{
  "best_metric": 0.03668233007192612,
  "best_model_checkpoint": "./susurro-model/checkpoint-3500",
  "epoch": 8.0,
  "eval_steps": 250,
  "global_step": 5000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.016,
      "grad_norm": 22.783082962036133,
      "learning_rate": 4.0000000000000003e-07,
      "loss": 1.4707,
      "step": 10
    },
    {
      "epoch": 0.032,
      "grad_norm": 15.918609619140625,
      "learning_rate": 8.000000000000001e-07,
      "loss": 1.1318,
      "step": 20
    },
    {
      "epoch": 0.048,
      "grad_norm": 6.734115123748779,
      "learning_rate": 1.2000000000000002e-06,
      "loss": 0.566,
      "step": 30
    },
    {
      "epoch": 0.064,
      "grad_norm": 4.893101692199707,
      "learning_rate": 1.6000000000000001e-06,
      "loss": 0.3064,
      "step": 40
    },
    {
      "epoch": 0.08,
      "grad_norm": 4.515320777893066,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.1913,
      "step": 50
    },
    {
      "epoch": 0.096,
      "grad_norm": 2.93345046043396,
      "learning_rate": 2.4000000000000003e-06,
      "loss": 0.0643,
      "step": 60
    },
    {
      "epoch": 0.112,
      "grad_norm": 1.9217101335525513,
      "learning_rate": 2.8000000000000003e-06,
      "loss": 0.0534,
      "step": 70
    },
    {
      "epoch": 0.128,
      "grad_norm": 2.2198150157928467,
      "learning_rate": 3.2000000000000003e-06,
      "loss": 0.0547,
      "step": 80
    },
    {
      "epoch": 0.144,
      "grad_norm": 1.7477046251296997,
      "learning_rate": 3.6000000000000003e-06,
      "loss": 0.0376,
      "step": 90
    },
    {
      "epoch": 0.16,
      "grad_norm": 2.2940073013305664,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.0414,
      "step": 100
    },
    {
      "epoch": 0.176,
      "grad_norm": 2.83594012260437,
      "learning_rate": 4.4e-06,
      "loss": 0.0501,
      "step": 110
    },
    {
      "epoch": 0.192,
      "grad_norm": 1.6579861640930176,
      "learning_rate": 4.800000000000001e-06,
      "loss": 0.0483,
      "step": 120
    },
    {
      "epoch": 0.208,
      "grad_norm": 2.768068313598633,
      "learning_rate": 5.2e-06,
      "loss": 0.0504,
      "step": 130
    },
    {
      "epoch": 0.224,
      "grad_norm": 1.8856732845306396,
      "learning_rate": 5.600000000000001e-06,
      "loss": 0.0452,
      "step": 140
    },
    {
      "epoch": 0.24,
      "grad_norm": 2.343675374984741,
      "learning_rate": 6e-06,
      "loss": 0.0527,
      "step": 150
    },
    {
      "epoch": 0.256,
      "grad_norm": 2.380032539367676,
      "learning_rate": 6.4000000000000006e-06,
      "loss": 0.0488,
      "step": 160
    },
    {
      "epoch": 0.272,
      "grad_norm": 2.847942352294922,
      "learning_rate": 6.800000000000001e-06,
      "loss": 0.0481,
      "step": 170
    },
    {
      "epoch": 0.288,
      "grad_norm": 2.505783796310425,
      "learning_rate": 7.2000000000000005e-06,
      "loss": 0.0525,
      "step": 180
    },
    {
      "epoch": 0.304,
      "grad_norm": 2.4004359245300293,
      "learning_rate": 7.600000000000001e-06,
      "loss": 0.0475,
      "step": 190
    },
    {
      "epoch": 0.32,
      "grad_norm": 2.5646846294403076,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.0541,
      "step": 200
    },
    {
      "epoch": 0.336,
      "grad_norm": 2.2978780269622803,
      "learning_rate": 8.400000000000001e-06,
      "loss": 0.056,
      "step": 210
    },
    {
      "epoch": 0.352,
      "grad_norm": 1.9104993343353271,
      "learning_rate": 8.8e-06,
      "loss": 0.0507,
      "step": 220
    },
    {
      "epoch": 0.368,
      "grad_norm": 3.259106397628784,
      "learning_rate": 9.200000000000002e-06,
      "loss": 0.0599,
      "step": 230
    },
    {
      "epoch": 0.384,
      "grad_norm": 1.9815133810043335,
      "learning_rate": 9.600000000000001e-06,
      "loss": 0.0531,
      "step": 240
    },
    {
      "epoch": 0.4,
      "grad_norm": 2.202136278152466,
      "learning_rate": 1e-05,
      "loss": 0.0494,
      "step": 250
    },
    {
      "epoch": 0.4,
      "eval_loss": 0.0524754524230957,
      "eval_runtime": 1455.6988,
      "eval_samples_per_second": 3.876,
      "eval_steps_per_second": 3.876,
      "step": 250
    },
    {
      "epoch": 0.416,
      "grad_norm": 2.5911672115325928,
      "learning_rate": 9.978947368421053e-06,
      "loss": 0.0512,
      "step": 260
    },
    {
      "epoch": 0.432,
      "grad_norm": 1.6327568292617798,
      "learning_rate": 9.957894736842106e-06,
      "loss": 0.0485,
      "step": 270
    },
    {
      "epoch": 0.448,
      "grad_norm": 2.290238380432129,
      "learning_rate": 9.936842105263159e-06,
      "loss": 0.0586,
      "step": 280
    },
    {
      "epoch": 0.464,
      "grad_norm": 2.7893710136413574,
      "learning_rate": 9.915789473684211e-06,
      "loss": 0.0597,
      "step": 290
    },
    {
      "epoch": 0.48,
      "grad_norm": 2.4395689964294434,
      "learning_rate": 9.894736842105264e-06,
      "loss": 0.0596,
      "step": 300
    },
    {
      "epoch": 0.496,
      "grad_norm": 2.9003653526306152,
      "learning_rate": 9.873684210526317e-06,
      "loss": 0.0551,
      "step": 310
    },
    {
      "epoch": 0.512,
      "grad_norm": 2.7941534519195557,
      "learning_rate": 9.85263157894737e-06,
      "loss": 0.0579,
      "step": 320
    },
    {
      "epoch": 0.528,
      "grad_norm": 2.705819845199585,
      "learning_rate": 9.831578947368422e-06,
      "loss": 0.059,
      "step": 330
    },
    {
      "epoch": 0.544,
      "grad_norm": 3.2800068855285645,
      "learning_rate": 9.810526315789475e-06,
      "loss": 0.0649,
      "step": 340
    },
    {
      "epoch": 0.56,
      "grad_norm": 1.8395154476165771,
      "learning_rate": 9.789473684210527e-06,
      "loss": 0.0487,
      "step": 350
    },
    {
      "epoch": 0.576,
      "grad_norm": 2.447082757949829,
      "learning_rate": 9.76842105263158e-06,
      "loss": 0.0552,
      "step": 360
    },
    {
      "epoch": 0.592,
      "grad_norm": 2.2035391330718994,
      "learning_rate": 9.747368421052633e-06,
      "loss": 0.0505,
      "step": 370
    },
    {
      "epoch": 0.608,
      "grad_norm": 3.010380506515503,
      "learning_rate": 9.726315789473685e-06,
      "loss": 0.0717,
      "step": 380
    },
    {
      "epoch": 0.624,
      "grad_norm": 2.2279305458068848,
      "learning_rate": 9.705263157894738e-06,
      "loss": 0.0574,
      "step": 390
    },
    {
      "epoch": 0.64,
      "grad_norm": 2.1078450679779053,
      "learning_rate": 9.68421052631579e-06,
      "loss": 0.0549,
      "step": 400
    },
    {
      "epoch": 0.656,
      "grad_norm": 1.9342560768127441,
      "learning_rate": 9.663157894736843e-06,
      "loss": 0.055,
      "step": 410
    },
    {
      "epoch": 0.672,
      "grad_norm": 2.2525532245635986,
      "learning_rate": 9.642105263157896e-06,
      "loss": 0.0562,
      "step": 420
    },
    {
      "epoch": 0.688,
      "grad_norm": 1.9181652069091797,
      "learning_rate": 9.621052631578947e-06,
      "loss": 0.0561,
      "step": 430
    },
    {
      "epoch": 0.704,
      "grad_norm": 2.088951826095581,
      "learning_rate": 9.600000000000001e-06,
      "loss": 0.0575,
      "step": 440
    },
    {
      "epoch": 0.72,
      "grad_norm": 2.951388120651245,
      "learning_rate": 9.578947368421054e-06,
      "loss": 0.0503,
      "step": 450
    },
    {
      "epoch": 0.736,
      "grad_norm": 2.5841259956359863,
      "learning_rate": 9.557894736842107e-06,
      "loss": 0.0613,
      "step": 460
    },
    {
      "epoch": 0.752,
      "grad_norm": 3.1597883701324463,
      "learning_rate": 9.53684210526316e-06,
      "loss": 0.0594,
      "step": 470
    },
    {
      "epoch": 0.768,
      "grad_norm": 1.4449310302734375,
      "learning_rate": 9.515789473684212e-06,
      "loss": 0.0565,
      "step": 480
    },
    {
      "epoch": 0.784,
      "grad_norm": 1.588442325592041,
      "learning_rate": 9.494736842105265e-06,
      "loss": 0.0505,
      "step": 490
    },
    {
      "epoch": 0.8,
      "grad_norm": 2.315497398376465,
      "learning_rate": 9.473684210526315e-06,
      "loss": 0.0571,
      "step": 500
    },
    {
      "epoch": 0.8,
      "eval_loss": 0.04724394157528877,
      "eval_runtime": 1481.9655,
      "eval_samples_per_second": 3.807,
      "eval_steps_per_second": 3.807,
      "step": 500
    },
    {
      "epoch": 0.816,
      "grad_norm": 3.124136209487915,
      "learning_rate": 9.454736842105265e-06,
      "loss": 0.0439,
      "step": 510
    },
    {
      "epoch": 0.832,
      "grad_norm": 2.687971353530884,
      "learning_rate": 9.433684210526318e-06,
      "loss": 0.0535,
      "step": 520
    },
    {
      "epoch": 0.848,
      "grad_norm": 2.442317485809326,
      "learning_rate": 9.412631578947369e-06,
      "loss": 0.0516,
      "step": 530
    },
    {
      "epoch": 0.864,
      "grad_norm": 2.6689493656158447,
      "learning_rate": 9.391578947368421e-06,
      "loss": 0.0565,
      "step": 540
    },
    {
      "epoch": 0.88,
      "grad_norm": 1.959835171699524,
      "learning_rate": 9.370526315789474e-06,
      "loss": 0.0575,
      "step": 550
    },
    {
      "epoch": 0.896,
      "grad_norm": 3.6005074977874756,
      "learning_rate": 9.349473684210526e-06,
      "loss": 0.0444,
      "step": 560
    },
    {
      "epoch": 0.912,
      "grad_norm": 2.1912620067596436,
      "learning_rate": 9.328421052631579e-06,
      "loss": 0.052,
      "step": 570
    },
    {
      "epoch": 0.928,
      "grad_norm": 2.148808240890503,
      "learning_rate": 9.307368421052634e-06,
      "loss": 0.0555,
      "step": 580
    },
    {
      "epoch": 0.944,
      "grad_norm": 2.872231960296631,
      "learning_rate": 9.286315789473686e-06,
      "loss": 0.054,
      "step": 590
    },
    {
      "epoch": 0.96,
      "grad_norm": 3.3271450996398926,
      "learning_rate": 9.265263157894737e-06,
      "loss": 0.0554,
      "step": 600
    },
    {
      "epoch": 0.976,
      "grad_norm": 2.058793067932129,
      "learning_rate": 9.24421052631579e-06,
      "loss": 0.0418,
      "step": 610
    },
    {
      "epoch": 0.992,
      "grad_norm": 2.0928821563720703,
      "learning_rate": 9.225263157894737e-06,
      "loss": 0.0525,
      "step": 620
    },
    {
      "epoch": 1.008,
      "grad_norm": 1.410493016242981,
      "learning_rate": 9.20421052631579e-06,
      "loss": 0.0448,
      "step": 630
    },
    {
      "epoch": 1.024,
      "grad_norm": 1.4376940727233887,
      "learning_rate": 9.183157894736843e-06,
      "loss": 0.0281,
      "step": 640
    },
    {
      "epoch": 1.04,
      "grad_norm": 1.678579568862915,
      "learning_rate": 9.162105263157895e-06,
      "loss": 0.0207,
      "step": 650
    },
    {
      "epoch": 1.056,
      "grad_norm": 2.18996524810791,
      "learning_rate": 9.141052631578948e-06,
      "loss": 0.0288,
      "step": 660
    },
    {
      "epoch": 1.072,
      "grad_norm": 1.0160465240478516,
      "learning_rate": 9.12e-06,
      "loss": 0.0199,
      "step": 670
    },
    {
      "epoch": 1.088,
      "grad_norm": 1.4741309881210327,
      "learning_rate": 9.098947368421053e-06,
      "loss": 0.0248,
      "step": 680
    },
    {
      "epoch": 1.104,
      "grad_norm": 1.5385643243789673,
      "learning_rate": 9.077894736842106e-06,
      "loss": 0.0286,
      "step": 690
    },
    {
      "epoch": 1.12,
      "grad_norm": 1.65945565700531,
      "learning_rate": 9.056842105263159e-06,
      "loss": 0.0331,
      "step": 700
    },
    {
      "epoch": 1.1360000000000001,
      "grad_norm": 2.443169593811035,
      "learning_rate": 9.035789473684211e-06,
      "loss": 0.0328,
      "step": 710
    },
    {
      "epoch": 1.152,
      "grad_norm": 1.950245976448059,
      "learning_rate": 9.014736842105264e-06,
      "loss": 0.0297,
      "step": 720
    },
    {
      "epoch": 1.168,
      "grad_norm": 1.0469032526016235,
      "learning_rate": 8.993684210526317e-06,
      "loss": 0.0282,
      "step": 730
    },
    {
      "epoch": 1.184,
      "grad_norm": 1.5326778888702393,
      "learning_rate": 8.97263157894737e-06,
      "loss": 0.0307,
      "step": 740
    },
    {
      "epoch": 1.2,
      "grad_norm": 0.976245105266571,
      "learning_rate": 8.951578947368422e-06,
      "loss": 0.0293,
      "step": 750
    },
    {
      "epoch": 1.2,
      "eval_loss": 0.04405752569437027,
      "eval_runtime": 1500.9354,
      "eval_samples_per_second": 3.759,
      "eval_steps_per_second": 3.759,
      "step": 750
    },
    {
      "epoch": 1.216,
      "grad_norm": 1.4788917303085327,
      "learning_rate": 8.930526315789475e-06,
      "loss": 0.0341,
      "step": 760
    },
    {
      "epoch": 1.232,
      "grad_norm": 1.398728847503662,
      "learning_rate": 8.909473684210527e-06,
      "loss": 0.0301,
      "step": 770
    },
    {
      "epoch": 1.248,
      "grad_norm": 1.7809382677078247,
      "learning_rate": 8.88842105263158e-06,
      "loss": 0.0265,
      "step": 780
    },
    {
      "epoch": 1.264,
      "grad_norm": 1.9043110609054565,
      "learning_rate": 8.867368421052633e-06,
      "loss": 0.0277,
      "step": 790
    },
    {
      "epoch": 1.28,
      "grad_norm": 2.286780595779419,
      "learning_rate": 8.846315789473685e-06,
      "loss": 0.0241,
      "step": 800
    },
    {
      "epoch": 1.296,
      "grad_norm": 2.867587089538574,
      "learning_rate": 8.825263157894738e-06,
      "loss": 0.0288,
      "step": 810
    },
    {
      "epoch": 1.312,
      "grad_norm": 1.6300053596496582,
      "learning_rate": 8.80421052631579e-06,
      "loss": 0.0229,
      "step": 820
    },
    {
      "epoch": 1.328,
      "grad_norm": 2.1260311603546143,
      "learning_rate": 8.783157894736842e-06,
      "loss": 0.034,
      "step": 830
    },
    {
      "epoch": 1.3439999999999999,
      "grad_norm": 1.2183325290679932,
      "learning_rate": 8.762105263157896e-06,
      "loss": 0.0261,
      "step": 840
    },
    {
      "epoch": 1.3599999999999999,
      "grad_norm": 1.4144467115402222,
      "learning_rate": 8.741052631578949e-06,
      "loss": 0.0258,
      "step": 850
    },
    {
      "epoch": 1.376,
      "grad_norm": 1.658300757408142,
      "learning_rate": 8.720000000000001e-06,
      "loss": 0.0286,
      "step": 860
    },
    {
      "epoch": 1.392,
      "grad_norm": 1.566178321838379,
      "learning_rate": 8.698947368421054e-06,
      "loss": 0.0259,
      "step": 870
    },
    {
      "epoch": 1.408,
      "grad_norm": 1.7910547256469727,
      "learning_rate": 8.677894736842107e-06,
      "loss": 0.024,
      "step": 880
    },
    {
      "epoch": 1.424,
      "grad_norm": 1.882262110710144,
      "learning_rate": 8.65684210526316e-06,
      "loss": 0.0312,
      "step": 890
    },
    {
      "epoch": 1.44,
      "grad_norm": 1.2737823724746704,
      "learning_rate": 8.63578947368421e-06,
      "loss": 0.0324,
      "step": 900
    },
    {
      "epoch": 1.456,
      "grad_norm": 1.954757809638977,
      "learning_rate": 8.614736842105263e-06,
      "loss": 0.0276,
      "step": 910
    },
    {
      "epoch": 1.472,
      "grad_norm": 2.4381039142608643,
      "learning_rate": 8.593684210526315e-06,
      "loss": 0.0258,
      "step": 920
    },
    {
      "epoch": 1.488,
      "grad_norm": 1.962517499923706,
      "learning_rate": 8.57263157894737e-06,
      "loss": 0.0357,
      "step": 930
    },
    {
      "epoch": 1.504,
      "grad_norm": 1.7667138576507568,
      "learning_rate": 8.551578947368422e-06,
      "loss": 0.0358,
      "step": 940
    },
    {
      "epoch": 1.52,
      "grad_norm": 1.6980652809143066,
      "learning_rate": 8.530526315789475e-06,
      "loss": 0.0247,
      "step": 950
    },
    {
      "epoch": 1.536,
      "grad_norm": 1.765106201171875,
      "learning_rate": 8.509473684210528e-06,
      "loss": 0.0289,
      "step": 960
    },
    {
      "epoch": 1.552,
      "grad_norm": 2.0965826511383057,
      "learning_rate": 8.488421052631579e-06,
      "loss": 0.0293,
      "step": 970
    },
    {
      "epoch": 1.568,
      "grad_norm": 1.2776910066604614,
      "learning_rate": 8.467368421052631e-06,
      "loss": 0.0248,
      "step": 980
    },
    {
      "epoch": 1.584,
      "grad_norm": 1.9160679578781128,
      "learning_rate": 8.446315789473684e-06,
      "loss": 0.0327,
      "step": 990
    },
    {
      "epoch": 1.6,
      "grad_norm": 2.084198474884033,
      "learning_rate": 8.425263157894737e-06,
      "loss": 0.0248,
      "step": 1000
    },
    {
      "epoch": 1.6,
      "eval_loss": 0.04179142042994499,
      "eval_runtime": 1520.1889,
      "eval_samples_per_second": 3.711,
      "eval_steps_per_second": 3.711,
      "step": 1000
    },
    {
      "epoch": 1.616,
      "grad_norm": 1.8233721256256104,
      "learning_rate": 8.404210526315791e-06,
      "loss": 0.0325,
      "step": 1010
    },
    {
      "epoch": 1.6320000000000001,
      "grad_norm": 2.6856882572174072,
      "learning_rate": 8.383157894736844e-06,
      "loss": 0.0345,
      "step": 1020
    },
    {
      "epoch": 1.6480000000000001,
      "grad_norm": 1.6194241046905518,
      "learning_rate": 8.362105263157896e-06,
      "loss": 0.0248,
      "step": 1030
    },
    {
      "epoch": 1.6640000000000001,
      "grad_norm": 2.0623207092285156,
      "learning_rate": 8.341052631578947e-06,
      "loss": 0.0297,
      "step": 1040
    },
    {
      "epoch": 1.6800000000000002,
      "grad_norm": 2.122952461242676,
      "learning_rate": 8.32e-06,
      "loss": 0.0346,
      "step": 1050
    },
    {
      "epoch": 1.696,
      "grad_norm": 1.4452091455459595,
      "learning_rate": 8.298947368421053e-06,
      "loss": 0.03,
      "step": 1060
    },
    {
      "epoch": 1.712,
      "grad_norm": 2.9357564449310303,
      "learning_rate": 8.277894736842105e-06,
      "loss": 0.026,
      "step": 1070
    },
    {
      "epoch": 1.728,
      "grad_norm": 1.0874323844909668,
      "learning_rate": 8.256842105263158e-06,
      "loss": 0.0294,
      "step": 1080
    },
    {
      "epoch": 1.744,
      "grad_norm": 1.1625168323516846,
      "learning_rate": 8.235789473684212e-06,
      "loss": 0.0291,
      "step": 1090
    },
    {
      "epoch": 1.76,
      "grad_norm": 1.3480169773101807,
      "learning_rate": 8.214736842105265e-06,
      "loss": 0.034,
      "step": 1100
    },
    {
      "epoch": 1.776,
      "grad_norm": 1.1702100038528442,
      "learning_rate": 8.193684210526316e-06,
      "loss": 0.029,
      "step": 1110
    },
    {
      "epoch": 1.792,
      "grad_norm": 2.0428359508514404,
      "learning_rate": 8.172631578947369e-06,
      "loss": 0.023,
      "step": 1120
    },
    {
      "epoch": 1.808,
      "grad_norm": 0.7537345886230469,
      "learning_rate": 8.151578947368421e-06,
      "loss": 0.03,
      "step": 1130
    },
    {
      "epoch": 1.8239999999999998,
      "grad_norm": 2.184401273727417,
      "learning_rate": 8.130526315789474e-06,
      "loss": 0.0327,
      "step": 1140
    },
    {
      "epoch": 1.8399999999999999,
      "grad_norm": 2.1545660495758057,
      "learning_rate": 8.109473684210527e-06,
      "loss": 0.0275,
      "step": 1150
    },
    {
      "epoch": 1.8559999999999999,
      "grad_norm": 2.535369396209717,
      "learning_rate": 8.08842105263158e-06,
      "loss": 0.0292,
      "step": 1160
    },
    {
      "epoch": 1.8719999999999999,
      "grad_norm": 1.8030229806900024,
      "learning_rate": 8.067368421052632e-06,
      "loss": 0.0264,
      "step": 1170
    },
    {
      "epoch": 1.888,
      "grad_norm": 1.8990155458450317,
      "learning_rate": 8.046315789473686e-06,
      "loss": 0.0365,
      "step": 1180
    },
    {
      "epoch": 1.904,
      "grad_norm": 1.4508986473083496,
      "learning_rate": 8.025263157894737e-06,
      "loss": 0.0251,
      "step": 1190
    },
    {
      "epoch": 1.92,
      "grad_norm": 1.0793194770812988,
      "learning_rate": 8.00421052631579e-06,
      "loss": 0.0274,
      "step": 1200
    },
    {
      "epoch": 1.936,
      "grad_norm": 1.9805344343185425,
      "learning_rate": 7.983157894736842e-06,
      "loss": 0.0299,
      "step": 1210
    },
    {
      "epoch": 1.952,
      "grad_norm": 2.258500576019287,
      "learning_rate": 7.962105263157895e-06,
      "loss": 0.0266,
      "step": 1220
    },
    {
      "epoch": 1.968,
      "grad_norm": 1.4639638662338257,
      "learning_rate": 7.941052631578948e-06,
      "loss": 0.026,
      "step": 1230
    },
    {
      "epoch": 1.984,
      "grad_norm": 1.3057489395141602,
      "learning_rate": 7.92e-06,
      "loss": 0.0305,
      "step": 1240
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.29533052444458,
      "learning_rate": 7.898947368421053e-06,
      "loss": 0.0265,
      "step": 1250
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.04163273423910141,
      "eval_runtime": 1542.8649,
      "eval_samples_per_second": 3.657,
      "eval_steps_per_second": 3.657,
      "step": 1250
    },
    {
      "epoch": 2.016,
      "grad_norm": 0.5275375843048096,
      "learning_rate": 7.877894736842106e-06,
      "loss": 0.0117,
      "step": 1260
    },
    {
      "epoch": 2.032,
      "grad_norm": 2.0590779781341553,
      "learning_rate": 7.856842105263158e-06,
      "loss": 0.0156,
      "step": 1270
    },
    {
      "epoch": 2.048,
      "grad_norm": 1.2595654726028442,
      "learning_rate": 7.835789473684211e-06,
      "loss": 0.0105,
      "step": 1280
    },
    {
      "epoch": 2.064,
      "grad_norm": 1.605797529220581,
      "learning_rate": 7.814736842105264e-06,
      "loss": 0.0176,
      "step": 1290
    },
    {
      "epoch": 2.08,
      "grad_norm": 1.2974120378494263,
      "learning_rate": 7.793684210526316e-06,
      "loss": 0.0196,
      "step": 1300
    },
    {
      "epoch": 2.096,
      "grad_norm": 3.662148952484131,
      "learning_rate": 7.772631578947369e-06,
      "loss": 0.0144,
      "step": 1310
    },
    {
      "epoch": 2.112,
      "grad_norm": 0.9286357164382935,
      "learning_rate": 7.751578947368422e-06,
      "loss": 0.0126,
      "step": 1320
    },
    {
      "epoch": 2.128,
      "grad_norm": 1.4938029050827026,
      "learning_rate": 7.730526315789474e-06,
      "loss": 0.0181,
      "step": 1330
    },
    {
      "epoch": 2.144,
      "grad_norm": 1.9004080295562744,
      "learning_rate": 7.709473684210527e-06,
      "loss": 0.0192,
      "step": 1340
    },
    {
      "epoch": 2.16,
      "grad_norm": 0.7749639749526978,
      "learning_rate": 7.68842105263158e-06,
      "loss": 0.0146,
      "step": 1350
    },
    {
      "epoch": 2.176,
      "grad_norm": 1.69952392578125,
      "learning_rate": 7.667368421052632e-06,
      "loss": 0.0169,
      "step": 1360
    },
    {
      "epoch": 2.192,
      "grad_norm": 0.6169717311859131,
      "learning_rate": 7.646315789473685e-06,
      "loss": 0.0102,
      "step": 1370
    },
    {
      "epoch": 2.208,
      "grad_norm": 3.5800750255584717,
      "learning_rate": 7.6252631578947376e-06,
      "loss": 0.014,
      "step": 1380
    },
    {
      "epoch": 2.224,
      "grad_norm": 0.6337082982063293,
      "learning_rate": 7.60421052631579e-06,
      "loss": 0.0161,
      "step": 1390
    },
    {
      "epoch": 2.24,
      "grad_norm": 0.7504271268844604,
      "learning_rate": 7.583157894736842e-06,
      "loss": 0.0145,
      "step": 1400
    },
    {
      "epoch": 2.2560000000000002,
      "grad_norm": 0.6736518740653992,
      "learning_rate": 7.562105263157895e-06,
      "loss": 0.0105,
      "step": 1410
    },
    {
      "epoch": 2.2720000000000002,
      "grad_norm": 1.1138242483139038,
      "learning_rate": 7.541052631578948e-06,
      "loss": 0.0136,
      "step": 1420
    },
    {
      "epoch": 2.288,
      "grad_norm": 1.6553155183792114,
      "learning_rate": 7.520000000000001e-06,
      "loss": 0.0135,
      "step": 1430
    },
    {
      "epoch": 2.304,
      "grad_norm": 1.3883057832717896,
      "learning_rate": 7.4989473684210535e-06,
      "loss": 0.0136,
      "step": 1440
    },
    {
      "epoch": 2.32,
      "grad_norm": 1.3742883205413818,
      "learning_rate": 7.477894736842106e-06,
      "loss": 0.0146,
      "step": 1450
    },
    {
      "epoch": 2.336,
      "grad_norm": 2.129136562347412,
      "learning_rate": 7.456842105263159e-06,
      "loss": 0.0142,
      "step": 1460
    },
    {
      "epoch": 2.352,
      "grad_norm": 1.771249532699585,
      "learning_rate": 7.435789473684211e-06,
      "loss": 0.0182,
      "step": 1470
    },
    {
      "epoch": 2.368,
      "grad_norm": 0.906812846660614,
      "learning_rate": 7.414736842105263e-06,
      "loss": 0.0127,
      "step": 1480
    },
    {
      "epoch": 2.384,
      "grad_norm": 1.0035991668701172,
      "learning_rate": 7.393684210526316e-06,
      "loss": 0.0148,
      "step": 1490
    },
    {
      "epoch": 2.4,
      "grad_norm": 1.0261139869689941,
      "learning_rate": 7.3726315789473694e-06,
      "loss": 0.0162,
      "step": 1500
    },
    {
      "epoch": 2.4,
      "eval_loss": 0.04060380160808563,
      "eval_runtime": 1542.3166,
      "eval_samples_per_second": 3.658,
      "eval_steps_per_second": 3.658,
      "step": 1500
    },
    {
      "epoch": 2.416,
      "grad_norm": 0.9582253098487854,
      "learning_rate": 7.351578947368422e-06,
      "loss": 0.0139,
      "step": 1510
    },
    {
      "epoch": 2.432,
      "grad_norm": 1.6380131244659424,
      "learning_rate": 7.330526315789475e-06,
      "loss": 0.0175,
      "step": 1520
    },
    {
      "epoch": 2.448,
      "grad_norm": 2.057828426361084,
      "learning_rate": 7.309473684210527e-06,
      "loss": 0.0168,
      "step": 1530
    },
    {
      "epoch": 2.464,
      "grad_norm": 1.6444942951202393,
      "learning_rate": 7.288421052631579e-06,
      "loss": 0.0151,
      "step": 1540
    },
    {
      "epoch": 2.48,
      "grad_norm": 0.5916880369186401,
      "learning_rate": 7.267368421052632e-06,
      "loss": 0.0184,
      "step": 1550
    },
    {
      "epoch": 2.496,
      "grad_norm": 1.4819905757904053,
      "learning_rate": 7.2463157894736845e-06,
      "loss": 0.0215,
      "step": 1560
    },
    {
      "epoch": 2.512,
      "grad_norm": 0.944874107837677,
      "learning_rate": 7.225263157894737e-06,
      "loss": 0.0146,
      "step": 1570
    },
    {
      "epoch": 2.528,
      "grad_norm": 1.1218395233154297,
      "learning_rate": 7.20421052631579e-06,
      "loss": 0.0155,
      "step": 1580
    },
    {
      "epoch": 2.544,
      "grad_norm": 1.7295416593551636,
      "learning_rate": 7.183157894736843e-06,
      "loss": 0.0149,
      "step": 1590
    },
    {
      "epoch": 2.56,
      "grad_norm": 2.064664602279663,
      "learning_rate": 7.162105263157896e-06,
      "loss": 0.0209,
      "step": 1600
    },
    {
      "epoch": 2.576,
      "grad_norm": 1.4919706583023071,
      "learning_rate": 7.141052631578948e-06,
      "loss": 0.0174,
      "step": 1610
    },
    {
      "epoch": 2.592,
      "grad_norm": 1.8185276985168457,
      "learning_rate": 7.1200000000000004e-06,
      "loss": 0.0153,
      "step": 1620
    },
    {
      "epoch": 2.608,
      "grad_norm": 1.0406450033187866,
      "learning_rate": 7.098947368421053e-06,
      "loss": 0.017,
      "step": 1630
    },
    {
      "epoch": 2.624,
      "grad_norm": 2.1542301177978516,
      "learning_rate": 7.077894736842106e-06,
      "loss": 0.0171,
      "step": 1640
    },
    {
      "epoch": 2.64,
      "grad_norm": 1.3869684934616089,
      "learning_rate": 7.056842105263158e-06,
      "loss": 0.0141,
      "step": 1650
    },
    {
      "epoch": 2.656,
      "grad_norm": 1.0020642280578613,
      "learning_rate": 7.035789473684211e-06,
      "loss": 0.0176,
      "step": 1660
    },
    {
      "epoch": 2.672,
      "grad_norm": 2.786790370941162,
      "learning_rate": 7.0147368421052646e-06,
      "loss": 0.0132,
      "step": 1670
    },
    {
      "epoch": 2.6879999999999997,
      "grad_norm": 1.0302237272262573,
      "learning_rate": 6.9936842105263155e-06,
      "loss": 0.0139,
      "step": 1680
    },
    {
      "epoch": 2.7039999999999997,
      "grad_norm": 1.0650746822357178,
      "learning_rate": 6.972631578947369e-06,
      "loss": 0.0137,
      "step": 1690
    },
    {
      "epoch": 2.7199999999999998,
      "grad_norm": 1.5508835315704346,
      "learning_rate": 6.951578947368422e-06,
      "loss": 0.0151,
      "step": 1700
    },
    {
      "epoch": 2.7359999999999998,
      "grad_norm": 0.7046643495559692,
      "learning_rate": 6.930526315789474e-06,
      "loss": 0.0127,
      "step": 1710
    },
    {
      "epoch": 2.752,
      "grad_norm": 1.3175015449523926,
      "learning_rate": 6.909473684210527e-06,
      "loss": 0.015,
      "step": 1720
    },
    {
      "epoch": 2.768,
      "grad_norm": 2.0063388347625732,
      "learning_rate": 6.88842105263158e-06,
      "loss": 0.0143,
      "step": 1730
    },
    {
      "epoch": 2.784,
      "grad_norm": 1.5595343112945557,
      "learning_rate": 6.867368421052632e-06,
      "loss": 0.0175,
      "step": 1740
    },
    {
      "epoch": 2.8,
      "grad_norm": 1.6928199529647827,
      "learning_rate": 6.846315789473684e-06,
      "loss": 0.0147,
      "step": 1750
    },
    {
      "epoch": 2.8,
      "eval_loss": 0.03860897570848465,
      "eval_runtime": 9405.2396,
      "eval_samples_per_second": 0.6,
      "eval_steps_per_second": 0.6,
      "step": 1750
    },
    {
      "epoch": 2.816,
      "grad_norm": 2.1976370811462402,
      "learning_rate": 6.825263157894737e-06,
      "loss": 0.0158,
      "step": 1760
    },
    {
      "epoch": 2.832,
      "grad_norm": 1.3669664859771729,
      "learning_rate": 6.804210526315789e-06,
      "loss": 0.0213,
      "step": 1770
    },
    {
      "epoch": 2.848,
      "grad_norm": 1.8074935674667358,
      "learning_rate": 6.783157894736843e-06,
      "loss": 0.018,
      "step": 1780
    },
    {
      "epoch": 2.864,
      "grad_norm": 0.8959200382232666,
      "learning_rate": 6.7621052631578956e-06,
      "loss": 0.0137,
      "step": 1790
    },
    {
      "epoch": 2.88,
      "grad_norm": 1.1795583963394165,
      "learning_rate": 6.741052631578948e-06,
      "loss": 0.0157,
      "step": 1800
    },
    {
      "epoch": 2.896,
      "grad_norm": 1.4148980379104614,
      "learning_rate": 6.720000000000001e-06,
      "loss": 0.0174,
      "step": 1810
    },
    {
      "epoch": 2.912,
      "grad_norm": 0.8638339042663574,
      "learning_rate": 6.698947368421053e-06,
      "loss": 0.0209,
      "step": 1820
    },
    {
      "epoch": 2.928,
      "grad_norm": 0.9411534667015076,
      "learning_rate": 6.677894736842105e-06,
      "loss": 0.0114,
      "step": 1830
    },
    {
      "epoch": 2.944,
      "grad_norm": 0.6806014180183411,
      "learning_rate": 6.656842105263158e-06,
      "loss": 0.02,
      "step": 1840
    },
    {
      "epoch": 2.96,
      "grad_norm": 0.9796069264411926,
      "learning_rate": 6.635789473684211e-06,
      "loss": 0.0116,
      "step": 1850
    },
    {
      "epoch": 2.976,
      "grad_norm": 3.364715814590454,
      "learning_rate": 6.614736842105264e-06,
      "loss": 0.0132,
      "step": 1860
    },
    {
      "epoch": 2.992,
      "grad_norm": 0.729645848274231,
      "learning_rate": 6.593684210526317e-06,
      "loss": 0.0147,
      "step": 1870
    },
    {
      "epoch": 3.008,
      "grad_norm": 2.3798277378082275,
      "learning_rate": 6.5726315789473695e-06,
      "loss": 0.0174,
      "step": 1880
    },
    {
      "epoch": 3.024,
      "grad_norm": 0.6777742505073547,
      "learning_rate": 6.551578947368421e-06,
      "loss": 0.0105,
      "step": 1890
    },
    {
      "epoch": 3.04,
      "grad_norm": 1.1942986249923706,
      "learning_rate": 6.530526315789474e-06,
      "loss": 0.0098,
      "step": 1900
    },
    {
      "epoch": 3.056,
      "grad_norm": 0.5124061107635498,
      "learning_rate": 6.509473684210527e-06,
      "loss": 0.0122,
      "step": 1910
    },
    {
      "epoch": 3.072,
      "grad_norm": 0.9624982476234436,
      "learning_rate": 6.488421052631579e-06,
      "loss": 0.0075,
      "step": 1920
    },
    {
      "epoch": 3.088,
      "grad_norm": 0.6776851415634155,
      "learning_rate": 6.467368421052632e-06,
      "loss": 0.0078,
      "step": 1930
    },
    {
      "epoch": 3.104,
      "grad_norm": 0.4030074179172516,
      "learning_rate": 6.4463157894736845e-06,
      "loss": 0.0078,
      "step": 1940
    },
    {
      "epoch": 3.12,
      "grad_norm": 0.5822834372520447,
      "learning_rate": 6.425263157894738e-06,
      "loss": 0.0058,
      "step": 1950
    },
    {
      "epoch": 3.136,
      "grad_norm": 0.23138459026813507,
      "learning_rate": 6.404210526315791e-06,
      "loss": 0.0062,
      "step": 1960
    },
    {
      "epoch": 3.152,
      "grad_norm": 1.4345389604568481,
      "learning_rate": 6.3831578947368425e-06,
      "loss": 0.009,
      "step": 1970
    },
    {
      "epoch": 3.168,
      "grad_norm": 1.3896571397781372,
      "learning_rate": 6.362105263157895e-06,
      "loss": 0.0091,
      "step": 1980
    },
    {
      "epoch": 3.184,
      "grad_norm": 0.7762678861618042,
      "learning_rate": 6.341052631578948e-06,
      "loss": 0.0078,
      "step": 1990
    },
    {
      "epoch": 3.2,
      "grad_norm": 0.7915733456611633,
      "learning_rate": 6.3200000000000005e-06,
      "loss": 0.0061,
      "step": 2000
    },
    {
      "epoch": 3.2,
      "eval_loss": 0.039765432476997375,
      "eval_runtime": 1647.5775,
      "eval_samples_per_second": 3.424,
      "eval_steps_per_second": 3.424,
      "step": 2000
    },
    {
      "epoch": 3.216,
      "grad_norm": 0.6915295124053955,
      "learning_rate": 6.298947368421053e-06,
      "loss": 0.0072,
      "step": 2010
    },
    {
      "epoch": 3.232,
      "grad_norm": 1.2451772689819336,
      "learning_rate": 6.277894736842106e-06,
      "loss": 0.0085,
      "step": 2020
    },
    {
      "epoch": 3.248,
      "grad_norm": 0.7011457085609436,
      "learning_rate": 6.256842105263159e-06,
      "loss": 0.0095,
      "step": 2030
    },
    {
      "epoch": 3.2640000000000002,
      "grad_norm": 0.3362152874469757,
      "learning_rate": 6.23578947368421e-06,
      "loss": 0.0068,
      "step": 2040
    },
    {
      "epoch": 3.2800000000000002,
      "grad_norm": 2.8105270862579346,
      "learning_rate": 6.214736842105264e-06,
      "loss": 0.0065,
      "step": 2050
    },
    {
      "epoch": 3.296,
      "grad_norm": 0.5386505722999573,
      "learning_rate": 6.193684210526316e-06,
      "loss": 0.0064,
      "step": 2060
    },
    {
      "epoch": 3.312,
      "grad_norm": 1.7395727634429932,
      "learning_rate": 6.172631578947369e-06,
      "loss": 0.0054,
      "step": 2070
    },
    {
      "epoch": 3.328,
      "grad_norm": 0.7333295345306396,
      "learning_rate": 6.151578947368422e-06,
      "loss": 0.0074,
      "step": 2080
    },
    {
      "epoch": 3.344,
      "grad_norm": 0.36293238401412964,
      "learning_rate": 6.130526315789474e-06,
      "loss": 0.0083,
      "step": 2090
    },
    {
      "epoch": 3.36,
      "grad_norm": 0.7955753207206726,
      "learning_rate": 6.109473684210527e-06,
      "loss": 0.0057,
      "step": 2100
    },
    {
      "epoch": 3.376,
      "grad_norm": 1.8287686109542847,
      "learning_rate": 6.088421052631579e-06,
      "loss": 0.0071,
      "step": 2110
    },
    {
      "epoch": 3.392,
      "grad_norm": 0.6369631886482239,
      "learning_rate": 6.0673684210526315e-06,
      "loss": 0.0073,
      "step": 2120
    },
    {
      "epoch": 3.408,
      "grad_norm": 0.9539582133293152,
      "learning_rate": 6.046315789473685e-06,
      "loss": 0.0114,
      "step": 2130
    },
    {
      "epoch": 3.424,
      "grad_norm": 0.8086543083190918,
      "learning_rate": 6.025263157894738e-06,
      "loss": 0.0099,
      "step": 2140
    },
    {
      "epoch": 3.44,
      "grad_norm": 0.40894851088523865,
      "learning_rate": 6.00421052631579e-06,
      "loss": 0.0066,
      "step": 2150
    },
    {
      "epoch": 3.456,
      "grad_norm": 0.39193499088287354,
      "learning_rate": 5.983157894736843e-06,
      "loss": 0.0044,
      "step": 2160
    },
    {
      "epoch": 3.472,
      "grad_norm": 1.0509319305419922,
      "learning_rate": 5.962105263157896e-06,
      "loss": 0.0081,
      "step": 2170
    },
    {
      "epoch": 3.488,
      "grad_norm": 2.615633964538574,
      "learning_rate": 5.941052631578947e-06,
      "loss": 0.0076,
      "step": 2180
    },
    {
      "epoch": 3.504,
      "grad_norm": 0.9589405059814453,
      "learning_rate": 5.92e-06,
      "loss": 0.0078,
      "step": 2190
    },
    {
      "epoch": 3.52,
      "grad_norm": 1.0844931602478027,
      "learning_rate": 5.898947368421053e-06,
      "loss": 0.0064,
      "step": 2200
    },
    {
      "epoch": 3.536,
      "grad_norm": 0.5187249183654785,
      "learning_rate": 5.8800000000000005e-06,
      "loss": 0.0094,
      "step": 2210
    },
    {
      "epoch": 3.552,
      "grad_norm": 1.4108637571334839,
      "learning_rate": 5.858947368421053e-06,
      "loss": 0.0075,
      "step": 2220
    },
    {
      "epoch": 3.568,
      "grad_norm": 0.7219350337982178,
      "learning_rate": 5.837894736842106e-06,
      "loss": 0.0091,
      "step": 2230
    },
    {
      "epoch": 3.584,
      "grad_norm": 0.8116509914398193,
      "learning_rate": 5.816842105263158e-06,
      "loss": 0.008,
      "step": 2240
    },
    {
      "epoch": 3.6,
      "grad_norm": 0.27414098381996155,
      "learning_rate": 5.795789473684211e-06,
      "loss": 0.0094,
      "step": 2250
    },
    {
      "epoch": 3.6,
      "eval_loss": 0.03814266622066498,
      "eval_runtime": 1450.6262,
      "eval_samples_per_second": 3.889,
      "eval_steps_per_second": 3.889,
      "step": 2250
    },
    {
      "epoch": 3.616,
      "grad_norm": 0.7917847633361816,
      "learning_rate": 5.774736842105264e-06,
      "loss": 0.0074,
      "step": 2260
    },
    {
      "epoch": 3.632,
      "grad_norm": 0.7660414576530457,
      "learning_rate": 5.7536842105263155e-06,
      "loss": 0.0081,
      "step": 2270
    },
    {
      "epoch": 3.648,
      "grad_norm": 0.9440973401069641,
      "learning_rate": 5.732631578947368e-06,
      "loss": 0.0072,
      "step": 2280
    },
    {
      "epoch": 3.664,
      "grad_norm": 1.79617440700531,
      "learning_rate": 5.711578947368422e-06,
      "loss": 0.0094,
      "step": 2290
    },
    {
      "epoch": 3.68,
      "grad_norm": 0.6188564896583557,
      "learning_rate": 5.690526315789474e-06,
      "loss": 0.0101,
      "step": 2300
    },
    {
      "epoch": 3.6959999999999997,
      "grad_norm": 1.2464138269424438,
      "learning_rate": 5.669473684210527e-06,
      "loss": 0.0073,
      "step": 2310
    },
    {
      "epoch": 3.7119999999999997,
      "grad_norm": 0.35902971029281616,
      "learning_rate": 5.64842105263158e-06,
      "loss": 0.007,
      "step": 2320
    },
    {
      "epoch": 3.7279999999999998,
      "grad_norm": 0.5012713670730591,
      "learning_rate": 5.627368421052632e-06,
      "loss": 0.0046,
      "step": 2330
    },
    {
      "epoch": 3.7439999999999998,
      "grad_norm": 1.280382752418518,
      "learning_rate": 5.606315789473684e-06,
      "loss": 0.0105,
      "step": 2340
    },
    {
      "epoch": 3.76,
      "grad_norm": 0.7401492595672607,
      "learning_rate": 5.585263157894737e-06,
      "loss": 0.0079,
      "step": 2350
    },
    {
      "epoch": 3.776,
      "grad_norm": 0.9203608632087708,
      "learning_rate": 5.5642105263157894e-06,
      "loss": 0.0052,
      "step": 2360
    },
    {
      "epoch": 3.792,
      "grad_norm": 1.1353886127471924,
      "learning_rate": 5.543157894736843e-06,
      "loss": 0.0087,
      "step": 2370
    },
    {
      "epoch": 3.808,
      "grad_norm": 1.4035215377807617,
      "learning_rate": 5.522105263157896e-06,
      "loss": 0.0077,
      "step": 2380
    },
    {
      "epoch": 3.824,
      "grad_norm": 0.429685115814209,
      "learning_rate": 5.501052631578948e-06,
      "loss": 0.0084,
      "step": 2390
    },
    {
      "epoch": 3.84,
      "grad_norm": 0.868377685546875,
      "learning_rate": 5.480000000000001e-06,
      "loss": 0.0114,
      "step": 2400
    },
    {
      "epoch": 3.856,
      "grad_norm": 0.9295462369918823,
      "learning_rate": 5.458947368421053e-06,
      "loss": 0.0066,
      "step": 2410
    },
    {
      "epoch": 3.872,
      "grad_norm": 1.1884615421295166,
      "learning_rate": 5.437894736842105e-06,
      "loss": 0.0083,
      "step": 2420
    },
    {
      "epoch": 3.888,
      "grad_norm": 0.9172056913375854,
      "learning_rate": 5.416842105263158e-06,
      "loss": 0.0071,
      "step": 2430
    },
    {
      "epoch": 3.904,
      "grad_norm": 0.7041633725166321,
      "learning_rate": 5.395789473684211e-06,
      "loss": 0.0069,
      "step": 2440
    },
    {
      "epoch": 3.92,
      "grad_norm": 0.7116276621818542,
      "learning_rate": 5.374736842105263e-06,
      "loss": 0.0076,
      "step": 2450
    },
    {
      "epoch": 3.936,
      "grad_norm": 0.7440093159675598,
      "learning_rate": 5.353684210526317e-06,
      "loss": 0.0076,
      "step": 2460
    },
    {
      "epoch": 3.952,
      "grad_norm": 1.1006033420562744,
      "learning_rate": 5.3326315789473695e-06,
      "loss": 0.0077,
      "step": 2470
    },
    {
      "epoch": 3.968,
      "grad_norm": 1.046329140663147,
      "learning_rate": 5.311578947368421e-06,
      "loss": 0.0089,
      "step": 2480
    },
    {
      "epoch": 3.984,
      "grad_norm": 0.9022756814956665,
      "learning_rate": 5.290526315789474e-06,
      "loss": 0.0073,
      "step": 2490
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.0019030570983887,
      "learning_rate": 5.269473684210527e-06,
      "loss": 0.0074,
      "step": 2500
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.03700412064790726,
      "eval_runtime": 1477.9736,
      "eval_samples_per_second": 3.817,
      "eval_steps_per_second": 3.817,
      "step": 2500
    },
    {
      "epoch": 4.016,
      "grad_norm": 1.2217806577682495,
      "learning_rate": 5.248421052631579e-06,
      "loss": 0.0037,
      "step": 2510
    },
    {
      "epoch": 4.032,
      "grad_norm": 0.7005161643028259,
      "learning_rate": 5.227368421052632e-06,
      "loss": 0.0047,
      "step": 2520
    },
    {
      "epoch": 4.048,
      "grad_norm": 0.3553329110145569,
      "learning_rate": 5.2063157894736846e-06,
      "loss": 0.0038,
      "step": 2530
    },
    {
      "epoch": 4.064,
      "grad_norm": 0.4705675542354584,
      "learning_rate": 5.185263157894738e-06,
      "loss": 0.0038,
      "step": 2540
    },
    {
      "epoch": 4.08,
      "grad_norm": 0.9150580763816833,
      "learning_rate": 5.164210526315791e-06,
      "loss": 0.0055,
      "step": 2550
    },
    {
      "epoch": 4.096,
      "grad_norm": 0.7440536618232727,
      "learning_rate": 5.1431578947368425e-06,
      "loss": 0.0064,
      "step": 2560
    },
    {
      "epoch": 4.112,
      "grad_norm": 0.16911202669143677,
      "learning_rate": 5.122105263157895e-06,
      "loss": 0.0033,
      "step": 2570
    },
    {
      "epoch": 4.128,
      "grad_norm": 0.20625616610050201,
      "learning_rate": 5.101052631578948e-06,
      "loss": 0.0029,
      "step": 2580
    },
    {
      "epoch": 4.144,
      "grad_norm": 1.1767489910125732,
      "learning_rate": 5.0800000000000005e-06,
      "loss": 0.0043,
      "step": 2590
    },
    {
      "epoch": 4.16,
      "grad_norm": 0.7702798247337341,
      "learning_rate": 5.058947368421053e-06,
      "loss": 0.002,
      "step": 2600
    },
    {
      "epoch": 4.176,
      "grad_norm": 0.9442148804664612,
      "learning_rate": 5.037894736842106e-06,
      "loss": 0.0049,
      "step": 2610
    },
    {
      "epoch": 4.192,
      "grad_norm": 0.439223974943161,
      "learning_rate": 5.0168421052631585e-06,
      "loss": 0.0033,
      "step": 2620
    },
    {
      "epoch": 4.208,
      "grad_norm": 0.2672136127948761,
      "learning_rate": 4.995789473684211e-06,
      "loss": 0.0055,
      "step": 2630
    },
    {
      "epoch": 4.224,
      "grad_norm": 0.1722344309091568,
      "learning_rate": 4.974736842105263e-06,
      "loss": 0.0049,
      "step": 2640
    },
    {
      "epoch": 4.24,
      "grad_norm": 0.8054834604263306,
      "learning_rate": 4.953684210526316e-06,
      "loss": 0.0028,
      "step": 2650
    },
    {
      "epoch": 4.256,
      "grad_norm": 0.38357770442962646,
      "learning_rate": 4.932631578947369e-06,
      "loss": 0.0022,
      "step": 2660
    },
    {
      "epoch": 4.272,
      "grad_norm": 0.7503048777580261,
      "learning_rate": 4.911578947368422e-06,
      "loss": 0.0021,
      "step": 2670
    },
    {
      "epoch": 4.288,
      "grad_norm": 0.95183265209198,
      "learning_rate": 4.8905263157894735e-06,
      "loss": 0.0035,
      "step": 2680
    },
    {
      "epoch": 4.304,
      "grad_norm": 0.9550835490226746,
      "learning_rate": 4.869473684210527e-06,
      "loss": 0.003,
      "step": 2690
    },
    {
      "epoch": 4.32,
      "grad_norm": 1.5608851909637451,
      "learning_rate": 4.84842105263158e-06,
      "loss": 0.0042,
      "step": 2700
    },
    {
      "epoch": 4.336,
      "grad_norm": 1.4856270551681519,
      "learning_rate": 4.8273684210526315e-06,
      "loss": 0.0048,
      "step": 2710
    },
    {
      "epoch": 4.352,
      "grad_norm": 0.11427481472492218,
      "learning_rate": 4.806315789473684e-06,
      "loss": 0.0022,
      "step": 2720
    },
    {
      "epoch": 4.368,
      "grad_norm": 1.5165557861328125,
      "learning_rate": 4.785263157894738e-06,
      "loss": 0.0037,
      "step": 2730
    },
    {
      "epoch": 4.384,
      "grad_norm": 0.30141228437423706,
      "learning_rate": 4.76421052631579e-06,
      "loss": 0.0026,
      "step": 2740
    },
    {
      "epoch": 4.4,
      "grad_norm": 0.4358047842979431,
      "learning_rate": 4.743157894736842e-06,
      "loss": 0.0046,
      "step": 2750
    },
    {
      "epoch": 4.4,
      "eval_loss": 0.0372794084250927,
      "eval_runtime": 1452.4236,
      "eval_samples_per_second": 3.885,
      "eval_steps_per_second": 3.885,
      "step": 2750
    },
    {
      "epoch": 4.416,
      "grad_norm": 0.8224077820777893,
      "learning_rate": 4.72421052631579e-06,
      "loss": 0.0046,
      "step": 2760
    },
    {
      "epoch": 4.432,
      "grad_norm": 1.4491019248962402,
      "learning_rate": 4.7031578947368425e-06,
      "loss": 0.0029,
      "step": 2770
    },
    {
      "epoch": 4.448,
      "grad_norm": 0.2398335337638855,
      "learning_rate": 4.682105263157895e-06,
      "loss": 0.0038,
      "step": 2780
    },
    {
      "epoch": 4.464,
      "grad_norm": 0.4416545629501343,
      "learning_rate": 4.661052631578948e-06,
      "loss": 0.0059,
      "step": 2790
    },
    {
      "epoch": 4.48,
      "grad_norm": 0.12480773776769638,
      "learning_rate": 4.6400000000000005e-06,
      "loss": 0.0036,
      "step": 2800
    },
    {
      "epoch": 4.496,
      "grad_norm": 0.9009574055671692,
      "learning_rate": 4.618947368421053e-06,
      "loss": 0.0044,
      "step": 2810
    },
    {
      "epoch": 4.5120000000000005,
      "grad_norm": 0.5740355253219604,
      "learning_rate": 4.597894736842106e-06,
      "loss": 0.0032,
      "step": 2820
    },
    {
      "epoch": 4.5280000000000005,
      "grad_norm": 0.6319159865379333,
      "learning_rate": 4.5768421052631584e-06,
      "loss": 0.0026,
      "step": 2830
    },
    {
      "epoch": 4.5440000000000005,
      "grad_norm": 0.710137665271759,
      "learning_rate": 4.55578947368421e-06,
      "loss": 0.0025,
      "step": 2840
    },
    {
      "epoch": 4.5600000000000005,
      "grad_norm": 0.3798495829105377,
      "learning_rate": 4.534736842105264e-06,
      "loss": 0.0032,
      "step": 2850
    },
    {
      "epoch": 4.576,
      "grad_norm": 0.09057649224996567,
      "learning_rate": 4.513684210526316e-06,
      "loss": 0.0019,
      "step": 2860
    },
    {
      "epoch": 4.592,
      "grad_norm": 0.10224127769470215,
      "learning_rate": 4.492631578947369e-06,
      "loss": 0.0024,
      "step": 2870
    },
    {
      "epoch": 4.608,
      "grad_norm": 0.45868900418281555,
      "learning_rate": 4.471578947368421e-06,
      "loss": 0.0033,
      "step": 2880
    },
    {
      "epoch": 4.624,
      "grad_norm": 0.12800952792167664,
      "learning_rate": 4.450526315789474e-06,
      "loss": 0.006,
      "step": 2890
    },
    {
      "epoch": 4.64,
      "grad_norm": 0.12955500185489655,
      "learning_rate": 4.429473684210527e-06,
      "loss": 0.0028,
      "step": 2900
    },
    {
      "epoch": 4.656,
      "grad_norm": 0.3282037675380707,
      "learning_rate": 4.408421052631579e-06,
      "loss": 0.0032,
      "step": 2910
    },
    {
      "epoch": 4.672,
      "grad_norm": 0.480395644903183,
      "learning_rate": 4.3873684210526315e-06,
      "loss": 0.0039,
      "step": 2920
    },
    {
      "epoch": 4.688,
      "grad_norm": 1.337486743927002,
      "learning_rate": 4.366315789473685e-06,
      "loss": 0.0053,
      "step": 2930
    },
    {
      "epoch": 4.704,
      "grad_norm": 0.39609357714653015,
      "learning_rate": 4.345263157894738e-06,
      "loss": 0.0026,
      "step": 2940
    },
    {
      "epoch": 4.72,
      "grad_norm": 0.2566475570201874,
      "learning_rate": 4.3242105263157894e-06,
      "loss": 0.0041,
      "step": 2950
    },
    {
      "epoch": 4.736,
      "grad_norm": 0.6467047929763794,
      "learning_rate": 4.303157894736842e-06,
      "loss": 0.0052,
      "step": 2960
    },
    {
      "epoch": 4.752,
      "grad_norm": 1.1100505590438843,
      "learning_rate": 4.282105263157896e-06,
      "loss": 0.0062,
      "step": 2970
    },
    {
      "epoch": 4.768,
      "grad_norm": 0.6272072196006775,
      "learning_rate": 4.261052631578947e-06,
      "loss": 0.0042,
      "step": 2980
    },
    {
      "epoch": 4.784,
      "grad_norm": 0.34918469190597534,
      "learning_rate": 4.24e-06,
      "loss": 0.0036,
      "step": 2990
    },
    {
      "epoch": 4.8,
      "grad_norm": 0.37302619218826294,
      "learning_rate": 4.218947368421053e-06,
      "loss": 0.0036,
      "step": 3000
    },
    {
      "epoch": 4.8,
      "eval_loss": 0.03739345446228981,
      "eval_runtime": 1465.397,
      "eval_samples_per_second": 3.85,
      "eval_steps_per_second": 3.85,
      "step": 3000
    },
    {
      "epoch": 4.816,
      "grad_norm": 0.5744602084159851,
      "learning_rate": 4.197894736842106e-06,
      "loss": 0.0061,
      "step": 3010
    },
    {
      "epoch": 4.832,
      "grad_norm": 1.2474104166030884,
      "learning_rate": 4.176842105263158e-06,
      "loss": 0.0035,
      "step": 3020
    },
    {
      "epoch": 4.848,
      "grad_norm": 0.25358426570892334,
      "learning_rate": 4.155789473684211e-06,
      "loss": 0.0039,
      "step": 3030
    },
    {
      "epoch": 4.864,
      "grad_norm": 1.7587385177612305,
      "learning_rate": 4.134736842105263e-06,
      "loss": 0.0043,
      "step": 3040
    },
    {
      "epoch": 4.88,
      "grad_norm": 0.5396108627319336,
      "learning_rate": 4.113684210526316e-06,
      "loss": 0.0039,
      "step": 3050
    },
    {
      "epoch": 4.896,
      "grad_norm": 0.2963917553424835,
      "learning_rate": 4.092631578947369e-06,
      "loss": 0.0037,
      "step": 3060
    },
    {
      "epoch": 4.912,
      "grad_norm": 0.6284441947937012,
      "learning_rate": 4.071578947368421e-06,
      "loss": 0.0042,
      "step": 3070
    },
    {
      "epoch": 4.928,
      "grad_norm": 0.1446247100830078,
      "learning_rate": 4.050526315789474e-06,
      "loss": 0.0054,
      "step": 3080
    },
    {
      "epoch": 4.944,
      "grad_norm": 0.5421654582023621,
      "learning_rate": 4.029473684210527e-06,
      "loss": 0.0021,
      "step": 3090
    },
    {
      "epoch": 4.96,
      "grad_norm": 0.10596631467342377,
      "learning_rate": 4.008421052631579e-06,
      "loss": 0.0032,
      "step": 3100
    },
    {
      "epoch": 4.976,
      "grad_norm": 0.36356231570243835,
      "learning_rate": 3.987368421052632e-06,
      "loss": 0.0029,
      "step": 3110
    },
    {
      "epoch": 4.992,
      "grad_norm": 0.13546526432037354,
      "learning_rate": 3.9663157894736846e-06,
      "loss": 0.0026,
      "step": 3120
    },
    {
      "epoch": 5.008,
      "grad_norm": 0.09022821485996246,
      "learning_rate": 3.945263157894737e-06,
      "loss": 0.0021,
      "step": 3130
    },
    {
      "epoch": 5.024,
      "grad_norm": 0.11343013495206833,
      "learning_rate": 3.92421052631579e-06,
      "loss": 0.0035,
      "step": 3140
    },
    {
      "epoch": 5.04,
      "grad_norm": 0.39134401082992554,
      "learning_rate": 3.9031578947368425e-06,
      "loss": 0.0018,
      "step": 3150
    },
    {
      "epoch": 5.056,
      "grad_norm": 0.05532646179199219,
      "learning_rate": 3.882105263157895e-06,
      "loss": 0.0015,
      "step": 3160
    },
    {
      "epoch": 5.072,
      "grad_norm": 0.8032190203666687,
      "learning_rate": 3.861052631578948e-06,
      "loss": 0.0011,
      "step": 3170
    },
    {
      "epoch": 5.088,
      "grad_norm": 0.05971115827560425,
      "learning_rate": 3.8400000000000005e-06,
      "loss": 0.0015,
      "step": 3180
    },
    {
      "epoch": 5.104,
      "grad_norm": 0.690596342086792,
      "learning_rate": 3.818947368421053e-06,
      "loss": 0.0024,
      "step": 3190
    },
    {
      "epoch": 5.12,
      "grad_norm": 0.15326976776123047,
      "learning_rate": 3.7978947368421054e-06,
      "loss": 0.0014,
      "step": 3200
    },
    {
      "epoch": 5.136,
      "grad_norm": 0.05565618351101875,
      "learning_rate": 3.776842105263158e-06,
      "loss": 0.0014,
      "step": 3210
    },
    {
      "epoch": 5.152,
      "grad_norm": 0.18008044362068176,
      "learning_rate": 3.755789473684211e-06,
      "loss": 0.0053,
      "step": 3220
    },
    {
      "epoch": 5.168,
      "grad_norm": 0.3844296932220459,
      "learning_rate": 3.7347368421052634e-06,
      "loss": 0.0018,
      "step": 3230
    },
    {
      "epoch": 5.184,
      "grad_norm": 0.15020981431007385,
      "learning_rate": 3.713684210526316e-06,
      "loss": 0.0013,
      "step": 3240
    },
    {
      "epoch": 5.2,
      "grad_norm": 0.2373928427696228,
      "learning_rate": 3.6926315789473687e-06,
      "loss": 0.0013,
      "step": 3250
    },
    {
      "epoch": 5.2,
      "eval_loss": 0.03558925539255142,
      "eval_runtime": 3288.1466,
      "eval_samples_per_second": 1.716,
      "eval_steps_per_second": 1.716,
      "step": 3250
    },
    {
      "epoch": 5.216,
      "grad_norm": 0.6427579522132874,
      "learning_rate": 3.6715789473684217e-06,
      "loss": 0.0024,
      "step": 3260
    },
    {
      "epoch": 5.232,
      "grad_norm": 0.09459923207759857,
      "learning_rate": 3.650526315789474e-06,
      "loss": 0.0014,
      "step": 3270
    },
    {
      "epoch": 5.248,
      "grad_norm": 0.4201672077178955,
      "learning_rate": 3.6294736842105266e-06,
      "loss": 0.0013,
      "step": 3280
    },
    {
      "epoch": 5.264,
      "grad_norm": 2.386017322540283,
      "learning_rate": 3.6084210526315793e-06,
      "loss": 0.0014,
      "step": 3290
    },
    {
      "epoch": 5.28,
      "grad_norm": 0.05354594066739082,
      "learning_rate": 3.5873684210526315e-06,
      "loss": 0.0012,
      "step": 3300
    },
    {
      "epoch": 5.296,
      "grad_norm": 0.4693075120449066,
      "learning_rate": 3.5663157894736846e-06,
      "loss": 0.0011,
      "step": 3310
    },
    {
      "epoch": 5.312,
      "grad_norm": 0.23063229024410248,
      "learning_rate": 3.5452631578947372e-06,
      "loss": 0.0011,
      "step": 3320
    },
    {
      "epoch": 5.328,
      "grad_norm": 0.6210919618606567,
      "learning_rate": 3.52421052631579e-06,
      "loss": 0.002,
      "step": 3330
    },
    {
      "epoch": 5.344,
      "grad_norm": 0.19046130776405334,
      "learning_rate": 3.503157894736842e-06,
      "loss": 0.0011,
      "step": 3340
    },
    {
      "epoch": 5.36,
      "grad_norm": 0.8910662531852722,
      "learning_rate": 3.482105263157895e-06,
      "loss": 0.0037,
      "step": 3350
    },
    {
      "epoch": 5.376,
      "grad_norm": 1.3482168912887573,
      "learning_rate": 3.461052631578948e-06,
      "loss": 0.0018,
      "step": 3360
    },
    {
      "epoch": 5.392,
      "grad_norm": 0.053552981466054916,
      "learning_rate": 3.44e-06,
      "loss": 0.001,
      "step": 3370
    },
    {
      "epoch": 5.408,
      "grad_norm": 0.165148064494133,
      "learning_rate": 3.4189473684210528e-06,
      "loss": 0.0018,
      "step": 3380
    },
    {
      "epoch": 5.424,
      "grad_norm": 0.1561366617679596,
      "learning_rate": 3.3978947368421054e-06,
      "loss": 0.0017,
      "step": 3390
    },
    {
      "epoch": 5.44,
      "grad_norm": 0.0557156503200531,
      "learning_rate": 3.3768421052631585e-06,
      "loss": 0.0011,
      "step": 3400
    },
    {
      "epoch": 5.456,
      "grad_norm": 0.045138657093048096,
      "learning_rate": 3.3557894736842107e-06,
      "loss": 0.0008,
      "step": 3410
    },
    {
      "epoch": 5.4719999999999995,
      "grad_norm": 0.6297252178192139,
      "learning_rate": 3.3347368421052634e-06,
      "loss": 0.0023,
      "step": 3420
    },
    {
      "epoch": 5.4879999999999995,
      "grad_norm": 0.431007444858551,
      "learning_rate": 3.313684210526316e-06,
      "loss": 0.0014,
      "step": 3430
    },
    {
      "epoch": 5.504,
      "grad_norm": 0.09312240779399872,
      "learning_rate": 3.292631578947369e-06,
      "loss": 0.002,
      "step": 3440
    },
    {
      "epoch": 5.52,
      "grad_norm": 0.19819211959838867,
      "learning_rate": 3.2715789473684213e-06,
      "loss": 0.001,
      "step": 3450
    },
    {
      "epoch": 5.536,
      "grad_norm": 0.09475424885749817,
      "learning_rate": 3.250526315789474e-06,
      "loss": 0.0019,
      "step": 3460
    },
    {
      "epoch": 5.552,
      "grad_norm": 0.034463245421648026,
      "learning_rate": 3.2294736842105266e-06,
      "loss": 0.0024,
      "step": 3470
    },
    {
      "epoch": 5.568,
      "grad_norm": 0.037909191101789474,
      "learning_rate": 3.208421052631579e-06,
      "loss": 0.0019,
      "step": 3480
    },
    {
      "epoch": 5.584,
      "grad_norm": 0.14103959500789642,
      "learning_rate": 3.187368421052632e-06,
      "loss": 0.0023,
      "step": 3490
    },
    {
      "epoch": 5.6,
      "grad_norm": 0.6614345908164978,
      "learning_rate": 3.1663157894736846e-06,
      "loss": 0.0014,
      "step": 3500
    },
    {
      "epoch": 5.6,
      "eval_loss": 0.03668233007192612,
      "eval_runtime": 1484.1952,
      "eval_samples_per_second": 3.801,
      "eval_steps_per_second": 3.801,
      "step": 3500
    },
    {
      "epoch": 5.616,
      "grad_norm": 0.051937129348516464,
      "learning_rate": 3.1452631578947373e-06,
      "loss": 0.0021,
      "step": 3510
    },
    {
      "epoch": 5.632,
      "grad_norm": 0.07696125656366348,
      "learning_rate": 3.1242105263157895e-06,
      "loss": 0.0017,
      "step": 3520
    },
    {
      "epoch": 5.648,
      "grad_norm": 0.2881506681442261,
      "learning_rate": 3.1031578947368426e-06,
      "loss": 0.0014,
      "step": 3530
    },
    {
      "epoch": 5.664,
      "grad_norm": 0.06252899765968323,
      "learning_rate": 3.0821052631578952e-06,
      "loss": 0.0017,
      "step": 3540
    },
    {
      "epoch": 5.68,
      "grad_norm": 0.4536062180995941,
      "learning_rate": 3.0610526315789475e-06,
      "loss": 0.0015,
      "step": 3550
    },
    {
      "epoch": 5.696,
      "grad_norm": 0.06475075334310532,
      "learning_rate": 3.04e-06,
      "loss": 0.0007,
      "step": 3560
    },
    {
      "epoch": 5.712,
      "grad_norm": 0.10876597464084625,
      "learning_rate": 3.0189473684210528e-06,
      "loss": 0.0012,
      "step": 3570
    },
    {
      "epoch": 5.728,
      "grad_norm": 0.04506557062268257,
      "learning_rate": 2.997894736842106e-06,
      "loss": 0.0012,
      "step": 3580
    },
    {
      "epoch": 5.744,
      "grad_norm": 0.08606194704771042,
      "learning_rate": 2.976842105263158e-06,
      "loss": 0.0008,
      "step": 3590
    },
    {
      "epoch": 5.76,
      "grad_norm": 0.10132581740617752,
      "learning_rate": 2.9557894736842107e-06,
      "loss": 0.0008,
      "step": 3600
    },
    {
      "epoch": 5.776,
      "grad_norm": 0.07705429196357727,
      "learning_rate": 2.9347368421052634e-06,
      "loss": 0.0006,
      "step": 3610
    },
    {
      "epoch": 5.792,
      "grad_norm": 0.09106363356113434,
      "learning_rate": 2.9136842105263156e-06,
      "loss": 0.0018,
      "step": 3620
    },
    {
      "epoch": 5.808,
      "grad_norm": 0.23617351055145264,
      "learning_rate": 2.8926315789473687e-06,
      "loss": 0.0019,
      "step": 3630
    },
    {
      "epoch": 5.824,
      "grad_norm": 0.20716822147369385,
      "learning_rate": 2.8715789473684214e-06,
      "loss": 0.0013,
      "step": 3640
    },
    {
      "epoch": 5.84,
      "grad_norm": 0.21701906621456146,
      "learning_rate": 2.850526315789474e-06,
      "loss": 0.0014,
      "step": 3650
    },
    {
      "epoch": 5.856,
      "grad_norm": 0.611148476600647,
      "learning_rate": 2.8294736842105262e-06,
      "loss": 0.0016,
      "step": 3660
    },
    {
      "epoch": 5.872,
      "grad_norm": 0.28972306847572327,
      "learning_rate": 2.8084210526315793e-06,
      "loss": 0.0021,
      "step": 3670
    },
    {
      "epoch": 5.888,
      "grad_norm": 0.13573560118675232,
      "learning_rate": 2.787368421052632e-06,
      "loss": 0.0023,
      "step": 3680
    },
    {
      "epoch": 5.904,
      "grad_norm": 0.15231332182884216,
      "learning_rate": 2.766315789473684e-06,
      "loss": 0.0017,
      "step": 3690
    },
    {
      "epoch": 5.92,
      "grad_norm": 0.24314993619918823,
      "learning_rate": 2.745263157894737e-06,
      "loss": 0.0015,
      "step": 3700
    },
    {
      "epoch": 5.936,
      "grad_norm": 0.2743528485298157,
      "learning_rate": 2.72421052631579e-06,
      "loss": 0.0011,
      "step": 3710
    },
    {
      "epoch": 5.952,
      "grad_norm": 0.20800799131393433,
      "learning_rate": 2.7031578947368426e-06,
      "loss": 0.0009,
      "step": 3720
    },
    {
      "epoch": 5.968,
      "grad_norm": 0.04552097246050835,
      "learning_rate": 2.682105263157895e-06,
      "loss": 0.0014,
      "step": 3730
    },
    {
      "epoch": 5.984,
      "grad_norm": 0.7857608199119568,
      "learning_rate": 2.6610526315789475e-06,
      "loss": 0.0022,
      "step": 3740
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.8046532273292542,
      "learning_rate": 2.64e-06,
      "loss": 0.0018,
      "step": 3750
    },
    {
      "epoch": 6.0,
      "eval_loss": 0.037111204117536545,
      "eval_runtime": 1495.9727,
      "eval_samples_per_second": 3.771,
      "eval_steps_per_second": 3.771,
      "step": 3750
    },
    {
      "epoch": 6.016,
      "grad_norm": 0.09064371883869171,
      "learning_rate": 2.6189473684210524e-06,
      "loss": 0.0005,
      "step": 3760
    },
    {
      "epoch": 6.032,
      "grad_norm": 0.148435577750206,
      "learning_rate": 2.5978947368421054e-06,
      "loss": 0.0008,
      "step": 3770
    },
    {
      "epoch": 6.048,
      "grad_norm": 0.03715546801686287,
      "learning_rate": 2.576842105263158e-06,
      "loss": 0.0007,
      "step": 3780
    },
    {
      "epoch": 6.064,
      "grad_norm": 0.024771861732006073,
      "learning_rate": 2.5557894736842108e-06,
      "loss": 0.0005,
      "step": 3790
    },
    {
      "epoch": 6.08,
      "grad_norm": 0.08074558526277542,
      "learning_rate": 2.534736842105263e-06,
      "loss": 0.0013,
      "step": 3800
    },
    {
      "epoch": 6.096,
      "grad_norm": 0.054898303002119064,
      "learning_rate": 2.513684210526316e-06,
      "loss": 0.0007,
      "step": 3810
    },
    {
      "epoch": 6.112,
      "grad_norm": 0.3437640070915222,
      "learning_rate": 2.4926315789473687e-06,
      "loss": 0.0007,
      "step": 3820
    },
    {
      "epoch": 6.128,
      "grad_norm": 0.12333539873361588,
      "learning_rate": 2.4715789473684214e-06,
      "loss": 0.0004,
      "step": 3830
    },
    {
      "epoch": 6.144,
      "grad_norm": 0.027326831594109535,
      "learning_rate": 2.4505263157894736e-06,
      "loss": 0.0005,
      "step": 3840
    },
    {
      "epoch": 6.16,
      "grad_norm": 0.025773270055651665,
      "learning_rate": 2.4294736842105267e-06,
      "loss": 0.0005,
      "step": 3850
    },
    {
      "epoch": 6.176,
      "grad_norm": 0.04546317830681801,
      "learning_rate": 2.408421052631579e-06,
      "loss": 0.0009,
      "step": 3860
    },
    {
      "epoch": 6.192,
      "grad_norm": 0.054939109832048416,
      "learning_rate": 2.387368421052632e-06,
      "loss": 0.0005,
      "step": 3870
    },
    {
      "epoch": 6.208,
      "grad_norm": 0.17456945776939392,
      "learning_rate": 2.3663157894736842e-06,
      "loss": 0.0004,
      "step": 3880
    },
    {
      "epoch": 6.224,
      "grad_norm": 0.04125133529305458,
      "learning_rate": 2.3452631578947373e-06,
      "loss": 0.0004,
      "step": 3890
    },
    {
      "epoch": 6.24,
      "grad_norm": 0.2721328139305115,
      "learning_rate": 2.3242105263157895e-06,
      "loss": 0.0005,
      "step": 3900
    },
    {
      "epoch": 6.256,
      "grad_norm": 0.20841915905475616,
      "learning_rate": 2.3031578947368426e-06,
      "loss": 0.0004,
      "step": 3910
    },
    {
      "epoch": 6.272,
      "grad_norm": 0.024725254625082016,
      "learning_rate": 2.282105263157895e-06,
      "loss": 0.0011,
      "step": 3920
    },
    {
      "epoch": 6.288,
      "grad_norm": 0.04614824429154396,
      "learning_rate": 2.2610526315789475e-06,
      "loss": 0.0004,
      "step": 3930
    },
    {
      "epoch": 6.304,
      "grad_norm": 0.02417997643351555,
      "learning_rate": 2.24e-06,
      "loss": 0.0004,
      "step": 3940
    },
    {
      "epoch": 6.32,
      "grad_norm": 0.025578554719686508,
      "learning_rate": 2.218947368421053e-06,
      "loss": 0.0005,
      "step": 3950
    },
    {
      "epoch": 6.336,
      "grad_norm": 0.024059904739260674,
      "learning_rate": 2.1978947368421055e-06,
      "loss": 0.0004,
      "step": 3960
    },
    {
      "epoch": 6.352,
      "grad_norm": 0.01979261450469494,
      "learning_rate": 2.176842105263158e-06,
      "loss": 0.0006,
      "step": 3970
    },
    {
      "epoch": 6.368,
      "grad_norm": 0.0502333827316761,
      "learning_rate": 2.1557894736842108e-06,
      "loss": 0.0007,
      "step": 3980
    },
    {
      "epoch": 6.384,
      "grad_norm": 0.390522837638855,
      "learning_rate": 2.1347368421052634e-06,
      "loss": 0.0005,
      "step": 3990
    },
    {
      "epoch": 6.4,
      "grad_norm": 0.21075351536273956,
      "learning_rate": 2.1136842105263157e-06,
      "loss": 0.0008,
      "step": 4000
    },
    {
      "epoch": 6.4,
      "eval_loss": 0.03717663884162903,
      "eval_runtime": 1496.5829,
      "eval_samples_per_second": 3.77,
      "eval_steps_per_second": 3.77,
      "step": 4000
    },
    {
      "epoch": 6.416,
      "grad_norm": 0.08990900218486786,
      "learning_rate": 2.0926315789473687e-06,
      "loss": 0.0004,
      "step": 4010
    },
    {
      "epoch": 6.432,
      "grad_norm": 0.11204410344362259,
      "learning_rate": 2.071578947368421e-06,
      "loss": 0.0003,
      "step": 4020
    },
    {
      "epoch": 6.448,
      "grad_norm": 0.038647376000881195,
      "learning_rate": 2.050526315789474e-06,
      "loss": 0.0007,
      "step": 4030
    },
    {
      "epoch": 6.464,
      "grad_norm": 0.01824771985411644,
      "learning_rate": 2.0294736842105263e-06,
      "loss": 0.0003,
      "step": 4040
    },
    {
      "epoch": 6.48,
      "grad_norm": 0.02977917715907097,
      "learning_rate": 2.0084210526315794e-06,
      "loss": 0.0004,
      "step": 4050
    },
    {
      "epoch": 6.496,
      "grad_norm": 0.05292543023824692,
      "learning_rate": 1.9873684210526316e-06,
      "loss": 0.0005,
      "step": 4060
    },
    {
      "epoch": 6.5120000000000005,
      "grad_norm": 0.011542899534106255,
      "learning_rate": 1.9663157894736842e-06,
      "loss": 0.0005,
      "step": 4070
    },
    {
      "epoch": 6.5280000000000005,
      "grad_norm": 0.5990496277809143,
      "learning_rate": 1.945263157894737e-06,
      "loss": 0.0005,
      "step": 4080
    },
    {
      "epoch": 6.5440000000000005,
      "grad_norm": 0.8402637839317322,
      "learning_rate": 1.9242105263157896e-06,
      "loss": 0.0004,
      "step": 4090
    },
    {
      "epoch": 6.5600000000000005,
      "grad_norm": 0.02681225724518299,
      "learning_rate": 1.9031578947368424e-06,
      "loss": 0.0005,
      "step": 4100
    },
    {
      "epoch": 6.576,
      "grad_norm": 0.05136146768927574,
      "learning_rate": 1.8821052631578949e-06,
      "loss": 0.0009,
      "step": 4110
    },
    {
      "epoch": 6.592,
      "grad_norm": 0.02351960726082325,
      "learning_rate": 1.8610526315789475e-06,
      "loss": 0.0004,
      "step": 4120
    },
    {
      "epoch": 6.608,
      "grad_norm": 0.0421082004904747,
      "learning_rate": 1.8400000000000002e-06,
      "loss": 0.0005,
      "step": 4130
    },
    {
      "epoch": 6.624,
      "grad_norm": 0.039992865175008774,
      "learning_rate": 1.8189473684210528e-06,
      "loss": 0.0007,
      "step": 4140
    },
    {
      "epoch": 6.64,
      "grad_norm": 0.044527649879455566,
      "learning_rate": 1.7978947368421055e-06,
      "loss": 0.0014,
      "step": 4150
    },
    {
      "epoch": 6.656,
      "grad_norm": 0.3133164942264557,
      "learning_rate": 1.776842105263158e-06,
      "loss": 0.0011,
      "step": 4160
    },
    {
      "epoch": 6.672,
      "grad_norm": 0.013667403720319271,
      "learning_rate": 1.7557894736842108e-06,
      "loss": 0.0018,
      "step": 4170
    },
    {
      "epoch": 6.688,
      "grad_norm": 0.06877803057432175,
      "learning_rate": 1.7347368421052632e-06,
      "loss": 0.0003,
      "step": 4180
    },
    {
      "epoch": 6.704,
      "grad_norm": 0.03892038017511368,
      "learning_rate": 1.713684210526316e-06,
      "loss": 0.0003,
      "step": 4190
    },
    {
      "epoch": 6.72,
      "grad_norm": 0.05400446802377701,
      "learning_rate": 1.6926315789473685e-06,
      "loss": 0.0003,
      "step": 4200
    },
    {
      "epoch": 6.736,
      "grad_norm": 0.04348497465252876,
      "learning_rate": 1.6715789473684212e-06,
      "loss": 0.0012,
      "step": 4210
    },
    {
      "epoch": 6.752,
      "grad_norm": 0.1150202602148056,
      "learning_rate": 1.6505263157894739e-06,
      "loss": 0.0004,
      "step": 4220
    },
    {
      "epoch": 6.768,
      "grad_norm": 0.022777294740080833,
      "learning_rate": 1.6294736842105263e-06,
      "loss": 0.0003,
      "step": 4230
    },
    {
      "epoch": 6.784,
      "grad_norm": 0.05009898915886879,
      "learning_rate": 1.6084210526315792e-06,
      "loss": 0.0004,
      "step": 4240
    },
    {
      "epoch": 6.8,
      "grad_norm": 0.012213055975735188,
      "learning_rate": 1.5873684210526316e-06,
      "loss": 0.0003,
      "step": 4250
    },
    {
      "epoch": 6.8,
      "eval_loss": 0.03691783547401428,
      "eval_runtime": 1507.6111,
      "eval_samples_per_second": 3.742,
      "eval_steps_per_second": 3.742,
      "step": 4250
    },
    {
      "epoch": 6.816,
      "grad_norm": 0.23179693520069122,
      "learning_rate": 1.5663157894736845e-06,
      "loss": 0.0005,
      "step": 4260
    },
    {
      "epoch": 6.832,
      "grad_norm": 0.03861673176288605,
      "learning_rate": 1.545263157894737e-06,
      "loss": 0.0004,
      "step": 4270
    },
    {
      "epoch": 6.848,
      "grad_norm": 0.1452307105064392,
      "learning_rate": 1.5242105263157898e-06,
      "loss": 0.0003,
      "step": 4280
    },
    {
      "epoch": 6.864,
      "grad_norm": 0.050570085644721985,
      "learning_rate": 1.5031578947368422e-06,
      "loss": 0.0003,
      "step": 4290
    },
    {
      "epoch": 6.88,
      "grad_norm": 0.08024540543556213,
      "learning_rate": 1.482105263157895e-06,
      "loss": 0.0004,
      "step": 4300
    },
    {
      "epoch": 6.896,
      "grad_norm": 0.02591465599834919,
      "learning_rate": 1.4610526315789475e-06,
      "loss": 0.0005,
      "step": 4310
    },
    {
      "epoch": 6.912,
      "grad_norm": 0.01832268387079239,
      "learning_rate": 1.44e-06,
      "loss": 0.0004,
      "step": 4320
    },
    {
      "epoch": 6.928,
      "grad_norm": 0.044344332069158554,
      "learning_rate": 1.4189473684210528e-06,
      "loss": 0.0004,
      "step": 4330
    },
    {
      "epoch": 6.944,
      "grad_norm": 0.03570440039038658,
      "learning_rate": 1.3978947368421053e-06,
      "loss": 0.0003,
      "step": 4340
    },
    {
      "epoch": 6.96,
      "grad_norm": 0.18772557377815247,
      "learning_rate": 1.3768421052631582e-06,
      "loss": 0.0004,
      "step": 4350
    },
    {
      "epoch": 6.976,
      "grad_norm": 0.05984479933977127,
      "learning_rate": 1.3557894736842106e-06,
      "loss": 0.0003,
      "step": 4360
    },
    {
      "epoch": 6.992,
      "grad_norm": 0.3860021233558655,
      "learning_rate": 1.3347368421052635e-06,
      "loss": 0.0006,
      "step": 4370
    },
    {
      "epoch": 7.008,
      "grad_norm": 0.015999499708414078,
      "learning_rate": 1.313684210526316e-06,
      "loss": 0.0003,
      "step": 4380
    },
    {
      "epoch": 7.024,
      "grad_norm": 0.027498619630932808,
      "learning_rate": 1.2926315789473683e-06,
      "loss": 0.0002,
      "step": 4390
    },
    {
      "epoch": 7.04,
      "grad_norm": 0.24971920251846313,
      "learning_rate": 1.2715789473684212e-06,
      "loss": 0.0004,
      "step": 4400
    },
    {
      "epoch": 7.056,
      "grad_norm": 0.011065910570323467,
      "learning_rate": 1.2505263157894737e-06,
      "loss": 0.0003,
      "step": 4410
    },
    {
      "epoch": 7.072,
      "grad_norm": 0.01698007434606552,
      "learning_rate": 1.2294736842105263e-06,
      "loss": 0.0008,
      "step": 4420
    },
    {
      "epoch": 7.088,
      "grad_norm": 0.02532590739428997,
      "learning_rate": 1.208421052631579e-06,
      "loss": 0.0003,
      "step": 4430
    },
    {
      "epoch": 7.104,
      "grad_norm": 0.01716386154294014,
      "learning_rate": 1.1873684210526316e-06,
      "loss": 0.0003,
      "step": 4440
    },
    {
      "epoch": 7.12,
      "grad_norm": 0.016450321301817894,
      "learning_rate": 1.1663157894736843e-06,
      "loss": 0.0002,
      "step": 4450
    },
    {
      "epoch": 7.136,
      "grad_norm": 0.012627181597054005,
      "learning_rate": 1.145263157894737e-06,
      "loss": 0.0003,
      "step": 4460
    },
    {
      "epoch": 7.152,
      "grad_norm": 0.05683906376361847,
      "learning_rate": 1.1242105263157896e-06,
      "loss": 0.0003,
      "step": 4470
    },
    {
      "epoch": 7.168,
      "grad_norm": 0.03171288222074509,
      "learning_rate": 1.1031578947368422e-06,
      "loss": 0.0002,
      "step": 4480
    },
    {
      "epoch": 7.184,
      "grad_norm": 0.02577533759176731,
      "learning_rate": 1.082105263157895e-06,
      "loss": 0.0003,
      "step": 4490
    },
    {
      "epoch": 7.2,
      "grad_norm": 0.019202925264835358,
      "learning_rate": 1.0610526315789473e-06,
      "loss": 0.0003,
      "step": 4500
    },
    {
      "epoch": 7.2,
      "eval_loss": 0.037218254059553146,
      "eval_runtime": 1593.609,
      "eval_samples_per_second": 3.54,
      "eval_steps_per_second": 3.54,
      "step": 4500
    },
    {
      "epoch": 7.216,
      "grad_norm": 0.01943960040807724,
      "learning_rate": 1.04e-06,
      "loss": 0.0003,
      "step": 4510
    },
    {
      "epoch": 7.232,
      "grad_norm": 0.012748222798109055,
      "learning_rate": 1.0189473684210527e-06,
      "loss": 0.0002,
      "step": 4520
    },
    {
      "epoch": 7.248,
      "grad_norm": 0.017005812376737595,
      "learning_rate": 9.978947368421053e-07,
      "loss": 0.0003,
      "step": 4530
    },
    {
      "epoch": 7.264,
      "grad_norm": 0.024691853672266006,
      "learning_rate": 9.76842105263158e-07,
      "loss": 0.0002,
      "step": 4540
    },
    {
      "epoch": 7.28,
      "grad_norm": 0.3427879512310028,
      "learning_rate": 9.557894736842106e-07,
      "loss": 0.0004,
      "step": 4550
    },
    {
      "epoch": 7.296,
      "grad_norm": 0.03403201699256897,
      "learning_rate": 9.347368421052633e-07,
      "loss": 0.0002,
      "step": 4560
    },
    {
      "epoch": 7.312,
      "grad_norm": 0.022941039875149727,
      "learning_rate": 9.136842105263159e-07,
      "loss": 0.0003,
      "step": 4570
    },
    {
      "epoch": 7.328,
      "grad_norm": 0.032287679612636566,
      "learning_rate": 8.926315789473685e-07,
      "loss": 0.0002,
      "step": 4580
    },
    {
      "epoch": 7.344,
      "grad_norm": 0.013065003789961338,
      "learning_rate": 8.715789473684211e-07,
      "loss": 0.0002,
      "step": 4590
    },
    {
      "epoch": 7.36,
      "grad_norm": 0.026764079928398132,
      "learning_rate": 8.505263157894737e-07,
      "loss": 0.0002,
      "step": 4600
    },
    {
      "epoch": 7.376,
      "grad_norm": 0.013478124514222145,
      "learning_rate": 8.294736842105263e-07,
      "loss": 0.0002,
      "step": 4610
    },
    {
      "epoch": 7.392,
      "grad_norm": 0.014998686499893665,
      "learning_rate": 8.08421052631579e-07,
      "loss": 0.0003,
      "step": 4620
    },
    {
      "epoch": 7.408,
      "grad_norm": 0.015325380489230156,
      "learning_rate": 7.873684210526316e-07,
      "loss": 0.0002,
      "step": 4630
    },
    {
      "epoch": 7.424,
      "grad_norm": 0.018431052565574646,
      "learning_rate": 7.663157894736843e-07,
      "loss": 0.0002,
      "step": 4640
    },
    {
      "epoch": 7.44,
      "grad_norm": 0.01385441143065691,
      "learning_rate": 7.45263157894737e-07,
      "loss": 0.0003,
      "step": 4650
    },
    {
      "epoch": 7.456,
      "grad_norm": 0.018942181020975113,
      "learning_rate": 7.242105263157895e-07,
      "loss": 0.0002,
      "step": 4660
    },
    {
      "epoch": 7.4719999999999995,
      "grad_norm": 0.036915648728609085,
      "learning_rate": 7.031578947368422e-07,
      "loss": 0.0002,
      "step": 4670
    },
    {
      "epoch": 7.4879999999999995,
      "grad_norm": 0.053086813539266586,
      "learning_rate": 6.821052631578948e-07,
      "loss": 0.0002,
      "step": 4680
    },
    {
      "epoch": 7.504,
      "grad_norm": 0.028192859143018723,
      "learning_rate": 6.610526315789475e-07,
      "loss": 0.0003,
      "step": 4690
    },
    {
      "epoch": 7.52,
      "grad_norm": 0.6953186988830566,
      "learning_rate": 6.4e-07,
      "loss": 0.0005,
      "step": 4700
    },
    {
      "epoch": 7.536,
      "grad_norm": 0.016007578000426292,
      "learning_rate": 6.189473684210527e-07,
      "loss": 0.0002,
      "step": 4710
    },
    {
      "epoch": 7.552,
      "grad_norm": 0.029922867193818092,
      "learning_rate": 5.978947368421053e-07,
      "loss": 0.0003,
      "step": 4720
    },
    {
      "epoch": 7.568,
      "grad_norm": 0.01892116479575634,
      "learning_rate": 5.76842105263158e-07,
      "loss": 0.0003,
      "step": 4730
    },
    {
      "epoch": 7.584,
      "grad_norm": 0.03329944983124733,
      "learning_rate": 5.557894736842106e-07,
      "loss": 0.0002,
      "step": 4740
    },
    {
      "epoch": 7.6,
      "grad_norm": 0.025351159274578094,
      "learning_rate": 5.347368421052633e-07,
      "loss": 0.0002,
      "step": 4750
    },
    {
      "epoch": 7.6,
      "eval_loss": 0.03762885555624962,
      "eval_runtime": 1462.6796,
      "eval_samples_per_second": 3.857,
      "eval_steps_per_second": 3.857,
      "step": 4750
    },
    {
      "epoch": 7.616,
      "grad_norm": 0.017339129000902176,
      "learning_rate": 5.136842105263158e-07,
      "loss": 0.0002,
      "step": 4760
    },
    {
      "epoch": 7.632,
      "grad_norm": 0.016879908740520477,
      "learning_rate": 4.926315789473685e-07,
      "loss": 0.0004,
      "step": 4770
    },
    {
      "epoch": 7.648,
      "grad_norm": 0.02509094960987568,
      "learning_rate": 4.715789473684211e-07,
      "loss": 0.0002,
      "step": 4780
    },
    {
      "epoch": 7.664,
      "grad_norm": 0.017198877409100533,
      "learning_rate": 4.505263157894737e-07,
      "loss": 0.0003,
      "step": 4790
    },
    {
      "epoch": 7.68,
      "grad_norm": 0.014106789603829384,
      "learning_rate": 4.2947368421052635e-07,
      "loss": 0.0002,
      "step": 4800
    },
    {
      "epoch": 7.696,
      "grad_norm": 0.03618380054831505,
      "learning_rate": 4.08421052631579e-07,
      "loss": 0.0002,
      "step": 4810
    },
    {
      "epoch": 7.712,
      "grad_norm": 0.013972431421279907,
      "learning_rate": 3.873684210526316e-07,
      "loss": 0.0002,
      "step": 4820
    },
    {
      "epoch": 7.728,
      "grad_norm": 0.01894882321357727,
      "learning_rate": 3.663157894736842e-07,
      "loss": 0.0002,
      "step": 4830
    },
    {
      "epoch": 7.744,
      "grad_norm": 0.01931534893810749,
      "learning_rate": 3.4526315789473686e-07,
      "loss": 0.0003,
      "step": 4840
    },
    {
      "epoch": 7.76,
      "grad_norm": 0.015359962359070778,
      "learning_rate": 3.242105263157895e-07,
      "loss": 0.0002,
      "step": 4850
    },
    {
      "epoch": 7.776,
      "grad_norm": 0.0260631013661623,
      "learning_rate": 3.031578947368421e-07,
      "loss": 0.0002,
      "step": 4860
    },
    {
      "epoch": 7.792,
      "grad_norm": 0.06113347038626671,
      "learning_rate": 2.821052631578948e-07,
      "loss": 0.0002,
      "step": 4870
    },
    {
      "epoch": 7.808,
      "grad_norm": 0.0257665254175663,
      "learning_rate": 2.610526315789474e-07,
      "loss": 0.0003,
      "step": 4880
    },
    {
      "epoch": 7.824,
      "grad_norm": 0.01588517054915428,
      "learning_rate": 2.4000000000000003e-07,
      "loss": 0.0002,
      "step": 4890
    },
    {
      "epoch": 7.84,
      "grad_norm": 0.012915355153381824,
      "learning_rate": 2.1894736842105263e-07,
      "loss": 0.0002,
      "step": 4900
    },
    {
      "epoch": 7.856,
      "grad_norm": 0.01570785976946354,
      "learning_rate": 1.978947368421053e-07,
      "loss": 0.0002,
      "step": 4910
    },
    {
      "epoch": 7.872,
      "grad_norm": 0.020730208605527878,
      "learning_rate": 1.768421052631579e-07,
      "loss": 0.0002,
      "step": 4920
    },
    {
      "epoch": 7.888,
      "grad_norm": 0.03537014126777649,
      "learning_rate": 1.5578947368421054e-07,
      "loss": 0.0003,
      "step": 4930
    },
    {
      "epoch": 7.904,
      "grad_norm": 0.022489851340651512,
      "learning_rate": 1.3473684210526317e-07,
      "loss": 0.0002,
      "step": 4940
    },
    {
      "epoch": 7.92,
      "grad_norm": 0.017718670889735222,
      "learning_rate": 1.136842105263158e-07,
      "loss": 0.0002,
      "step": 4950
    },
    {
      "epoch": 7.936,
      "grad_norm": 0.016573410481214523,
      "learning_rate": 9.263157894736844e-08,
      "loss": 0.0007,
      "step": 4960
    },
    {
      "epoch": 7.952,
      "grad_norm": 0.008749530650675297,
      "learning_rate": 7.157894736842106e-08,
      "loss": 0.0003,
      "step": 4970
    },
    {
      "epoch": 7.968,
      "grad_norm": 0.04577890783548355,
      "learning_rate": 5.0526315789473687e-08,
      "loss": 0.0003,
      "step": 4980
    },
    {
      "epoch": 7.984,
      "grad_norm": 0.061525970697402954,
      "learning_rate": 2.9473684210526318e-08,
      "loss": 0.0002,
      "step": 4990
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.029113447293639183,
      "learning_rate": 8.421052631578948e-09,
      "loss": 0.0002,
      "step": 5000
    },
    {
      "epoch": 8.0,
      "eval_loss": 0.03780149295926094,
      "eval_runtime": 1480.7553,
      "eval_samples_per_second": 3.81,
      "eval_steps_per_second": 3.81,
      "step": 5000
    }
  ],
  "logging_steps": 10,
  "max_steps": 5000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 8,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.727921844224e+20,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}