{
  "best_metric": 0.020064357668161392,
  "best_model_checkpoint": "./vit-base-mask-finetuned/checkpoint-3500",
  "epoch": 4.0,
  "global_step": 3608,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "learning_rate": 0.00019944567627494457,
      "loss": 0.3036,
      "step": 10
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00019889135254988916,
      "loss": 0.0608,
      "step": 20
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00019833702882483372,
      "loss": 0.279,
      "step": 30
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00019778270509977829,
      "loss": 0.0541,
      "step": 40
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00019722838137472285,
      "loss": 0.0914,
      "step": 50
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0001966740576496674,
      "loss": 0.054,
      "step": 60
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00019611973392461197,
      "loss": 0.0358,
      "step": 70
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00019556541019955653,
      "loss": 0.1087,
      "step": 80
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00019501108647450112,
      "loss": 0.0825,
      "step": 90
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00019445676274944569,
      "loss": 0.1983,
      "step": 100
    },
    {
      "epoch": 0.11,
      "eval_accuracy": 0.8654060066740823,
      "eval_loss": 0.4739435613155365,
      "eval_runtime": 55.7772,
      "eval_samples_per_second": 64.471,
      "eval_steps_per_second": 8.068,
      "step": 100
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00019390243902439025,
      "loss": 0.0989,
      "step": 110
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00019334811529933484,
      "loss": 0.0927,
      "step": 120
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0001927937915742794,
      "loss": 0.0285,
      "step": 130
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00019223946784922396,
      "loss": 0.0902,
      "step": 140
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00019168514412416852,
      "loss": 0.1752,
      "step": 150
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00019113082039911309,
      "loss": 0.1193,
      "step": 160
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00019057649667405765,
      "loss": 0.0224,
      "step": 170
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0001900221729490022,
      "loss": 0.0087,
      "step": 180
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0001894678492239468,
      "loss": 0.1051,
      "step": 190
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00018891352549889136,
      "loss": 0.067,
      "step": 200
    },
    {
      "epoch": 0.22,
      "eval_accuracy": 0.9883203559510567,
      "eval_loss": 0.052814140915870667,
      "eval_runtime": 60.4371,
      "eval_samples_per_second": 59.5,
      "eval_steps_per_second": 7.446,
      "step": 200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00018835920177383592,
      "loss": 0.0395,
      "step": 210
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0001878048780487805,
      "loss": 0.0902,
      "step": 220
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00018725055432372508,
      "loss": 0.0561,
      "step": 230
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00018669623059866964,
      "loss": 0.0737,
      "step": 240
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0001861419068736142,
      "loss": 0.1694,
      "step": 250
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00018558758314855876,
      "loss": 0.0941,
      "step": 260
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00018503325942350332,
      "loss": 0.0564,
      "step": 270
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0001844789356984479,
      "loss": 0.1086,
      "step": 280
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00018392461197339248,
      "loss": 0.0907,
      "step": 290
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00018337028824833704,
      "loss": 0.0313,
      "step": 300
    },
    {
      "epoch": 0.33,
      "eval_accuracy": 0.9872080088987765,
      "eval_loss": 0.04620526358485222,
      "eval_runtime": 58.7722,
      "eval_samples_per_second": 61.185,
      "eval_steps_per_second": 7.657,
      "step": 300
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0001828159645232816,
      "loss": 0.0853,
      "step": 310
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00018226164079822616,
      "loss": 0.0177,
      "step": 320
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00018170731707317075,
      "loss": 0.0603,
      "step": 330
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00018115299334811531,
      "loss": 0.122,
      "step": 340
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00018059866962305988,
      "loss": 0.0216,
      "step": 350
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00018004434589800444,
      "loss": 0.0723,
      "step": 360
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000179490022172949,
      "loss": 0.1421,
      "step": 370
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00017893569844789356,
      "loss": 0.0695,
      "step": 380
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00017838137472283815,
      "loss": 0.0655,
      "step": 390
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00017782705099778271,
      "loss": 0.0628,
      "step": 400
    },
    {
      "epoch": 0.44,
      "eval_accuracy": 0.9763626251390434,
      "eval_loss": 0.08295725286006927,
      "eval_runtime": 58.2224,
      "eval_samples_per_second": 61.763,
      "eval_steps_per_second": 7.729,
      "step": 400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00017727272727272728,
      "loss": 0.0909,
      "step": 410
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00017671840354767184,
      "loss": 0.0186,
      "step": 420
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00017616407982261643,
      "loss": 0.0316,
      "step": 430
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000175609756097561,
      "loss": 0.0313,
      "step": 440
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00017505543237250555,
      "loss": 0.0066,
      "step": 450
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00017450110864745014,
      "loss": 0.0602,
      "step": 460
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00017394678492239468,
      "loss": 0.0727,
      "step": 470
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00017339246119733924,
      "loss": 0.0624,
      "step": 480
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0001728381374722838,
      "loss": 0.0064,
      "step": 490
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0001722838137472284,
      "loss": 0.0042,
      "step": 500
    },
    {
      "epoch": 0.55,
      "eval_accuracy": 0.9924916573971079,
      "eval_loss": 0.03670423477888107,
      "eval_runtime": 59.4878,
      "eval_samples_per_second": 60.449,
      "eval_steps_per_second": 7.565,
      "step": 500
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00017172949002217295,
      "loss": 0.0343,
      "step": 510
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00017117516629711752,
      "loss": 0.1144,
      "step": 520
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0001706208425720621,
      "loss": 0.0439,
      "step": 530
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00017006651884700667,
      "loss": 0.0852,
      "step": 540
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00016951219512195123,
      "loss": 0.0988,
      "step": 550
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00016895787139689582,
      "loss": 0.0594,
      "step": 560
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00016840354767184035,
      "loss": 0.0483,
      "step": 570
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00016784922394678492,
      "loss": 0.011,
      "step": 580
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00016729490022172948,
      "loss": 0.0241,
      "step": 590
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016674057649667407,
      "loss": 0.0881,
      "step": 600
    },
    {
      "epoch": 0.67,
      "eval_accuracy": 0.9866518353726362,
      "eval_loss": 0.05000607296824455,
      "eval_runtime": 58.7478,
      "eval_samples_per_second": 61.211,
      "eval_steps_per_second": 7.66,
      "step": 600
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016618625277161863,
      "loss": 0.0463,
      "step": 610
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001656319290465632,
      "loss": 0.0324,
      "step": 620
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00016507760532150778,
      "loss": 0.1116,
      "step": 630
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00016452328159645234,
      "loss": 0.0725,
      "step": 640
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001639689578713969,
      "loss": 0.0402,
      "step": 650
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00016341463414634147,
      "loss": 0.0205,
      "step": 660
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00016286031042128606,
      "loss": 0.0759,
      "step": 670
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001623059866962306,
      "loss": 0.0658,
      "step": 680
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016175166297117515,
      "loss": 0.0316,
      "step": 690
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00016119733924611974,
      "loss": 0.0047,
      "step": 700
    },
    {
      "epoch": 0.78,
      "eval_accuracy": 0.9899888765294772,
      "eval_loss": 0.04604267328977585,
      "eval_runtime": 58.9442,
      "eval_samples_per_second": 61.007,
      "eval_steps_per_second": 7.634,
      "step": 700
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001606430155210643,
      "loss": 0.0808,
      "step": 710
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00016008869179600887,
      "loss": 0.0872,
      "step": 720
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00015953436807095346,
      "loss": 0.0779,
      "step": 730
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00015898004434589802,
      "loss": 0.0614,
      "step": 740
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00015842572062084258,
      "loss": 0.0226,
      "step": 750
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00015787139689578714,
      "loss": 0.0344,
      "step": 760
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00015731707317073173,
      "loss": 0.0044,
      "step": 770
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001567627494456763,
      "loss": 0.0934,
      "step": 780
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.00015620842572062083,
      "loss": 0.0315,
      "step": 790
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.00015565410199556542,
      "loss": 0.0755,
      "step": 800
    },
    {
      "epoch": 0.89,
      "eval_accuracy": 0.9888765294771968,
      "eval_loss": 0.04636429622769356,
      "eval_runtime": 58.4439,
      "eval_samples_per_second": 61.529,
      "eval_steps_per_second": 7.7,
      "step": 800
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.00015509977827050998,
      "loss": 0.0159,
      "step": 810
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.00015454545454545454,
      "loss": 0.1032,
      "step": 820
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0001539911308203991,
      "loss": 0.0535,
      "step": 830
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0001534368070953437,
      "loss": 0.1643,
      "step": 840
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.00015288248337028826,
      "loss": 0.0909,
      "step": 850
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.00015232815964523282,
      "loss": 0.1228,
      "step": 860
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0001517738359201774,
      "loss": 0.0528,
      "step": 870
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.00015121951219512197,
      "loss": 0.0133,
      "step": 880
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.00015066518847006653,
      "loss": 0.0948,
      "step": 890
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0001501108647450111,
      "loss": 0.0258,
      "step": 900
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9880422691879867,
      "eval_loss": 0.037437278777360916,
      "eval_runtime": 58.2201,
      "eval_samples_per_second": 61.766,
      "eval_steps_per_second": 7.729,
      "step": 900
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00014955654101995566,
      "loss": 0.0117,
      "step": 910
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00014900221729490022,
      "loss": 0.0229,
      "step": 920
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00014844789356984478,
      "loss": 0.0058,
      "step": 930
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00014789356984478937,
      "loss": 0.0059,
      "step": 940
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00014733924611973393,
      "loss": 0.0039,
      "step": 950
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0001467849223946785,
      "loss": 0.0459,
      "step": 960
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00014623059866962309,
      "loss": 0.0564,
      "step": 970
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00014567627494456765,
      "loss": 0.0832,
      "step": 980
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0001451219512195122,
      "loss": 0.0419,
      "step": 990
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.00014456762749445675,
      "loss": 0.0564,
      "step": 1000
    },
    {
      "epoch": 1.11,
      "eval_accuracy": 0.9874860956618465,
      "eval_loss": 0.05776744335889816,
      "eval_runtime": 58.8261,
      "eval_samples_per_second": 61.129,
      "eval_steps_per_second": 7.65,
      "step": 1000
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.00014401330376940133,
      "loss": 0.0534,
      "step": 1010
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0001434589800443459,
      "loss": 0.1231,
      "step": 1020
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.00014290465631929046,
      "loss": 0.087,
      "step": 1030
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.00014235033259423505,
      "loss": 0.0739,
      "step": 1040
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0001417960088691796,
      "loss": 0.0574,
      "step": 1050
    },
    {
      "epoch": 1.18,
      "learning_rate": 0.00014124168514412417,
      "loss": 0.1169,
      "step": 1060
    },
    {
      "epoch": 1.19,
      "learning_rate": 0.00014068736141906876,
      "loss": 0.017,
      "step": 1070
    },
    {
      "epoch": 1.2,
      "learning_rate": 0.00014013303769401332,
      "loss": 0.0614,
      "step": 1080
    },
    {
      "epoch": 1.21,
      "learning_rate": 0.0001395787139689579,
      "loss": 0.0093,
      "step": 1090
    },
    {
      "epoch": 1.22,
      "learning_rate": 0.00013902439024390245,
      "loss": 0.0453,
      "step": 1100
    },
    {
      "epoch": 1.22,
      "eval_accuracy": 0.9844271412680756,
      "eval_loss": 0.05230843648314476,
      "eval_runtime": 59.332,
      "eval_samples_per_second": 60.608,
      "eval_steps_per_second": 7.584,
      "step": 1100
    },
    {
      "epoch": 1.23,
      "learning_rate": 0.000138470066518847,
      "loss": 0.0061,
      "step": 1110
    },
    {
      "epoch": 1.24,
      "learning_rate": 0.00013791574279379157,
      "loss": 0.0329,
      "step": 1120
    },
    {
      "epoch": 1.25,
      "learning_rate": 0.00013736141906873614,
      "loss": 0.0105,
      "step": 1130
    },
    {
      "epoch": 1.26,
      "learning_rate": 0.00013680709534368072,
      "loss": 0.0327,
      "step": 1140
    },
    {
      "epoch": 1.27,
      "learning_rate": 0.0001362527716186253,
      "loss": 0.0884,
      "step": 1150
    },
    {
      "epoch": 1.29,
      "learning_rate": 0.00013569844789356985,
      "loss": 0.0061,
      "step": 1160
    },
    {
      "epoch": 1.3,
      "learning_rate": 0.00013514412416851444,
      "loss": 0.0739,
      "step": 1170
    },
    {
      "epoch": 1.31,
      "learning_rate": 0.000134589800443459,
      "loss": 0.0185,
      "step": 1180
    },
    {
      "epoch": 1.32,
      "learning_rate": 0.00013403547671840356,
      "loss": 0.0176,
      "step": 1190
    },
    {
      "epoch": 1.33,
      "learning_rate": 0.00013348115299334812,
      "loss": 0.0656,
      "step": 1200
    },
    {
      "epoch": 1.33,
      "eval_accuracy": 0.9805339265850945,
      "eval_loss": 0.08926890790462494,
      "eval_runtime": 59.14,
      "eval_samples_per_second": 60.805,
      "eval_steps_per_second": 7.609,
      "step": 1200
    },
    {
      "epoch": 1.34,
      "learning_rate": 0.0001329268292682927,
      "loss": 0.0047,
      "step": 1210
    },
    {
      "epoch": 1.35,
      "learning_rate": 0.00013237250554323725,
      "loss": 0.0416,
      "step": 1220
    },
    {
      "epoch": 1.36,
      "learning_rate": 0.0001318181818181818,
      "loss": 0.0418,
      "step": 1230
    },
    {
      "epoch": 1.37,
      "learning_rate": 0.0001312638580931264,
      "loss": 0.127,
      "step": 1240
    },
    {
      "epoch": 1.39,
      "learning_rate": 0.00013070953436807096,
      "loss": 0.0412,
      "step": 1250
    },
    {
      "epoch": 1.4,
      "learning_rate": 0.00013015521064301553,
      "loss": 0.0963,
      "step": 1260
    },
    {
      "epoch": 1.41,
      "learning_rate": 0.0001296008869179601,
      "loss": 0.0722,
      "step": 1270
    },
    {
      "epoch": 1.42,
      "learning_rate": 0.00012904656319290468,
      "loss": 0.0046,
      "step": 1280
    },
    {
      "epoch": 1.43,
      "learning_rate": 0.00012849223946784924,
      "loss": 0.0526,
      "step": 1290
    },
    {
      "epoch": 1.44,
      "learning_rate": 0.0001279379157427938,
      "loss": 0.0103,
      "step": 1300
    },
    {
      "epoch": 1.44,
      "eval_accuracy": 0.9911012235817576,
      "eval_loss": 0.03496430441737175,
      "eval_runtime": 58.9191,
      "eval_samples_per_second": 61.033,
      "eval_steps_per_second": 7.638,
      "step": 1300
    },
    {
      "epoch": 1.45,
      "learning_rate": 0.00012738359201773836,
      "loss": 0.0196,
      "step": 1310
    },
    {
      "epoch": 1.46,
      "learning_rate": 0.00012682926829268293,
      "loss": 0.0037,
      "step": 1320
    },
    {
      "epoch": 1.47,
      "learning_rate": 0.0001262749445676275,
      "loss": 0.0045,
      "step": 1330
    },
    {
      "epoch": 1.49,
      "learning_rate": 0.00012572062084257208,
      "loss": 0.0038,
      "step": 1340
    },
    {
      "epoch": 1.5,
      "learning_rate": 0.00012516629711751664,
      "loss": 0.0048,
      "step": 1350
    },
    {
      "epoch": 1.51,
      "learning_rate": 0.0001246119733924612,
      "loss": 0.0692,
      "step": 1360
    },
    {
      "epoch": 1.52,
      "learning_rate": 0.00012405764966740576,
      "loss": 0.0357,
      "step": 1370
    },
    {
      "epoch": 1.53,
      "learning_rate": 0.00012350332594235035,
      "loss": 0.0088,
      "step": 1380
    },
    {
      "epoch": 1.54,
      "learning_rate": 0.00012294900221729492,
      "loss": 0.0126,
      "step": 1390
    },
    {
      "epoch": 1.55,
      "learning_rate": 0.00012239467849223948,
      "loss": 0.0057,
      "step": 1400
    },
    {
      "epoch": 1.55,
      "eval_accuracy": 0.9916573971078977,
      "eval_loss": 0.02847210131585598,
      "eval_runtime": 59.5202,
      "eval_samples_per_second": 60.416,
      "eval_steps_per_second": 7.56,
      "step": 1400
    },
    {
      "epoch": 1.56,
      "learning_rate": 0.00012184035476718405,
      "loss": 0.0051,
      "step": 1410
    },
    {
      "epoch": 1.57,
      "learning_rate": 0.00012128603104212862,
      "loss": 0.0015,
      "step": 1420
    },
    {
      "epoch": 1.59,
      "learning_rate": 0.00012073170731707318,
      "loss": 0.0022,
      "step": 1430
    },
    {
      "epoch": 1.6,
      "learning_rate": 0.00012017738359201774,
      "loss": 0.0014,
      "step": 1440
    },
    {
      "epoch": 1.61,
      "learning_rate": 0.00011962305986696232,
      "loss": 0.0796,
      "step": 1450
    },
    {
      "epoch": 1.62,
      "learning_rate": 0.00011906873614190688,
      "loss": 0.0088,
      "step": 1460
    },
    {
      "epoch": 1.63,
      "learning_rate": 0.00011851441241685144,
      "loss": 0.0083,
      "step": 1470
    },
    {
      "epoch": 1.64,
      "learning_rate": 0.00011796008869179602,
      "loss": 0.0016,
      "step": 1480
    },
    {
      "epoch": 1.65,
      "learning_rate": 0.00011740576496674058,
      "loss": 0.0449,
      "step": 1490
    },
    {
      "epoch": 1.66,
      "learning_rate": 0.00011685144124168514,
      "loss": 0.0543,
      "step": 1500
    },
    {
      "epoch": 1.66,
      "eval_accuracy": 0.989154616240267,
      "eval_loss": 0.033378347754478455,
      "eval_runtime": 58.4631,
      "eval_samples_per_second": 61.509,
      "eval_steps_per_second": 7.697,
      "step": 1500
    },
    {
      "epoch": 1.67,
      "learning_rate": 0.00011629711751662973,
      "loss": 0.0251,
      "step": 1510
    },
    {
      "epoch": 1.69,
      "learning_rate": 0.00011574279379157429,
      "loss": 0.008,
      "step": 1520
    },
    {
      "epoch": 1.7,
      "learning_rate": 0.00011518847006651885,
      "loss": 0.0442,
      "step": 1530
    },
    {
      "epoch": 1.71,
      "learning_rate": 0.00011463414634146342,
      "loss": 0.0382,
      "step": 1540
    },
    {
      "epoch": 1.72,
      "learning_rate": 0.00011407982261640799,
      "loss": 0.0136,
      "step": 1550
    },
    {
      "epoch": 1.73,
      "learning_rate": 0.00011352549889135255,
      "loss": 0.0023,
      "step": 1560
    },
    {
      "epoch": 1.74,
      "learning_rate": 0.00011297117516629712,
      "loss": 0.0022,
      "step": 1570
    },
    {
      "epoch": 1.75,
      "learning_rate": 0.0001124168514412417,
      "loss": 0.0844,
      "step": 1580
    },
    {
      "epoch": 1.76,
      "learning_rate": 0.00011186252771618625,
      "loss": 0.01,
      "step": 1590
    },
    {
      "epoch": 1.77,
      "learning_rate": 0.00011130820399113082,
      "loss": 0.078,
      "step": 1600
    },
    {
      "epoch": 1.77,
      "eval_accuracy": 0.9916573971078977,
      "eval_loss": 0.0301674697548151,
      "eval_runtime": 58.0308,
      "eval_samples_per_second": 61.967,
      "eval_steps_per_second": 7.755,
      "step": 1600
    },
    {
      "epoch": 1.78,
      "learning_rate": 0.00011075388026607538,
      "loss": 0.0041,
      "step": 1610
    },
    {
      "epoch": 1.8,
      "learning_rate": 0.00011019955654101997,
      "loss": 0.014,
      "step": 1620
    },
    {
      "epoch": 1.81,
      "learning_rate": 0.00010964523281596453,
      "loss": 0.0286,
      "step": 1630
    },
    {
      "epoch": 1.82,
      "learning_rate": 0.00010909090909090909,
      "loss": 0.0142,
      "step": 1640
    },
    {
      "epoch": 1.83,
      "learning_rate": 0.00010853658536585367,
      "loss": 0.0351,
      "step": 1650
    },
    {
      "epoch": 1.84,
      "learning_rate": 0.00010798226164079823,
      "loss": 0.0525,
      "step": 1660
    },
    {
      "epoch": 1.85,
      "learning_rate": 0.00010742793791574279,
      "loss": 0.0086,
      "step": 1670
    },
    {
      "epoch": 1.86,
      "learning_rate": 0.00010687361419068738,
      "loss": 0.0449,
      "step": 1680
    },
    {
      "epoch": 1.87,
      "learning_rate": 0.00010631929046563194,
      "loss": 0.0341,
      "step": 1690
    },
    {
      "epoch": 1.88,
      "learning_rate": 0.00010576496674057649,
      "loss": 0.1053,
      "step": 1700
    },
    {
      "epoch": 1.88,
      "eval_accuracy": 0.9880422691879867,
      "eval_loss": 0.03492213040590286,
      "eval_runtime": 57.9028,
      "eval_samples_per_second": 62.104,
      "eval_steps_per_second": 7.772,
      "step": 1700
    },
    {
      "epoch": 1.9,
      "learning_rate": 0.00010521064301552105,
      "loss": 0.0068,
      "step": 1710
    },
    {
      "epoch": 1.91,
      "learning_rate": 0.00010465631929046564,
      "loss": 0.0121,
      "step": 1720
    },
    {
      "epoch": 1.92,
      "learning_rate": 0.0001041019955654102,
      "loss": 0.0037,
      "step": 1730
    },
    {
      "epoch": 1.93,
      "learning_rate": 0.00010354767184035477,
      "loss": 0.0164,
      "step": 1740
    },
    {
      "epoch": 1.94,
      "learning_rate": 0.00010299334811529934,
      "loss": 0.0016,
      "step": 1750
    },
    {
      "epoch": 1.95,
      "learning_rate": 0.0001024390243902439,
      "loss": 0.0593,
      "step": 1760
    },
    {
      "epoch": 1.96,
      "learning_rate": 0.00010188470066518847,
      "loss": 0.0313,
      "step": 1770
    },
    {
      "epoch": 1.97,
      "learning_rate": 0.00010133037694013303,
      "loss": 0.0124,
      "step": 1780
    },
    {
      "epoch": 1.98,
      "learning_rate": 0.00010077605321507762,
      "loss": 0.0029,
      "step": 1790
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00010022172949002218,
      "loss": 0.071,
      "step": 1800
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9902669632925473,
      "eval_loss": 0.034953419119119644,
      "eval_runtime": 58.2314,
      "eval_samples_per_second": 61.754,
      "eval_steps_per_second": 7.728,
      "step": 1800
    },
    {
      "epoch": 2.01,
      "learning_rate": 9.966740576496674e-05,
      "loss": 0.0145,
      "step": 1810
    },
    {
      "epoch": 2.02,
      "learning_rate": 9.911308203991131e-05,
      "loss": 0.0018,
      "step": 1820
    },
    {
      "epoch": 2.03,
      "learning_rate": 9.855875831485588e-05,
      "loss": 0.0021,
      "step": 1830
    },
    {
      "epoch": 2.04,
      "learning_rate": 9.800443458980046e-05,
      "loss": 0.0016,
      "step": 1840
    },
    {
      "epoch": 2.05,
      "learning_rate": 9.745011086474502e-05,
      "loss": 0.0025,
      "step": 1850
    },
    {
      "epoch": 2.06,
      "learning_rate": 9.689578713968958e-05,
      "loss": 0.0013,
      "step": 1860
    },
    {
      "epoch": 2.07,
      "learning_rate": 9.634146341463415e-05,
      "loss": 0.0441,
      "step": 1870
    },
    {
      "epoch": 2.08,
      "learning_rate": 9.578713968957872e-05,
      "loss": 0.0723,
      "step": 1880
    },
    {
      "epoch": 2.1,
      "learning_rate": 9.523281596452328e-05,
      "loss": 0.0099,
      "step": 1890
    },
    {
      "epoch": 2.11,
      "learning_rate": 9.467849223946786e-05,
      "loss": 0.0454,
      "step": 1900
    },
    {
      "epoch": 2.11,
      "eval_accuracy": 0.9899888765294772,
      "eval_loss": 0.043182216584682465,
      "eval_runtime": 58.2341,
      "eval_samples_per_second": 61.751,
      "eval_steps_per_second": 7.727,
      "step": 1900
    },
    {
      "epoch": 2.12,
      "learning_rate": 9.412416851441242e-05,
      "loss": 0.0451,
      "step": 1910
    },
    {
      "epoch": 2.13,
      "learning_rate": 9.356984478935698e-05,
      "loss": 0.0121,
      "step": 1920
    },
    {
      "epoch": 2.14,
      "learning_rate": 9.301552106430156e-05,
      "loss": 0.0117,
      "step": 1930
    },
    {
      "epoch": 2.15,
      "learning_rate": 9.246119733924612e-05,
      "loss": 0.0178,
      "step": 1940
    },
    {
      "epoch": 2.16,
      "learning_rate": 9.19068736141907e-05,
      "loss": 0.002,
      "step": 1950
    },
    {
      "epoch": 2.17,
      "learning_rate": 9.135254988913526e-05,
      "loss": 0.0023,
      "step": 1960
    },
    {
      "epoch": 2.18,
      "learning_rate": 9.079822616407982e-05,
      "loss": 0.0159,
      "step": 1970
    },
    {
      "epoch": 2.2,
      "learning_rate": 9.02439024390244e-05,
      "loss": 0.0051,
      "step": 1980
    },
    {
      "epoch": 2.21,
      "learning_rate": 8.968957871396896e-05,
      "loss": 0.0199,
      "step": 1990
    },
    {
      "epoch": 2.22,
      "learning_rate": 8.913525498891354e-05,
      "loss": 0.0465,
      "step": 2000
    },
    {
      "epoch": 2.22,
      "eval_accuracy": 0.985539488320356,
      "eval_loss": 0.05088553577661514,
      "eval_runtime": 58.1787,
      "eval_samples_per_second": 61.81,
      "eval_steps_per_second": 7.735,
      "step": 2000
    },
    {
      "epoch": 2.23,
      "learning_rate": 8.85809312638581e-05,
      "loss": 0.0223,
      "step": 2010
    },
    {
      "epoch": 2.24,
      "learning_rate": 8.802660753880266e-05,
      "loss": 0.012,
      "step": 2020
    },
    {
      "epoch": 2.25,
      "learning_rate": 8.747228381374724e-05,
      "loss": 0.0025,
      "step": 2030
    },
    {
      "epoch": 2.26,
      "learning_rate": 8.69179600886918e-05,
      "loss": 0.0413,
      "step": 2040
    },
    {
      "epoch": 2.27,
      "learning_rate": 8.636363636363637e-05,
      "loss": 0.0428,
      "step": 2050
    },
    {
      "epoch": 2.28,
      "learning_rate": 8.580931263858094e-05,
      "loss": 0.0134,
      "step": 2060
    },
    {
      "epoch": 2.29,
      "learning_rate": 8.52549889135255e-05,
      "loss": 0.063,
      "step": 2070
    },
    {
      "epoch": 2.31,
      "learning_rate": 8.470066518847007e-05,
      "loss": 0.0095,
      "step": 2080
    },
    {
      "epoch": 2.32,
      "learning_rate": 8.414634146341464e-05,
      "loss": 0.0495,
      "step": 2090
    },
    {
      "epoch": 2.33,
      "learning_rate": 8.359201773835921e-05,
      "loss": 0.0024,
      "step": 2100
    },
    {
      "epoch": 2.33,
      "eval_accuracy": 0.9933259176863182,
      "eval_loss": 0.022791525349020958,
      "eval_runtime": 58.8696,
      "eval_samples_per_second": 61.084,
      "eval_steps_per_second": 7.644,
      "step": 2100
    },
    {
      "epoch": 2.34,
      "learning_rate": 8.303769401330377e-05,
      "loss": 0.0253,
      "step": 2110
    },
    {
      "epoch": 2.35,
      "learning_rate": 8.248337028824834e-05,
      "loss": 0.0022,
      "step": 2120
    },
    {
      "epoch": 2.36,
      "learning_rate": 8.192904656319291e-05,
      "loss": 0.0478,
      "step": 2130
    },
    {
      "epoch": 2.37,
      "learning_rate": 8.137472283813747e-05,
      "loss": 0.0149,
      "step": 2140
    },
    {
      "epoch": 2.38,
      "learning_rate": 8.082039911308205e-05,
      "loss": 0.0143,
      "step": 2150
    },
    {
      "epoch": 2.39,
      "learning_rate": 8.026607538802661e-05,
      "loss": 0.0273,
      "step": 2160
    },
    {
      "epoch": 2.41,
      "learning_rate": 7.971175166297117e-05,
      "loss": 0.0021,
      "step": 2170
    },
    {
      "epoch": 2.42,
      "learning_rate": 7.915742793791575e-05,
      "loss": 0.0036,
      "step": 2180
    },
    {
      "epoch": 2.43,
      "learning_rate": 7.860310421286031e-05,
      "loss": 0.0025,
      "step": 2190
    },
    {
      "epoch": 2.44,
      "learning_rate": 7.804878048780489e-05,
      "loss": 0.0039,
      "step": 2200
    },
    {
      "epoch": 2.44,
      "eval_accuracy": 0.993047830923248,
      "eval_loss": 0.0303057748824358,
      "eval_runtime": 58.3751,
      "eval_samples_per_second": 61.602,
      "eval_steps_per_second": 7.709,
      "step": 2200
    },
    {
      "epoch": 2.45,
      "learning_rate": 7.749445676274945e-05,
      "loss": 0.0417,
      "step": 2210
    },
    {
      "epoch": 2.46,
      "learning_rate": 7.694013303769401e-05,
      "loss": 0.0094,
      "step": 2220
    },
    {
      "epoch": 2.47,
      "learning_rate": 7.638580931263859e-05,
      "loss": 0.0402,
      "step": 2230
    },
    {
      "epoch": 2.48,
      "learning_rate": 7.583148558758315e-05,
      "loss": 0.0081,
      "step": 2240
    },
    {
      "epoch": 2.49,
      "learning_rate": 7.527716186252773e-05,
      "loss": 0.0083,
      "step": 2250
    },
    {
      "epoch": 2.51,
      "learning_rate": 7.472283813747229e-05,
      "loss": 0.0277,
      "step": 2260
    },
    {
      "epoch": 2.52,
      "learning_rate": 7.416851441241685e-05,
      "loss": 0.0028,
      "step": 2270
    },
    {
      "epoch": 2.53,
      "learning_rate": 7.361419068736141e-05,
      "loss": 0.0051,
      "step": 2280
    },
    {
      "epoch": 2.54,
      "learning_rate": 7.305986696230599e-05,
      "loss": 0.009,
      "step": 2290
    },
    {
      "epoch": 2.55,
      "learning_rate": 7.250554323725056e-05,
      "loss": 0.0111,
      "step": 2300
    },
    {
      "epoch": 2.55,
      "eval_accuracy": 0.9924916573971079,
      "eval_loss": 0.031308963894844055,
      "eval_runtime": 58.2011,
      "eval_samples_per_second": 61.786,
      "eval_steps_per_second": 7.732,
      "step": 2300
    },
    {
      "epoch": 2.56,
      "learning_rate": 7.195121951219513e-05,
      "loss": 0.0012,
      "step": 2310
    },
    {
      "epoch": 2.57,
      "learning_rate": 7.139689578713969e-05,
      "loss": 0.0015,
      "step": 2320
    },
    {
      "epoch": 2.58,
      "learning_rate": 7.084257206208425e-05,
      "loss": 0.0401,
      "step": 2330
    },
    {
      "epoch": 2.59,
      "learning_rate": 7.028824833702883e-05,
      "loss": 0.0013,
      "step": 2340
    },
    {
      "epoch": 2.61,
      "learning_rate": 6.97339246119734e-05,
      "loss": 0.0013,
      "step": 2350
    },
    {
      "epoch": 2.62,
      "learning_rate": 6.917960088691796e-05,
      "loss": 0.0025,
      "step": 2360
    },
    {
      "epoch": 2.63,
      "learning_rate": 6.862527716186254e-05,
      "loss": 0.0657,
      "step": 2370
    },
    {
      "epoch": 2.64,
      "learning_rate": 6.807095343680709e-05,
      "loss": 0.0433,
      "step": 2380
    },
    {
      "epoch": 2.65,
      "learning_rate": 6.751662971175166e-05,
      "loss": 0.0048,
      "step": 2390
    },
    {
      "epoch": 2.66,
      "learning_rate": 6.696230598669624e-05,
      "loss": 0.0082,
      "step": 2400
    },
    {
      "epoch": 2.66,
      "eval_accuracy": 0.9913793103448276,
      "eval_loss": 0.03090614266693592,
      "eval_runtime": 58.3701,
      "eval_samples_per_second": 61.607,
      "eval_steps_per_second": 7.709,
      "step": 2400
    },
    {
      "epoch": 2.67,
      "learning_rate": 6.64079822616408e-05,
      "loss": 0.0016,
      "step": 2410
    },
    {
      "epoch": 2.68,
      "learning_rate": 6.585365853658538e-05,
      "loss": 0.0014,
      "step": 2420
    },
    {
      "epoch": 2.69,
      "learning_rate": 6.529933481152993e-05,
      "loss": 0.0038,
      "step": 2430
    },
    {
      "epoch": 2.71,
      "learning_rate": 6.47450110864745e-05,
      "loss": 0.0011,
      "step": 2440
    },
    {
      "epoch": 2.72,
      "learning_rate": 6.419068736141908e-05,
      "loss": 0.0099,
      "step": 2450
    },
    {
      "epoch": 2.73,
      "learning_rate": 6.363636363636364e-05,
      "loss": 0.031,
      "step": 2460
    },
    {
      "epoch": 2.74,
      "learning_rate": 6.308203991130822e-05,
      "loss": 0.041,
      "step": 2470
    },
    {
      "epoch": 2.75,
      "learning_rate": 6.252771618625277e-05,
      "loss": 0.0017,
      "step": 2480
    },
    {
      "epoch": 2.76,
      "learning_rate": 6.197339246119734e-05,
      "loss": 0.0151,
      "step": 2490
    },
    {
      "epoch": 2.77,
      "learning_rate": 6.14190687361419e-05,
      "loss": 0.0296,
      "step": 2500
    },
    {
      "epoch": 2.77,
      "eval_accuracy": 0.9947163515016685,
      "eval_loss": 0.024169722571969032,
      "eval_runtime": 57.9558,
      "eval_samples_per_second": 62.047,
      "eval_steps_per_second": 7.765,
      "step": 2500
    },
    {
      "epoch": 2.78,
      "learning_rate": 6.086474501108648e-05,
      "loss": 0.0013,
      "step": 2510
    },
    {
      "epoch": 2.79,
      "learning_rate": 6.031042128603105e-05,
      "loss": 0.0013,
      "step": 2520
    },
    {
      "epoch": 2.8,
      "learning_rate": 5.975609756097561e-05,
      "loss": 0.0015,
      "step": 2530
    },
    {
      "epoch": 2.82,
      "learning_rate": 5.920177383592018e-05,
      "loss": 0.0167,
      "step": 2540
    },
    {
      "epoch": 2.83,
      "learning_rate": 5.864745011086474e-05,
      "loss": 0.018,
      "step": 2550
    },
    {
      "epoch": 2.84,
      "learning_rate": 5.809312638580932e-05,
      "loss": 0.0134,
      "step": 2560
    },
    {
      "epoch": 2.85,
      "learning_rate": 5.7538802660753886e-05,
      "loss": 0.0086,
      "step": 2570
    },
    {
      "epoch": 2.86,
      "learning_rate": 5.698447893569845e-05,
      "loss": 0.0029,
      "step": 2580
    },
    {
      "epoch": 2.87,
      "learning_rate": 5.6430155210643024e-05,
      "loss": 0.0014,
      "step": 2590
    },
    {
      "epoch": 2.88,
      "learning_rate": 5.587583148558758e-05,
      "loss": 0.0011,
      "step": 2600
    },
    {
      "epoch": 2.88,
      "eval_accuracy": 0.9924916573971079,
      "eval_loss": 0.028317071497440338,
      "eval_runtime": 58.1403,
      "eval_samples_per_second": 61.85,
      "eval_steps_per_second": 7.74,
      "step": 2600
    },
    {
      "epoch": 2.89,
      "learning_rate": 5.5321507760532155e-05,
      "loss": 0.0009,
      "step": 2610
    },
    {
      "epoch": 2.9,
      "learning_rate": 5.4767184035476724e-05,
      "loss": 0.001,
      "step": 2620
    },
    {
      "epoch": 2.92,
      "learning_rate": 5.4212860310421286e-05,
      "loss": 0.0014,
      "step": 2630
    },
    {
      "epoch": 2.93,
      "learning_rate": 5.365853658536586e-05,
      "loss": 0.0014,
      "step": 2640
    },
    {
      "epoch": 2.94,
      "learning_rate": 5.310421286031042e-05,
      "loss": 0.0118,
      "step": 2650
    },
    {
      "epoch": 2.95,
      "learning_rate": 5.254988913525499e-05,
      "loss": 0.0011,
      "step": 2660
    },
    {
      "epoch": 2.96,
      "learning_rate": 5.1995565410199555e-05,
      "loss": 0.0098,
      "step": 2670
    },
    {
      "epoch": 2.97,
      "learning_rate": 5.1441241685144124e-05,
      "loss": 0.0087,
      "step": 2680
    },
    {
      "epoch": 2.98,
      "learning_rate": 5.08869179600887e-05,
      "loss": 0.0011,
      "step": 2690
    },
    {
      "epoch": 2.99,
      "learning_rate": 5.033259423503326e-05,
      "loss": 0.0139,
      "step": 2700
    },
    {
      "epoch": 2.99,
      "eval_accuracy": 0.9941601779755284,
      "eval_loss": 0.022749271243810654,
      "eval_runtime": 58.1506,
      "eval_samples_per_second": 61.839,
      "eval_steps_per_second": 7.739,
      "step": 2700
    },
    {
      "epoch": 3.0,
      "learning_rate": 4.977827050997783e-05,
      "loss": 0.0036,
      "step": 2710
    },
    {
      "epoch": 3.02,
      "learning_rate": 4.92239467849224e-05,
      "loss": 0.0012,
      "step": 2720
    },
    {
      "epoch": 3.03,
      "learning_rate": 4.866962305986696e-05,
      "loss": 0.0123,
      "step": 2730
    },
    {
      "epoch": 3.04,
      "learning_rate": 4.811529933481153e-05,
      "loss": 0.0008,
      "step": 2740
    },
    {
      "epoch": 3.05,
      "learning_rate": 4.75609756097561e-05,
      "loss": 0.0008,
      "step": 2750
    },
    {
      "epoch": 3.06,
      "learning_rate": 4.700665188470067e-05,
      "loss": 0.0235,
      "step": 2760
    },
    {
      "epoch": 3.07,
      "learning_rate": 4.645232815964524e-05,
      "loss": 0.0009,
      "step": 2770
    },
    {
      "epoch": 3.08,
      "learning_rate": 4.58980044345898e-05,
      "loss": 0.0119,
      "step": 2780
    },
    {
      "epoch": 3.09,
      "learning_rate": 4.534368070953437e-05,
      "loss": 0.0452,
      "step": 2790
    },
    {
      "epoch": 3.1,
      "learning_rate": 4.478935698447894e-05,
      "loss": 0.0008,
      "step": 2800
    },
    {
      "epoch": 3.1,
      "eval_accuracy": 0.9927697441601779,
      "eval_loss": 0.02841453067958355,
      "eval_runtime": 58.4898,
      "eval_samples_per_second": 61.481,
      "eval_steps_per_second": 7.694,
      "step": 2800
    },
    {
      "epoch": 3.12,
      "learning_rate": 4.42350332594235e-05,
      "loss": 0.005,
      "step": 2810
    },
    {
      "epoch": 3.13,
      "learning_rate": 4.3680709534368077e-05,
      "loss": 0.0389,
      "step": 2820
    },
    {
      "epoch": 3.14,
      "learning_rate": 4.312638580931264e-05,
      "loss": 0.0019,
      "step": 2830
    },
    {
      "epoch": 3.15,
      "learning_rate": 4.257206208425721e-05,
      "loss": 0.0065,
      "step": 2840
    },
    {
      "epoch": 3.16,
      "learning_rate": 4.201773835920178e-05,
      "loss": 0.0025,
      "step": 2850
    },
    {
      "epoch": 3.17,
      "learning_rate": 4.146341463414634e-05,
      "loss": 0.0018,
      "step": 2860
    },
    {
      "epoch": 3.18,
      "learning_rate": 4.0909090909090915e-05,
      "loss": 0.0026,
      "step": 2870
    },
    {
      "epoch": 3.19,
      "learning_rate": 4.035476718403548e-05,
      "loss": 0.0008,
      "step": 2880
    },
    {
      "epoch": 3.2,
      "learning_rate": 3.9800443458980046e-05,
      "loss": 0.0009,
      "step": 2890
    },
    {
      "epoch": 3.22,
      "learning_rate": 3.9246119733924615e-05,
      "loss": 0.0042,
      "step": 2900
    },
    {
      "epoch": 3.22,
      "eval_accuracy": 0.9938820912124583,
      "eval_loss": 0.025445684790611267,
      "eval_runtime": 58.549,
      "eval_samples_per_second": 61.419,
      "eval_steps_per_second": 7.686,
      "step": 2900
    },
    {
      "epoch": 3.23,
      "learning_rate": 3.869179600886918e-05,
      "loss": 0.0066,
      "step": 2910
    },
    {
      "epoch": 3.24,
      "learning_rate": 3.8137472283813746e-05,
      "loss": 0.0436,
      "step": 2920
    },
    {
      "epoch": 3.25,
      "learning_rate": 3.758314855875832e-05,
      "loss": 0.0023,
      "step": 2930
    },
    {
      "epoch": 3.26,
      "learning_rate": 3.7028824833702884e-05,
      "loss": 0.0008,
      "step": 2940
    },
    {
      "epoch": 3.27,
      "learning_rate": 3.647450110864745e-05,
      "loss": 0.0008,
      "step": 2950
    },
    {
      "epoch": 3.28,
      "learning_rate": 3.5920177383592015e-05,
      "loss": 0.0087,
      "step": 2960
    },
    {
      "epoch": 3.29,
      "learning_rate": 3.5365853658536584e-05,
      "loss": 0.044,
      "step": 2970
    },
    {
      "epoch": 3.3,
      "learning_rate": 3.481152993348116e-05,
      "loss": 0.018,
      "step": 2980
    },
    {
      "epoch": 3.31,
      "learning_rate": 3.425720620842572e-05,
      "loss": 0.001,
      "step": 2990
    },
    {
      "epoch": 3.33,
      "learning_rate": 3.370288248337029e-05,
      "loss": 0.0032,
      "step": 3000
    },
    {
      "epoch": 3.33,
      "eval_accuracy": 0.9933259176863182,
      "eval_loss": 0.023396195843815804,
      "eval_runtime": 58.2141,
      "eval_samples_per_second": 61.772,
      "eval_steps_per_second": 7.73,
      "step": 3000
    },
    {
      "epoch": 3.34,
      "learning_rate": 3.314855875831486e-05,
      "loss": 0.0019,
      "step": 3010
    },
    {
      "epoch": 3.35,
      "learning_rate": 3.259423503325942e-05,
      "loss": 0.0009,
      "step": 3020
    },
    {
      "epoch": 3.36,
      "learning_rate": 3.203991130820399e-05,
      "loss": 0.0009,
      "step": 3030
    },
    {
      "epoch": 3.37,
      "learning_rate": 3.148558758314856e-05,
      "loss": 0.0037,
      "step": 3040
    },
    {
      "epoch": 3.38,
      "learning_rate": 3.093126385809313e-05,
      "loss": 0.0279,
      "step": 3050
    },
    {
      "epoch": 3.39,
      "learning_rate": 3.0376940133037695e-05,
      "loss": 0.0009,
      "step": 3060
    },
    {
      "epoch": 3.4,
      "learning_rate": 2.9822616407982264e-05,
      "loss": 0.0018,
      "step": 3070
    },
    {
      "epoch": 3.41,
      "learning_rate": 2.926829268292683e-05,
      "loss": 0.0009,
      "step": 3080
    },
    {
      "epoch": 3.43,
      "learning_rate": 2.8713968957871395e-05,
      "loss": 0.0467,
      "step": 3090
    },
    {
      "epoch": 3.44,
      "learning_rate": 2.8159645232815967e-05,
      "loss": 0.0142,
      "step": 3100
    },
    {
      "epoch": 3.44,
      "eval_accuracy": 0.9933259176863182,
      "eval_loss": 0.024244721978902817,
      "eval_runtime": 58.0241,
      "eval_samples_per_second": 61.974,
      "eval_steps_per_second": 7.755,
      "step": 3100
    },
    {
      "epoch": 3.45,
      "learning_rate": 2.7605321507760533e-05,
      "loss": 0.0008,
      "step": 3110
    },
    {
      "epoch": 3.46,
      "learning_rate": 2.7050997782705102e-05,
      "loss": 0.0037,
      "step": 3120
    },
    {
      "epoch": 3.47,
      "learning_rate": 2.6496674057649668e-05,
      "loss": 0.0009,
      "step": 3130
    },
    {
      "epoch": 3.48,
      "learning_rate": 2.5942350332594233e-05,
      "loss": 0.0008,
      "step": 3140
    },
    {
      "epoch": 3.49,
      "learning_rate": 2.5388026607538806e-05,
      "loss": 0.0249,
      "step": 3150
    },
    {
      "epoch": 3.5,
      "learning_rate": 2.483370288248337e-05,
      "loss": 0.0011,
      "step": 3160
    },
    {
      "epoch": 3.51,
      "learning_rate": 2.427937915742794e-05,
      "loss": 0.0007,
      "step": 3170
    },
    {
      "epoch": 3.53,
      "learning_rate": 2.3725055432372506e-05,
      "loss": 0.0007,
      "step": 3180
    },
    {
      "epoch": 3.54,
      "learning_rate": 2.3170731707317075e-05,
      "loss": 0.0007,
      "step": 3190
    },
    {
      "epoch": 3.55,
      "learning_rate": 2.261640798226164e-05,
      "loss": 0.0434,
      "step": 3200
    },
    {
      "epoch": 3.55,
      "eval_accuracy": 0.993047830923248,
      "eval_loss": 0.026953959837555885,
      "eval_runtime": 58.4633,
      "eval_samples_per_second": 61.509,
      "eval_steps_per_second": 7.697,
      "step": 3200
    },
    {
      "epoch": 3.56,
      "learning_rate": 2.206208425720621e-05,
      "loss": 0.0008,
      "step": 3210
    },
    {
      "epoch": 3.57,
      "learning_rate": 2.150776053215078e-05,
      "loss": 0.0009,
      "step": 3220
    },
    {
      "epoch": 3.58,
      "learning_rate": 2.0953436807095344e-05,
      "loss": 0.0009,
      "step": 3230
    },
    {
      "epoch": 3.59,
      "learning_rate": 2.0399113082039913e-05,
      "loss": 0.001,
      "step": 3240
    },
    {
      "epoch": 3.6,
      "learning_rate": 1.9844789356984482e-05,
      "loss": 0.0039,
      "step": 3250
    },
    {
      "epoch": 3.61,
      "learning_rate": 1.9290465631929047e-05,
      "loss": 0.0015,
      "step": 3260
    },
    {
      "epoch": 3.63,
      "learning_rate": 1.8736141906873613e-05,
      "loss": 0.0013,
      "step": 3270
    },
    {
      "epoch": 3.64,
      "learning_rate": 1.8181818181818182e-05,
      "loss": 0.0009,
      "step": 3280
    },
    {
      "epoch": 3.65,
      "learning_rate": 1.762749445676275e-05,
      "loss": 0.0063,
      "step": 3290
    },
    {
      "epoch": 3.66,
      "learning_rate": 1.707317073170732e-05,
      "loss": 0.0196,
      "step": 3300
    },
    {
      "epoch": 3.66,
      "eval_accuracy": 0.9936040044493882,
      "eval_loss": 0.022605180740356445,
      "eval_runtime": 57.9368,
      "eval_samples_per_second": 62.068,
      "eval_steps_per_second": 7.767,
      "step": 3300
    },
    {
      "epoch": 3.67,
      "learning_rate": 1.6518847006651886e-05,
      "loss": 0.0068,
      "step": 3310
    },
    {
      "epoch": 3.68,
      "learning_rate": 1.596452328159645e-05,
      "loss": 0.0104,
      "step": 3320
    },
    {
      "epoch": 3.69,
      "learning_rate": 1.541019955654102e-05,
      "loss": 0.0197,
      "step": 3330
    },
    {
      "epoch": 3.7,
      "learning_rate": 1.485587583148559e-05,
      "loss": 0.003,
      "step": 3340
    },
    {
      "epoch": 3.71,
      "learning_rate": 1.4301552106430155e-05,
      "loss": 0.0011,
      "step": 3350
    },
    {
      "epoch": 3.73,
      "learning_rate": 1.3747228381374724e-05,
      "loss": 0.0181,
      "step": 3360
    },
    {
      "epoch": 3.74,
      "learning_rate": 1.3192904656319291e-05,
      "loss": 0.002,
      "step": 3370
    },
    {
      "epoch": 3.75,
      "learning_rate": 1.2638580931263858e-05,
      "loss": 0.0007,
      "step": 3380
    },
    {
      "epoch": 3.76,
      "learning_rate": 1.2084257206208427e-05,
      "loss": 0.0007,
      "step": 3390
    },
    {
      "epoch": 3.77,
      "learning_rate": 1.1529933481152993e-05,
      "loss": 0.0007,
      "step": 3400
    },
    {
      "epoch": 3.77,
      "eval_accuracy": 0.9944382647385984,
      "eval_loss": 0.02052154392004013,
      "eval_runtime": 57.8787,
      "eval_samples_per_second": 62.13,
      "eval_steps_per_second": 7.775,
      "step": 3400
    },
    {
      "epoch": 3.78,
      "learning_rate": 1.0975609756097562e-05,
      "loss": 0.0008,
      "step": 3410
    },
    {
      "epoch": 3.79,
      "learning_rate": 1.042128603104213e-05,
      "loss": 0.0007,
      "step": 3420
    },
    {
      "epoch": 3.8,
      "learning_rate": 9.866962305986696e-06,
      "loss": 0.0083,
      "step": 3430
    },
    {
      "epoch": 3.81,
      "learning_rate": 9.312638580931264e-06,
      "loss": 0.0012,
      "step": 3440
    },
    {
      "epoch": 3.82,
      "learning_rate": 8.758314855875833e-06,
      "loss": 0.0018,
      "step": 3450
    },
    {
      "epoch": 3.84,
      "learning_rate": 8.2039911308204e-06,
      "loss": 0.0419,
      "step": 3460
    },
    {
      "epoch": 3.85,
      "learning_rate": 7.649667405764967e-06,
      "loss": 0.0008,
      "step": 3470
    },
    {
      "epoch": 3.86,
      "learning_rate": 7.095343680709535e-06,
      "loss": 0.0009,
      "step": 3480
    },
    {
      "epoch": 3.87,
      "learning_rate": 6.541019955654103e-06,
      "loss": 0.0071,
      "step": 3490
    },
    {
      "epoch": 3.88,
      "learning_rate": 5.98669623059867e-06,
      "loss": 0.0042,
      "step": 3500
    },
    {
      "epoch": 3.88,
      "eval_accuracy": 0.9941601779755284,
      "eval_loss": 0.020064357668161392,
      "eval_runtime": 55.4082,
      "eval_samples_per_second": 64.9,
      "eval_steps_per_second": 8.122,
      "step": 3500
    },
    {
      "epoch": 3.89,
      "learning_rate": 5.432372505543237e-06,
      "loss": 0.0426,
      "step": 3510
    },
    {
      "epoch": 3.9,
      "learning_rate": 4.8780487804878055e-06,
      "loss": 0.0028,
      "step": 3520
    },
    {
      "epoch": 3.91,
      "learning_rate": 4.323725055432373e-06,
      "loss": 0.002,
      "step": 3530
    },
    {
      "epoch": 3.92,
      "learning_rate": 3.7694013303769405e-06,
      "loss": 0.0012,
      "step": 3540
    },
    {
      "epoch": 3.94,
      "learning_rate": 3.2150776053215078e-06,
      "loss": 0.0008,
      "step": 3550
    },
    {
      "epoch": 3.95,
      "learning_rate": 2.6607538802660755e-06,
      "loss": 0.0013,
      "step": 3560
    },
    {
      "epoch": 3.96,
      "learning_rate": 2.106430155210643e-06,
      "loss": 0.0025,
      "step": 3570
    },
    {
      "epoch": 3.97,
      "learning_rate": 1.5521064301552107e-06,
      "loss": 0.0033,
      "step": 3580
    },
    {
      "epoch": 3.98,
      "learning_rate": 9.977827050997782e-07,
      "loss": 0.0083,
      "step": 3590
    },
    {
      "epoch": 3.99,
      "learning_rate": 4.434589800443459e-07,
      "loss": 0.0008,
      "step": 3600
    },
    {
      "epoch": 3.99,
      "eval_accuracy": 0.9947163515016685,
      "eval_loss": 0.02037554606795311,
      "eval_runtime": 55.7966,
      "eval_samples_per_second": 64.448,
      "eval_steps_per_second": 8.065,
      "step": 3600
    },
    {
      "epoch": 4.0,
      "step": 3608,
      "total_flos": 4.4722177046373335e+18,
      "train_loss": 0.03198626538376714,
      "train_runtime": 5538.163,
      "train_samples_per_second": 10.421,
      "train_steps_per_second": 0.651
    }
  ],
  "max_steps": 3608,
  "num_train_epochs": 4,
  "total_flos": 4.4722177046373335e+18,
  "trial_name": null,
  "trial_params": null
}