate": 2.9360465116279072e-05,
+      "loss": 0.0419,
+      "step": 659
+    },
+    {
+      "epoch": 15.348837209302326,
+      "grad_norm": 1.2438925504684448,
+      "learning_rate": 2.921511627906977e-05,
+      "loss": 0.027,
+      "step": 660
+    },
+    {
+      "epoch": 15.348837209302326,
+      "eval_accuracy": 0.898360655737705,
+      "eval_f1": 0.8980801776455497,
+      "eval_loss": 0.36847957968711853,
+      "eval_precision": 0.8985798152464819,
+      "eval_recall": 0.8977556646851038,
+      "eval_runtime": 0.0739,
+      "eval_samples_per_second": 4125.669,
+      "eval_steps_per_second": 67.634,
+      "step": 660
+    },
+    {
+      "epoch": 15.372093023255815,
+      "grad_norm": 2.8159120082855225,
+      "learning_rate": 2.9069767441860467e-05,
+      "loss": 0.0665,
+      "step": 661
+    },
+    {
+      "epoch": 15.395348837209303,
+      "grad_norm": 3.529680013656616,
+      "learning_rate": 2.8924418604651166e-05,
+      "loss": 0.0238,
+      "step": 662
+    },
+    {
+      "epoch": 15.418604651162791,
+      "grad_norm": 0.9973799586296082,
+      "learning_rate": 2.8779069767441864e-05,
+      "loss": 0.0051,
+      "step": 663
+    },
+    {
+      "epoch": 15.44186046511628,
+      "grad_norm": 2.763955593109131,
+      "learning_rate": 2.863372093023256e-05,
+      "loss": 0.0107,
+      "step": 664
+    },
+    {
+      "epoch": 15.465116279069768,
+      "grad_norm": 3.6373209953308105,
+      "learning_rate": 2.848837209302326e-05,
+      "loss": 0.0466,
+      "step": 665
+    },
+    {
+      "epoch": 15.488372093023255,
+      "grad_norm": 3.262023448944092,
+      "learning_rate": 2.8343023255813954e-05,
+      "loss": 0.025,
+      "step": 666
+    },
+    {
+      "epoch": 15.511627906976745,
+      "grad_norm": 0.7062492370605469,
+      "learning_rate": 2.8197674418604653e-05,
+      "loss": 0.0067,
+      "step": 667
+    },
+    {
+      "epoch": 15.534883720930232,
+      "grad_norm": 2.9254820346832275,
+      "learning_rate": 2.805232558139535e-05,
+      "loss": 0.0394,
+      "step": 668
+    },
+    {
+      "epoch": 15.55813953488372,
+      "grad_norm": 1.5290218591690063,
+      "learning_rate": 2.7906976744186048e-05,
+      "loss": 0.0268,
+      "step": 669
+    },
+    {
+      "epoch": 15.581395348837209,
+      "grad_norm": 3.9581668376922607,
+      "learning_rate": 2.7761627906976746e-05,
+      "loss": 0.0683,
+      "step": 670
+    },
+    {
+      "epoch": 15.604651162790697,
+      "grad_norm": 3.121964454650879,
+      "learning_rate": 2.7616279069767442e-05,
+      "loss": 0.0307,
+      "step": 671
+    },
+    {
+      "epoch": 15.627906976744185,
+      "grad_norm": 4.223913669586182,
+      "learning_rate": 2.747093023255814e-05,
+      "loss": 0.0778,
+      "step": 672
+    },
+    {
+      "epoch": 15.651162790697674,
+      "grad_norm": 1.8996158838272095,
+      "learning_rate": 2.7325581395348836e-05,
+      "loss": 0.0149,
+      "step": 673
+    },
+    {
+      "epoch": 15.674418604651162,
+      "grad_norm": 2.5367538928985596,
+      "learning_rate": 2.7180232558139535e-05,
+      "loss": 0.0238,
+      "step": 674
+    },
+    {
+      "epoch": 15.69767441860465,
+      "grad_norm": 1.5632944107055664,
+      "learning_rate": 2.703488372093023e-05,
+      "loss": 0.0159,
+      "step": 675
+    },
+    {
+      "epoch": 15.720930232558139,
+      "grad_norm": 2.031212329864502,
+      "learning_rate": 2.688953488372093e-05,
+      "loss": 0.0265,
+      "step": 676
+    },
+    {
+      "epoch": 15.744186046511627,
+      "grad_norm": 1.0653328895568848,
+      "learning_rate": 2.674418604651163e-05,
+      "loss": 0.0146,
+      "step": 677
+    },
+    {
+      "epoch": 15.767441860465116,
+      "grad_norm": 2.078573226928711,
+      "learning_rate": 2.6598837209302324e-05,
+      "loss": 0.0314,
+      "step": 678
+    },
+    {
+      "epoch": 15.790697674418604,
+      "grad_norm": 1.004683017730713,
+      "learning_rate": 2.6453488372093026e-05,
+      "loss": 0.0079,
+      "step": 679
+    },
+    {
+      "epoch": 15.813953488372093,
+      "grad_norm": 0.6423361897468567,
+      "learning_rate": 2.6308139534883725e-05,
+      "loss": 0.0265,
+      "step": 680
+    },
+    {
+      "epoch": 15.813953488372093,
+      "eval_accuracy": 0.898360655737705,
+      "eval_f1": 0.8981460534962133,
+      "eval_loss": 0.3592439889907837,
+      "eval_precision": 0.8982758620689655,
+      "eval_recall": 0.8980356681313001,
+      "eval_runtime": 0.0743,
+      "eval_samples_per_second": 4106.888,
+      "eval_steps_per_second": 67.326,
+      "step": 680
+    },
+    {
+      "epoch": 15.837209302325581,
+      "grad_norm": 2.9641597270965576,
+      "learning_rate": 2.616279069767442e-05,
+      "loss": 0.0753,
+      "step": 681
+    },
+    {
+      "epoch": 15.86046511627907,
+      "grad_norm": 5.338639259338379,
+      "learning_rate": 2.601744186046512e-05,
+      "loss": 0.0325,
+      "step": 682
+    },
+    {
+      "epoch": 15.883720930232558,
+      "grad_norm": 1.9736541509628296,
+      "learning_rate": 2.5872093023255818e-05,
+      "loss": 0.0236,
+      "step": 683
+    },
+    {
+      "epoch": 15.906976744186046,
+      "grad_norm": 0.5987362265586853,
+      "learning_rate": 2.5726744186046514e-05,
+      "loss": 0.0056,
+      "step": 684
+    },
+    {
+      "epoch": 15.930232558139535,
+      "grad_norm": 1.9034494161605835,
+      "learning_rate": 2.5581395348837212e-05,
+      "loss": 0.095,
+      "step": 685
+    },
+    {
+      "epoch": 15.953488372093023,
+      "grad_norm": 0.7597903609275818,
+      "learning_rate": 2.5436046511627908e-05,
+      "loss": 0.0111,
+      "step": 686
+    },
+    {
+      "epoch": 15.976744186046512,
+      "grad_norm": 2.3135693073272705,
+      "learning_rate": 2.5290697674418607e-05,
+      "loss": 0.0792,
+      "step": 687
+    },
+    {
+      "epoch": 16.0,
+      "grad_norm": 0.07970131933689117,
+      "learning_rate": 2.5145348837209302e-05,
+      "loss": 0.0024,
+      "step": 688
+    },
+    {
+      "epoch": 16.023255813953487,
+      "grad_norm": 1.8181146383285522,
+      "learning_rate": 2.5e-05,
+      "loss": 0.0181,
+      "step": 689
+    },
+    {
+      "epoch": 16.046511627906977,
+      "grad_norm": 4.323759078979492,
+      "learning_rate": 2.48546511627907e-05,
+      "loss": 0.0386,
+      "step": 690
+    },
+    {
+      "epoch": 16.069767441860463,
+      "grad_norm": 3.518137216567993,
+      "learning_rate": 2.4709302325581396e-05,
+      "loss": 0.0794,
+      "step": 691
+    },
+    {
+      "epoch": 16.093023255813954,
+      "grad_norm": 0.384950190782547,
+      "learning_rate": 2.4563953488372094e-05,
+      "loss": 0.0042,
+      "step": 692
+    },
+    {
+      "epoch": 16.11627906976744,
+      "grad_norm": 0.4682580232620239,
+      "learning_rate": 2.441860465116279e-05,
+      "loss": 0.0078,
+      "step": 693
+    },
+    {
+      "epoch": 16.13953488372093,
+      "grad_norm": 1.6920162439346313,
+      "learning_rate": 2.427325581395349e-05,
+      "loss": 0.0181,
+      "step": 694
+    },
+    {
+      "epoch": 16.162790697674417,
+      "grad_norm": 3.120163679122925,
+      "learning_rate": 2.4127906976744188e-05,
+      "loss": 0.0414,
+      "step": 695
+    },
+    {
+      "epoch": 16.186046511627907,
+      "grad_norm": 0.8569607734680176,
+      "learning_rate": 2.3982558139534887e-05,
+      "loss": 0.0063,
+      "step": 696
+    },
+    {
+      "epoch": 16.209302325581394,
+      "grad_norm": 1.5792855024337769,
+      "learning_rate": 2.3837209302325582e-05,
+      "loss": 0.0233,
+      "step": 697
+    },
+    {
+      "epoch": 16.232558139534884,
+      "grad_norm": 11.798049926757812,
+      "learning_rate": 2.369186046511628e-05,
+      "loss": 0.031,
+      "step": 698
+    },
+    {
+      "epoch": 16.25581395348837,
+      "grad_norm": 2.0980348587036133,
+      "learning_rate": 2.354651162790698e-05,
+      "loss": 0.0345,
+      "step": 699
+    },
+    {
+      "epoch": 16.27906976744186,
+      "grad_norm": 1.9684696197509766,
+      "learning_rate": 2.3401162790697675e-05,
+      "loss": 0.0109,
+      "step": 700
+    },
+    {
+      "epoch": 16.27906976744186,
+      "eval_accuracy": 0.898360655737705,
+      "eval_f1": 0.8979217273954117,
+      "eval_loss": 0.36475130915641785,
+      "eval_precision": 0.8996056167114501,
+      "eval_recall": 0.8971956577927114,
+      "eval_runtime": 0.0743,
+      "eval_samples_per_second": 4105.148,
+      "eval_steps_per_second": 67.298,
+      "step": 700
+    },
+    {
+      "epoch": 16.302325581395348,
+      "grad_norm": 1.2527034282684326,
+      "learning_rate": 2.3255813953488374e-05,
+      "loss": 0.0088,
+      "step": 701
+    },
+    {
+      "epoch": 16.325581395348838,
+      "grad_norm": 4.517597198486328,
+      "learning_rate": 2.311046511627907e-05,
+      "loss": 0.05,
+      "step": 702
+    },
+    {
+      "epoch": 16.348837209302324,
+      "grad_norm": 2.9931998252868652,
+      "learning_rate": 2.296511627906977e-05,
+      "loss": 0.0852,
+      "step": 703
+    },
+    {
+      "epoch": 16.372093023255815,
+      "grad_norm": 1.5361837148666382,
+      "learning_rate": 2.2819767441860464e-05,
+      "loss": 0.0205,
+      "step": 704
+    },
+    {
+      "epoch": 16.3953488372093,
+      "grad_norm": 12.71114730834961,
+      "learning_rate": 2.2674418604651163e-05,
+      "loss": 0.0643,
+      "step": 705
+    },
+    {
+      "epoch": 16.41860465116279,
+      "grad_norm": 0.1732826828956604,
+      "learning_rate": 2.252906976744186e-05,
+      "loss": 0.0028,
+      "step": 706
+    },
+    {
+      "epoch": 16.441860465116278,
+      "grad_norm": 5.568197250366211,
+      "learning_rate": 2.238372093023256e-05,
+      "loss": 0.0147,
+      "step": 707
+    },
+    {
+      "epoch": 16.46511627906977,
+      "grad_norm": 3.462963581085205,
+      "learning_rate": 2.2238372093023256e-05,
+      "loss": 0.0639,
+      "step": 708
+    },
+    {
+      "epoch": 16.488372093023255,
+      "grad_norm": 0.7707027792930603,
+      "learning_rate": 2.2093023255813955e-05,
+      "loss": 0.0052,
+      "step": 709
+    },
+    {
+      "epoch": 16.511627906976745,
+      "grad_norm": 2.6119136810302734,
+      "learning_rate": 2.1947674418604654e-05,
+      "loss": 0.0679,
+      "step": 710
+    },
+    {
+      "epoch": 16.53488372093023,
+      "grad_norm": 4.328090190887451,
+      "learning_rate": 2.180232558139535e-05,
+      "loss": 0.0336,
+      "step": 711
+    },
+    {
+      "epoch": 16.558139534883722,
+      "grad_norm": 3.516861915588379,
+      "learning_rate": 2.1656976744186048e-05,
+      "loss": 0.1004,
+      "step": 712
+    },
+    {
+      "epoch": 16.58139534883721,
+      "grad_norm": 0.8900988698005676,
+      "learning_rate": 2.1511627906976744e-05,
+      "loss": 0.0083,
+      "step": 713
+    },
+    {
+      "epoch": 16.6046511627907,
+      "grad_norm": 4.924454689025879,
+      "learning_rate": 2.1366279069767442e-05,
+      "loss": 0.1108,
+      "step": 714
+    },
+    {
+      "epoch": 16.627906976744185,
+      "grad_norm": 4.6815619468688965,
+      "learning_rate": 2.1220930232558138e-05,
+      "loss": 0.0611,
+      "step": 715
+    },
+    {
+      "epoch": 16.651162790697676,
+      "grad_norm": 7.625583171844482,
+      "learning_rate": 2.107558139534884e-05,
+      "loss": 0.0698,
+      "step": 716
+    },
+    {
+      "epoch": 16.674418604651162,
+      "grad_norm": 2.8963229656219482,
+      "learning_rate": 2.0930232558139536e-05,
+      "loss": 0.0226,
+      "step": 717
+    },
+    {
+      "epoch": 16.697674418604652,
+      "grad_norm": 1.349528193473816,
+      "learning_rate": 2.0784883720930235e-05,
+      "loss": 0.0074,
+      "step": 718
+    },
+    {
+      "epoch": 16.72093023255814,
+      "grad_norm": 0.9660943746566772,
+      "learning_rate": 2.0639534883720933e-05,
+      "loss": 0.0042,
+      "step": 719
+    },
+    {
+      "epoch": 16.74418604651163,
+      "grad_norm": 0.205625519156456,
+      "learning_rate": 2.049418604651163e-05,
+      "loss": 0.0036,
+      "step": 720
+    },
+    {
+      "epoch": 16.74418604651163,
+      "eval_accuracy": 0.9081967213114754,
+      "eval_f1": 0.9080296372878436,
+      "eval_loss": 0.3279436528682709,
+      "eval_precision": 0.9080296372878436,
+      "eval_recall": 0.9080296372878436,
+      "eval_runtime": 0.0765,
+      "eval_samples_per_second": 3988.597,
+      "eval_steps_per_second": 65.387,
+      "step": 720
+    },
+    {
+      "epoch": 16.767441860465116,
+      "grad_norm": 9.84315299987793,
+      "learning_rate": 2.0348837209302328e-05,
+      "loss": 0.1307,
+      "step": 721
+    },
+    {
+      "epoch": 16.790697674418606,
+      "grad_norm": 11.633748054504395,
+      "learning_rate": 2.0203488372093023e-05,
+      "loss": 0.065,
+      "step": 722
+    },
+    {
+      "epoch": 16.813953488372093,
+      "grad_norm": 2.6331350803375244,
+      "learning_rate": 2.0058139534883722e-05,
+      "loss": 0.0336,
+      "step": 723
+    },
+    {
+      "epoch": 16.837209302325583,
+      "grad_norm": 6.747828006744385,
+      "learning_rate": 1.9912790697674418e-05,
+      "loss": 0.0478,
+      "step": 724
+    },
+    {
+      "epoch": 16.86046511627907,
+      "grad_norm": 1.7129838466644287,
+      "learning_rate": 1.9767441860465116e-05,
+      "loss": 0.0403,
+      "step": 725
+    },
+    {
+      "epoch": 16.88372093023256,
+      "grad_norm": 1.0863773822784424,
+      "learning_rate": 1.9622093023255815e-05,
+      "loss": 0.0105,
+      "step": 726
+    },
+    {
+      "epoch": 16.906976744186046,
+      "grad_norm": 1.2210698127746582,
+      "learning_rate": 1.9476744186046514e-05,
+      "loss": 0.0114,
+      "step": 727
+    },
+    {
+      "epoch": 16.930232558139537,
+      "grad_norm": 3.351280689239502,
+      "learning_rate": 1.933139534883721e-05,
+      "loss": 0.0353,
+      "step": 728
+    },
+    {
+      "epoch": 16.953488372093023,
+      "grad_norm": 1.9887293577194214,
+      "learning_rate": 1.918604651162791e-05,
+      "loss": 0.0208,
+      "step": 729
+    },
+    {
+      "epoch": 16.97674418604651,
+      "grad_norm": 0.7091693878173828,
+      "learning_rate": 1.9040697674418607e-05,
+      "loss": 0.0074,
+      "step": 730
+    },
+    {
+      "epoch": 17.0,
+      "grad_norm": 0.6858041882514954,
+      "learning_rate": 1.8895348837209303e-05,
+      "loss": 0.0051,
+      "step": 731
+    },
+    {
+      "epoch": 17.023255813953487,
+      "grad_norm": 4.898313045501709,
+      "learning_rate": 1.8750000000000002e-05,
+      "loss": 0.1197,
+      "step": 732
+    },
+    {
+      "epoch": 17.046511627906977,
+      "grad_norm": 2.349663496017456,
+      "learning_rate": 1.8604651162790697e-05,
+      "loss": 0.0244,
+      "step": 733
+    },
+    {
+      "epoch": 17.069767441860463,
+      "grad_norm": 2.8203916549682617,
+      "learning_rate": 1.8459302325581396e-05,
+      "loss": 0.0801,
+      "step": 734
+    },
+    {
+      "epoch": 17.093023255813954,
+      "grad_norm": 4.241973400115967,
+      "learning_rate": 1.831395348837209e-05,
+      "loss": 0.0303,
+      "step": 735
+    },
+    {
+      "epoch": 17.11627906976744,
+      "grad_norm": 7.765374660491943,
+      "learning_rate": 1.816860465116279e-05,
+      "loss": 0.0758,
+      "step": 736
+    },
+    {
+      "epoch": 17.13953488372093,
+      "grad_norm": 7.214727401733398,
+      "learning_rate": 1.802325581395349e-05,
+      "loss": 0.0461,
+      "step": 737
+    },
+    {
+      "epoch": 17.162790697674417,
+      "grad_norm": 7.209427833557129,
+      "learning_rate": 1.7877906976744188e-05,
+      "loss": 0.0354,
+      "step": 738
+    },
+    {
+      "epoch": 17.186046511627907,
+      "grad_norm": 4.5369744300842285,
+      "learning_rate": 1.7732558139534887e-05,
+      "loss": 0.0223,
+      "step": 739
+    },
+    {
+      "epoch": 17.209302325581394,
+      "grad_norm": 0.8520297408103943,
+      "learning_rate": 1.7587209302325583e-05,
+      "loss": 0.0115,
+      "step": 740
+    },
+    {
+      "epoch": 17.209302325581394,
+      "eval_accuracy": 0.8950819672131147,
+      "eval_f1": 0.8945822172297591,
+      "eval_loss": 0.35655906796455383,
+      "eval_precision": 0.8966414996094767,
+      "eval_recall": 0.8937710002584647,
+      "eval_runtime": 0.0741,
+      "eval_samples_per_second": 4118.417,
+      "eval_steps_per_second": 67.515,
+      "step": 740
+    },
+    {
+      "epoch": 17.232558139534884,
+      "grad_norm": 2.845405101776123,
+      "learning_rate": 1.744186046511628e-05,
+      "loss": 0.0416,
+      "step": 741
+    },
+    {
+      "epoch": 17.25581395348837,
+      "grad_norm": 12.433396339416504,
+      "learning_rate": 1.7296511627906977e-05,
+      "loss": 0.0335,
+      "step": 742
+    },
+    {
+      "epoch": 17.27906976744186,
+      "grad_norm": 0.4390711486339569,
+      "learning_rate": 1.7151162790697676e-05,
+      "loss": 0.0062,
+      "step": 743
+    },
+    {
+      "epoch": 17.302325581395348,
+      "grad_norm": 2.333735942840576,
+      "learning_rate": 1.700581395348837e-05,
+      "loss": 0.0303,
+      "step": 744
+    },
+    {
+      "epoch": 17.325581395348838,
+      "grad_norm": 0.1029660701751709,
+      "learning_rate": 1.686046511627907e-05,
+      "loss": 0.0023,
+      "step": 745
+    },
+    {
+      "epoch": 17.348837209302324,
+      "grad_norm": 4.543797016143799,
+      "learning_rate": 1.671511627906977e-05,
+      "loss": 0.0634,
+      "step": 746
+    },
+    {
+      "epoch": 17.372093023255815,
+      "grad_norm": 4.657381534576416,
+      "learning_rate": 1.6569767441860464e-05,
+      "loss": 0.0199,
+      "step": 747
+    },
+    {
+      "epoch": 17.3953488372093,
+      "grad_norm": 6.3823018074035645,
+      "learning_rate": 1.6424418604651163e-05,
+      "loss": 0.1186,
+      "step": 748
+    },
+    {
+      "epoch": 17.41860465116279,
+      "grad_norm": 1.975292682647705,
+      "learning_rate": 1.6279069767441862e-05,
+      "loss": 0.0948,
+      "step": 749
+    },
+    {
+      "epoch": 17.441860465116278,
+      "grad_norm": 4.323884010314941,
+      "learning_rate": 1.613372093023256e-05,
+      "loss": 0.1327,
+      "step": 750
+    },
+    {
+      "epoch": 17.46511627906977,
+      "grad_norm": 1.2401987314224243,
+      "learning_rate": 1.5988372093023257e-05,
+      "loss": 0.0251,
+      "step": 751
+    },
+    {
+      "epoch": 17.488372093023255,
+      "grad_norm": 0.8695264458656311,
+      "learning_rate": 1.5843023255813955e-05,
+      "loss": 0.005,
+      "step": 752
+    },
+    {
+      "epoch": 17.511627906976745,
+      "grad_norm": 2.0256264209747314,
+      "learning_rate": 1.569767441860465e-05,
+      "loss": 0.0133,
+      "step": 753
+    },
+    {
+      "epoch": 17.53488372093023,
+      "grad_norm": 0.6134868264198303,
+      "learning_rate": 1.555232558139535e-05,
+      "loss": 0.0103,
+      "step": 754
+    },
+    {
+      "epoch": 17.558139534883722,
+      "grad_norm": 4.545815467834473,
+      "learning_rate": 1.5406976744186045e-05,
+      "loss": 0.0191,
+      "step": 755
+    },
+    {
+      "epoch": 17.58139534883721,
+      "grad_norm": 3.8483781814575195,
+      "learning_rate": 1.5261627906976744e-05,
+      "loss": 0.0311,
+      "step": 756
+    },
+    {
+      "epoch": 17.6046511627907,
+      "grad_norm": 2.32814359664917,
+      "learning_rate": 1.5116279069767441e-05,
+      "loss": 0.023,
+      "step": 757
+    },
+    {
+      "epoch": 17.627906976744185,
+      "grad_norm": 0.3794184625148773,
+      "learning_rate": 1.4970930232558138e-05,
+      "loss": 0.0054,
+      "step": 758
+    },
+    {
+      "epoch": 17.651162790697676,
+      "grad_norm": 1.3837960958480835,
+      "learning_rate": 1.4825581395348839e-05,
+      "loss": 0.0342,
+      "step": 759
+    },
+    {
+      "epoch": 17.674418604651162,
+      "grad_norm": 1.8951282501220703,
+      "learning_rate": 1.4680232558139536e-05,
+      "loss": 0.0618,
+      "step": 760
+    },
+    {
+      "epoch": 17.674418604651162,
+      "eval_accuracy": 0.898360655737705,
+      "eval_f1": 0.8980054152598138,
+      "eval_loss": 0.3718366026878357,
+      "eval_precision": 0.8990226604393703,
+      "eval_recall": 0.8974756612389075,
+      "eval_runtime": 0.0766,
+      "eval_samples_per_second": 3981.633,
+      "eval_steps_per_second": 65.273,
+      "step": 760
+    },
+    {
+      "epoch": 17.697674418604652,
+      "grad_norm": 1.703765630722046,
+      "learning_rate": 1.4534883720930233e-05,
+      "loss": 0.0089,
+      "step": 761
+    },
+    {
+      "epoch": 17.72093023255814,
+      "grad_norm": 2.194350004196167,
+      "learning_rate": 1.4389534883720932e-05,
+      "loss": 0.0156,
+      "step": 762
+    },
+    {
+      "epoch": 17.74418604651163,
+      "grad_norm": 1.5929670333862305,
+      "learning_rate": 1.424418604651163e-05,
+      "loss": 0.0093,
+      "step": 763
+    },
+    {
+      "epoch": 17.767441860465116,
+      "grad_norm": 1.1527395248413086,
+      "learning_rate": 1.4098837209302327e-05,
+      "loss": 0.0087,
+      "step": 764
+    },
+    {
+      "epoch": 17.790697674418606,
+      "grad_norm": 0.9619439244270325,
+      "learning_rate": 1.3953488372093024e-05,
+      "loss": 0.0101,
+      "step": 765
+    },
+    {
+      "epoch": 17.813953488372093,
+      "grad_norm": 2.041372537612915,
+      "learning_rate": 1.3808139534883721e-05,
+      "loss": 0.0259,
+      "step": 766
+    },
+    {
+      "epoch": 17.837209302325583,
+      "grad_norm": 4.107051849365234,
+      "learning_rate": 1.3662790697674418e-05,
+      "loss": 0.0195,
+      "step": 767
+    },
+    {
+      "epoch": 17.86046511627907,
+      "grad_norm": 0.8670945763587952,
+      "learning_rate": 1.3517441860465115e-05,
+      "loss": 0.0076,
+      "step": 768
+    },
+    {
+      "epoch": 17.88372093023256,
+      "grad_norm": 2.5873591899871826,
+      "learning_rate": 1.3372093023255814e-05,
+      "loss": 0.0325,
+      "step": 769
+    },
+    {
+      "epoch": 17.906976744186046,
+      "grad_norm": 3.630141496658325,
+      "learning_rate": 1.3226744186046513e-05,
+      "loss": 0.0261,
+      "step": 770
+    },
+    {
+      "epoch": 17.930232558139537,
+      "grad_norm": 0.1693820059299469,
+      "learning_rate": 1.308139534883721e-05,
+      "loss": 0.0036,
+      "step": 771
+    },
+    {
+      "epoch": 17.953488372093023,
+      "grad_norm": 0.5527888536453247,
+      "learning_rate": 1.2936046511627909e-05,
+      "loss": 0.0051,
+      "step": 772
+    },
+    {
+      "epoch": 17.97674418604651,
+      "grad_norm": 0.8153883814811707,
+      "learning_rate": 1.2790697674418606e-05,
+      "loss": 0.0095,
+      "step": 773
+    },
+    {
+      "epoch": 18.0,
+      "grad_norm": 0.2584695518016815,
+      "learning_rate": 1.2645348837209303e-05,
+      "loss": 0.0036,
+      "step": 774
+    },
+    {
+      "epoch": 18.023255813953487,
+      "grad_norm": 0.5313040018081665,
+      "learning_rate": 1.25e-05,
+      "loss": 0.0071,
+      "step": 775
+    },
+    {
+      "epoch": 18.046511627906977,
+      "grad_norm": 1.0279793739318848,
+      "learning_rate": 1.2354651162790698e-05,
+      "loss": 0.01,
+      "step": 776
+    },
+    {
+      "epoch": 18.069767441860463,
+      "grad_norm": 1.0528727769851685,
+      "learning_rate": 1.2209302325581395e-05,
+      "loss": 0.0168,
+      "step": 777
+    },
+    {
+      "epoch": 18.093023255813954,
+      "grad_norm": 1.2885291576385498,
+      "learning_rate": 1.2063953488372094e-05,
+      "loss": 0.0077,
+      "step": 778
+    },
+    {
+      "epoch": 18.11627906976744,
+      "grad_norm": 0.3943243622779846,
+      "learning_rate": 1.1918604651162791e-05,
+      "loss": 0.0026,
+      "step": 779
+    },
+    {
+      "epoch": 18.13953488372093,
+      "grad_norm": 1.1616085767745972,
+      "learning_rate": 1.177325581395349e-05,
+      "loss": 0.0061,
+      "step": 780
+    },
+    {
+      "epoch": 18.13953488372093,
+      "eval_accuracy": 0.898360655737705,
+      "eval_f1": 0.8981460534962133,
+      "eval_loss": 0.37966448068618774,
+      "eval_precision": 0.8982758620689655,
+      "eval_recall": 0.8980356681313001,
+      "eval_runtime": 0.0743,
+      "eval_samples_per_second": 4106.506,
+      "eval_steps_per_second": 67.32,
+      "step": 780
+    },
+    {
+      "epoch": 18.162790697674417,
+      "grad_norm": 0.11722904443740845,
+      "learning_rate": 1.1627906976744187e-05,
+      "loss": 0.0023,
+      "step": 781
+    },
+    {
+      "epoch": 18.186046511627907,
+      "grad_norm": 1.2253535985946655,
+      "learning_rate": 1.1482558139534884e-05,
+      "loss": 0.0182,
+      "step": 782
+    },
+    {
+      "epoch": 18.209302325581394,
+      "grad_norm": 0.4131571054458618,
+      "learning_rate": 1.1337209302325581e-05,
+      "loss": 0.0038,
+      "step": 783
+    },
+    {
+      "epoch": 18.232558139534884,
+      "grad_norm": 0.5466241240501404,
+      "learning_rate": 1.119186046511628e-05,
+      "loss": 0.0033,
+      "step": 784
+    },
+    {
+      "epoch": 18.25581395348837,
+      "grad_norm": 0.8692965507507324,
+      "learning_rate": 1.1046511627906977e-05,
+      "loss": 0.0079,
+      "step": 785
+    },
+    {
+      "epoch": 18.27906976744186,
+      "grad_norm": 3.9168105125427246,
+      "learning_rate": 1.0901162790697675e-05,
+      "loss": 0.0861,
+      "step": 786
+    },
+    {
+      "epoch": 18.302325581395348,
+      "grad_norm": 1.2862012386322021,
+      "learning_rate": 1.0755813953488372e-05,
+      "loss": 0.0081,
+      "step": 787
+    },
+    {
+      "epoch": 18.325581395348838,
+      "grad_norm": 6.316368103027344,
+      "learning_rate": 1.0610465116279069e-05,
+      "loss": 0.044,
+      "step": 788
+    },
+    {
+      "epoch": 18.348837209302324,
+      "grad_norm": 7.892980098724365,
+      "learning_rate": 1.0465116279069768e-05,
+      "loss": 0.1037,
+      "step": 789
+    },
+    {
+      "epoch": 18.372093023255815,
+      "grad_norm": 7.4783172607421875,
+      "learning_rate": 1.0319767441860467e-05,
+      "loss": 0.0339,
+      "step": 790
+    },
+    {
+      "epoch": 18.3953488372093,
+      "grad_norm": 0.7634483575820923,
+      "learning_rate": 1.0174418604651164e-05,
+      "loss": 0.0039,
+      "step": 791
+    },
+    {
+      "epoch": 18.41860465116279,
+      "grad_norm": 3.145096778869629,
+      "learning_rate": 1.0029069767441861e-05,
+      "loss": 0.0739,
+      "step": 792
+    },
+    {
+      "epoch": 18.441860465116278,
+      "grad_norm": 6.179558277130127,
+      "learning_rate": 9.883720930232558e-06,
+      "loss": 0.0333,
+      "step": 793
+    },
+    {
+      "epoch": 18.46511627906977,
+      "grad_norm": 0.10819542407989502,
+      "learning_rate": 9.738372093023257e-06,
+      "loss": 0.0023,
+      "step": 794
+    },
+    {
+      "epoch": 18.488372093023255,
+      "grad_norm": 11.038155555725098,
+      "learning_rate": 9.593023255813954e-06,
+      "loss": 0.0816,
+      "step": 795
+    },
+    {
+      "epoch": 18.511627906976745,
+      "grad_norm": 4.89321756362915,
+      "learning_rate": 9.447674418604651e-06,
+      "loss": 0.0669,
+      "step": 796
+    },
+    {
+      "epoch": 18.53488372093023,
+      "grad_norm": 0.6758410930633545,
+      "learning_rate": 9.302325581395349e-06,
+      "loss": 0.0124,
+      "step": 797
+    },
+    {
+      "epoch": 18.558139534883722,
+      "grad_norm": 2.737079381942749,
+      "learning_rate": 9.156976744186046e-06,
+      "loss": 0.0134,
+      "step": 798
+    },
+    {
+      "epoch": 18.58139534883721,
+      "grad_norm": 5.82487154006958,
+      "learning_rate": 9.011627906976745e-06,
+      "loss": 0.0826,
+      "step": 799
+    },
+    {
+      "epoch": 18.6046511627907,
+      "grad_norm": 0.12569262087345123,
+      "learning_rate": 8.866279069767444e-06,
+      "loss": 0.0021,
+      "step": 800
+    },
+    {
+      "epoch": 18.6046511627907,
+      "eval_accuracy": 0.8950819672131147,
+      "eval_f1": 0.8945822172297591,
+      "eval_loss": 0.41386428475379944,
+      "eval_precision": 0.8966414996094767,
+      "eval_recall": 0.8937710002584647,
+      "eval_runtime": 0.074,
+      "eval_samples_per_second": 4121.655,
+      "eval_steps_per_second": 67.568,
+      "step": 800
+    },
+    {
+      "epoch": 18.627906976744185,
+      "grad_norm": 7.690613269805908,
+      "learning_rate": 8.72093023255814e-06,
+      "loss": 0.1334,
+      "step": 801
+    },
+    {
+      "epoch": 18.651162790697676,
+      "grad_norm": 7.942806720733643,
+      "learning_rate": 8.575581395348838e-06,
+      "loss": 0.0349,
+      "step": 802
+    },
+    {
+      "epoch": 18.674418604651162,
+      "grad_norm": 1.4096626043319702,
+      "learning_rate": 8.430232558139535e-06,
+      "loss": 0.0086,
+      "step": 803
+    },
+    {
+      "epoch": 18.697674418604652,
+      "grad_norm": 2.184845209121704,
+      "learning_rate": 8.284883720930232e-06,
+      "loss": 0.021,
+      "step": 804
+    },
+    {
+      "epoch": 18.72093023255814,
+      "grad_norm": 5.123547554016113,
+      "learning_rate": 8.139534883720931e-06,
+      "loss": 0.0714,
+      "step": 805
+    },
+    {
+      "epoch": 18.74418604651163,
+      "grad_norm": 2.141044855117798,
+      "learning_rate": 7.994186046511628e-06,
+      "loss": 0.0227,
+      "step": 806
+    },
+    {
+      "epoch": 18.767441860465116,
+      "grad_norm": 0.7044129967689514,
+      "learning_rate": 7.848837209302325e-06,
+      "loss": 0.0064,
+      "step": 807
+    },
+    {
+      "epoch": 18.790697674418606,
+      "grad_norm": 0.9102310538291931,
+      "learning_rate": 7.703488372093023e-06,
+      "loss": 0.0056,
+      "step": 808
+    },
+    {
+      "epoch": 18.813953488372093,
+      "grad_norm": 6.990550518035889,
+      "learning_rate": 7.558139534883721e-06,
+      "loss": 0.0359,
+      "step": 809
+    },
+    {
+      "epoch": 18.837209302325583,
+      "grad_norm": 0.36822423338890076,
+      "learning_rate": 7.4127906976744195e-06,
+      "loss": 0.0036,
+      "step": 810
+    },
+    {
+      "epoch": 18.86046511627907,
+      "grad_norm": 1.0232497453689575,
+      "learning_rate": 7.267441860465117e-06,
+      "loss": 0.0077,
+      "step": 811
+    },
+    {
+      "epoch": 18.88372093023256,
+      "grad_norm": 2.3726754188537598,
+      "learning_rate": 7.122093023255815e-06,
+      "loss": 0.0625,
+      "step": 812
+    },
+    {
+      "epoch": 18.906976744186046,
+      "grad_norm": 1.6161715984344482,
+      "learning_rate": 6.976744186046512e-06,
+      "loss": 0.015,
+      "step": 813
+    },
+    {
+      "epoch": 18.930232558139537,
+      "grad_norm": 1.230698823928833,
+      "learning_rate": 6.831395348837209e-06,
+      "loss": 0.0595,
+      "step": 814
+    },
+    {
+      "epoch": 18.953488372093023,
+      "grad_norm": 4.402613162994385,
+      "learning_rate": 6.686046511627907e-06,
+      "loss": 0.0847,
+      "step": 815
+    },
+    {
+      "epoch": 18.97674418604651,
+      "grad_norm": 6.6045966148376465,
+      "learning_rate": 6.540697674418605e-06,
+      "loss": 0.036,
+      "step": 816
+    },
+    {
+      "epoch": 19.0,
+      "grad_norm": 0.6431266069412231,
+      "learning_rate": 6.395348837209303e-06,
+      "loss": 0.0043,
+      "step": 817
+    },
+    {
+      "epoch": 19.023255813953487,
+      "grad_norm": 0.4433208703994751,
+      "learning_rate": 6.25e-06,
+      "loss": 0.0032,
+      "step": 818
+    },
+    {
+      "epoch": 19.046511627906977,
+      "grad_norm": 3.6785125732421875,
+      "learning_rate": 6.1046511627906975e-06,
+      "loss": 0.1244,
+      "step": 819
+    },
+    {
+      "epoch": 19.069767441860463,
+      "grad_norm": 2.5291504859924316,
+      "learning_rate": 5.9593023255813955e-06,
+      "loss": 0.0071,
+      "step": 820
+    },
+    {
+      "epoch": 19.069767441860463,
+      "eval_accuracy": 0.9049180327868852,
+      "eval_f1": 0.9046556500555143,
+      "eval_loss": 0.3841802775859833,
+      "eval_precision": 0.9051627384960719,
+      "eval_recall": 0.9043249763074007,
+      "eval_runtime": 0.0744,
+      "eval_samples_per_second": 4097.049,
+      "eval_steps_per_second": 67.165,
+      "step": 820
+    },
+    {
+      "epoch": 19.093023255813954,
+      "grad_norm": 8.134071350097656,
+      "learning_rate": 5.8139534883720935e-06,
+      "loss": 0.0308,
+      "step": 821
+    },
+    {
+      "epoch": 19.11627906976744,
+      "grad_norm": 0.23603279888629913,
+      "learning_rate": 5.668604651162791e-06,
+      "loss": 0.0028,
+      "step": 822
+    },
+    {
+      "epoch": 19.13953488372093,
+      "grad_norm": 5.337473392486572,
+      "learning_rate": 5.523255813953489e-06,
+      "loss": 0.034,
+      "step": 823
+    },
+    {
+      "epoch": 19.162790697674417,
+      "grad_norm": 4.512674808502197,
+      "learning_rate": 5.377906976744186e-06,
+      "loss": 0.0775,
+      "step": 824
+    },
+    {
+      "epoch": 19.186046511627907,
+      "grad_norm": 1.7216310501098633,
+      "learning_rate": 5.232558139534884e-06,
+      "loss": 0.0056,
+      "step": 825
+    },
+    {
+      "epoch": 19.209302325581394,
+      "grad_norm": 1.9352085590362549,
+      "learning_rate": 5.087209302325582e-06,
+      "loss": 0.0423,
+      "step": 826
+    },
+    {
+      "epoch": 19.232558139534884,
+      "grad_norm": 10.692374229431152,
+      "learning_rate": 4.941860465116279e-06,
+      "loss": 0.0782,
+      "step": 827
+    },
+    {
+      "epoch": 19.25581395348837,
+      "grad_norm": 0.10650653392076492,
+      "learning_rate": 4.796511627906977e-06,
+      "loss": 0.0017,
+      "step": 828
+    },
+    {
+      "epoch": 19.27906976744186,
+      "grad_norm": 0.14889536798000336,
+      "learning_rate": 4.651162790697674e-06,
+      "loss": 0.0024,
+      "step": 829
+    },
+    {
+      "epoch": 19.302325581395348,
+      "grad_norm": 5.617801666259766,
+      "learning_rate": 4.505813953488372e-06,
+      "loss": 0.0121,
+      "step": 830
+    },
+    {
+      "epoch": 19.325581395348838,
+      "grad_norm": 2.859057903289795,
+      "learning_rate": 4.36046511627907e-06,
+      "loss": 0.034,
+      "step": 831
+    },
+    {
+      "epoch": 19.348837209302324,
+      "grad_norm": 0.27750277519226074,
+      "learning_rate": 4.2151162790697675e-06,
+      "loss": 0.0044,
+      "step": 832
+    },
+    {
+      "epoch": 19.372093023255815,
+      "grad_norm": 0.04883375018835068,
+      "learning_rate": 4.0697674418604655e-06,
+      "loss": 0.0018,
+      "step": 833
+    },
+    {
+      "epoch": 19.3953488372093,
+      "grad_norm": 1.9595685005187988,
+      "learning_rate": 3.924418604651163e-06,
+      "loss": 0.011,
+      "step": 834
+    },
+    {
+      "epoch": 19.41860465116279,
+      "grad_norm": 1.273353099822998,
+      "learning_rate": 3.7790697674418603e-06,
+      "loss": 0.0214,
+      "step": 835
+    },
+    {
+      "epoch": 19.441860465116278,
+      "grad_norm": 6.320394039154053,
+      "learning_rate": 3.6337209302325583e-06,
+      "loss": 0.081,
+      "step": 836
+    },
+    {
+      "epoch": 19.46511627906977,
+      "grad_norm": 2.93029522895813,
+      "learning_rate": 3.488372093023256e-06,
+      "loss": 0.0256,
+      "step": 837
+    },
+    {
+      "epoch": 19.488372093023255,
+      "grad_norm": 0.4745703339576721,
+      "learning_rate": 3.3430232558139535e-06,
+      "loss": 0.0033,
+      "step": 838
+    },
+    {
+      "epoch": 19.511627906976745,
+      "grad_norm": 4.16102933883667,
+      "learning_rate": 3.1976744186046516e-06,
+      "loss": 0.0482,
+      "step": 839
+    },
+    {
+      "epoch": 19.53488372093023,
+      "grad_norm": 0.2739373743534088,
+      "learning_rate": 3.0523255813953487e-06,
+      "loss": 0.0045,
+      "step": 840
+    },
+    {
+      "epoch": 19.53488372093023,
+      "eval_accuracy": 0.9049180327868852,
+      "eval_f1": 0.9046556500555143,
+      "eval_loss": 0.38903769850730896,
+      "eval_precision": 0.9051627384960719,
+      "eval_recall": 0.9043249763074007,
+      "eval_runtime": 0.0754,
+      "eval_samples_per_second": 4042.913,
+      "eval_steps_per_second": 66.277,
+      "step": 840
+    },
+    {
+      "epoch": 19.558139534883722,
+      "grad_norm": 0.87278813123703,
+      "learning_rate": 2.9069767441860468e-06,
+      "loss": 0.0116,
+      "step": 841
+    },
+    {
+      "epoch": 19.58139534883721,
+      "grad_norm": 0.7536394596099854,
+      "learning_rate": 2.7616279069767444e-06,
+      "loss": 0.0036,
+      "step": 842
+    },
+    {
+      "epoch": 19.6046511627907,
+      "grad_norm": 4.64411735534668,
+      "learning_rate": 2.616279069767442e-06,
+      "loss": 0.0712,
+      "step": 843
+    },
+    {
+      "epoch": 19.627906976744185,
+      "grad_norm": 0.5402886271476746,
+      "learning_rate": 2.4709302325581396e-06,
+      "loss": 0.0079,
+      "step": 844
+    },
+    {
+      "epoch": 19.651162790697676,
+      "grad_norm": 0.591332733631134,
+      "learning_rate": 2.325581395348837e-06,
+      "loss": 0.0028,
+      "step": 845
+    },
+    {
+      "epoch": 19.674418604651162,
+      "grad_norm": 1.511197805404663,
+      "learning_rate": 2.180232558139535e-06,
+      "loss": 0.0216,
+      "step": 846
+    },
+    {
+      "epoch": 19.697674418604652,
+      "grad_norm": 2.847646951675415,
+      "learning_rate": 2.0348837209302328e-06,
+      "loss": 0.0246,
+      "step": 847
+    },
+    {
+      "epoch": 19.72093023255814,
+      "grad_norm": 0.33959656953811646,
+      "learning_rate": 1.8895348837209302e-06,
+      "loss": 0.0034,
+      "step": 848
+    },
+    {
+      "epoch": 19.74418604651163,
+      "grad_norm": 1.8157447576522827,
+      "learning_rate": 1.744186046511628e-06,
+      "loss": 0.0436,
+      "step": 849
+    },
+    {
+      "epoch": 19.767441860465116,
+      "grad_norm": 0.3644404113292694,
+      "learning_rate": 1.5988372093023258e-06,
+      "loss": 0.0038,
+      "step": 850
+    },
+    {
+      "epoch": 19.790697674418606,
+      "grad_norm": 2.5243899822235107,
+      "learning_rate": 1.4534883720930234e-06,
+      "loss": 0.0115,
+      "step": 851
+    },
+    {
+      "epoch": 19.813953488372093,
+      "grad_norm": 5.748671054840088,
+      "learning_rate": 1.308139534883721e-06,
+      "loss": 0.04,
+      "step": 852
+    },
+    {
+      "epoch": 19.837209302325583,
+      "grad_norm": 0.5189786553382874,
+      "learning_rate": 1.1627906976744186e-06,
+      "loss": 0.0049,
+      "step": 853
+    },
+    {
+      "epoch": 19.86046511627907,
+      "grad_norm": 2.05354380607605,
+      "learning_rate": 1.0174418604651164e-06,
+      "loss": 0.0591,
+      "step": 854
+    },
+    {
+      "epoch": 19.88372093023256,
+      "grad_norm": 2.982142925262451,
+      "learning_rate": 8.72093023255814e-07,
+      "loss": 0.0224,
+      "step": 855
+    },
+    {
+      "epoch": 19.906976744186046,
+      "grad_norm": 2.0666937828063965,
+      "learning_rate": 7.267441860465117e-07,
+      "loss": 0.0217,
+      "step": 856
+    },
+    {
+      "epoch": 19.930232558139537,
+      "grad_norm": 4.0683722496032715,
+      "learning_rate": 5.813953488372093e-07,
+      "loss": 0.108,
+      "step": 857
+    },
+    {
+      "epoch": 19.953488372093023,
+      "grad_norm": 10.890898704528809,
+      "learning_rate": 4.36046511627907e-07,
+      "loss": 0.069,
+      "step": 858
+    },
+    {
+      "epoch": 19.97674418604651,
+      "grad_norm": 0.7343159914016724,
+      "learning_rate": 2.9069767441860464e-07,
+      "loss": 0.0071,
+      "step": 859
+    },
+    {
+      "epoch": 20.0,
+      "grad_norm": 1.6608028411865234,
+      "learning_rate": 1.4534883720930232e-07,
+      "loss": 0.0411,
+      "step": 860
+    },
+    {
+      "epoch": 20.0,
+      "eval_accuracy": 0.8950819672131147,
+      "eval_f1": 0.8946732280065612,
+      "eval_loss": 0.3955073356628418,
+      "eval_precision": 0.895995670995671,
+      "eval_recall": 0.894051003704661,
+      "eval_runtime": 0.076,
+      "eval_samples_per_second": 4014.381,
+      "eval_steps_per_second": 65.81,
+      "step": 860
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 860,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 20,
+  "save_steps": 200,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 68001814271952.0,
+  "train_batch_size": 64,
+  "trial_name": null,
+  "trial_params": null
+}