{
  "best_global_step": 3900,
  "best_metric": 0.6421848255039291,
  "best_model_checkpoint": "./experiments/qwen3-0.6b-router-lr1e-5-ep2-batch20-20250917-11:04/checkpoint-3900",
  "epoch": 1.597908192911098,
  "eval_steps": 100,
  "global_step": 5500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0002905287623474724,
      "grad_norm": 422.0,
      "learning_rate": 0.0,
      "loss": 4.1917,
      "step": 1
    },
    {
      "epoch": 0.002905287623474724,
      "grad_norm": 260.0,
      "learning_rate": 9.000000000000001e-07,
      "loss": 3.1155,
      "step": 10
    },
    {
      "epoch": 0.005810575246949448,
      "grad_norm": 362.0,
      "learning_rate": 1.9000000000000002e-06,
      "loss": 3.6823,
      "step": 20
    },
    {
      "epoch": 0.008715862870424172,
      "grad_norm": 362.0,
      "learning_rate": 2.9e-06,
      "loss": 3.5894,
      "step": 30
    },
    {
      "epoch": 0.011621150493898896,
      "grad_norm": 286.0,
      "learning_rate": 3.900000000000001e-06,
      "loss": 2.7935,
      "step": 40
    },
    {
      "epoch": 0.01452643811737362,
      "grad_norm": 108.5,
      "learning_rate": 4.9000000000000005e-06,
      "loss": 2.1466,
      "step": 50
    },
    {
      "epoch": 0.017431725740848343,
      "grad_norm": 171.0,
      "learning_rate": 5.9e-06,
      "loss": 2.174,
      "step": 60
    },
    {
      "epoch": 0.02033701336432307,
      "grad_norm": 182.0,
      "learning_rate": 6.9e-06,
      "loss": 1.9269,
      "step": 70
    },
    {
      "epoch": 0.023242300987797792,
      "grad_norm": 152.0,
      "learning_rate": 7.9e-06,
      "loss": 2.0153,
      "step": 80
    },
    {
      "epoch": 0.026147588611272515,
      "grad_norm": 124.5,
      "learning_rate": 8.900000000000001e-06,
      "loss": 1.7756,
      "step": 90
    },
    {
      "epoch": 0.02905287623474724,
      "grad_norm": 220.0,
      "learning_rate": 9.9e-06,
      "loss": 1.7598,
      "step": 100
    },
    {
      "epoch": 0.02905287623474724,
      "eval_accuracy": 0.5466593042517945,
      "eval_f1": 0.5420321197439284,
      "eval_loss": 0.8443426489830017,
      "eval_precision": 0.5556366418233831,
      "eval_recall": 0.5466593042517945,
      "eval_runtime": 17.2355,
      "eval_samples_per_second": 105.074,
      "eval_steps_per_second": 10.56,
      "step": 100
    },
    {
      "epoch": 0.031958163858221963,
      "grad_norm": 180.0,
      "learning_rate": 9.986733490566039e-06,
      "loss": 1.6999,
      "step": 110
    },
    {
      "epoch": 0.034863451481696686,
      "grad_norm": 98.0,
      "learning_rate": 9.971992924528303e-06,
      "loss": 1.6424,
      "step": 120
    },
    {
      "epoch": 0.03776873910517141,
      "grad_norm": 92.5,
      "learning_rate": 9.957252358490566e-06,
      "loss": 1.4874,
      "step": 130
    },
    {
      "epoch": 0.04067402672864614,
      "grad_norm": 99.5,
      "learning_rate": 9.942511792452831e-06,
      "loss": 1.668,
      "step": 140
    },
    {
      "epoch": 0.04357931435212086,
      "grad_norm": 89.5,
      "learning_rate": 9.927771226415096e-06,
      "loss": 1.4386,
      "step": 150
    },
    {
      "epoch": 0.046484601975595584,
      "grad_norm": 78.5,
      "learning_rate": 9.91303066037736e-06,
      "loss": 1.3427,
      "step": 160
    },
    {
      "epoch": 0.04938988959907031,
      "grad_norm": 77.5,
      "learning_rate": 9.898290094339623e-06,
      "loss": 1.7596,
      "step": 170
    },
    {
      "epoch": 0.05229517722254503,
      "grad_norm": 136.0,
      "learning_rate": 9.883549528301888e-06,
      "loss": 1.639,
      "step": 180
    },
    {
      "epoch": 0.05520046484601976,
      "grad_norm": 78.5,
      "learning_rate": 9.868808962264151e-06,
      "loss": 1.6126,
      "step": 190
    },
    {
      "epoch": 0.05810575246949448,
      "grad_norm": 104.5,
      "learning_rate": 9.854068396226416e-06,
      "loss": 1.4535,
      "step": 200
    },
    {
      "epoch": 0.05810575246949448,
      "eval_accuracy": 0.5737161789066814,
      "eval_f1": 0.5535126986816034,
      "eval_loss": 0.7697113752365112,
      "eval_precision": 0.5771129416178113,
      "eval_recall": 0.5737161789066814,
      "eval_runtime": 23.2381,
      "eval_samples_per_second": 77.932,
      "eval_steps_per_second": 7.832,
      "step": 200
    },
    {
      "epoch": 0.061011040092969204,
      "grad_norm": 67.0,
      "learning_rate": 9.83932783018868e-06,
      "loss": 1.4115,
      "step": 210
    },
    {
      "epoch": 0.06391632771644393,
      "grad_norm": 103.5,
      "learning_rate": 9.824587264150945e-06,
      "loss": 1.5474,
      "step": 220
    },
    {
      "epoch": 0.06682161533991865,
      "grad_norm": 70.5,
      "learning_rate": 9.809846698113208e-06,
      "loss": 1.3732,
      "step": 230
    },
    {
      "epoch": 0.06972690296339337,
      "grad_norm": 79.5,
      "learning_rate": 9.795106132075473e-06,
      "loss": 1.5937,
      "step": 240
    },
    {
      "epoch": 0.0726321905868681,
      "grad_norm": 160.0,
      "learning_rate": 9.780365566037736e-06,
      "loss": 1.5243,
      "step": 250
    },
    {
      "epoch": 0.07553747821034282,
      "grad_norm": 86.5,
      "learning_rate": 9.765625e-06,
      "loss": 1.5532,
      "step": 260
    },
    {
      "epoch": 0.07844276583381755,
      "grad_norm": 87.0,
      "learning_rate": 9.750884433962265e-06,
      "loss": 1.4715,
      "step": 270
    },
    {
      "epoch": 0.08134805345729228,
      "grad_norm": 72.0,
      "learning_rate": 9.73614386792453e-06,
      "loss": 1.5738,
      "step": 280
    },
    {
      "epoch": 0.084253341080767,
      "grad_norm": 60.0,
      "learning_rate": 9.721403301886794e-06,
      "loss": 1.4415,
      "step": 290
    },
    {
      "epoch": 0.08715862870424172,
      "grad_norm": 79.5,
      "learning_rate": 9.706662735849057e-06,
      "loss": 1.4364,
      "step": 300
    },
    {
      "epoch": 0.08715862870424172,
      "eval_accuracy": 0.5808945334069575,
      "eval_f1": 0.579545757594244,
      "eval_loss": 0.7065893411636353,
      "eval_precision": 0.5868008140090218,
      "eval_recall": 0.5808945334069575,
      "eval_runtime": 25.9239,
      "eval_samples_per_second": 69.858,
      "eval_steps_per_second": 7.021,
      "step": 300
    },
    {
      "epoch": 0.09006391632771645,
      "grad_norm": 50.0,
      "learning_rate": 9.691922169811322e-06,
      "loss": 1.323,
      "step": 310
    },
    {
      "epoch": 0.09296920395119117,
      "grad_norm": 77.0,
      "learning_rate": 9.677181603773585e-06,
      "loss": 1.4155,
      "step": 320
    },
    {
      "epoch": 0.09587449157466589,
      "grad_norm": 64.5,
      "learning_rate": 9.66244103773585e-06,
      "loss": 1.4485,
      "step": 330
    },
    {
      "epoch": 0.09877977919814061,
      "grad_norm": 115.0,
      "learning_rate": 9.647700471698114e-06,
      "loss": 1.5113,
      "step": 340
    },
    {
      "epoch": 0.10168506682161534,
      "grad_norm": 94.0,
      "learning_rate": 9.632959905660379e-06,
      "loss": 1.4558,
      "step": 350
    },
    {
      "epoch": 0.10459035444509006,
      "grad_norm": 74.5,
      "learning_rate": 9.618219339622642e-06,
      "loss": 1.385,
      "step": 360
    },
    {
      "epoch": 0.10749564206856478,
      "grad_norm": 58.0,
      "learning_rate": 9.603478773584906e-06,
      "loss": 1.4507,
      "step": 370
    },
    {
      "epoch": 0.11040092969203952,
      "grad_norm": 53.75,
      "learning_rate": 9.588738207547171e-06,
      "loss": 1.4854,
      "step": 380
    },
    {
      "epoch": 0.11330621731551424,
      "grad_norm": 67.5,
      "learning_rate": 9.573997641509436e-06,
      "loss": 1.4493,
      "step": 390
    },
    {
      "epoch": 0.11621150493898896,
      "grad_norm": 82.5,
      "learning_rate": 9.559257075471699e-06,
      "loss": 1.332,
      "step": 400
    },
    {
      "epoch": 0.11621150493898896,
      "eval_accuracy": 0.6046383213694092,
      "eval_f1": 0.5982589248765446,
      "eval_loss": 0.6897996068000793,
      "eval_precision": 0.6052451933014279,
      "eval_recall": 0.6046383213694092,
      "eval_runtime": 21.176,
      "eval_samples_per_second": 85.522,
      "eval_steps_per_second": 8.595,
      "step": 400
    },
    {
      "epoch": 0.11911679256246369,
      "grad_norm": 56.5,
      "learning_rate": 9.544516509433963e-06,
      "loss": 1.3926,
      "step": 410
    },
    {
      "epoch": 0.12202208018593841,
      "grad_norm": 95.5,
      "learning_rate": 9.529775943396226e-06,
      "loss": 1.4383,
      "step": 420
    },
    {
      "epoch": 0.12492736780941313,
      "grad_norm": 105.0,
      "learning_rate": 9.515035377358491e-06,
      "loss": 1.4494,
      "step": 430
    },
    {
      "epoch": 0.12783265543288785,
      "grad_norm": 41.25,
      "learning_rate": 9.500294811320756e-06,
      "loss": 1.3664,
      "step": 440
    },
    {
      "epoch": 0.13073794305636258,
      "grad_norm": 89.0,
      "learning_rate": 9.48555424528302e-06,
      "loss": 1.34,
      "step": 450
    },
    {
      "epoch": 0.1336432306798373,
      "grad_norm": 50.0,
      "learning_rate": 9.470813679245285e-06,
      "loss": 1.354,
      "step": 460
    },
    {
      "epoch": 0.13654851830331202,
      "grad_norm": 54.25,
      "learning_rate": 9.456073113207548e-06,
      "loss": 1.3451,
      "step": 470
    },
    {
      "epoch": 0.13945380592678674,
      "grad_norm": 51.75,
      "learning_rate": 9.441332547169812e-06,
      "loss": 1.4787,
      "step": 480
    },
    {
      "epoch": 0.14235909355026147,
      "grad_norm": 43.0,
      "learning_rate": 9.426591981132075e-06,
      "loss": 1.3808,
      "step": 490
    },
    {
      "epoch": 0.1452643811737362,
      "grad_norm": 62.5,
      "learning_rate": 9.41185141509434e-06,
      "loss": 1.4256,
      "step": 500
    },
    {
      "epoch": 0.1452643811737362,
      "eval_accuracy": 0.6112644947542794,
      "eval_f1": 0.6088118132920133,
      "eval_loss": 0.6750182509422302,
      "eval_precision": 0.6105845735531097,
      "eval_recall": 0.6112644947542794,
      "eval_runtime": 25.2545,
      "eval_samples_per_second": 71.71,
      "eval_steps_per_second": 7.207,
      "step": 500
    },
    {
      "epoch": 0.1481696687972109,
      "grad_norm": 66.0,
      "learning_rate": 9.397110849056605e-06,
      "loss": 1.2969,
      "step": 510
    },
    {
      "epoch": 0.15107495642068564,
      "grad_norm": 118.5,
      "learning_rate": 9.38237028301887e-06,
      "loss": 1.3989,
      "step": 520
    },
    {
      "epoch": 0.15398024404416036,
      "grad_norm": 49.25,
      "learning_rate": 9.367629716981132e-06,
      "loss": 1.4072,
      "step": 530
    },
    {
      "epoch": 0.1568855316676351,
      "grad_norm": 50.5,
      "learning_rate": 9.352889150943397e-06,
      "loss": 1.3402,
      "step": 540
    },
    {
      "epoch": 0.15979081929110983,
      "grad_norm": 55.5,
      "learning_rate": 9.33814858490566e-06,
      "loss": 1.4138,
      "step": 550
    },
    {
      "epoch": 0.16269610691458455,
      "grad_norm": 46.25,
      "learning_rate": 9.323408018867925e-06,
      "loss": 1.3791,
      "step": 560
    },
    {
      "epoch": 0.16560139453805928,
      "grad_norm": 126.5,
      "learning_rate": 9.30866745283019e-06,
      "loss": 1.4886,
      "step": 570
    },
    {
      "epoch": 0.168506682161534,
      "grad_norm": 35.0,
      "learning_rate": 9.293926886792454e-06,
      "loss": 1.3154,
      "step": 580
    },
    {
      "epoch": 0.17141196978500872,
      "grad_norm": 87.5,
      "learning_rate": 9.279186320754717e-06,
      "loss": 1.3715,
      "step": 590
    },
    {
      "epoch": 0.17431725740848344,
      "grad_norm": 97.0,
      "learning_rate": 9.264445754716982e-06,
      "loss": 1.4323,
      "step": 600
    },
    {
      "epoch": 0.17431725740848344,
      "eval_accuracy": 0.6007730535615682,
      "eval_f1": 0.6010027767957538,
      "eval_loss": 0.6762142777442932,
      "eval_precision": 0.6018480565642976,
      "eval_recall": 0.6007730535615682,
      "eval_runtime": 19.088,
      "eval_samples_per_second": 94.876,
      "eval_steps_per_second": 9.535,
      "step": 600
    },
    {
      "epoch": 0.17722254503195817,
      "grad_norm": 48.25,
      "learning_rate": 9.249705188679246e-06,
      "loss": 1.2719,
      "step": 610
    },
    {
      "epoch": 0.1801278326554329,
      "grad_norm": 125.0,
      "learning_rate": 9.234964622641511e-06,
      "loss": 1.3752,
      "step": 620
    },
    {
      "epoch": 0.1830331202789076,
      "grad_norm": 83.5,
      "learning_rate": 9.220224056603776e-06,
      "loss": 1.3115,
      "step": 630
    },
    {
      "epoch": 0.18593840790238234,
      "grad_norm": 42.25,
      "learning_rate": 9.205483490566038e-06,
      "loss": 1.4751,
      "step": 640
    },
    {
      "epoch": 0.18884369552585706,
      "grad_norm": 116.0,
      "learning_rate": 9.190742924528303e-06,
      "loss": 1.4054,
      "step": 650
    },
    {
      "epoch": 0.19174898314933178,
      "grad_norm": 59.75,
      "learning_rate": 9.176002358490566e-06,
      "loss": 1.2389,
      "step": 660
    },
    {
      "epoch": 0.1946542707728065,
      "grad_norm": 77.5,
      "learning_rate": 9.16126179245283e-06,
      "loss": 1.4555,
      "step": 670
    },
    {
      "epoch": 0.19755955839628123,
      "grad_norm": 28.625,
      "learning_rate": 9.146521226415095e-06,
      "loss": 1.2733,
      "step": 680
    },
    {
      "epoch": 0.20046484601975595,
      "grad_norm": 74.0,
      "learning_rate": 9.13178066037736e-06,
      "loss": 1.4384,
      "step": 690
    },
    {
      "epoch": 0.20337013364323067,
      "grad_norm": 58.75,
      "learning_rate": 9.117040094339623e-06,
      "loss": 1.3355,
      "step": 700
    },
    {
      "epoch": 0.20337013364323067,
      "eval_accuracy": 0.5969077857537273,
      "eval_f1": 0.5967480017714881,
      "eval_loss": 0.6716373562812805,
      "eval_precision": 0.6002412035566317,
      "eval_recall": 0.5969077857537273,
      "eval_runtime": 24.8517,
      "eval_samples_per_second": 72.872,
      "eval_steps_per_second": 7.323,
      "step": 700
    },
    {
      "epoch": 0.2062754212667054,
      "grad_norm": 53.5,
      "learning_rate": 9.102299528301888e-06,
      "loss": 1.371,
      "step": 710
    },
    {
      "epoch": 0.20918070889018012,
      "grad_norm": 71.5,
      "learning_rate": 9.08755896226415e-06,
      "loss": 1.3885,
      "step": 720
    },
    {
      "epoch": 0.21208599651365484,
      "grad_norm": 57.75,
      "learning_rate": 9.072818396226415e-06,
      "loss": 1.286,
      "step": 730
    },
    {
      "epoch": 0.21499128413712956,
      "grad_norm": 89.5,
      "learning_rate": 9.05807783018868e-06,
      "loss": 1.3166,
      "step": 740
    },
    {
      "epoch": 0.2178965717606043,
      "grad_norm": 49.0,
      "learning_rate": 9.043337264150945e-06,
      "loss": 1.446,
      "step": 750
    },
    {
      "epoch": 0.22080185938407904,
      "grad_norm": 48.25,
      "learning_rate": 9.028596698113208e-06,
      "loss": 1.3413,
      "step": 760
    },
    {
      "epoch": 0.22370714700755376,
      "grad_norm": 70.5,
      "learning_rate": 9.013856132075472e-06,
      "loss": 1.4005,
      "step": 770
    },
    {
      "epoch": 0.22661243463102848,
      "grad_norm": 85.0,
      "learning_rate": 8.999115566037737e-06,
      "loss": 1.3292,
      "step": 780
    },
    {
      "epoch": 0.2295177222545032,
      "grad_norm": 44.25,
      "learning_rate": 8.984375000000002e-06,
      "loss": 1.36,
      "step": 790
    },
    {
      "epoch": 0.23242300987797793,
      "grad_norm": 49.5,
      "learning_rate": 8.969634433962266e-06,
      "loss": 1.3268,
      "step": 800
    },
    {
      "epoch": 0.23242300987797793,
      "eval_accuracy": 0.609055770292656,
      "eval_f1": 0.6092814247668411,
      "eval_loss": 0.6692959666252136,
      "eval_precision": 0.610225869236403,
      "eval_recall": 0.609055770292656,
      "eval_runtime": 24.1313,
      "eval_samples_per_second": 75.048,
      "eval_steps_per_second": 7.542,
      "step": 800
    },
    {
      "epoch": 0.23532829750145265,
      "grad_norm": 63.75,
      "learning_rate": 8.95489386792453e-06,
      "loss": 1.3547,
      "step": 810
    },
    {
      "epoch": 0.23823358512492737,
      "grad_norm": 76.5,
      "learning_rate": 8.940153301886794e-06,
      "loss": 1.3306,
      "step": 820
    },
    {
      "epoch": 0.2411388727484021,
      "grad_norm": 105.5,
      "learning_rate": 8.925412735849057e-06,
      "loss": 1.3161,
      "step": 830
    },
    {
      "epoch": 0.24404416037187682,
      "grad_norm": 104.0,
      "learning_rate": 8.910672169811321e-06,
      "loss": 1.3919,
      "step": 840
    },
    {
      "epoch": 0.24694944799535154,
      "grad_norm": 94.0,
      "learning_rate": 8.895931603773586e-06,
      "loss": 1.3846,
      "step": 850
    },
    {
      "epoch": 0.24985473561882626,
      "grad_norm": 62.75,
      "learning_rate": 8.88119103773585e-06,
      "loss": 1.3407,
      "step": 860
    },
    {
      "epoch": 0.252760023242301,
      "grad_norm": 73.5,
      "learning_rate": 8.866450471698114e-06,
      "loss": 1.2485,
      "step": 870
    },
    {
      "epoch": 0.2556653108657757,
      "grad_norm": 66.5,
      "learning_rate": 8.851709905660378e-06,
      "loss": 1.4223,
      "step": 880
    },
    {
      "epoch": 0.25857059848925046,
      "grad_norm": 44.0,
      "learning_rate": 8.836969339622641e-06,
      "loss": 1.3466,
      "step": 890
    },
    {
      "epoch": 0.26147588611272515,
      "grad_norm": 37.0,
      "learning_rate": 8.822228773584906e-06,
      "loss": 1.3865,
      "step": 900
    },
    {
      "epoch": 0.26147588611272515,
      "eval_accuracy": 0.6107123136388736,
      "eval_f1": 0.6104421517436964,
      "eval_loss": 0.6648093461990356,
      "eval_precision": 0.6103362392904956,
      "eval_recall": 0.6107123136388736,
      "eval_runtime": 26.8405,
      "eval_samples_per_second": 67.473,
      "eval_steps_per_second": 6.781,
      "step": 900
    },
    {
      "epoch": 0.2643811737361999,
      "grad_norm": 54.75,
      "learning_rate": 8.80748820754717e-06,
      "loss": 1.4233,
      "step": 910
    },
    {
      "epoch": 0.2672864613596746,
      "grad_norm": 63.75,
      "learning_rate": 8.792747641509435e-06,
      "loss": 1.3615,
      "step": 920
    },
    {
      "epoch": 0.27019174898314935,
      "grad_norm": 55.5,
      "learning_rate": 8.778007075471698e-06,
      "loss": 1.3773,
      "step": 930
    },
    {
      "epoch": 0.27309703660662404,
      "grad_norm": 49.25,
      "learning_rate": 8.763266509433963e-06,
      "loss": 1.4202,
      "step": 940
    },
    {
      "epoch": 0.2760023242300988,
      "grad_norm": 59.25,
      "learning_rate": 8.748525943396226e-06,
      "loss": 1.2954,
      "step": 950
    },
    {
      "epoch": 0.2789076118535735,
      "grad_norm": 43.0,
      "learning_rate": 8.73378537735849e-06,
      "loss": 1.3296,
      "step": 960
    },
    {
      "epoch": 0.28181289947704824,
      "grad_norm": 77.5,
      "learning_rate": 8.719044811320755e-06,
      "loss": 1.364,
      "step": 970
    },
    {
      "epoch": 0.28471818710052293,
      "grad_norm": 88.5,
      "learning_rate": 8.70430424528302e-06,
      "loss": 1.4751,
      "step": 980
    },
    {
      "epoch": 0.2876234747239977,
      "grad_norm": 46.0,
      "learning_rate": 8.689563679245284e-06,
      "loss": 1.471,
      "step": 990
    },
    {
      "epoch": 0.2905287623474724,
      "grad_norm": 44.75,
      "learning_rate": 8.674823113207547e-06,
      "loss": 1.3453,
      "step": 1000
    },
    {
      "epoch": 0.2905287623474724,
      "eval_accuracy": 0.6062948647156268,
      "eval_f1": 0.5996489127755472,
      "eval_loss": 0.6624994874000549,
      "eval_precision": 0.6071509218682795,
      "eval_recall": 0.6062948647156268,
      "eval_runtime": 27.0761,
      "eval_samples_per_second": 66.885,
      "eval_steps_per_second": 6.722,
      "step": 1000
    },
    {
      "epoch": 0.29343404997094713,
      "grad_norm": 52.25,
      "learning_rate": 8.660082547169812e-06,
      "loss": 1.3341,
      "step": 1010
    },
    {
      "epoch": 0.2963393375944218,
      "grad_norm": 34.25,
      "learning_rate": 8.645341981132077e-06,
      "loss": 1.341,
      "step": 1020
    },
    {
      "epoch": 0.2992446252178966,
      "grad_norm": 50.75,
      "learning_rate": 8.630601415094341e-06,
      "loss": 1.3216,
      "step": 1030
    },
    {
      "epoch": 0.30214991284137127,
      "grad_norm": 35.75,
      "learning_rate": 8.615860849056604e-06,
      "loss": 1.3186,
      "step": 1040
    },
    {
      "epoch": 0.305055200464846,
      "grad_norm": 30.0,
      "learning_rate": 8.601120283018869e-06,
      "loss": 1.3138,
      "step": 1050
    },
    {
      "epoch": 0.3079604880883207,
      "grad_norm": 44.5,
      "learning_rate": 8.586379716981132e-06,
      "loss": 1.2962,
      "step": 1060
    },
    {
      "epoch": 0.31086577571179547,
      "grad_norm": 46.5,
      "learning_rate": 8.571639150943397e-06,
      "loss": 1.3494,
      "step": 1070
    },
    {
      "epoch": 0.3137710633352702,
      "grad_norm": 64.5,
      "learning_rate": 8.556898584905661e-06,
      "loss": 1.2946,
      "step": 1080
    },
    {
      "epoch": 0.3166763509587449,
      "grad_norm": 87.0,
      "learning_rate": 8.542158018867926e-06,
      "loss": 1.2792,
      "step": 1090
    },
    {
      "epoch": 0.31958163858221966,
      "grad_norm": 37.75,
      "learning_rate": 8.527417452830189e-06,
      "loss": 1.3412,
      "step": 1100
    },
    {
      "epoch": 0.31958163858221966,
      "eval_accuracy": 0.6250690226394258,
      "eval_f1": 0.6234416564785212,
      "eval_loss": 0.6603183150291443,
      "eval_precision": 0.6244569183256099,
      "eval_recall": 0.6250690226394258,
      "eval_runtime": 23.7723,
      "eval_samples_per_second": 76.181,
      "eval_steps_per_second": 7.656,
      "step": 1100
    },
    {
      "epoch": 0.32248692620569436,
      "grad_norm": 48.75,
      "learning_rate": 8.512676886792454e-06,
      "loss": 1.3886,
      "step": 1110
    },
    {
      "epoch": 0.3253922138291691,
      "grad_norm": 46.25,
      "learning_rate": 8.497936320754717e-06,
      "loss": 1.4322,
      "step": 1120
    },
    {
      "epoch": 0.3282975014526438,
      "grad_norm": 58.75,
      "learning_rate": 8.483195754716981e-06,
      "loss": 1.3484,
      "step": 1130
    },
    {
      "epoch": 0.33120278907611855,
      "grad_norm": 42.0,
      "learning_rate": 8.468455188679246e-06,
      "loss": 1.364,
      "step": 1140
    },
    {
      "epoch": 0.33410807669959325,
      "grad_norm": 58.75,
      "learning_rate": 8.45371462264151e-06,
      "loss": 1.4051,
      "step": 1150
    },
    {
      "epoch": 0.337013364323068,
      "grad_norm": 45.0,
      "learning_rate": 8.438974056603775e-06,
      "loss": 1.3447,
      "step": 1160
    },
    {
      "epoch": 0.3399186519465427,
      "grad_norm": 45.5,
      "learning_rate": 8.424233490566038e-06,
      "loss": 1.3961,
      "step": 1170
    },
    {
      "epoch": 0.34282393957001744,
      "grad_norm": 68.5,
      "learning_rate": 8.409492924528303e-06,
      "loss": 1.3798,
      "step": 1180
    },
    {
      "epoch": 0.34572922719349214,
      "grad_norm": 55.25,
      "learning_rate": 8.394752358490566e-06,
      "loss": 1.3101,
      "step": 1190
    },
    {
      "epoch": 0.3486345148169669,
      "grad_norm": 37.75,
      "learning_rate": 8.380011792452832e-06,
      "loss": 1.3918,
      "step": 1200
    },
    {
      "epoch": 0.3486345148169669,
      "eval_accuracy": 0.6024295969077857,
      "eval_f1": 0.6026431424760783,
      "eval_loss": 0.6649277210235596,
      "eval_precision": 0.6032326068057581,
      "eval_recall": 0.6024295969077857,
      "eval_runtime": 17.6379,
      "eval_samples_per_second": 102.677,
      "eval_steps_per_second": 10.319,
      "step": 1200
    },
    {
      "epoch": 0.3515398024404416,
      "grad_norm": 30.625,
      "learning_rate": 8.365271226415095e-06,
      "loss": 1.3555,
      "step": 1210
    },
    {
      "epoch": 0.35444509006391633,
      "grad_norm": 33.75,
      "learning_rate": 8.35053066037736e-06,
      "loss": 1.2323,
      "step": 1220
    },
    {
      "epoch": 0.35735037768739103,
      "grad_norm": 30.625,
      "learning_rate": 8.335790094339623e-06,
      "loss": 1.3533,
      "step": 1230
    },
    {
      "epoch": 0.3602556653108658,
      "grad_norm": 101.0,
      "learning_rate": 8.321049528301887e-06,
      "loss": 1.3743,
      "step": 1240
    },
    {
      "epoch": 0.3631609529343405,
      "grad_norm": 38.5,
      "learning_rate": 8.306308962264152e-06,
      "loss": 1.3463,
      "step": 1250
    },
    {
      "epoch": 0.3660662405578152,
      "grad_norm": 48.0,
      "learning_rate": 8.291568396226417e-06,
      "loss": 1.2404,
      "step": 1260
    },
    {
      "epoch": 0.3689715281812899,
      "grad_norm": 101.0,
      "learning_rate": 8.27682783018868e-06,
      "loss": 1.3486,
      "step": 1270
    },
    {
      "epoch": 0.37187681580476467,
      "grad_norm": 30.875,
      "learning_rate": 8.262087264150944e-06,
      "loss": 1.3636,
      "step": 1280
    },
    {
      "epoch": 0.3747821034282394,
      "grad_norm": 34.25,
      "learning_rate": 8.247346698113207e-06,
      "loss": 1.2899,
      "step": 1290
    },
    {
      "epoch": 0.3776873910517141,
      "grad_norm": 60.5,
      "learning_rate": 8.232606132075472e-06,
      "loss": 1.2714,
      "step": 1300
    },
    {
      "epoch": 0.3776873910517141,
      "eval_accuracy": 0.6272777471010491,
      "eval_f1": 0.6221601134611754,
      "eval_loss": 0.6585363745689392,
      "eval_precision": 0.6285042745519395,
      "eval_recall": 0.6272777471010491,
      "eval_runtime": 26.4941,
      "eval_samples_per_second": 68.355,
      "eval_steps_per_second": 6.869,
      "step": 1300
    },
    {
      "epoch": 0.38059267867518887,
      "grad_norm": 41.5,
      "learning_rate": 8.217865566037737e-06,
      "loss": 1.3589,
      "step": 1310
    },
    {
      "epoch": 0.38349796629866356,
      "grad_norm": 65.5,
      "learning_rate": 8.203125000000001e-06,
      "loss": 1.3503,
      "step": 1320
    },
    {
      "epoch": 0.3864032539221383,
      "grad_norm": 76.5,
      "learning_rate": 8.188384433962266e-06,
      "loss": 1.2889,
      "step": 1330
    },
    {
      "epoch": 0.389308541545613,
      "grad_norm": 83.0,
      "learning_rate": 8.173643867924529e-06,
      "loss": 1.4435,
      "step": 1340
    },
    {
      "epoch": 0.39221382916908776,
      "grad_norm": 46.5,
      "learning_rate": 8.158903301886793e-06,
      "loss": 1.3266,
      "step": 1350
    },
    {
      "epoch": 0.39511911679256245,
      "grad_norm": 50.75,
      "learning_rate": 8.144162735849056e-06,
      "loss": 1.2778,
      "step": 1360
    },
    {
      "epoch": 0.3980244044160372,
      "grad_norm": 59.0,
      "learning_rate": 8.129422169811321e-06,
      "loss": 1.3124,
      "step": 1370
    },
    {
      "epoch": 0.4009296920395119,
      "grad_norm": 63.5,
      "learning_rate": 8.114681603773586e-06,
      "loss": 1.2415,
      "step": 1380
    },
    {
      "epoch": 0.40383497966298665,
      "grad_norm": 67.5,
      "learning_rate": 8.09994103773585e-06,
      "loss": 1.3139,
      "step": 1390
    },
    {
      "epoch": 0.40674026728646134,
      "grad_norm": 45.0,
      "learning_rate": 8.085200471698113e-06,
      "loss": 1.2793,
      "step": 1400
    },
    {
      "epoch": 0.40674026728646134,
      "eval_accuracy": 0.6123688569850911,
      "eval_f1": 0.6125918781610761,
      "eval_loss": 0.6618815660476685,
      "eval_precision": 0.6136048525390123,
      "eval_recall": 0.6123688569850911,
      "eval_runtime": 25.4254,
      "eval_samples_per_second": 71.228,
      "eval_steps_per_second": 7.158,
      "step": 1400
    },
    {
      "epoch": 0.4096455549099361,
      "grad_norm": 135.0,
      "learning_rate": 8.070459905660378e-06,
      "loss": 1.4438,
      "step": 1410
    },
    {
      "epoch": 0.4125508425334108,
      "grad_norm": 55.75,
      "learning_rate": 8.055719339622643e-06,
      "loss": 1.3361,
      "step": 1420
    },
    {
      "epoch": 0.41545613015688554,
      "grad_norm": 97.0,
      "learning_rate": 8.040978773584907e-06,
      "loss": 1.2857,
      "step": 1430
    },
    {
      "epoch": 0.41836141778036023,
      "grad_norm": 97.0,
      "learning_rate": 8.02623820754717e-06,
      "loss": 1.2567,
      "step": 1440
    },
    {
      "epoch": 0.421266705403835,
      "grad_norm": 41.0,
      "learning_rate": 8.011497641509435e-06,
      "loss": 1.3682,
      "step": 1450
    },
    {
      "epoch": 0.4241719930273097,
      "grad_norm": 74.5,
      "learning_rate": 7.996757075471698e-06,
      "loss": 1.2763,
      "step": 1460
    },
    {
      "epoch": 0.42707728065078443,
      "grad_norm": 37.0,
      "learning_rate": 7.982016509433963e-06,
      "loss": 1.3787,
      "step": 1470
    },
    {
      "epoch": 0.4299825682742591,
      "grad_norm": 32.25,
      "learning_rate": 7.967275943396227e-06,
      "loss": 1.2549,
      "step": 1480
    },
    {
      "epoch": 0.4328878558977339,
      "grad_norm": 84.5,
      "learning_rate": 7.952535377358492e-06,
      "loss": 1.3206,
      "step": 1490
    },
    {
      "epoch": 0.4357931435212086,
      "grad_norm": 69.0,
      "learning_rate": 7.937794811320757e-06,
      "loss": 1.41,
      "step": 1500
    },
    {
      "epoch": 0.4357931435212086,
      "eval_accuracy": 0.627829928216455,
      "eval_f1": 0.6269606515260162,
      "eval_loss": 0.652948796749115,
      "eval_precision": 0.6272111466552719,
      "eval_recall": 0.627829928216455,
      "eval_runtime": 21.9245,
      "eval_samples_per_second": 82.602,
      "eval_steps_per_second": 8.301,
      "step": 1500
    },
    {
      "epoch": 0.4386984311446833,
      "grad_norm": 38.5,
      "learning_rate": 7.92305424528302e-06,
      "loss": 1.372,
      "step": 1510
    },
    {
      "epoch": 0.44160371876815807,
      "grad_norm": 64.5,
      "learning_rate": 7.908313679245284e-06,
      "loss": 1.3912,
      "step": 1520
    },
    {
      "epoch": 0.44450900639163277,
      "grad_norm": 62.5,
      "learning_rate": 7.893573113207547e-06,
      "loss": 1.3164,
      "step": 1530
    },
    {
      "epoch": 0.4474142940151075,
      "grad_norm": 40.75,
      "learning_rate": 7.878832547169812e-06,
      "loss": 1.3526,
      "step": 1540
    },
    {
      "epoch": 0.4503195816385822,
      "grad_norm": 36.25,
      "learning_rate": 7.864091981132076e-06,
      "loss": 1.3417,
      "step": 1550
    },
    {
      "epoch": 0.45322486926205696,
      "grad_norm": 38.5,
      "learning_rate": 7.849351415094341e-06,
      "loss": 1.4355,
      "step": 1560
    },
    {
      "epoch": 0.45613015688553166,
      "grad_norm": 44.75,
      "learning_rate": 7.834610849056604e-06,
      "loss": 1.3182,
      "step": 1570
    },
    {
      "epoch": 0.4590354445090064,
      "grad_norm": 32.5,
      "learning_rate": 7.819870283018869e-06,
      "loss": 1.2961,
      "step": 1580
    },
    {
      "epoch": 0.4619407321324811,
      "grad_norm": 45.75,
      "learning_rate": 7.805129716981132e-06,
      "loss": 1.2738,
      "step": 1590
    },
    {
      "epoch": 0.46484601975595585,
      "grad_norm": 38.5,
      "learning_rate": 7.790389150943396e-06,
      "loss": 1.3503,
      "step": 1600
    },
    {
      "epoch": 0.46484601975595585,
      "eval_accuracy": 0.6223081170623964,
      "eval_f1": 0.6190936146728951,
      "eval_loss": 0.6521285772323608,
      "eval_precision": 0.6221827660719746,
      "eval_recall": 0.6223081170623964,
      "eval_runtime": 25.4159,
      "eval_samples_per_second": 71.255,
      "eval_steps_per_second": 7.161,
      "step": 1600
    },
    {
      "epoch": 0.46775130737943055,
      "grad_norm": 58.0,
      "learning_rate": 7.775648584905661e-06,
      "loss": 1.2569,
      "step": 1610
    },
    {
      "epoch": 0.4706565950029053,
      "grad_norm": 87.5,
      "learning_rate": 7.760908018867926e-06,
      "loss": 1.3135,
      "step": 1620
    },
    {
      "epoch": 0.47356188262638,
      "grad_norm": 98.0,
      "learning_rate": 7.746167452830189e-06,
      "loss": 1.3008,
      "step": 1630
    },
    {
      "epoch": 0.47646717024985474,
      "grad_norm": 31.25,
      "learning_rate": 7.731426886792453e-06,
      "loss": 1.2975,
      "step": 1640
    },
    {
      "epoch": 0.47937245787332944,
      "grad_norm": 103.5,
      "learning_rate": 7.716686320754718e-06,
      "loss": 1.2854,
      "step": 1650
    },
    {
      "epoch": 0.4822777454968042,
      "grad_norm": 41.25,
      "learning_rate": 7.701945754716983e-06,
      "loss": 1.2613,
      "step": 1660
    },
    {
      "epoch": 0.4851830331202789,
      "grad_norm": 47.25,
      "learning_rate": 7.687205188679246e-06,
      "loss": 1.3469,
      "step": 1670
    },
    {
      "epoch": 0.48808832074375363,
      "grad_norm": 36.5,
      "learning_rate": 7.67246462264151e-06,
      "loss": 1.3195,
      "step": 1680
    },
    {
      "epoch": 0.49099360836722833,
      "grad_norm": 77.5,
      "learning_rate": 7.657724056603775e-06,
      "loss": 1.3431,
      "step": 1690
    },
    {
      "epoch": 0.4938988959907031,
      "grad_norm": 38.25,
      "learning_rate": 7.642983490566038e-06,
      "loss": 1.3423,
      "step": 1700
    },
    {
      "epoch": 0.4938988959907031,
      "eval_accuracy": 0.6300386526780785,
      "eval_f1": 0.6163155948087964,
      "eval_loss": 0.656129777431488,
      "eval_precision": 0.6398336533169895,
      "eval_recall": 0.6300386526780785,
      "eval_runtime": 17.7356,
      "eval_samples_per_second": 102.111,
      "eval_steps_per_second": 10.262,
      "step": 1700
    },
    {
      "epoch": 0.49680418361417783,
      "grad_norm": 63.0,
      "learning_rate": 7.628242924528303e-06,
      "loss": 1.2512,
      "step": 1710
    },
    {
      "epoch": 0.4997094712376525,
      "grad_norm": 26.875,
      "learning_rate": 7.613502358490566e-06,
      "loss": 1.2956,
      "step": 1720
    },
    {
      "epoch": 0.5026147588611273,
      "grad_norm": 52.25,
      "learning_rate": 7.598761792452831e-06,
      "loss": 1.4073,
      "step": 1730
    },
    {
      "epoch": 0.505520046484602,
      "grad_norm": 84.0,
      "learning_rate": 7.584021226415095e-06,
      "loss": 1.2805,
      "step": 1740
    },
    {
      "epoch": 0.5084253341080767,
      "grad_norm": 53.75,
      "learning_rate": 7.569280660377359e-06,
      "loss": 1.3359,
      "step": 1750
    },
    {
      "epoch": 0.5113306217315514,
      "grad_norm": 27.125,
      "learning_rate": 7.554540094339623e-06,
      "loss": 1.2798,
      "step": 1760
    },
    {
      "epoch": 0.5142359093550262,
      "grad_norm": 55.5,
      "learning_rate": 7.539799528301888e-06,
      "loss": 1.3784,
      "step": 1770
    },
    {
      "epoch": 0.5171411969785009,
      "grad_norm": 72.5,
      "learning_rate": 7.525058962264152e-06,
      "loss": 1.3408,
      "step": 1780
    },
    {
      "epoch": 0.5200464846019756,
      "grad_norm": 50.5,
      "learning_rate": 7.510318396226416e-06,
      "loss": 1.3178,
      "step": 1790
    },
    {
      "epoch": 0.5229517722254503,
      "grad_norm": 49.5,
      "learning_rate": 7.495577830188679e-06,
      "loss": 1.3394,
      "step": 1800
    },
    {
      "epoch": 0.5229517722254503,
      "eval_accuracy": 0.6178906681391496,
      "eval_f1": 0.6181047441367624,
      "eval_loss": 0.6562695503234863,
      "eval_precision": 0.61880686931104,
      "eval_recall": 0.6178906681391496,
      "eval_runtime": 23.6625,
      "eval_samples_per_second": 76.535,
      "eval_steps_per_second": 7.691,
      "step": 1800
    },
    {
      "epoch": 0.5258570598489251,
      "grad_norm": 34.75,
      "learning_rate": 7.480837264150944e-06,
      "loss": 1.2929,
      "step": 1810
    },
    {
      "epoch": 0.5287623474723998,
      "grad_norm": 45.5,
      "learning_rate": 7.466096698113208e-06,
      "loss": 1.3049,
      "step": 1820
    },
    {
      "epoch": 0.5316676350958744,
      "grad_norm": 48.75,
      "learning_rate": 7.451356132075472e-06,
      "loss": 1.2783,
      "step": 1830
    },
    {
      "epoch": 0.5345729227193492,
      "grad_norm": 83.0,
      "learning_rate": 7.436615566037736e-06,
      "loss": 1.2851,
      "step": 1840
    },
    {
      "epoch": 0.537478210342824,
      "grad_norm": 83.5,
      "learning_rate": 7.421875000000001e-06,
      "loss": 1.3286,
      "step": 1850
    },
    {
      "epoch": 0.5403834979662987,
      "grad_norm": 40.25,
      "learning_rate": 7.4071344339622655e-06,
      "loss": 1.3461,
      "step": 1860
    },
    {
      "epoch": 0.5432887855897733,
      "grad_norm": 38.75,
      "learning_rate": 7.3923938679245284e-06,
      "loss": 1.2203,
      "step": 1870
    },
    {
      "epoch": 0.5461940732132481,
      "grad_norm": 46.75,
      "learning_rate": 7.377653301886793e-06,
      "loss": 1.2875,
      "step": 1880
    },
    {
      "epoch": 0.5490993608367228,
      "grad_norm": 28.125,
      "learning_rate": 7.362912735849057e-06,
      "loss": 1.3201,
      "step": 1890
    },
    {
      "epoch": 0.5520046484601976,
      "grad_norm": 61.0,
      "learning_rate": 7.3481721698113216e-06,
      "loss": 1.2116,
      "step": 1900
    },
    {
      "epoch": 0.5520046484601976,
      "eval_accuracy": 0.6217559359469906,
      "eval_f1": 0.6041253940543542,
      "eval_loss": 0.6534772515296936,
      "eval_precision": 0.6342473819530806,
      "eval_recall": 0.6217559359469906,
      "eval_runtime": 22.658,
      "eval_samples_per_second": 79.928,
      "eval_steps_per_second": 8.032,
      "step": 1900
    },
    {
      "epoch": 0.5549099360836722,
      "grad_norm": 83.0,
      "learning_rate": 7.333431603773585e-06,
      "loss": 1.2667,
      "step": 1910
    },
    {
      "epoch": 0.557815223707147,
      "grad_norm": 42.5,
      "learning_rate": 7.31869103773585e-06,
      "loss": 1.277,
      "step": 1920
    },
    {
      "epoch": 0.5607205113306217,
      "grad_norm": 53.0,
      "learning_rate": 7.303950471698114e-06,
      "loss": 1.3001,
      "step": 1930
    },
    {
      "epoch": 0.5636257989540965,
      "grad_norm": 44.75,
      "learning_rate": 7.2892099056603785e-06,
      "loss": 1.2845,
      "step": 1940
    },
    {
      "epoch": 0.5665310865775712,
      "grad_norm": 47.0,
      "learning_rate": 7.2744693396226415e-06,
      "loss": 1.35,
      "step": 1950
    },
    {
      "epoch": 0.5694363742010459,
      "grad_norm": 58.75,
      "learning_rate": 7.259728773584906e-06,
      "loss": 1.3458,
      "step": 1960
    },
    {
      "epoch": 0.5723416618245206,
      "grad_norm": 42.0,
      "learning_rate": 7.24498820754717e-06,
      "loss": 1.3163,
      "step": 1970
    },
    {
      "epoch": 0.5752469494479954,
      "grad_norm": 81.0,
      "learning_rate": 7.230247641509435e-06,
      "loss": 1.3481,
      "step": 1980
    },
    {
      "epoch": 0.5781522370714701,
      "grad_norm": 57.75,
      "learning_rate": 7.215507075471698e-06,
      "loss": 1.1711,
      "step": 1990
    },
    {
      "epoch": 0.5810575246949448,
      "grad_norm": 33.5,
      "learning_rate": 7.200766509433963e-06,
      "loss": 1.3834,
      "step": 2000
    },
    {
      "epoch": 0.5810575246949448,
      "eval_accuracy": 0.6355604638321369,
      "eval_f1": 0.6342784909550675,
      "eval_loss": 0.650558352470398,
      "eval_precision": 0.6350165949733959,
      "eval_recall": 0.6355604638321369,
      "eval_runtime": 21.9148,
      "eval_samples_per_second": 82.638,
      "eval_steps_per_second": 8.305,
      "step": 2000
    },
    {
      "epoch": 0.5839628123184195,
      "grad_norm": 46.0,
      "learning_rate": 7.186025943396227e-06,
      "loss": 1.3424,
      "step": 2010
    },
    {
      "epoch": 0.5868680999418943,
      "grad_norm": 30.0,
      "learning_rate": 7.1712853773584915e-06,
      "loss": 1.2667,
      "step": 2020
    },
    {
      "epoch": 0.589773387565369,
      "grad_norm": 35.0,
      "learning_rate": 7.156544811320756e-06,
      "loss": 1.2963,
      "step": 2030
    },
    {
      "epoch": 0.5926786751888437,
      "grad_norm": 26.25,
      "learning_rate": 7.141804245283019e-06,
      "loss": 1.2806,
      "step": 2040
    },
    {
      "epoch": 0.5955839628123184,
      "grad_norm": 35.5,
      "learning_rate": 7.127063679245284e-06,
      "loss": 1.2645,
      "step": 2050
    },
    {
      "epoch": 0.5984892504357932,
      "grad_norm": 54.75,
      "learning_rate": 7.112323113207548e-06,
      "loss": 1.3138,
      "step": 2060
    },
    {
      "epoch": 0.6013945380592679,
      "grad_norm": 38.0,
      "learning_rate": 7.097582547169812e-06,
      "loss": 1.2769,
      "step": 2070
    },
    {
      "epoch": 0.6042998256827425,
      "grad_norm": 50.75,
      "learning_rate": 7.082841981132076e-06,
      "loss": 1.3197,
      "step": 2080
    },
    {
      "epoch": 0.6072051133062173,
      "grad_norm": 52.5,
      "learning_rate": 7.068101415094341e-06,
      "loss": 1.363,
      "step": 2090
    },
    {
      "epoch": 0.610110400929692,
      "grad_norm": 29.5,
      "learning_rate": 7.0533608490566045e-06,
      "loss": 1.2125,
      "step": 2100
    },
    {
      "epoch": 0.610110400929692,
      "eval_accuracy": 0.6300386526780785,
      "eval_f1": 0.621290753411316,
      "eval_loss": 0.6497138738632202,
      "eval_precision": 0.6345540000240397,
      "eval_recall": 0.6300386526780785,
      "eval_runtime": 27.7006,
      "eval_samples_per_second": 65.378,
      "eval_steps_per_second": 6.57,
      "step": 2100
    },
    {
      "epoch": 0.6130156885531668,
      "grad_norm": 56.25,
      "learning_rate": 7.038620283018869e-06,
      "loss": 1.3201,
      "step": 2110
    },
    {
      "epoch": 0.6159209761766414,
      "grad_norm": 57.0,
      "learning_rate": 7.023879716981132e-06,
      "loss": 1.444,
      "step": 2120
    },
    {
      "epoch": 0.6188262638001162,
      "grad_norm": 52.5,
      "learning_rate": 7.009139150943397e-06,
      "loss": 1.3887,
      "step": 2130
    },
    {
      "epoch": 0.6217315514235909,
      "grad_norm": 51.25,
      "learning_rate": 6.994398584905661e-06,
      "loss": 1.3022,
      "step": 2140
    },
    {
      "epoch": 0.6246368390470657,
      "grad_norm": 25.25,
      "learning_rate": 6.979658018867925e-06,
      "loss": 1.3132,
      "step": 2150
    },
    {
      "epoch": 0.6275421266705404,
      "grad_norm": 45.25,
      "learning_rate": 6.964917452830189e-06,
      "loss": 1.2461,
      "step": 2160
    },
    {
      "epoch": 0.6304474142940151,
      "grad_norm": 27.125,
      "learning_rate": 6.950176886792454e-06,
      "loss": 1.2606,
      "step": 2170
    },
    {
      "epoch": 0.6333527019174898,
      "grad_norm": 52.0,
      "learning_rate": 6.935436320754717e-06,
      "loss": 1.2901,
      "step": 2180
    },
    {
      "epoch": 0.6362579895409646,
      "grad_norm": 32.75,
      "learning_rate": 6.920695754716981e-06,
      "loss": 1.3803,
      "step": 2190
    },
    {
      "epoch": 0.6391632771644393,
      "grad_norm": 34.25,
      "learning_rate": 6.905955188679245e-06,
      "loss": 1.2805,
      "step": 2200
    },
    {
      "epoch": 0.6391632771644393,
      "eval_accuracy": 0.6311430149088901,
      "eval_f1": 0.6298455029666443,
      "eval_loss": 0.648868203163147,
      "eval_precision": 0.630553601839031,
      "eval_recall": 0.6311430149088901,
      "eval_runtime": 23.3959,
      "eval_samples_per_second": 77.407,
      "eval_steps_per_second": 7.779,
      "step": 2200
    },
    {
      "epoch": 0.642068564787914,
      "grad_norm": 45.25,
      "learning_rate": 6.89121462264151e-06,
      "loss": 1.3036,
      "step": 2210
    },
    {
      "epoch": 0.6449738524113887,
      "grad_norm": 54.0,
      "learning_rate": 6.8764740566037745e-06,
      "loss": 1.1919,
      "step": 2220
    },
    {
      "epoch": 0.6478791400348635,
      "grad_norm": 37.25,
      "learning_rate": 6.861733490566038e-06,
      "loss": 1.4039,
      "step": 2230
    },
    {
      "epoch": 0.6507844276583382,
      "grad_norm": 71.0,
      "learning_rate": 6.846992924528303e-06,
      "loss": 1.3153,
      "step": 2240
    },
    {
      "epoch": 0.6536897152818129,
      "grad_norm": 85.0,
      "learning_rate": 6.832252358490567e-06,
      "loss": 1.3351,
      "step": 2250
    },
    {
      "epoch": 0.6565950029052876,
      "grad_norm": 54.0,
      "learning_rate": 6.817511792452831e-06,
      "loss": 1.2602,
      "step": 2260
    },
    {
      "epoch": 0.6595002905287624,
      "grad_norm": 101.5,
      "learning_rate": 6.802771226415094e-06,
      "loss": 1.2256,
      "step": 2270
    },
    {
      "epoch": 0.6624055781522371,
      "grad_norm": 80.5,
      "learning_rate": 6.788030660377359e-06,
      "loss": 1.3267,
      "step": 2280
    },
    {
      "epoch": 0.6653108657757117,
      "grad_norm": 48.25,
      "learning_rate": 6.773290094339623e-06,
      "loss": 1.356,
      "step": 2290
    },
    {
      "epoch": 0.6682161533991865,
      "grad_norm": 33.25,
      "learning_rate": 6.7585495283018875e-06,
      "loss": 1.3723,
      "step": 2300
    },
    {
      "epoch": 0.6682161533991865,
      "eval_accuracy": 0.6322473771397018,
      "eval_f1": 0.6307364826527131,
      "eval_loss": 0.646483302116394,
      "eval_precision": 0.6317087816843026,
      "eval_recall": 0.6322473771397018,
      "eval_runtime": 17.5393,
      "eval_samples_per_second": 103.254,
      "eval_steps_per_second": 10.377,
      "step": 2300
    },
    {
      "epoch": 0.6711214410226612,
      "grad_norm": 100.0,
      "learning_rate": 6.743808962264151e-06,
      "loss": 1.3795,
      "step": 2310
    },
    {
      "epoch": 0.674026728646136,
      "grad_norm": 32.0,
      "learning_rate": 6.729068396226416e-06,
      "loss": 1.3164,
      "step": 2320
    },
    {
      "epoch": 0.6769320162696106,
      "grad_norm": 57.25,
      "learning_rate": 6.71432783018868e-06,
      "loss": 1.3095,
      "step": 2330
    },
    {
      "epoch": 0.6798373038930854,
      "grad_norm": 43.5,
      "learning_rate": 6.699587264150944e-06,
      "loss": 1.3305,
      "step": 2340
    },
    {
      "epoch": 0.6827425915165601,
      "grad_norm": 59.5,
      "learning_rate": 6.684846698113207e-06,
      "loss": 1.3211,
      "step": 2350
    },
    {
      "epoch": 0.6856478791400349,
      "grad_norm": 38.75,
      "learning_rate": 6.670106132075472e-06,
      "loss": 1.2753,
      "step": 2360
    },
    {
      "epoch": 0.6885531667635096,
      "grad_norm": 102.5,
      "learning_rate": 6.655365566037736e-06,
      "loss": 1.2837,
      "step": 2370
    },
    {
      "epoch": 0.6914584543869843,
      "grad_norm": 34.25,
      "learning_rate": 6.6406250000000005e-06,
      "loss": 1.2707,
      "step": 2380
    },
    {
      "epoch": 0.694363742010459,
      "grad_norm": 59.75,
      "learning_rate": 6.625884433962265e-06,
      "loss": 1.2989,
      "step": 2390
    },
    {
      "epoch": 0.6972690296339338,
      "grad_norm": 58.0,
      "learning_rate": 6.611143867924529e-06,
      "loss": 1.3785,
      "step": 2400
    },
    {
      "epoch": 0.6972690296339338,
      "eval_accuracy": 0.635008282716731,
      "eval_f1": 0.633801429811518,
      "eval_loss": 0.6450901627540588,
      "eval_precision": 0.6344474468599942,
      "eval_recall": 0.635008282716731,
      "eval_runtime": 27.9994,
      "eval_samples_per_second": 64.68,
      "eval_steps_per_second": 6.5,
      "step": 2400
    },
    {
      "epoch": 0.7001743172574085,
      "grad_norm": 62.25,
      "learning_rate": 6.596403301886794e-06,
      "loss": 1.3779,
      "step": 2410
    },
    {
      "epoch": 0.7030796048808832,
      "grad_norm": 35.25,
      "learning_rate": 6.581662735849057e-06,
      "loss": 1.3529,
      "step": 2420
    },
    {
      "epoch": 0.7059848925043579,
      "grad_norm": 34.25,
      "learning_rate": 6.566922169811322e-06,
      "loss": 1.2754,
      "step": 2430
    },
    {
      "epoch": 0.7088901801278327,
      "grad_norm": 52.0,
      "learning_rate": 6.552181603773585e-06,
      "loss": 1.3055,
      "step": 2440
    },
    {
      "epoch": 0.7117954677513074,
      "grad_norm": 43.0,
      "learning_rate": 6.53744103773585e-06,
      "loss": 1.2766,
      "step": 2450
    },
    {
      "epoch": 0.7147007553747821,
      "grad_norm": 28.75,
      "learning_rate": 6.5227004716981135e-06,
      "loss": 1.293,
      "step": 2460
    },
    {
      "epoch": 0.7176060429982568,
      "grad_norm": 33.25,
      "learning_rate": 6.507959905660378e-06,
      "loss": 1.324,
      "step": 2470
    },
    {
      "epoch": 0.7205113306217316,
      "grad_norm": 26.5,
      "learning_rate": 6.493219339622642e-06,
      "loss": 1.2177,
      "step": 2480
    },
    {
      "epoch": 0.7234166182452063,
      "grad_norm": 65.5,
      "learning_rate": 6.478478773584907e-06,
      "loss": 1.3648,
      "step": 2490
    },
    {
      "epoch": 0.726321905868681,
      "grad_norm": 58.0,
      "learning_rate": 6.4637382075471696e-06,
      "loss": 1.2685,
      "step": 2500
    },
    {
      "epoch": 0.726321905868681,
      "eval_accuracy": 0.6068470458310325,
      "eval_f1": 0.6061765727272004,
      "eval_loss": 0.6521552801132202,
      "eval_precision": 0.6118061263264858,
      "eval_recall": 0.6068470458310325,
      "eval_runtime": 23.9512,
      "eval_samples_per_second": 75.612,
      "eval_steps_per_second": 7.599,
      "step": 2500
    },
    {
      "epoch": 0.7292271934921557,
      "grad_norm": 73.5,
      "learning_rate": 6.448997641509434e-06,
      "loss": 1.2741,
      "step": 2510
    },
    {
      "epoch": 0.7321324811156305,
      "grad_norm": 25.125,
      "learning_rate": 6.434257075471698e-06,
      "loss": 1.3256,
      "step": 2520
    },
    {
      "epoch": 0.7350377687391052,
      "grad_norm": 48.5,
      "learning_rate": 6.419516509433963e-06,
      "loss": 1.3056,
      "step": 2530
    },
    {
      "epoch": 0.7379430563625798,
      "grad_norm": 41.0,
      "learning_rate": 6.4047759433962265e-06,
      "loss": 1.2686,
      "step": 2540
    },
    {
      "epoch": 0.7408483439860546,
      "grad_norm": 47.0,
      "learning_rate": 6.390035377358491e-06,
      "loss": 1.3235,
      "step": 2550
    },
    {
      "epoch": 0.7437536316095293,
      "grad_norm": 62.25,
      "learning_rate": 6.375294811320756e-06,
      "loss": 1.3393,
      "step": 2560
    },
    {
      "epoch": 0.7466589192330041,
      "grad_norm": 42.75,
      "learning_rate": 6.36055424528302e-06,
      "loss": 1.2951,
      "step": 2570
    },
    {
      "epoch": 0.7495642068564788,
      "grad_norm": 59.5,
      "learning_rate": 6.345813679245284e-06,
      "loss": 1.3177,
      "step": 2580
    },
    {
      "epoch": 0.7524694944799535,
      "grad_norm": 56.25,
      "learning_rate": 6.331073113207547e-06,
      "loss": 1.1857,
      "step": 2590
    },
    {
      "epoch": 0.7553747821034282,
      "grad_norm": 83.5,
      "learning_rate": 6.316332547169812e-06,
      "loss": 1.3168,
      "step": 2600
    },
    {
      "epoch": 0.7553747821034282,
      "eval_accuracy": 0.6267255659856433,
      "eval_f1": 0.6263898768518164,
      "eval_loss": 0.6455510854721069,
      "eval_precision": 0.626312724736393,
      "eval_recall": 0.6267255659856433,
      "eval_runtime": 22.9922,
      "eval_samples_per_second": 78.766,
      "eval_steps_per_second": 7.916,
      "step": 2600
    },
    {
      "epoch": 0.758280069726903,
      "grad_norm": 65.5,
      "learning_rate": 6.301591981132076e-06,
      "loss": 1.2172,
      "step": 2610
    },
    {
      "epoch": 0.7611853573503777,
      "grad_norm": 44.5,
      "learning_rate": 6.28685141509434e-06,
      "loss": 1.3155,
      "step": 2620
    },
    {
      "epoch": 0.7640906449738524,
      "grad_norm": 36.0,
      "learning_rate": 6.272110849056604e-06,
      "loss": 1.3665,
      "step": 2630
    },
    {
      "epoch": 0.7669959325973271,
      "grad_norm": 44.75,
      "learning_rate": 6.257370283018869e-06,
      "loss": 1.2973,
      "step": 2640
    },
    {
      "epoch": 0.7699012202208019,
      "grad_norm": 63.25,
      "learning_rate": 6.242629716981133e-06,
      "loss": 1.2979,
      "step": 2650
    },
    {
      "epoch": 0.7728065078442766,
      "grad_norm": 33.5,
      "learning_rate": 6.227889150943397e-06,
      "loss": 1.3267,
      "step": 2660
    },
    {
      "epoch": 0.7757117954677513,
      "grad_norm": 46.5,
      "learning_rate": 6.21314858490566e-06,
      "loss": 1.3511,
      "step": 2670
    },
    {
      "epoch": 0.778617083091226,
      "grad_norm": 39.0,
      "learning_rate": 6.198408018867925e-06,
      "loss": 1.3943,
      "step": 2680
    },
    {
      "epoch": 0.7815223707147008,
      "grad_norm": 34.5,
      "learning_rate": 6.183667452830189e-06,
      "loss": 1.2841,
      "step": 2690
    },
    {
      "epoch": 0.7844276583381755,
      "grad_norm": 41.5,
      "learning_rate": 6.168926886792453e-06,
      "loss": 1.3644,
      "step": 2700
    },
    {
      "epoch": 0.7844276583381755,
      "eval_accuracy": 0.6344561016013253,
      "eval_f1": 0.6324778464405518,
      "eval_loss": 0.643452525138855,
      "eval_precision": 0.6340971076510771,
      "eval_recall": 0.6344561016013253,
      "eval_runtime": 26.9644,
      "eval_samples_per_second": 67.163,
      "eval_steps_per_second": 6.75,
      "step": 2700
    },
    {
      "epoch": 0.7873329459616502,
      "grad_norm": 64.5,
      "learning_rate": 6.154186320754717e-06,
      "loss": 1.3015,
      "step": 2710
    },
    {
      "epoch": 0.7902382335851249,
      "grad_norm": 93.0,
      "learning_rate": 6.139445754716982e-06,
      "loss": 1.3085,
      "step": 2720
    },
    {
      "epoch": 0.7931435212085997,
      "grad_norm": 34.5,
      "learning_rate": 6.124705188679246e-06,
      "loss": 1.2709,
      "step": 2730
    },
    {
      "epoch": 0.7960488088320744,
      "grad_norm": 36.5,
      "learning_rate": 6.10996462264151e-06,
      "loss": 1.2619,
      "step": 2740
    },
    {
      "epoch": 0.798954096455549,
      "grad_norm": 58.75,
      "learning_rate": 6.095224056603775e-06,
      "loss": 1.2363,
      "step": 2750
    },
    {
      "epoch": 0.8018593840790238,
      "grad_norm": 54.75,
      "learning_rate": 6.080483490566038e-06,
      "loss": 1.3241,
      "step": 2760
    },
    {
      "epoch": 0.8047646717024985,
      "grad_norm": 62.0,
      "learning_rate": 6.0657429245283026e-06,
      "loss": 1.2894,
      "step": 2770
    },
    {
      "epoch": 0.8076699593259733,
      "grad_norm": 38.0,
      "learning_rate": 6.051002358490566e-06,
      "loss": 1.2915,
      "step": 2780
    },
    {
      "epoch": 0.810575246949448,
      "grad_norm": 29.25,
      "learning_rate": 6.036261792452831e-06,
      "loss": 1.3298,
      "step": 2790
    },
    {
      "epoch": 0.8134805345729227,
      "grad_norm": 44.25,
      "learning_rate": 6.021521226415095e-06,
      "loss": 1.2547,
      "step": 2800
    },
    {
      "epoch": 0.8134805345729227,
      "eval_accuracy": 0.6289342904472667,
      "eval_f1": 0.6291307580361674,
      "eval_loss": 0.6478989124298096,
      "eval_precision": 0.6305301619963926,
      "eval_recall": 0.6289342904472667,
      "eval_runtime": 19.0947,
      "eval_samples_per_second": 94.843,
      "eval_steps_per_second": 9.531,
      "step": 2800
    },
    {
      "epoch": 0.8163858221963974,
      "grad_norm": 64.5,
      "learning_rate": 6.0067806603773595e-06,
      "loss": 1.2978,
      "step": 2810
    },
    {
      "epoch": 0.8192911098198722,
      "grad_norm": 36.5,
      "learning_rate": 5.9920400943396225e-06,
      "loss": 1.3289,
      "step": 2820
    },
    {
      "epoch": 0.8221963974433469,
      "grad_norm": 78.0,
      "learning_rate": 5.977299528301887e-06,
      "loss": 1.2113,
      "step": 2830
    },
    {
      "epoch": 0.8251016850668216,
      "grad_norm": 48.0,
      "learning_rate": 5.962558962264151e-06,
      "loss": 1.3046,
      "step": 2840
    },
    {
      "epoch": 0.8280069726902963,
      "grad_norm": 46.25,
      "learning_rate": 5.9478183962264156e-06,
      "loss": 1.2813,
      "step": 2850
    },
    {
      "epoch": 0.8309122603137711,
      "grad_norm": 61.5,
      "learning_rate": 5.933077830188679e-06,
      "loss": 1.3428,
      "step": 2860
    },
    {
      "epoch": 0.8338175479372458,
      "grad_norm": 25.375,
      "learning_rate": 5.918337264150944e-06,
      "loss": 1.2289,
      "step": 2870
    },
    {
      "epoch": 0.8367228355607205,
      "grad_norm": 62.75,
      "learning_rate": 5.903596698113208e-06,
      "loss": 1.3327,
      "step": 2880
    },
    {
      "epoch": 0.8396281231841952,
      "grad_norm": 69.0,
      "learning_rate": 5.8888561320754725e-06,
      "loss": 1.3704,
      "step": 2890
    },
    {
      "epoch": 0.84253341080767,
      "grad_norm": 48.5,
      "learning_rate": 5.8741155660377355e-06,
      "loss": 1.2572,
      "step": 2900
    },
    {
      "epoch": 0.84253341080767,
      "eval_accuracy": 0.6311430149088901,
      "eval_f1": 0.6297924211046073,
      "eval_loss": 0.6452035307884216,
      "eval_precision": 0.6305609562115095,
      "eval_recall": 0.6311430149088901,
      "eval_runtime": 22.0587,
      "eval_samples_per_second": 82.099,
      "eval_steps_per_second": 8.251,
      "step": 2900
    },
    {
      "epoch": 0.8454386984311447,
      "grad_norm": 56.75,
      "learning_rate": 5.859375e-06,
      "loss": 1.324,
      "step": 2910
    },
    {
      "epoch": 0.8483439860546194,
      "grad_norm": 37.0,
      "learning_rate": 5.844634433962265e-06,
      "loss": 1.3031,
      "step": 2920
    },
    {
      "epoch": 0.8512492736780941,
      "grad_norm": 87.0,
      "learning_rate": 5.829893867924529e-06,
      "loss": 1.4083,
      "step": 2930
    },
    {
      "epoch": 0.8541545613015689,
      "grad_norm": 35.25,
      "learning_rate": 5.815153301886793e-06,
      "loss": 1.3622,
      "step": 2940
    },
    {
      "epoch": 0.8570598489250436,
      "grad_norm": 31.375,
      "learning_rate": 5.800412735849057e-06,
      "loss": 1.3026,
      "step": 2950
    },
    {
      "epoch": 0.8599651365485182,
      "grad_norm": 37.25,
      "learning_rate": 5.785672169811322e-06,
      "loss": 1.2841,
      "step": 2960
    },
    {
      "epoch": 0.862870424171993,
      "grad_norm": 41.5,
      "learning_rate": 5.7709316037735855e-06,
      "loss": 1.2886,
      "step": 2970
    },
    {
      "epoch": 0.8657757117954678,
      "grad_norm": 33.5,
      "learning_rate": 5.75619103773585e-06,
      "loss": 1.2448,
      "step": 2980
    },
    {
      "epoch": 0.8686809994189425,
      "grad_norm": 38.0,
      "learning_rate": 5.741450471698113e-06,
      "loss": 1.3723,
      "step": 2990
    },
    {
      "epoch": 0.8715862870424173,
      "grad_norm": 43.5,
      "learning_rate": 5.726709905660378e-06,
      "loss": 1.3656,
      "step": 3000
    },
    {
      "epoch": 0.8715862870424173,
      "eval_accuracy": 0.6200993926007731,
      "eval_f1": 0.6196755505318609,
      "eval_loss": 0.6481147408485413,
      "eval_precision": 0.624547308512698,
      "eval_recall": 0.6200993926007731,
      "eval_runtime": 22.5073,
      "eval_samples_per_second": 80.463,
      "eval_steps_per_second": 8.086,
      "step": 3000
    },
    {
      "epoch": 0.8744915746658919,
      "grad_norm": 64.0,
      "learning_rate": 5.711969339622642e-06,
      "loss": 1.2059,
      "step": 3010
    },
    {
      "epoch": 0.8773968622893666,
      "grad_norm": 37.5,
      "learning_rate": 5.697228773584906e-06,
      "loss": 1.4072,
      "step": 3020
    },
    {
      "epoch": 0.8803021499128414,
      "grad_norm": 54.75,
      "learning_rate": 5.68248820754717e-06,
      "loss": 1.3369,
      "step": 3030
    },
    {
      "epoch": 0.8832074375363161,
      "grad_norm": 59.25,
      "learning_rate": 5.667747641509435e-06,
      "loss": 1.3416,
      "step": 3040
    },
    {
      "epoch": 0.8861127251597908,
      "grad_norm": 38.5,
      "learning_rate": 5.6530070754716985e-06,
      "loss": 1.3036,
      "step": 3050
    },
    {
      "epoch": 0.8890180127832655,
      "grad_norm": 28.875,
      "learning_rate": 5.638266509433963e-06,
      "loss": 1.3368,
      "step": 3060
    },
    {
      "epoch": 0.8919233004067403,
      "grad_norm": 38.0,
      "learning_rate": 5.623525943396226e-06,
      "loss": 1.2894,
      "step": 3070
    },
    {
      "epoch": 0.894828588030215,
      "grad_norm": 35.75,
      "learning_rate": 5.608785377358491e-06,
      "loss": 1.3777,
      "step": 3080
    },
    {
      "epoch": 0.8977338756536897,
      "grad_norm": 31.25,
      "learning_rate": 5.5940448113207555e-06,
      "loss": 1.2222,
      "step": 3090
    },
    {
      "epoch": 0.9006391632771644,
      "grad_norm": 48.5,
      "learning_rate": 5.579304245283019e-06,
      "loss": 1.3222,
      "step": 3100
    },
    {
      "epoch": 0.9006391632771644,
      "eval_accuracy": 0.6327995582551077,
      "eval_f1": 0.6309066331655309,
      "eval_loss": 0.6468316316604614,
      "eval_precision": 0.6323790829666001,
      "eval_recall": 0.6327995582551077,
      "eval_runtime": 22.1011,
      "eval_samples_per_second": 81.942,
      "eval_steps_per_second": 8.235,
      "step": 3100
    },
    {
      "epoch": 0.9035444509006392,
      "grad_norm": 95.0,
      "learning_rate": 5.564563679245284e-06,
      "loss": 1.468,
      "step": 3110
    },
    {
      "epoch": 0.9064497385241139,
      "grad_norm": 65.0,
      "learning_rate": 5.549823113207548e-06,
      "loss": 1.3194,
      "step": 3120
    },
    {
      "epoch": 0.9093550261475886,
      "grad_norm": 46.0,
      "learning_rate": 5.535082547169812e-06,
      "loss": 1.2986,
      "step": 3130
    },
    {
      "epoch": 0.9122603137710633,
      "grad_norm": 57.25,
      "learning_rate": 5.520341981132075e-06,
      "loss": 1.2991,
      "step": 3140
    },
    {
      "epoch": 0.9151656013945381,
      "grad_norm": 45.0,
      "learning_rate": 5.505601415094341e-06,
      "loss": 1.2372,
      "step": 3150
    },
    {
      "epoch": 0.9180708890180128,
      "grad_norm": 47.25,
      "learning_rate": 5.490860849056604e-06,
      "loss": 1.3458,
      "step": 3160
    },
    {
      "epoch": 0.9209761766414875,
      "grad_norm": 49.0,
      "learning_rate": 5.4761202830188685e-06,
      "loss": 1.3117,
      "step": 3170
    },
    {
      "epoch": 0.9238814642649622,
      "grad_norm": 48.5,
      "learning_rate": 5.461379716981132e-06,
      "loss": 1.2655,
      "step": 3180
    },
    {
      "epoch": 0.926786751888437,
      "grad_norm": 33.75,
      "learning_rate": 5.446639150943397e-06,
      "loss": 1.2547,
      "step": 3190
    },
    {
      "epoch": 0.9296920395119117,
      "grad_norm": 54.75,
      "learning_rate": 5.431898584905661e-06,
      "loss": 1.2231,
      "step": 3200
    },
    {
      "epoch": 0.9296920395119117,
      "eval_accuracy": 0.6366648260629486,
      "eval_f1": 0.6312223269046636,
      "eval_loss": 0.6476677060127258,
      "eval_precision": 0.6387207626644001,
      "eval_recall": 0.6366648260629486,
      "eval_runtime": 22.0118,
      "eval_samples_per_second": 82.274,
      "eval_steps_per_second": 8.268,
      "step": 3200
    },
    {
      "epoch": 0.9325973271353865,
      "grad_norm": 43.5,
      "learning_rate": 5.417158018867925e-06,
      "loss": 1.2574,
      "step": 3210
    },
    {
      "epoch": 0.9355026147588611,
      "grad_norm": 37.0,
      "learning_rate": 5.402417452830188e-06,
      "loss": 1.2772,
      "step": 3220
    },
    {
      "epoch": 0.9384079023823358,
      "grad_norm": 38.75,
      "learning_rate": 5.387676886792453e-06,
      "loss": 1.2504,
      "step": 3230
    },
    {
      "epoch": 0.9413131900058106,
      "grad_norm": 66.0,
      "learning_rate": 5.372936320754717e-06,
      "loss": 1.3723,
      "step": 3240
    },
    {
      "epoch": 0.9442184776292853,
      "grad_norm": 110.0,
      "learning_rate": 5.3581957547169815e-06,
      "loss": 1.3505,
      "step": 3250
    },
    {
      "epoch": 0.94712376525276,
      "grad_norm": 42.0,
      "learning_rate": 5.343455188679245e-06,
      "loss": 1.2563,
      "step": 3260
    },
    {
      "epoch": 0.9500290528762347,
      "grad_norm": 50.0,
      "learning_rate": 5.32871462264151e-06,
      "loss": 1.2599,
      "step": 3270
    },
    {
      "epoch": 0.9529343404997095,
      "grad_norm": 46.5,
      "learning_rate": 5.313974056603775e-06,
      "loss": 1.4014,
      "step": 3280
    },
    {
      "epoch": 0.9558396281231842,
      "grad_norm": 49.5,
      "learning_rate": 5.299233490566038e-06,
      "loss": 1.2596,
      "step": 3290
    },
    {
      "epoch": 0.9587449157466589,
      "grad_norm": 46.75,
      "learning_rate": 5.284492924528303e-06,
      "loss": 1.3213,
      "step": 3300
    },
    {
      "epoch": 0.9587449157466589,
      "eval_accuracy": 0.6327995582551077,
      "eval_f1": 0.6278569180798685,
      "eval_loss": 0.6428092122077942,
      "eval_precision": 0.6342108267486769,
      "eval_recall": 0.6327995582551077,
      "eval_runtime": 21.9512,
      "eval_samples_per_second": 82.501,
      "eval_steps_per_second": 8.291,
      "step": 3300
    },
    {
      "epoch": 0.9616502033701336,
      "grad_norm": 69.0,
      "learning_rate": 5.269752358490566e-06,
      "loss": 1.2727,
      "step": 3310
    },
    {
      "epoch": 0.9645554909936084,
      "grad_norm": 62.75,
      "learning_rate": 5.255011792452831e-06,
      "loss": 1.2832,
      "step": 3320
    },
    {
      "epoch": 0.9674607786170831,
      "grad_norm": 45.25,
      "learning_rate": 5.2402712264150945e-06,
      "loss": 1.3935,
      "step": 3330
    },
    {
      "epoch": 0.9703660662405578,
      "grad_norm": 75.5,
      "learning_rate": 5.225530660377359e-06,
      "loss": 1.3141,
      "step": 3340
    },
    {
      "epoch": 0.9732713538640325,
      "grad_norm": 48.25,
      "learning_rate": 5.210790094339623e-06,
      "loss": 1.3165,
      "step": 3350
    },
    {
      "epoch": 0.9761766414875073,
      "grad_norm": 40.25,
      "learning_rate": 5.196049528301888e-06,
      "loss": 1.2481,
      "step": 3360
    },
    {
      "epoch": 0.979081929110982,
      "grad_norm": 99.5,
      "learning_rate": 5.181308962264151e-06,
      "loss": 1.3187,
      "step": 3370
    },
    {
      "epoch": 0.9819872167344567,
      "grad_norm": 27.125,
      "learning_rate": 5.166568396226416e-06,
      "loss": 1.293,
      "step": 3380
    },
    {
      "epoch": 0.9848925043579314,
      "grad_norm": 36.0,
      "learning_rate": 5.151827830188679e-06,
      "loss": 1.3806,
      "step": 3390
    },
    {
      "epoch": 0.9877977919814062,
      "grad_norm": 66.5,
      "learning_rate": 5.137087264150944e-06,
      "loss": 1.3186,
      "step": 3400
    },
    {
      "epoch": 0.9877977919814062,
      "eval_accuracy": 0.6294864715626726,
      "eval_f1": 0.6285987096192257,
      "eval_loss": 0.6412285566329956,
      "eval_precision": 0.6288758322854454,
      "eval_recall": 0.6294864715626726,
      "eval_runtime": 16.5438,
      "eval_samples_per_second": 109.467,
      "eval_steps_per_second": 11.001,
      "step": 3400
    },
    {
      "epoch": 0.9907030796048809,
      "grad_norm": 35.75,
      "learning_rate": 5.1223466981132075e-06,
      "loss": 1.2888,
      "step": 3410
    },
    {
      "epoch": 0.9936083672283557,
      "grad_norm": 29.0,
      "learning_rate": 5.107606132075472e-06,
      "loss": 1.2755,
      "step": 3420
    },
    {
      "epoch": 0.9965136548518303,
      "grad_norm": 39.0,
      "learning_rate": 5.092865566037736e-06,
      "loss": 1.187,
      "step": 3430
    },
    {
      "epoch": 0.999418942475305,
      "grad_norm": 69.5,
      "learning_rate": 5.078125000000001e-06,
      "loss": 1.2494,
      "step": 3440
    },
    {
      "epoch": 1.0023242300987798,
      "grad_norm": 37.25,
      "learning_rate": 5.063384433962265e-06,
      "loss": 1.3013,
      "step": 3450
    },
    {
      "epoch": 1.0052295177222546,
      "grad_norm": 38.0,
      "learning_rate": 5.048643867924528e-06,
      "loss": 1.2494,
      "step": 3460
    },
    {
      "epoch": 1.0081348053457293,
      "grad_norm": 94.5,
      "learning_rate": 5.033903301886794e-06,
      "loss": 1.2126,
      "step": 3470
    },
    {
      "epoch": 1.011040092969204,
      "grad_norm": 42.5,
      "learning_rate": 5.019162735849057e-06,
      "loss": 1.2814,
      "step": 3480
    },
    {
      "epoch": 1.0139453805926786,
      "grad_norm": 31.0,
      "learning_rate": 5.004422169811321e-06,
      "loss": 1.1684,
      "step": 3490
    },
    {
      "epoch": 1.0168506682161533,
      "grad_norm": 27.625,
      "learning_rate": 4.989681603773585e-06,
      "loss": 1.2539,
      "step": 3500
    },
    {
      "epoch": 1.0168506682161533,
      "eval_accuracy": 0.6256212037548315,
      "eval_f1": 0.6249995872703885,
      "eval_loss": 0.6409505605697632,
      "eval_precision": 0.6250521871698147,
      "eval_recall": 0.6256212037548315,
      "eval_runtime": 22.8239,
      "eval_samples_per_second": 79.347,
      "eval_steps_per_second": 7.974,
      "step": 3500
    },
    {
      "epoch": 1.019755955839628,
      "grad_norm": 48.5,
      "learning_rate": 4.974941037735849e-06,
      "loss": 1.3855,
      "step": 3510
    },
    {
      "epoch": 1.0226612434631028,
      "grad_norm": 63.5,
      "learning_rate": 4.960200471698114e-06,
      "loss": 1.2272,
      "step": 3520
    },
    {
      "epoch": 1.0255665310865776,
      "grad_norm": 67.0,
      "learning_rate": 4.945459905660378e-06,
      "loss": 1.2468,
      "step": 3530
    },
    {
      "epoch": 1.0284718187100523,
      "grad_norm": 48.5,
      "learning_rate": 4.930719339622642e-06,
      "loss": 1.2765,
      "step": 3540
    },
    {
      "epoch": 1.031377106333527,
      "grad_norm": 34.0,
      "learning_rate": 4.915978773584906e-06,
      "loss": 1.2537,
      "step": 3550
    },
    {
      "epoch": 1.0342823939570018,
      "grad_norm": 41.0,
      "learning_rate": 4.9012382075471706e-06,
      "loss": 1.2763,
      "step": 3560
    },
    {
      "epoch": 1.0371876815804764,
      "grad_norm": 47.25,
      "learning_rate": 4.886497641509434e-06,
      "loss": 1.211,
      "step": 3570
    },
    {
      "epoch": 1.0400929692039511,
      "grad_norm": 69.5,
      "learning_rate": 4.871757075471699e-06,
      "loss": 1.253,
      "step": 3580
    },
    {
      "epoch": 1.0429982568274259,
      "grad_norm": 51.5,
      "learning_rate": 4.857016509433963e-06,
      "loss": 1.198,
      "step": 3590
    },
    {
      "epoch": 1.0459035444509006,
      "grad_norm": 57.5,
      "learning_rate": 4.842275943396227e-06,
      "loss": 1.281,
      "step": 3600
    },
    {
      "epoch": 1.0459035444509006,
      "eval_accuracy": 0.631695196024296,
      "eval_f1": 0.6319043862620346,
      "eval_loss": 0.6441466212272644,
      "eval_precision": 0.6330383376944125,
      "eval_recall": 0.631695196024296,
      "eval_runtime": 22.3581,
      "eval_samples_per_second": 81.0,
      "eval_steps_per_second": 8.14,
      "step": 3600
    },
    {
      "epoch": 1.0488088320743754,
      "grad_norm": 41.5,
      "learning_rate": 4.827535377358491e-06,
      "loss": 1.238,
      "step": 3610
    },
    {
      "epoch": 1.0517141196978501,
      "grad_norm": 73.5,
      "learning_rate": 4.812794811320755e-06,
      "loss": 1.3438,
      "step": 3620
    },
    {
      "epoch": 1.0546194073213249,
      "grad_norm": 46.0,
      "learning_rate": 4.798054245283019e-06,
      "loss": 1.152,
      "step": 3630
    },
    {
      "epoch": 1.0575246949447996,
      "grad_norm": 42.75,
      "learning_rate": 4.7833136792452836e-06,
      "loss": 1.233,
      "step": 3640
    },
    {
      "epoch": 1.0604299825682744,
      "grad_norm": 39.25,
      "learning_rate": 4.768573113207547e-06,
      "loss": 1.258,
      "step": 3650
    },
    {
      "epoch": 1.063335270191749,
      "grad_norm": 82.5,
      "learning_rate": 4.753832547169811e-06,
      "loss": 1.2224,
      "step": 3660
    },
    {
      "epoch": 1.0662405578152236,
      "grad_norm": 32.25,
      "learning_rate": 4.739091981132076e-06,
      "loss": 1.2568,
      "step": 3670
    },
    {
      "epoch": 1.0691458454386984,
      "grad_norm": 69.5,
      "learning_rate": 4.72435141509434e-06,
      "loss": 1.2814,
      "step": 3680
    },
    {
      "epoch": 1.0720511330621731,
      "grad_norm": 44.25,
      "learning_rate": 4.709610849056604e-06,
      "loss": 1.348,
      "step": 3690
    },
    {
      "epoch": 1.074956420685648,
      "grad_norm": 40.0,
      "learning_rate": 4.694870283018868e-06,
      "loss": 1.1893,
      "step": 3700
    },
    {
      "epoch": 1.074956420685648,
      "eval_accuracy": 0.6355604638321369,
      "eval_f1": 0.6355406367530252,
      "eval_loss": 0.6429353952407837,
      "eval_precision": 0.6355223595138046,
      "eval_recall": 0.6355604638321369,
      "eval_runtime": 22.5961,
      "eval_samples_per_second": 80.147,
      "eval_steps_per_second": 8.054,
      "step": 3700
    },
    {
      "epoch": 1.0778617083091226,
      "grad_norm": 44.5,
      "learning_rate": 4.680129716981133e-06,
      "loss": 1.2232,
      "step": 3710
    },
    {
      "epoch": 1.0807669959325974,
      "grad_norm": 40.75,
      "learning_rate": 4.6653891509433966e-06,
      "loss": 1.1672,
      "step": 3720
    },
    {
      "epoch": 1.0836722835560721,
      "grad_norm": 40.5,
      "learning_rate": 4.650648584905661e-06,
      "loss": 1.2764,
      "step": 3730
    },
    {
      "epoch": 1.0865775711795467,
      "grad_norm": 36.0,
      "learning_rate": 4.635908018867925e-06,
      "loss": 1.2449,
      "step": 3740
    },
    {
      "epoch": 1.0894828588030214,
      "grad_norm": 69.0,
      "learning_rate": 4.621167452830189e-06,
      "loss": 1.2499,
      "step": 3750
    },
    {
      "epoch": 1.0923881464264962,
      "grad_norm": 39.5,
      "learning_rate": 4.6064268867924535e-06,
      "loss": 1.1284,
      "step": 3760
    },
    {
      "epoch": 1.095293434049971,
      "grad_norm": 29.0,
      "learning_rate": 4.591686320754717e-06,
      "loss": 1.1788,
      "step": 3770
    },
    {
      "epoch": 1.0981987216734457,
      "grad_norm": 28.25,
      "learning_rate": 4.576945754716982e-06,
      "loss": 1.2809,
      "step": 3780
    },
    {
      "epoch": 1.1011040092969204,
      "grad_norm": 39.75,
      "learning_rate": 4.562205188679246e-06,
      "loss": 1.2236,
      "step": 3790
    },
    {
      "epoch": 1.1040092969203952,
      "grad_norm": 55.75,
      "learning_rate": 4.54746462264151e-06,
      "loss": 1.2188,
      "step": 3800
    },
    {
      "epoch": 1.1040092969203952,
      "eval_accuracy": 0.6245168415240199,
      "eval_f1": 0.6229741864922596,
      "eval_loss": 0.6412906646728516,
      "eval_precision": 0.6238833289814343,
      "eval_recall": 0.6245168415240199,
      "eval_runtime": 21.6276,
      "eval_samples_per_second": 83.736,
      "eval_steps_per_second": 8.415,
      "step": 3800
    },
    {
      "epoch": 1.10691458454387,
      "grad_norm": 69.0,
      "learning_rate": 4.532724056603774e-06,
      "loss": 1.2715,
      "step": 3810
    },
    {
      "epoch": 1.1098198721673445,
      "grad_norm": 42.25,
      "learning_rate": 4.517983490566038e-06,
      "loss": 1.2385,
      "step": 3820
    },
    {
      "epoch": 1.1127251597908192,
      "grad_norm": 38.25,
      "learning_rate": 4.503242924528302e-06,
      "loss": 1.3215,
      "step": 3830
    },
    {
      "epoch": 1.115630447414294,
      "grad_norm": 51.25,
      "learning_rate": 4.4885023584905665e-06,
      "loss": 1.2579,
      "step": 3840
    },
    {
      "epoch": 1.1185357350377687,
      "grad_norm": 25.375,
      "learning_rate": 4.47376179245283e-06,
      "loss": 1.2701,
      "step": 3850
    },
    {
      "epoch": 1.1214410226612435,
      "grad_norm": 37.5,
      "learning_rate": 4.459021226415094e-06,
      "loss": 1.2051,
      "step": 3860
    },
    {
      "epoch": 1.1243463102847182,
      "grad_norm": 46.25,
      "learning_rate": 4.444280660377359e-06,
      "loss": 1.2782,
      "step": 3870
    },
    {
      "epoch": 1.127251597908193,
      "grad_norm": 48.0,
      "learning_rate": 4.429540094339623e-06,
      "loss": 1.2045,
      "step": 3880
    },
    {
      "epoch": 1.1301568855316677,
      "grad_norm": 58.0,
      "learning_rate": 4.414799528301887e-06,
      "loss": 1.2268,
      "step": 3890
    },
    {
      "epoch": 1.1330621731551425,
      "grad_norm": 33.25,
      "learning_rate": 4.400058962264152e-06,
      "loss": 1.2652,
      "step": 3900
    },
    {
      "epoch": 1.1330621731551425,
      "eval_accuracy": 0.6432909994478189,
      "eval_f1": 0.6421848255039291,
      "eval_loss": 0.6396148800849915,
      "eval_precision": 0.6427959924584501,
      "eval_recall": 0.6432909994478189,
      "eval_runtime": 17.5222,
      "eval_samples_per_second": 103.355,
      "eval_steps_per_second": 10.387,
      "step": 3900
    },
    {
      "epoch": 1.135967460778617,
      "grad_norm": 37.5,
      "learning_rate": 4.385318396226416e-06,
      "loss": 1.2721,
      "step": 3910
    },
    {
      "epoch": 1.1388727484020917,
      "grad_norm": 50.5,
      "learning_rate": 4.3705778301886795e-06,
      "loss": 1.1884,
      "step": 3920
    },
    {
      "epoch": 1.1417780360255665,
      "grad_norm": 97.5,
      "learning_rate": 4.355837264150944e-06,
      "loss": 1.197,
      "step": 3930
    },
    {
      "epoch": 1.1446833236490412,
      "grad_norm": 58.0,
      "learning_rate": 4.341096698113208e-06,
      "loss": 1.2079,
      "step": 3940
    },
    {
      "epoch": 1.147588611272516,
      "grad_norm": 43.75,
      "learning_rate": 4.326356132075472e-06,
      "loss": 1.2688,
      "step": 3950
    },
    {
      "epoch": 1.1504938988959907,
      "grad_norm": 57.5,
      "learning_rate": 4.3116155660377365e-06,
      "loss": 1.2103,
      "step": 3960
    },
    {
      "epoch": 1.1533991865194655,
      "grad_norm": 76.0,
      "learning_rate": 4.296875e-06,
      "loss": 1.2674,
      "step": 3970
    },
    {
      "epoch": 1.1563044741429402,
      "grad_norm": 90.0,
      "learning_rate": 4.282134433962264e-06,
      "loss": 1.2474,
      "step": 3980
    },
    {
      "epoch": 1.159209761766415,
      "grad_norm": 50.75,
      "learning_rate": 4.267393867924529e-06,
      "loss": 1.2753,
      "step": 3990
    },
    {
      "epoch": 1.1621150493898895,
      "grad_norm": 37.0,
      "learning_rate": 4.2526533018867925e-06,
      "loss": 1.2627,
      "step": 4000
    },
    {
      "epoch": 1.1621150493898895,
      "eval_accuracy": 0.631695196024296,
      "eval_f1": 0.63172355603383,
      "eval_loss": 0.640481173992157,
      "eval_precision": 0.6317553813298517,
      "eval_recall": 0.631695196024296,
      "eval_runtime": 21.5816,
      "eval_samples_per_second": 83.914,
      "eval_steps_per_second": 8.433,
      "step": 4000
    },
    {
      "epoch": 1.1650203370133643,
      "grad_norm": 51.0,
      "learning_rate": 4.237912735849057e-06,
      "loss": 1.3051,
      "step": 4010
    },
    {
      "epoch": 1.167925624636839,
      "grad_norm": 39.5,
      "learning_rate": 4.223172169811321e-06,
      "loss": 1.2549,
      "step": 4020
    },
    {
      "epoch": 1.1708309122603138,
      "grad_norm": 40.5,
      "learning_rate": 4.208431603773585e-06,
      "loss": 1.1884,
      "step": 4030
    },
    {
      "epoch": 1.1737361998837885,
      "grad_norm": 25.125,
      "learning_rate": 4.1936910377358495e-06,
      "loss": 1.1919,
      "step": 4040
    },
    {
      "epoch": 1.1766414875072633,
      "grad_norm": 65.5,
      "learning_rate": 4.178950471698113e-06,
      "loss": 1.2309,
      "step": 4050
    },
    {
      "epoch": 1.179546775130738,
      "grad_norm": 34.0,
      "learning_rate": 4.164209905660378e-06,
      "loss": 1.1595,
      "step": 4060
    },
    {
      "epoch": 1.1824520627542126,
      "grad_norm": 60.0,
      "learning_rate": 4.149469339622642e-06,
      "loss": 1.229,
      "step": 4070
    },
    {
      "epoch": 1.1853573503776873,
      "grad_norm": 35.75,
      "learning_rate": 4.134728773584906e-06,
      "loss": 1.237,
      "step": 4080
    },
    {
      "epoch": 1.188262638001162,
      "grad_norm": 43.0,
      "learning_rate": 4.11998820754717e-06,
      "loss": 1.2608,
      "step": 4090
    },
    {
      "epoch": 1.1911679256246368,
      "grad_norm": 95.0,
      "learning_rate": 4.105247641509435e-06,
      "loss": 1.2385,
      "step": 4100
    },
    {
      "epoch": 1.1911679256246368,
      "eval_accuracy": 0.6344561016013253,
      "eval_f1": 0.6330641107526459,
      "eval_loss": 0.6400484442710876,
      "eval_precision": 0.6339200503566697,
      "eval_recall": 0.6344561016013253,
      "eval_runtime": 21.1516,
      "eval_samples_per_second": 85.62,
      "eval_steps_per_second": 8.605,
      "step": 4100
    },
    {
      "epoch": 1.1940732132481116,
      "grad_norm": 60.0,
      "learning_rate": 4.090507075471699e-06,
      "loss": 1.2141,
      "step": 4110
    },
    {
      "epoch": 1.1969785008715863,
      "grad_norm": 73.0,
      "learning_rate": 4.0757665094339625e-06,
      "loss": 1.2815,
      "step": 4120
    },
    {
      "epoch": 1.199883788495061,
      "grad_norm": 31.25,
      "learning_rate": 4.061025943396227e-06,
      "loss": 1.2256,
      "step": 4130
    },
    {
      "epoch": 1.2027890761185358,
      "grad_norm": 63.0,
      "learning_rate": 4.046285377358491e-06,
      "loss": 1.3107,
      "step": 4140
    },
    {
      "epoch": 1.2056943637420106,
      "grad_norm": 40.0,
      "learning_rate": 4.031544811320755e-06,
      "loss": 1.24,
      "step": 4150
    },
    {
      "epoch": 1.208599651365485,
      "grad_norm": 57.5,
      "learning_rate": 4.016804245283019e-06,
      "loss": 1.2925,
      "step": 4160
    },
    {
      "epoch": 1.2115049389889598,
      "grad_norm": 32.0,
      "learning_rate": 4.002063679245283e-06,
      "loss": 1.2477,
      "step": 4170
    },
    {
      "epoch": 1.2144102266124346,
      "grad_norm": 35.75,
      "learning_rate": 3.987323113207547e-06,
      "loss": 1.2127,
      "step": 4180
    },
    {
      "epoch": 1.2173155142359093,
      "grad_norm": 46.25,
      "learning_rate": 3.972582547169812e-06,
      "loss": 1.2459,
      "step": 4190
    },
    {
      "epoch": 1.220220801859384,
      "grad_norm": 62.5,
      "learning_rate": 3.9578419811320755e-06,
      "loss": 1.2525,
      "step": 4200
    },
    {
      "epoch": 1.220220801859384,
      "eval_accuracy": 0.6366648260629486,
      "eval_f1": 0.6323882931612419,
      "eval_loss": 0.6401476263999939,
      "eval_precision": 0.6377675688323987,
      "eval_recall": 0.6366648260629486,
      "eval_runtime": 21.7359,
      "eval_samples_per_second": 83.318,
      "eval_steps_per_second": 8.373,
      "step": 4200
    },
    {
      "epoch": 1.2231260894828588,
      "grad_norm": 32.5,
      "learning_rate": 3.94310141509434e-06,
      "loss": 1.2414,
      "step": 4210
    },
    {
      "epoch": 1.2260313771063336,
      "grad_norm": 44.0,
      "learning_rate": 3.928360849056604e-06,
      "loss": 1.2214,
      "step": 4220
    },
    {
      "epoch": 1.2289366647298083,
      "grad_norm": 47.25,
      "learning_rate": 3.913620283018868e-06,
      "loss": 1.2093,
      "step": 4230
    },
    {
      "epoch": 1.231841952353283,
      "grad_norm": 28.375,
      "learning_rate": 3.898879716981132e-06,
      "loss": 1.1936,
      "step": 4240
    },
    {
      "epoch": 1.2347472399767576,
      "grad_norm": 51.25,
      "learning_rate": 3.884139150943397e-06,
      "loss": 1.2721,
      "step": 4250
    },
    {
      "epoch": 1.2376525276002324,
      "grad_norm": 33.75,
      "learning_rate": 3.869398584905661e-06,
      "loss": 1.2279,
      "step": 4260
    },
    {
      "epoch": 1.2405578152237071,
      "grad_norm": 43.75,
      "learning_rate": 3.854658018867925e-06,
      "loss": 1.2586,
      "step": 4270
    },
    {
      "epoch": 1.2434631028471819,
      "grad_norm": 51.25,
      "learning_rate": 3.839917452830189e-06,
      "loss": 1.1615,
      "step": 4280
    },
    {
      "epoch": 1.2463683904706566,
      "grad_norm": 69.0,
      "learning_rate": 3.825176886792453e-06,
      "loss": 1.2211,
      "step": 4290
    },
    {
      "epoch": 1.2492736780941314,
      "grad_norm": 56.75,
      "learning_rate": 3.8104363207547174e-06,
      "loss": 1.259,
      "step": 4300
    },
    {
      "epoch": 1.2492736780941314,
      "eval_accuracy": 0.6294864715626726,
      "eval_f1": 0.6256227167383146,
      "eval_loss": 0.6409916877746582,
      "eval_precision": 0.6299778534682698,
      "eval_recall": 0.6294864715626726,
      "eval_runtime": 22.2688,
      "eval_samples_per_second": 81.325,
      "eval_steps_per_second": 8.173,
      "step": 4300
    },
    {
      "epoch": 1.2521789657176061,
      "grad_norm": 41.25,
      "learning_rate": 3.7956957547169816e-06,
      "loss": 1.1971,
      "step": 4310
    },
    {
      "epoch": 1.2550842533410806,
      "grad_norm": 70.5,
      "learning_rate": 3.7809551886792454e-06,
      "loss": 1.1621,
      "step": 4320
    },
    {
      "epoch": 1.2579895409645556,
      "grad_norm": 31.0,
      "learning_rate": 3.7662146226415097e-06,
      "loss": 1.2817,
      "step": 4330
    },
    {
      "epoch": 1.2608948285880301,
      "grad_norm": 45.5,
      "learning_rate": 3.751474056603774e-06,
      "loss": 1.2521,
      "step": 4340
    },
    {
      "epoch": 1.263800116211505,
      "grad_norm": 47.75,
      "learning_rate": 3.736733490566038e-06,
      "loss": 1.2734,
      "step": 4350
    },
    {
      "epoch": 1.2667054038349796,
      "grad_norm": 48.25,
      "learning_rate": 3.721992924528302e-06,
      "loss": 1.2468,
      "step": 4360
    },
    {
      "epoch": 1.2696106914584544,
      "grad_norm": 33.25,
      "learning_rate": 3.707252358490566e-06,
      "loss": 1.1994,
      "step": 4370
    },
    {
      "epoch": 1.2725159790819291,
      "grad_norm": 45.25,
      "learning_rate": 3.6925117924528304e-06,
      "loss": 1.1964,
      "step": 4380
    },
    {
      "epoch": 1.275421266705404,
      "grad_norm": 93.0,
      "learning_rate": 3.6777712264150946e-06,
      "loss": 1.204,
      "step": 4390
    },
    {
      "epoch": 1.2783265543288787,
      "grad_norm": 35.5,
      "learning_rate": 3.6630306603773584e-06,
      "loss": 1.1888,
      "step": 4400
    },
    {
      "epoch": 1.2783265543288787,
      "eval_accuracy": 0.6327995582551077,
      "eval_f1": 0.6273511093523488,
      "eval_loss": 0.6400877833366394,
      "eval_precision": 0.6346113195792686,
      "eval_recall": 0.6327995582551077,
      "eval_runtime": 22.3949,
      "eval_samples_per_second": 80.867,
      "eval_steps_per_second": 8.127,
      "step": 4400
    },
    {
      "epoch": 1.2812318419523532,
      "grad_norm": 53.25,
      "learning_rate": 3.6482900943396227e-06,
      "loss": 1.2274,
      "step": 4410
    },
    {
      "epoch": 1.284137129575828,
      "grad_norm": 107.5,
      "learning_rate": 3.6335495283018873e-06,
      "loss": 1.2562,
      "step": 4420
    },
    {
      "epoch": 1.2870424171993027,
      "grad_norm": 40.75,
      "learning_rate": 3.6188089622641515e-06,
      "loss": 1.227,
      "step": 4430
    },
    {
      "epoch": 1.2899477048227774,
      "grad_norm": 60.25,
      "learning_rate": 3.6040683962264158e-06,
      "loss": 1.3077,
      "step": 4440
    },
    {
      "epoch": 1.2928529924462522,
      "grad_norm": 50.0,
      "learning_rate": 3.5893278301886796e-06,
      "loss": 1.2209,
      "step": 4450
    },
    {
      "epoch": 1.295758280069727,
      "grad_norm": 42.25,
      "learning_rate": 3.574587264150944e-06,
      "loss": 1.2701,
      "step": 4460
    },
    {
      "epoch": 1.2986635676932017,
      "grad_norm": 63.75,
      "learning_rate": 3.559846698113208e-06,
      "loss": 1.2894,
      "step": 4470
    },
    {
      "epoch": 1.3015688553166764,
      "grad_norm": 49.75,
      "learning_rate": 3.545106132075472e-06,
      "loss": 1.2274,
      "step": 4480
    },
    {
      "epoch": 1.3044741429401512,
      "grad_norm": 69.0,
      "learning_rate": 3.530365566037736e-06,
      "loss": 1.2479,
      "step": 4490
    },
    {
      "epoch": 1.3073794305636257,
      "grad_norm": 45.75,
      "learning_rate": 3.5156250000000003e-06,
      "loss": 1.2385,
      "step": 4500
    },
    {
      "epoch": 1.3073794305636257,
      "eval_accuracy": 0.6410822749861955,
      "eval_f1": 0.6406300387406191,
      "eval_loss": 0.6401379108428955,
      "eval_precision": 0.6406304536792782,
      "eval_recall": 0.6410822749861955,
      "eval_runtime": 17.5048,
      "eval_samples_per_second": 103.458,
      "eval_steps_per_second": 10.397,
      "step": 4500
    },
    {
      "epoch": 1.3102847181871005,
      "grad_norm": 51.25,
      "learning_rate": 3.5008844339622646e-06,
      "loss": 1.2097,
      "step": 4510
    },
    {
      "epoch": 1.3131900058105752,
      "grad_norm": 45.0,
      "learning_rate": 3.4861438679245284e-06,
      "loss": 1.2962,
      "step": 4520
    },
    {
      "epoch": 1.31609529343405,
      "grad_norm": 33.25,
      "learning_rate": 3.4714033018867926e-06,
      "loss": 1.2895,
      "step": 4530
    },
    {
      "epoch": 1.3190005810575247,
      "grad_norm": 41.0,
      "learning_rate": 3.456662735849057e-06,
      "loss": 1.2459,
      "step": 4540
    },
    {
      "epoch": 1.3219058686809995,
      "grad_norm": 38.5,
      "learning_rate": 3.441922169811321e-06,
      "loss": 1.2178,
      "step": 4550
    },
    {
      "epoch": 1.3248111563044742,
      "grad_norm": 39.0,
      "learning_rate": 3.427181603773585e-06,
      "loss": 1.1922,
      "step": 4560
    },
    {
      "epoch": 1.3277164439279487,
      "grad_norm": 63.25,
      "learning_rate": 3.412441037735849e-06,
      "loss": 1.2432,
      "step": 4570
    },
    {
      "epoch": 1.3306217315514237,
      "grad_norm": 57.5,
      "learning_rate": 3.3977004716981133e-06,
      "loss": 1.2522,
      "step": 4580
    },
    {
      "epoch": 1.3335270191748982,
      "grad_norm": 45.0,
      "learning_rate": 3.382959905660378e-06,
      "loss": 1.2405,
      "step": 4590
    },
    {
      "epoch": 1.336432306798373,
      "grad_norm": 49.75,
      "learning_rate": 3.3682193396226422e-06,
      "loss": 1.3357,
      "step": 4600
    },
    {
      "epoch": 1.336432306798373,
      "eval_accuracy": 0.6361126449475428,
      "eval_f1": 0.6358309002139024,
      "eval_loss": 0.6416363716125488,
      "eval_precision": 0.6357545053467629,
      "eval_recall": 0.6361126449475428,
      "eval_runtime": 21.4148,
      "eval_samples_per_second": 84.568,
      "eval_steps_per_second": 8.499,
      "step": 4600
    },
    {
      "epoch": 1.3393375944218477,
      "grad_norm": 36.75,
      "learning_rate": 3.353478773584906e-06,
      "loss": 1.1695,
      "step": 4610
    },
    {
      "epoch": 1.3422428820453225,
      "grad_norm": 44.0,
      "learning_rate": 3.3387382075471703e-06,
      "loss": 1.2259,
      "step": 4620
    },
    {
      "epoch": 1.3451481696687972,
      "grad_norm": 33.0,
      "learning_rate": 3.3239976415094345e-06,
      "loss": 1.2222,
      "step": 4630
    },
    {
      "epoch": 1.348053457292272,
      "grad_norm": 53.75,
      "learning_rate": 3.3092570754716983e-06,
      "loss": 1.1198,
      "step": 4640
    },
    {
      "epoch": 1.3509587449157467,
      "grad_norm": 39.25,
      "learning_rate": 3.2945165094339625e-06,
      "loss": 1.2235,
      "step": 4650
    },
    {
      "epoch": 1.3538640325392213,
      "grad_norm": 72.5,
      "learning_rate": 3.2797759433962268e-06,
      "loss": 1.2233,
      "step": 4660
    },
    {
      "epoch": 1.356769320162696,
      "grad_norm": 38.25,
      "learning_rate": 3.265035377358491e-06,
      "loss": 1.2763,
      "step": 4670
    },
    {
      "epoch": 1.3596746077861708,
      "grad_norm": 44.75,
      "learning_rate": 3.250294811320755e-06,
      "loss": 1.288,
      "step": 4680
    },
    {
      "epoch": 1.3625798954096455,
      "grad_norm": 41.0,
      "learning_rate": 3.235554245283019e-06,
      "loss": 1.2626,
      "step": 4690
    },
    {
      "epoch": 1.3654851830331203,
      "grad_norm": 80.0,
      "learning_rate": 3.2208136792452833e-06,
      "loss": 1.2926,
      "step": 4700
    },
    {
      "epoch": 1.3654851830331203,
      "eval_accuracy": 0.6372170071783545,
      "eval_f1": 0.6362579450911782,
      "eval_loss": 0.6420822739601135,
      "eval_precision": 0.6366568185144541,
      "eval_recall": 0.6372170071783545,
      "eval_runtime": 21.7934,
      "eval_samples_per_second": 83.099,
      "eval_steps_per_second": 8.351,
      "step": 4700
    },
    {
      "epoch": 1.368390470656595,
      "grad_norm": 43.25,
      "learning_rate": 3.2060731132075475e-06,
      "loss": 1.246,
      "step": 4710
    },
    {
      "epoch": 1.3712957582800698,
      "grad_norm": 39.5,
      "learning_rate": 3.1913325471698113e-06,
      "loss": 1.2012,
      "step": 4720
    },
    {
      "epoch": 1.3742010459035445,
      "grad_norm": 30.375,
      "learning_rate": 3.1765919811320755e-06,
      "loss": 1.2389,
      "step": 4730
    },
    {
      "epoch": 1.3771063335270193,
      "grad_norm": 42.75,
      "learning_rate": 3.1618514150943398e-06,
      "loss": 1.2816,
      "step": 4740
    },
    {
      "epoch": 1.3800116211504938,
      "grad_norm": 53.75,
      "learning_rate": 3.1471108490566036e-06,
      "loss": 1.2284,
      "step": 4750
    },
    {
      "epoch": 1.3829169087739686,
      "grad_norm": 60.0,
      "learning_rate": 3.132370283018868e-06,
      "loss": 1.2905,
      "step": 4760
    },
    {
      "epoch": 1.3858221963974433,
      "grad_norm": 58.75,
      "learning_rate": 3.1176297169811325e-06,
      "loss": 1.3362,
      "step": 4770
    },
    {
      "epoch": 1.388727484020918,
      "grad_norm": 73.5,
      "learning_rate": 3.1028891509433967e-06,
      "loss": 1.216,
      "step": 4780
    },
    {
      "epoch": 1.3916327716443928,
      "grad_norm": 34.0,
      "learning_rate": 3.088148584905661e-06,
      "loss": 1.1791,
      "step": 4790
    },
    {
      "epoch": 1.3945380592678676,
      "grad_norm": 47.5,
      "learning_rate": 3.0734080188679248e-06,
      "loss": 1.1769,
      "step": 4800
    },
    {
      "epoch": 1.3945380592678676,
      "eval_accuracy": 0.6294864715626726,
      "eval_f1": 0.6242980933823017,
      "eval_loss": 0.6427155137062073,
      "eval_precision": 0.6308927063460188,
      "eval_recall": 0.6294864715626726,
      "eval_runtime": 21.4015,
      "eval_samples_per_second": 84.62,
      "eval_steps_per_second": 8.504,
      "step": 4800
    },
    {
      "epoch": 1.3974433468913423,
      "grad_norm": 50.5,
      "learning_rate": 3.058667452830189e-06,
      "loss": 1.2116,
      "step": 4810
    },
    {
      "epoch": 1.4003486345148168,
      "grad_norm": 96.5,
      "learning_rate": 3.0439268867924532e-06,
      "loss": 1.266,
      "step": 4820
    },
    {
      "epoch": 1.4032539221382918,
      "grad_norm": 38.5,
      "learning_rate": 3.0291863207547174e-06,
      "loss": 1.2151,
      "step": 4830
    },
    {
      "epoch": 1.4061592097617663,
      "grad_norm": 41.0,
      "learning_rate": 3.0144457547169813e-06,
      "loss": 1.1757,
      "step": 4840
    },
    {
      "epoch": 1.409064497385241,
      "grad_norm": 79.0,
      "learning_rate": 2.9997051886792455e-06,
      "loss": 1.2929,
      "step": 4850
    },
    {
      "epoch": 1.4119697850087158,
      "grad_norm": 42.5,
      "learning_rate": 2.9849646226415097e-06,
      "loss": 1.1804,
      "step": 4860
    },
    {
      "epoch": 1.4148750726321906,
      "grad_norm": 39.75,
      "learning_rate": 2.970224056603774e-06,
      "loss": 1.2443,
      "step": 4870
    },
    {
      "epoch": 1.4177803602556653,
      "grad_norm": 45.25,
      "learning_rate": 2.9554834905660378e-06,
      "loss": 1.2595,
      "step": 4880
    },
    {
      "epoch": 1.42068564787914,
      "grad_norm": 35.0,
      "learning_rate": 2.940742924528302e-06,
      "loss": 1.2393,
      "step": 4890
    },
    {
      "epoch": 1.4235909355026148,
      "grad_norm": 50.25,
      "learning_rate": 2.9260023584905662e-06,
      "loss": 1.226,
      "step": 4900
    },
    {
      "epoch": 1.4235909355026148,
      "eval_accuracy": 0.6355604638321369,
      "eval_f1": 0.6343806261841474,
      "eval_loss": 0.6406518220901489,
      "eval_precision": 0.6350017249086662,
      "eval_recall": 0.6355604638321369,
      "eval_runtime": 20.0566,
      "eval_samples_per_second": 90.295,
      "eval_steps_per_second": 9.074,
      "step": 4900
    },
    {
      "epoch": 1.4264962231260894,
      "grad_norm": 61.75,
      "learning_rate": 2.9112617924528305e-06,
      "loss": 1.2209,
      "step": 4910
    },
    {
      "epoch": 1.4294015107495643,
      "grad_norm": 65.0,
      "learning_rate": 2.8965212264150943e-06,
      "loss": 1.2144,
      "step": 4920
    },
    {
      "epoch": 1.4323067983730389,
      "grad_norm": 175.0,
      "learning_rate": 2.8817806603773585e-06,
      "loss": 1.3651,
      "step": 4930
    },
    {
      "epoch": 1.4352120859965136,
      "grad_norm": 43.5,
      "learning_rate": 2.8670400943396227e-06,
      "loss": 1.2507,
      "step": 4940
    },
    {
      "epoch": 1.4381173736199884,
      "grad_norm": 48.75,
      "learning_rate": 2.8522995283018874e-06,
      "loss": 1.232,
      "step": 4950
    },
    {
      "epoch": 1.4410226612434631,
      "grad_norm": 86.0,
      "learning_rate": 2.837558962264151e-06,
      "loss": 1.2307,
      "step": 4960
    },
    {
      "epoch": 1.4439279488669379,
      "grad_norm": 64.5,
      "learning_rate": 2.8228183962264154e-06,
      "loss": 1.1689,
      "step": 4970
    },
    {
      "epoch": 1.4468332364904126,
      "grad_norm": 42.75,
      "learning_rate": 2.8080778301886797e-06,
      "loss": 1.3197,
      "step": 4980
    },
    {
      "epoch": 1.4497385241138874,
      "grad_norm": 30.5,
      "learning_rate": 2.793337264150944e-06,
      "loss": 1.2356,
      "step": 4990
    },
    {
      "epoch": 1.452643811737362,
      "grad_norm": 57.25,
      "learning_rate": 2.7785966981132077e-06,
      "loss": 1.185,
      "step": 5000
    },
    {
      "epoch": 1.452643811737362,
      "eval_accuracy": 0.6355604638321369,
      "eval_f1": 0.6347092433340812,
      "eval_loss": 0.6409675478935242,
      "eval_precision": 0.634991904463308,
      "eval_recall": 0.6355604638321369,
      "eval_runtime": 21.5778,
      "eval_samples_per_second": 83.929,
      "eval_steps_per_second": 8.435,
      "step": 5000
    },
    {
      "epoch": 1.4555490993608367,
      "grad_norm": 32.75,
      "learning_rate": 2.763856132075472e-06,
      "loss": 1.2552,
      "step": 5010
    },
    {
      "epoch": 1.4584543869843114,
      "grad_norm": 44.5,
      "learning_rate": 2.749115566037736e-06,
      "loss": 1.1698,
      "step": 5020
    },
    {
      "epoch": 1.4613596746077862,
      "grad_norm": 47.5,
      "learning_rate": 2.7343750000000004e-06,
      "loss": 1.2106,
      "step": 5030
    },
    {
      "epoch": 1.464264962231261,
      "grad_norm": 47.75,
      "learning_rate": 2.719634433962264e-06,
      "loss": 1.2006,
      "step": 5040
    },
    {
      "epoch": 1.4671702498547357,
      "grad_norm": 52.25,
      "learning_rate": 2.7048938679245284e-06,
      "loss": 1.2336,
      "step": 5050
    },
    {
      "epoch": 1.4700755374782104,
      "grad_norm": 74.5,
      "learning_rate": 2.6901533018867927e-06,
      "loss": 1.2123,
      "step": 5060
    },
    {
      "epoch": 1.472980825101685,
      "grad_norm": 35.0,
      "learning_rate": 2.675412735849057e-06,
      "loss": 1.289,
      "step": 5070
    },
    {
      "epoch": 1.47588611272516,
      "grad_norm": 73.0,
      "learning_rate": 2.6606721698113207e-06,
      "loss": 1.2429,
      "step": 5080
    },
    {
      "epoch": 1.4787914003486344,
      "grad_norm": 53.0,
      "learning_rate": 2.645931603773585e-06,
      "loss": 1.2213,
      "step": 5090
    },
    {
      "epoch": 1.4816966879721092,
      "grad_norm": 59.0,
      "learning_rate": 2.631191037735849e-06,
      "loss": 1.2239,
      "step": 5100
    },
    {
      "epoch": 1.4816966879721092,
      "eval_accuracy": 0.6339039204859194,
      "eval_f1": 0.6286765449836439,
      "eval_loss": 0.643553614616394,
      "eval_precision": 0.6356054506308846,
      "eval_recall": 0.6339039204859194,
      "eval_runtime": 16.6265,
      "eval_samples_per_second": 108.923,
      "eval_steps_per_second": 10.946,
      "step": 5100
    },
    {
      "epoch": 1.484601975595584,
      "grad_norm": 41.75,
      "learning_rate": 2.616450471698113e-06,
      "loss": 1.3158,
      "step": 5110
    },
    {
      "epoch": 1.4875072632190587,
      "grad_norm": 45.5,
      "learning_rate": 2.6017099056603776e-06,
      "loss": 1.2863,
      "step": 5120
    },
    {
      "epoch": 1.4904125508425334,
      "grad_norm": 35.5,
      "learning_rate": 2.586969339622642e-06,
      "loss": 1.2767,
      "step": 5130
    },
    {
      "epoch": 1.4933178384660082,
      "grad_norm": 55.0,
      "learning_rate": 2.572228773584906e-06,
      "loss": 1.2172,
      "step": 5140
    },
    {
      "epoch": 1.496223126089483,
      "grad_norm": 43.0,
      "learning_rate": 2.5574882075471703e-06,
      "loss": 1.2822,
      "step": 5150
    },
    {
      "epoch": 1.4991284137129575,
      "grad_norm": 37.25,
      "learning_rate": 2.542747641509434e-06,
      "loss": 1.2316,
      "step": 5160
    },
    {
      "epoch": 1.5020337013364324,
      "grad_norm": 71.0,
      "learning_rate": 2.5280070754716984e-06,
      "loss": 1.207,
      "step": 5170
    },
    {
      "epoch": 1.504938988959907,
      "grad_norm": 57.5,
      "learning_rate": 2.5132665094339626e-06,
      "loss": 1.2247,
      "step": 5180
    },
    {
      "epoch": 1.5078442765833817,
      "grad_norm": 42.75,
      "learning_rate": 2.498525943396227e-06,
      "loss": 1.2729,
      "step": 5190
    },
    {
      "epoch": 1.5107495642068565,
      "grad_norm": 75.0,
      "learning_rate": 2.4837853773584906e-06,
      "loss": 1.1602,
      "step": 5200
    },
    {
      "epoch": 1.5107495642068565,
      "eval_accuracy": 0.635008282716731,
      "eval_f1": 0.6342624572051594,
      "eval_loss": 0.6412646770477295,
      "eval_precision": 0.634448953777689,
      "eval_recall": 0.635008282716731,
      "eval_runtime": 22.4012,
      "eval_samples_per_second": 80.844,
      "eval_steps_per_second": 8.125,
      "step": 5200
    },
    {
      "epoch": 1.5136548518303312,
      "grad_norm": 24.875,
      "learning_rate": 2.469044811320755e-06,
      "loss": 1.2224,
      "step": 5210
    },
    {
      "epoch": 1.516560139453806,
      "grad_norm": 46.25,
      "learning_rate": 2.454304245283019e-06,
      "loss": 1.2277,
      "step": 5220
    },
    {
      "epoch": 1.5194654270772805,
      "grad_norm": 49.25,
      "learning_rate": 2.4395636792452833e-06,
      "loss": 1.3351,
      "step": 5230
    },
    {
      "epoch": 1.5223707147007555,
      "grad_norm": 48.25,
      "learning_rate": 2.424823113207547e-06,
      "loss": 1.2759,
      "step": 5240
    },
    {
      "epoch": 1.52527600232423,
      "grad_norm": 43.75,
      "learning_rate": 2.4100825471698114e-06,
      "loss": 1.205,
      "step": 5250
    },
    {
      "epoch": 1.528181289947705,
      "grad_norm": 42.5,
      "learning_rate": 2.3953419811320756e-06,
      "loss": 1.2956,
      "step": 5260
    },
    {
      "epoch": 1.5310865775711795,
      "grad_norm": 48.25,
      "learning_rate": 2.38060141509434e-06,
      "loss": 1.2831,
      "step": 5270
    },
    {
      "epoch": 1.5339918651946542,
      "grad_norm": 73.5,
      "learning_rate": 2.365860849056604e-06,
      "loss": 1.1228,
      "step": 5280
    },
    {
      "epoch": 1.536897152818129,
      "grad_norm": 88.0,
      "learning_rate": 2.3511202830188683e-06,
      "loss": 1.2236,
      "step": 5290
    },
    {
      "epoch": 1.5398024404416037,
      "grad_norm": 46.25,
      "learning_rate": 2.336379716981132e-06,
      "loss": 1.2955,
      "step": 5300
    },
    {
      "epoch": 1.5398024404416037,
      "eval_accuracy": 0.6339039204859194,
      "eval_f1": 0.6328417856318803,
      "eval_loss": 0.6400341391563416,
      "eval_precision": 0.6333204074415757,
      "eval_recall": 0.6339039204859194,
      "eval_runtime": 22.3787,
      "eval_samples_per_second": 80.925,
      "eval_steps_per_second": 8.133,
      "step": 5300
    },
    {
      "epoch": 1.5427077280650785,
      "grad_norm": 83.0,
      "learning_rate": 2.3216391509433964e-06,
      "loss": 1.2296,
      "step": 5310
    },
    {
      "epoch": 1.545613015688553,
      "grad_norm": 39.75,
      "learning_rate": 2.3068985849056606e-06,
      "loss": 1.2541,
      "step": 5320
    },
    {
      "epoch": 1.548518303312028,
      "grad_norm": 49.75,
      "learning_rate": 2.2921580188679244e-06,
      "loss": 1.2598,
      "step": 5330
    },
    {
      "epoch": 1.5514235909355025,
      "grad_norm": 55.5,
      "learning_rate": 2.277417452830189e-06,
      "loss": 1.1931,
      "step": 5340
    },
    {
      "epoch": 1.5543288785589775,
      "grad_norm": 49.0,
      "learning_rate": 2.2626768867924533e-06,
      "loss": 1.2074,
      "step": 5350
    },
    {
      "epoch": 1.557234166182452,
      "grad_norm": 55.75,
      "learning_rate": 2.247936320754717e-06,
      "loss": 1.271,
      "step": 5360
    },
    {
      "epoch": 1.5601394538059268,
      "grad_norm": 56.25,
      "learning_rate": 2.2331957547169813e-06,
      "loss": 1.3754,
      "step": 5370
    },
    {
      "epoch": 1.5630447414294015,
      "grad_norm": 61.25,
      "learning_rate": 2.2184551886792456e-06,
      "loss": 1.3038,
      "step": 5380
    },
    {
      "epoch": 1.5659500290528763,
      "grad_norm": 85.0,
      "learning_rate": 2.20371462264151e-06,
      "loss": 1.2226,
      "step": 5390
    },
    {
      "epoch": 1.568855316676351,
      "grad_norm": 51.25,
      "learning_rate": 2.1889740566037736e-06,
      "loss": 1.1983,
      "step": 5400
    },
    {
      "epoch": 1.568855316676351,
      "eval_accuracy": 0.6361126449475428,
      "eval_f1": 0.6339171740406839,
      "eval_loss": 0.6405659317970276,
      "eval_precision": 0.635879225143538,
      "eval_recall": 0.6361126449475428,
      "eval_runtime": 23.9319,
      "eval_samples_per_second": 75.673,
      "eval_steps_per_second": 7.605,
      "step": 5400
    },
    {
      "epoch": 1.5717606042998256,
      "grad_norm": 53.0,
      "learning_rate": 2.174233490566038e-06,
      "loss": 1.2049,
      "step": 5410
    },
    {
      "epoch": 1.5746658919233005,
      "grad_norm": 42.5,
      "learning_rate": 2.159492924528302e-06,
      "loss": 1.1668,
      "step": 5420
    },
    {
      "epoch": 1.577571179546775,
      "grad_norm": 81.0,
      "learning_rate": 2.1447523584905663e-06,
      "loss": 1.2764,
      "step": 5430
    },
    {
      "epoch": 1.5804764671702498,
      "grad_norm": 34.25,
      "learning_rate": 2.1300117924528305e-06,
      "loss": 1.2099,
      "step": 5440
    },
    {
      "epoch": 1.5833817547937246,
      "grad_norm": 32.0,
      "learning_rate": 2.1152712264150948e-06,
      "loss": 1.2476,
      "step": 5450
    },
    {
      "epoch": 1.5862870424171993,
      "grad_norm": 46.5,
      "learning_rate": 2.1005306603773586e-06,
      "loss": 1.2006,
      "step": 5460
    },
    {
      "epoch": 1.589192330040674,
      "grad_norm": 56.75,
      "learning_rate": 2.085790094339623e-06,
      "loss": 1.2122,
      "step": 5470
    },
    {
      "epoch": 1.5920976176641486,
      "grad_norm": 40.5,
      "learning_rate": 2.071049528301887e-06,
      "loss": 1.2338,
      "step": 5480
    },
    {
      "epoch": 1.5950029052876236,
      "grad_norm": 74.5,
      "learning_rate": 2.0563089622641513e-06,
      "loss": 1.2353,
      "step": 5490
    },
    {
      "epoch": 1.597908192911098,
      "grad_norm": 35.75,
      "learning_rate": 2.041568396226415e-06,
      "loss": 1.1418,
      "step": 5500
    },
    {
      "epoch": 1.597908192911098,
      "eval_accuracy": 0.639425731639978,
      "eval_f1": 0.637184135032376,
      "eval_loss": 0.6405666470527649,
      "eval_precision": 0.639288173918868,
      "eval_recall": 0.639425731639978,
      "eval_runtime": 21.7694,
      "eval_samples_per_second": 83.19,
      "eval_steps_per_second": 8.36,
      "step": 5500
    }
  ],
  "logging_steps": 10,
  "max_steps": 6884,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.9676824584855552e+17,
  "train_batch_size": 10,
  "trial_name": null,
  "trial_params": null
}