{
  "best_metric": 0.18984103202819824,
  "best_model_checkpoint": "/kaggle/working/model/checkpoint-1900",
  "epoch": 4.0,
  "eval_steps": 100,
  "global_step": 3852,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "learning_rate": 0.00019948078920041537,
      "loss": 0.6119,
      "step": 10
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00019896157840083075,
      "loss": 0.556,
      "step": 20
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00019844236760124614,
      "loss": 0.4675,
      "step": 30
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0001979231568016615,
      "loss": 0.4472,
      "step": 40
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00019740394600207685,
      "loss": 0.3713,
      "step": 50
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0001968847352024922,
      "loss": 0.339,
      "step": 60
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0001963655244029076,
      "loss": 0.4431,
      "step": 70
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00019584631360332295,
      "loss": 0.4784,
      "step": 80
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00019532710280373834,
      "loss": 0.4167,
      "step": 90
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0001948078920041537,
      "loss": 0.2732,
      "step": 100
    },
    {
      "epoch": 0.1,
      "eval_accuracy": 0.8460839266777761,
      "eval_loss": 0.39686205983161926,
      "eval_runtime": 106.8457,
      "eval_samples_per_second": 61.781,
      "eval_steps_per_second": 7.731,
      "step": 100
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00019428868120456908,
      "loss": 0.5051,
      "step": 110
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00019376947040498444,
      "loss": 0.306,
      "step": 120
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0001932502596053998,
      "loss": 0.3317,
      "step": 130
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00019273104880581515,
      "loss": 0.4264,
      "step": 140
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00019221183800623054,
      "loss": 0.4038,
      "step": 150
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00019169262720664592,
      "loss": 0.4715,
      "step": 160
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00019117341640706128,
      "loss": 0.3561,
      "step": 170
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00019065420560747664,
      "loss": 0.3471,
      "step": 180
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00019013499480789202,
      "loss": 0.3163,
      "step": 190
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00018961578400830738,
      "loss": 0.2784,
      "step": 200
    },
    {
      "epoch": 0.21,
      "eval_accuracy": 0.8579003181336161,
      "eval_loss": 0.37141430377960205,
      "eval_runtime": 106.5876,
      "eval_samples_per_second": 61.93,
      "eval_steps_per_second": 7.749,
      "step": 200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00018909657320872276,
      "loss": 0.4061,
      "step": 210
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00018857736240913812,
      "loss": 0.2339,
      "step": 220
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00018805815160955348,
      "loss": 0.3515,
      "step": 230
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00018753894080996886,
      "loss": 0.2553,
      "step": 240
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00018701973001038422,
      "loss": 0.4057,
      "step": 250
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00018650051921079958,
      "loss": 0.2917,
      "step": 260
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00018598130841121496,
      "loss": 0.384,
      "step": 270
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00018546209761163035,
      "loss": 0.3622,
      "step": 280
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0001849428868120457,
      "loss": 0.3697,
      "step": 290
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00018442367601246106,
      "loss": 0.301,
      "step": 300
    },
    {
      "epoch": 0.31,
      "eval_accuracy": 0.8376003635812755,
      "eval_loss": 0.35040077567100525,
      "eval_runtime": 106.1282,
      "eval_samples_per_second": 62.198,
      "eval_steps_per_second": 7.783,
      "step": 300
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00018390446521287642,
      "loss": 0.3128,
      "step": 310
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0001833852544132918,
      "loss": 0.2876,
      "step": 320
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0001828660436137072,
      "loss": 0.2879,
      "step": 330
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00018234683281412255,
      "loss": 0.2336,
      "step": 340
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0001818276220145379,
      "loss": 0.3879,
      "step": 350
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0001813084112149533,
      "loss": 0.333,
      "step": 360
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00018078920041536865,
      "loss": 0.3831,
      "step": 370
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000180269989615784,
      "loss": 0.2895,
      "step": 380
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0001797507788161994,
      "loss": 0.3258,
      "step": 390
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00017923156801661475,
      "loss": 0.2372,
      "step": 400
    },
    {
      "epoch": 0.42,
      "eval_accuracy": 0.8812301166489925,
      "eval_loss": 0.3391087055206299,
      "eval_runtime": 106.4818,
      "eval_samples_per_second": 61.992,
      "eval_steps_per_second": 7.757,
      "step": 400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00017871235721703013,
      "loss": 0.4028,
      "step": 410
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0001781931464174455,
      "loss": 0.244,
      "step": 420
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00017767393561786085,
      "loss": 0.2848,
      "step": 430
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00017715472481827623,
      "loss": 0.3357,
      "step": 440
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00017663551401869162,
      "loss": 0.3361,
      "step": 450
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00017611630321910697,
      "loss": 0.3148,
      "step": 460
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00017559709241952233,
      "loss": 0.3378,
      "step": 470
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0001750778816199377,
      "loss": 0.3758,
      "step": 480
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00017455867082035307,
      "loss": 0.2882,
      "step": 490
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00017403946002076843,
      "loss": 0.3136,
      "step": 500
    },
    {
      "epoch": 0.52,
      "eval_accuracy": 0.8966823208604757,
      "eval_loss": 0.2559354901313782,
      "eval_runtime": 106.8533,
      "eval_samples_per_second": 61.776,
      "eval_steps_per_second": 7.73,
      "step": 500
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00017352024922118382,
      "loss": 0.2467,
      "step": 510
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00017300103842159917,
      "loss": 0.2569,
      "step": 520
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00017248182762201456,
      "loss": 0.2846,
      "step": 530
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00017196261682242992,
      "loss": 0.3017,
      "step": 540
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00017144340602284527,
      "loss": 0.2795,
      "step": 550
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00017092419522326063,
      "loss": 0.3047,
      "step": 560
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00017040498442367602,
      "loss": 0.3351,
      "step": 570
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001698857736240914,
      "loss": 0.2825,
      "step": 580
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00016936656282450676,
      "loss": 0.2019,
      "step": 590
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00016884735202492212,
      "loss": 0.3517,
      "step": 600
    },
    {
      "epoch": 0.62,
      "eval_accuracy": 0.8397212543554007,
      "eval_loss": 0.41408056020736694,
      "eval_runtime": 104.998,
      "eval_samples_per_second": 62.868,
      "eval_steps_per_second": 7.867,
      "step": 600
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001683281412253375,
      "loss": 0.3487,
      "step": 610
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00016780893042575286,
      "loss": 0.2473,
      "step": 620
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00016728971962616824,
      "loss": 0.2432,
      "step": 630
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001667705088265836,
      "loss": 0.3265,
      "step": 640
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016625129802699896,
      "loss": 0.2954,
      "step": 650
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00016573208722741434,
      "loss": 0.2389,
      "step": 660
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001652128764278297,
      "loss": 0.2444,
      "step": 670
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00016469366562824509,
      "loss": 0.2776,
      "step": 680
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00016417445482866044,
      "loss": 0.3284,
      "step": 690
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00016365524402907583,
      "loss": 0.3312,
      "step": 700
    },
    {
      "epoch": 0.73,
      "eval_accuracy": 0.8841084684138767,
      "eval_loss": 0.3043135702610016,
      "eval_runtime": 108.7426,
      "eval_samples_per_second": 60.703,
      "eval_steps_per_second": 7.596,
      "step": 700
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00016313603322949118,
      "loss": 0.1868,
      "step": 710
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00016261682242990654,
      "loss": 0.3254,
      "step": 720
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001620976116303219,
      "loss": 0.3054,
      "step": 730
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016157840083073728,
      "loss": 0.1782,
      "step": 740
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00016105919003115267,
      "loss": 0.2538,
      "step": 750
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00016053997923156803,
      "loss": 0.2088,
      "step": 760
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00016002076843198338,
      "loss": 0.3186,
      "step": 770
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00015950155763239877,
      "loss": 0.3071,
      "step": 780
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00015898234683281413,
      "loss": 0.2838,
      "step": 790
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001584631360332295,
      "loss": 0.2515,
      "step": 800
    },
    {
      "epoch": 0.83,
      "eval_accuracy": 0.9062263293440388,
      "eval_loss": 0.2541460394859314,
      "eval_runtime": 106.4027,
      "eval_samples_per_second": 62.038,
      "eval_steps_per_second": 7.763,
      "step": 800
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00015794392523364487,
      "loss": 0.3887,
      "step": 810
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00015742471443406023,
      "loss": 0.2235,
      "step": 820
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001569055036344756,
      "loss": 0.2313,
      "step": 830
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.00015638629283489097,
      "loss": 0.274,
      "step": 840
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.00015586708203530633,
      "loss": 0.2223,
      "step": 850
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0001553478712357217,
      "loss": 0.3155,
      "step": 860
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0001548286604361371,
      "loss": 0.2737,
      "step": 870
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.00015430944963655245,
      "loss": 0.2564,
      "step": 880
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0001537902388369678,
      "loss": 0.2587,
      "step": 890
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.00015327102803738317,
      "loss": 0.2854,
      "step": 900
    },
    {
      "epoch": 0.93,
      "eval_accuracy": 0.9006211180124224,
      "eval_loss": 0.2561000883579254,
      "eval_runtime": 107.0006,
      "eval_samples_per_second": 61.691,
      "eval_steps_per_second": 7.72,
      "step": 900
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.00015275181723779855,
      "loss": 0.2348,
      "step": 910
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.00015223260643821394,
      "loss": 0.2356,
      "step": 920
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0001517133956386293,
      "loss": 0.2324,
      "step": 930
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.00015119418483904465,
      "loss": 0.265,
      "step": 940
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.00015067497403946004,
      "loss": 0.2554,
      "step": 950
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0001501557632398754,
      "loss": 0.2845,
      "step": 960
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00014963655244029075,
      "loss": 0.3135,
      "step": 970
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00014911734164070614,
      "loss": 0.1335,
      "step": 980
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0001485981308411215,
      "loss": 0.3072,
      "step": 990
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00014807892004153688,
      "loss": 0.2594,
      "step": 1000
    },
    {
      "epoch": 1.04,
      "eval_accuracy": 0.9019845477957885,
      "eval_loss": 0.26812365651130676,
      "eval_runtime": 106.2371,
      "eval_samples_per_second": 62.135,
      "eval_steps_per_second": 7.775,
      "step": 1000
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00014755970924195224,
      "loss": 0.1768,
      "step": 1010
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0001470404984423676,
      "loss": 0.187,
      "step": 1020
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00014652128764278298,
      "loss": 0.1335,
      "step": 1030
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00014600207684319837,
      "loss": 0.1969,
      "step": 1040
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00014548286604361372,
      "loss": 0.228,
      "step": 1050
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.00014496365524402908,
      "loss": 0.1665,
      "step": 1060
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.00014444444444444444,
      "loss": 0.1743,
      "step": 1070
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.00014392523364485982,
      "loss": 0.228,
      "step": 1080
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.00014340602284527518,
      "loss": 0.1356,
      "step": 1090
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.00014288681204569056,
      "loss": 0.177,
      "step": 1100
    },
    {
      "epoch": 1.14,
      "eval_accuracy": 0.8772913194970459,
      "eval_loss": 0.3406391143798828,
      "eval_runtime": 105.7842,
      "eval_samples_per_second": 62.401,
      "eval_steps_per_second": 7.808,
      "step": 1100
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.00014236760124610592,
      "loss": 0.1793,
      "step": 1110
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0001418483904465213,
      "loss": 0.2743,
      "step": 1120
    },
    {
      "epoch": 1.17,
      "learning_rate": 0.00014132917964693666,
      "loss": 0.3143,
      "step": 1130
    },
    {
      "epoch": 1.18,
      "learning_rate": 0.00014080996884735202,
      "loss": 0.1482,
      "step": 1140
    },
    {
      "epoch": 1.19,
      "learning_rate": 0.00014029075804776738,
      "loss": 0.1937,
      "step": 1150
    },
    {
      "epoch": 1.2,
      "learning_rate": 0.00013977154724818276,
      "loss": 0.1386,
      "step": 1160
    },
    {
      "epoch": 1.21,
      "learning_rate": 0.00013925233644859815,
      "loss": 0.1427,
      "step": 1170
    },
    {
      "epoch": 1.23,
      "learning_rate": 0.0001387331256490135,
      "loss": 0.1565,
      "step": 1180
    },
    {
      "epoch": 1.24,
      "learning_rate": 0.00013821391484942886,
      "loss": 0.2196,
      "step": 1190
    },
    {
      "epoch": 1.25,
      "learning_rate": 0.00013769470404984425,
      "loss": 0.2717,
      "step": 1200
    },
    {
      "epoch": 1.25,
      "eval_accuracy": 0.917133767610968,
      "eval_loss": 0.22659926116466522,
      "eval_runtime": 105.225,
      "eval_samples_per_second": 62.732,
      "eval_steps_per_second": 7.85,
      "step": 1200
    },
    {
      "epoch": 1.26,
      "learning_rate": 0.0001371754932502596,
      "loss": 0.2137,
      "step": 1210
    },
    {
      "epoch": 1.27,
      "learning_rate": 0.000136656282450675,
      "loss": 0.1641,
      "step": 1220
    },
    {
      "epoch": 1.28,
      "learning_rate": 0.00013613707165109035,
      "loss": 0.1665,
      "step": 1230
    },
    {
      "epoch": 1.29,
      "learning_rate": 0.0001356178608515057,
      "loss": 0.1631,
      "step": 1240
    },
    {
      "epoch": 1.3,
      "learning_rate": 0.0001350986500519211,
      "loss": 0.2048,
      "step": 1250
    },
    {
      "epoch": 1.31,
      "learning_rate": 0.00013457943925233645,
      "loss": 0.1899,
      "step": 1260
    },
    {
      "epoch": 1.32,
      "learning_rate": 0.0001340602284527518,
      "loss": 0.1521,
      "step": 1270
    },
    {
      "epoch": 1.33,
      "learning_rate": 0.0001335410176531672,
      "loss": 0.1784,
      "step": 1280
    },
    {
      "epoch": 1.34,
      "learning_rate": 0.00013302180685358258,
      "loss": 0.122,
      "step": 1290
    },
    {
      "epoch": 1.35,
      "learning_rate": 0.00013250259605399793,
      "loss": 0.2197,
      "step": 1300
    },
    {
      "epoch": 1.35,
      "eval_accuracy": 0.9236479321314952,
      "eval_loss": 0.20801472663879395,
      "eval_runtime": 105.552,
      "eval_samples_per_second": 62.538,
      "eval_steps_per_second": 7.826,
      "step": 1300
    },
    {
      "epoch": 1.36,
      "learning_rate": 0.0001319833852544133,
      "loss": 0.0729,
      "step": 1310
    },
    {
      "epoch": 1.37,
      "learning_rate": 0.00013146417445482865,
      "loss": 0.2226,
      "step": 1320
    },
    {
      "epoch": 1.38,
      "learning_rate": 0.00013094496365524403,
      "loss": 0.1275,
      "step": 1330
    },
    {
      "epoch": 1.39,
      "learning_rate": 0.00013042575285565942,
      "loss": 0.2397,
      "step": 1340
    },
    {
      "epoch": 1.4,
      "learning_rate": 0.00012990654205607478,
      "loss": 0.153,
      "step": 1350
    },
    {
      "epoch": 1.41,
      "learning_rate": 0.00012938733125649013,
      "loss": 0.178,
      "step": 1360
    },
    {
      "epoch": 1.42,
      "learning_rate": 0.00012886812045690552,
      "loss": 0.2325,
      "step": 1370
    },
    {
      "epoch": 1.43,
      "learning_rate": 0.00012834890965732088,
      "loss": 0.2731,
      "step": 1380
    },
    {
      "epoch": 1.44,
      "learning_rate": 0.00012782969885773626,
      "loss": 0.1804,
      "step": 1390
    },
    {
      "epoch": 1.45,
      "learning_rate": 0.00012731048805815162,
      "loss": 0.155,
      "step": 1400
    },
    {
      "epoch": 1.45,
      "eval_accuracy": 0.9236479321314952,
      "eval_loss": 0.20478595793247223,
      "eval_runtime": 104.1144,
      "eval_samples_per_second": 63.401,
      "eval_steps_per_second": 7.934,
      "step": 1400
    },
    {
      "epoch": 1.46,
      "learning_rate": 0.00012679127725856698,
      "loss": 0.1645,
      "step": 1410
    },
    {
      "epoch": 1.47,
      "learning_rate": 0.00012627206645898236,
      "loss": 0.2205,
      "step": 1420
    },
    {
      "epoch": 1.48,
      "learning_rate": 0.00012575285565939772,
      "loss": 0.1507,
      "step": 1430
    },
    {
      "epoch": 1.5,
      "learning_rate": 0.00012523364485981308,
      "loss": 0.4007,
      "step": 1440
    },
    {
      "epoch": 1.51,
      "learning_rate": 0.00012471443406022846,
      "loss": 0.1829,
      "step": 1450
    },
    {
      "epoch": 1.52,
      "learning_rate": 0.00012419522326064384,
      "loss": 0.1755,
      "step": 1460
    },
    {
      "epoch": 1.53,
      "learning_rate": 0.0001236760124610592,
      "loss": 0.0943,
      "step": 1470
    },
    {
      "epoch": 1.54,
      "learning_rate": 0.00012315680166147456,
      "loss": 0.1359,
      "step": 1480
    },
    {
      "epoch": 1.55,
      "learning_rate": 0.00012263759086188992,
      "loss": 0.2615,
      "step": 1490
    },
    {
      "epoch": 1.56,
      "learning_rate": 0.0001221183800623053,
      "loss": 0.2657,
      "step": 1500
    },
    {
      "epoch": 1.56,
      "eval_accuracy": 0.9256173307074685,
      "eval_loss": 0.2036730796098709,
      "eval_runtime": 104.5196,
      "eval_samples_per_second": 63.156,
      "eval_steps_per_second": 7.903,
      "step": 1500
    },
    {
      "epoch": 1.57,
      "learning_rate": 0.00012159916926272067,
      "loss": 0.1653,
      "step": 1510
    },
    {
      "epoch": 1.58,
      "learning_rate": 0.00012107995846313604,
      "loss": 0.2066,
      "step": 1520
    },
    {
      "epoch": 1.59,
      "learning_rate": 0.00012056074766355142,
      "loss": 0.1706,
      "step": 1530
    },
    {
      "epoch": 1.6,
      "learning_rate": 0.00012004153686396677,
      "loss": 0.2103,
      "step": 1540
    },
    {
      "epoch": 1.61,
      "learning_rate": 0.00011952232606438214,
      "loss": 0.2336,
      "step": 1550
    },
    {
      "epoch": 1.62,
      "learning_rate": 0.0001190031152647975,
      "loss": 0.1792,
      "step": 1560
    },
    {
      "epoch": 1.63,
      "learning_rate": 0.00011848390446521289,
      "loss": 0.0966,
      "step": 1570
    },
    {
      "epoch": 1.64,
      "learning_rate": 0.00011796469366562826,
      "loss": 0.1713,
      "step": 1580
    },
    {
      "epoch": 1.65,
      "learning_rate": 0.00011744548286604362,
      "loss": 0.1289,
      "step": 1590
    },
    {
      "epoch": 1.66,
      "learning_rate": 0.00011692627206645899,
      "loss": 0.118,
      "step": 1600
    },
    {
      "epoch": 1.66,
      "eval_accuracy": 0.9095591577033783,
      "eval_loss": 0.26160144805908203,
      "eval_runtime": 104.9104,
      "eval_samples_per_second": 62.92,
      "eval_steps_per_second": 7.873,
      "step": 1600
    },
    {
      "epoch": 1.67,
      "learning_rate": 0.00011640706126687434,
      "loss": 0.1872,
      "step": 1610
    },
    {
      "epoch": 1.68,
      "learning_rate": 0.00011588785046728972,
      "loss": 0.152,
      "step": 1620
    },
    {
      "epoch": 1.69,
      "learning_rate": 0.0001153686396677051,
      "loss": 0.1337,
      "step": 1630
    },
    {
      "epoch": 1.7,
      "learning_rate": 0.00011484942886812047,
      "loss": 0.1553,
      "step": 1640
    },
    {
      "epoch": 1.71,
      "learning_rate": 0.00011433021806853583,
      "loss": 0.1428,
      "step": 1650
    },
    {
      "epoch": 1.72,
      "learning_rate": 0.0001138110072689512,
      "loss": 0.1911,
      "step": 1660
    },
    {
      "epoch": 1.73,
      "learning_rate": 0.00011329179646936656,
      "loss": 0.1263,
      "step": 1670
    },
    {
      "epoch": 1.74,
      "learning_rate": 0.00011277258566978193,
      "loss": 0.1414,
      "step": 1680
    },
    {
      "epoch": 1.75,
      "learning_rate": 0.00011225337487019731,
      "loss": 0.2111,
      "step": 1690
    },
    {
      "epoch": 1.77,
      "learning_rate": 0.00011173416407061268,
      "loss": 0.1823,
      "step": 1700
    },
    {
      "epoch": 1.77,
      "eval_accuracy": 0.9241024087259506,
      "eval_loss": 0.21581591665744781,
      "eval_runtime": 104.5278,
      "eval_samples_per_second": 63.151,
      "eval_steps_per_second": 7.902,
      "step": 1700
    },
    {
      "epoch": 1.78,
      "learning_rate": 0.00011121495327102804,
      "loss": 0.1266,
      "step": 1710
    },
    {
      "epoch": 1.79,
      "learning_rate": 0.00011069574247144341,
      "loss": 0.1765,
      "step": 1720
    },
    {
      "epoch": 1.8,
      "learning_rate": 0.00011017653167185877,
      "loss": 0.1983,
      "step": 1730
    },
    {
      "epoch": 1.81,
      "learning_rate": 0.00010965732087227414,
      "loss": 0.253,
      "step": 1740
    },
    {
      "epoch": 1.82,
      "learning_rate": 0.00010913811007268953,
      "loss": 0.1542,
      "step": 1750
    },
    {
      "epoch": 1.83,
      "learning_rate": 0.00010861889927310488,
      "loss": 0.1171,
      "step": 1760
    },
    {
      "epoch": 1.84,
      "learning_rate": 0.00010809968847352026,
      "loss": 0.1701,
      "step": 1770
    },
    {
      "epoch": 1.85,
      "learning_rate": 0.00010758047767393563,
      "loss": 0.1689,
      "step": 1780
    },
    {
      "epoch": 1.86,
      "learning_rate": 0.00010706126687435098,
      "loss": 0.1405,
      "step": 1790
    },
    {
      "epoch": 1.87,
      "learning_rate": 0.00010654205607476636,
      "loss": 0.2175,
      "step": 1800
    },
    {
      "epoch": 1.87,
      "eval_accuracy": 0.9181942129980306,
      "eval_loss": 0.21594561636447906,
      "eval_runtime": 103.9435,
      "eval_samples_per_second": 63.506,
      "eval_steps_per_second": 7.947,
      "step": 1800
    },
    {
      "epoch": 1.88,
      "learning_rate": 0.00010602284527518174,
      "loss": 0.1923,
      "step": 1810
    },
    {
      "epoch": 1.89,
      "learning_rate": 0.0001055036344755971,
      "loss": 0.2476,
      "step": 1820
    },
    {
      "epoch": 1.9,
      "learning_rate": 0.00010498442367601247,
      "loss": 0.2243,
      "step": 1830
    },
    {
      "epoch": 1.91,
      "learning_rate": 0.00010446521287642783,
      "loss": 0.2611,
      "step": 1840
    },
    {
      "epoch": 1.92,
      "learning_rate": 0.0001039460020768432,
      "loss": 0.2186,
      "step": 1850
    },
    {
      "epoch": 1.93,
      "learning_rate": 0.00010342679127725856,
      "loss": 0.141,
      "step": 1860
    },
    {
      "epoch": 1.94,
      "learning_rate": 0.00010290758047767395,
      "loss": 0.1086,
      "step": 1870
    },
    {
      "epoch": 1.95,
      "learning_rate": 0.00010238836967808931,
      "loss": 0.1948,
      "step": 1880
    },
    {
      "epoch": 1.96,
      "learning_rate": 0.00010186915887850468,
      "loss": 0.166,
      "step": 1890
    },
    {
      "epoch": 1.97,
      "learning_rate": 0.00010134994807892004,
      "loss": 0.143,
      "step": 1900
    },
    {
      "epoch": 1.97,
      "eval_accuracy": 0.9242539009241024,
      "eval_loss": 0.18984103202819824,
      "eval_runtime": 107.0187,
      "eval_samples_per_second": 61.681,
      "eval_steps_per_second": 7.718,
      "step": 1900
    },
    {
      "epoch": 1.98,
      "learning_rate": 0.00010083073727933541,
      "loss": 0.1538,
      "step": 1910
    },
    {
      "epoch": 1.99,
      "learning_rate": 0.00010031152647975077,
      "loss": 0.1682,
      "step": 1920
    },
    {
      "epoch": 2.0,
      "learning_rate": 9.979231568016615e-05,
      "loss": 0.09,
      "step": 1930
    },
    {
      "epoch": 2.01,
      "learning_rate": 9.927310488058152e-05,
      "loss": 0.0338,
      "step": 1940
    },
    {
      "epoch": 2.02,
      "learning_rate": 9.87538940809969e-05,
      "loss": 0.1026,
      "step": 1950
    },
    {
      "epoch": 2.04,
      "learning_rate": 9.823468328141225e-05,
      "loss": 0.1165,
      "step": 1960
    },
    {
      "epoch": 2.05,
      "learning_rate": 9.771547248182762e-05,
      "loss": 0.0652,
      "step": 1970
    },
    {
      "epoch": 2.06,
      "learning_rate": 9.7196261682243e-05,
      "loss": 0.1643,
      "step": 1980
    },
    {
      "epoch": 2.07,
      "learning_rate": 9.667705088265837e-05,
      "loss": 0.0903,
      "step": 1990
    },
    {
      "epoch": 2.08,
      "learning_rate": 9.615784008307374e-05,
      "loss": 0.1051,
      "step": 2000
    },
    {
      "epoch": 2.08,
      "eval_accuracy": 0.9225874867444327,
      "eval_loss": 0.23080797493457794,
      "eval_runtime": 105.3241,
      "eval_samples_per_second": 62.673,
      "eval_steps_per_second": 7.842,
      "step": 2000
    },
    {
      "epoch": 2.09,
      "learning_rate": 9.56386292834891e-05,
      "loss": 0.1135,
      "step": 2010
    },
    {
      "epoch": 2.1,
      "learning_rate": 9.511941848390447e-05,
      "loss": 0.0586,
      "step": 2020
    },
    {
      "epoch": 2.11,
      "learning_rate": 9.460020768431984e-05,
      "loss": 0.1016,
      "step": 2030
    },
    {
      "epoch": 2.12,
      "learning_rate": 9.408099688473521e-05,
      "loss": 0.0808,
      "step": 2040
    },
    {
      "epoch": 2.13,
      "learning_rate": 9.356178608515057e-05,
      "loss": 0.0502,
      "step": 2050
    },
    {
      "epoch": 2.14,
      "learning_rate": 9.304257528556595e-05,
      "loss": 0.0712,
      "step": 2060
    },
    {
      "epoch": 2.15,
      "learning_rate": 9.252336448598131e-05,
      "loss": 0.0938,
      "step": 2070
    },
    {
      "epoch": 2.16,
      "learning_rate": 9.200415368639668e-05,
      "loss": 0.0539,
      "step": 2080
    },
    {
      "epoch": 2.17,
      "learning_rate": 9.148494288681205e-05,
      "loss": 0.1137,
      "step": 2090
    },
    {
      "epoch": 2.18,
      "learning_rate": 9.096573208722742e-05,
      "loss": 0.1963,
      "step": 2100
    },
    {
      "epoch": 2.18,
      "eval_accuracy": 0.9204665959703076,
      "eval_loss": 0.23536920547485352,
      "eval_runtime": 106.2404,
      "eval_samples_per_second": 62.133,
      "eval_steps_per_second": 7.775,
      "step": 2100
    },
    {
      "epoch": 2.19,
      "learning_rate": 9.044652128764278e-05,
      "loss": 0.1132,
      "step": 2110
    },
    {
      "epoch": 2.2,
      "learning_rate": 8.992731048805816e-05,
      "loss": 0.0318,
      "step": 2120
    },
    {
      "epoch": 2.21,
      "learning_rate": 8.940809968847352e-05,
      "loss": 0.0562,
      "step": 2130
    },
    {
      "epoch": 2.22,
      "learning_rate": 8.888888888888889e-05,
      "loss": 0.0991,
      "step": 2140
    },
    {
      "epoch": 2.23,
      "learning_rate": 8.836967808930426e-05,
      "loss": 0.1263,
      "step": 2150
    },
    {
      "epoch": 2.24,
      "learning_rate": 8.785046728971964e-05,
      "loss": 0.093,
      "step": 2160
    },
    {
      "epoch": 2.25,
      "learning_rate": 8.733125649013499e-05,
      "loss": 0.0992,
      "step": 2170
    },
    {
      "epoch": 2.26,
      "learning_rate": 8.681204569055036e-05,
      "loss": 0.0787,
      "step": 2180
    },
    {
      "epoch": 2.27,
      "learning_rate": 8.629283489096574e-05,
      "loss": 0.1306,
      "step": 2190
    },
    {
      "epoch": 2.28,
      "learning_rate": 8.577362409138109e-05,
      "loss": 0.0524,
      "step": 2200
    },
    {
      "epoch": 2.28,
      "eval_accuracy": 0.9281926980760491,
      "eval_loss": 0.22982539236545563,
      "eval_runtime": 105.4663,
      "eval_samples_per_second": 62.589,
      "eval_steps_per_second": 7.832,
      "step": 2200
    },
    {
      "epoch": 2.29,
      "learning_rate": 8.525441329179648e-05,
      "loss": 0.1902,
      "step": 2210
    },
    {
      "epoch": 2.31,
      "learning_rate": 8.473520249221184e-05,
      "loss": 0.0755,
      "step": 2220
    },
    {
      "epoch": 2.32,
      "learning_rate": 8.42159916926272e-05,
      "loss": 0.0826,
      "step": 2230
    },
    {
      "epoch": 2.33,
      "learning_rate": 8.369678089304258e-05,
      "loss": 0.1252,
      "step": 2240
    },
    {
      "epoch": 2.34,
      "learning_rate": 8.317757009345795e-05,
      "loss": 0.0848,
      "step": 2250
    },
    {
      "epoch": 2.35,
      "learning_rate": 8.265835929387332e-05,
      "loss": 0.0134,
      "step": 2260
    },
    {
      "epoch": 2.36,
      "learning_rate": 8.213914849428869e-05,
      "loss": 0.1297,
      "step": 2270
    },
    {
      "epoch": 2.37,
      "learning_rate": 8.161993769470405e-05,
      "loss": 0.0633,
      "step": 2280
    },
    {
      "epoch": 2.38,
      "learning_rate": 8.110072689511943e-05,
      "loss": 0.0625,
      "step": 2290
    },
    {
      "epoch": 2.39,
      "learning_rate": 8.058151609553479e-05,
      "loss": 0.097,
      "step": 2300
    },
    {
      "epoch": 2.39,
      "eval_accuracy": 0.9241024087259506,
      "eval_loss": 0.24952168762683868,
      "eval_runtime": 105.3776,
      "eval_samples_per_second": 62.641,
      "eval_steps_per_second": 7.838,
      "step": 2300
    },
    {
      "epoch": 2.4,
      "learning_rate": 8.006230529595016e-05,
      "loss": 0.0439,
      "step": 2310
    },
    {
      "epoch": 2.41,
      "learning_rate": 7.954309449636553e-05,
      "loss": 0.0743,
      "step": 2320
    },
    {
      "epoch": 2.42,
      "learning_rate": 7.90238836967809e-05,
      "loss": 0.1643,
      "step": 2330
    },
    {
      "epoch": 2.43,
      "learning_rate": 7.850467289719626e-05,
      "loss": 0.1348,
      "step": 2340
    },
    {
      "epoch": 2.44,
      "learning_rate": 7.798546209761163e-05,
      "loss": 0.1049,
      "step": 2350
    },
    {
      "epoch": 2.45,
      "learning_rate": 7.7466251298027e-05,
      "loss": 0.1646,
      "step": 2360
    },
    {
      "epoch": 2.46,
      "learning_rate": 7.694704049844238e-05,
      "loss": 0.123,
      "step": 2370
    },
    {
      "epoch": 2.47,
      "learning_rate": 7.642782969885775e-05,
      "loss": 0.0567,
      "step": 2380
    },
    {
      "epoch": 2.48,
      "learning_rate": 7.59086188992731e-05,
      "loss": 0.1242,
      "step": 2390
    },
    {
      "epoch": 2.49,
      "learning_rate": 7.538940809968848e-05,
      "loss": 0.0744,
      "step": 2400
    },
    {
      "epoch": 2.49,
      "eval_accuracy": 0.9194061505832449,
      "eval_loss": 0.24930374324321747,
      "eval_runtime": 104.785,
      "eval_samples_per_second": 62.996,
      "eval_steps_per_second": 7.883,
      "step": 2400
    },
    {
      "epoch": 2.5,
      "learning_rate": 7.487019730010385e-05,
      "loss": 0.0935,
      "step": 2410
    },
    {
      "epoch": 2.51,
      "learning_rate": 7.435098650051922e-05,
      "loss": 0.0632,
      "step": 2420
    },
    {
      "epoch": 2.52,
      "learning_rate": 7.383177570093458e-05,
      "loss": 0.0975,
      "step": 2430
    },
    {
      "epoch": 2.53,
      "learning_rate": 7.331256490134996e-05,
      "loss": 0.0355,
      "step": 2440
    },
    {
      "epoch": 2.54,
      "learning_rate": 7.279335410176532e-05,
      "loss": 0.0485,
      "step": 2450
    },
    {
      "epoch": 2.55,
      "learning_rate": 7.227414330218069e-05,
      "loss": 0.0503,
      "step": 2460
    },
    {
      "epoch": 2.56,
      "learning_rate": 7.175493250259606e-05,
      "loss": 0.0745,
      "step": 2470
    },
    {
      "epoch": 2.58,
      "learning_rate": 7.123572170301143e-05,
      "loss": 0.1093,
      "step": 2480
    },
    {
      "epoch": 2.59,
      "learning_rate": 7.071651090342679e-05,
      "loss": 0.0554,
      "step": 2490
    },
    {
      "epoch": 2.6,
      "learning_rate": 7.019730010384217e-05,
      "loss": 0.0744,
      "step": 2500
    },
    {
      "epoch": 2.6,
      "eval_accuracy": 0.9322829874261476,
      "eval_loss": 0.2428741455078125,
      "eval_runtime": 104.9683,
      "eval_samples_per_second": 62.886,
      "eval_steps_per_second": 7.869,
      "step": 2500
    },
    {
      "epoch": 2.61,
      "learning_rate": 6.967808930425753e-05,
      "loss": 0.0998,
      "step": 2510
    },
    {
      "epoch": 2.62,
      "learning_rate": 6.91588785046729e-05,
      "loss": 0.0561,
      "step": 2520
    },
    {
      "epoch": 2.63,
      "learning_rate": 6.863966770508827e-05,
      "loss": 0.0452,
      "step": 2530
    },
    {
      "epoch": 2.64,
      "learning_rate": 6.812045690550364e-05,
      "loss": 0.1821,
      "step": 2540
    },
    {
      "epoch": 2.65,
      "learning_rate": 6.7601246105919e-05,
      "loss": 0.0598,
      "step": 2550
    },
    {
      "epoch": 2.66,
      "learning_rate": 6.708203530633437e-05,
      "loss": 0.0722,
      "step": 2560
    },
    {
      "epoch": 2.67,
      "learning_rate": 6.656282450674974e-05,
      "loss": 0.0532,
      "step": 2570
    },
    {
      "epoch": 2.68,
      "learning_rate": 6.60436137071651e-05,
      "loss": 0.0644,
      "step": 2580
    },
    {
      "epoch": 2.69,
      "learning_rate": 6.552440290758049e-05,
      "loss": 0.1235,
      "step": 2590
    },
    {
      "epoch": 2.7,
      "learning_rate": 6.500519210799584e-05,
      "loss": 0.0345,
      "step": 2600
    },
    {
      "epoch": 2.7,
      "eval_accuracy": 0.9251628541130131,
      "eval_loss": 0.2586543560028076,
      "eval_runtime": 104.6212,
      "eval_samples_per_second": 63.094,
      "eval_steps_per_second": 7.895,
      "step": 2600
    },
    {
      "epoch": 2.71,
      "learning_rate": 6.448598130841122e-05,
      "loss": 0.1686,
      "step": 2610
    },
    {
      "epoch": 2.72,
      "learning_rate": 6.396677050882659e-05,
      "loss": 0.0775,
      "step": 2620
    },
    {
      "epoch": 2.73,
      "learning_rate": 6.344755970924196e-05,
      "loss": 0.0956,
      "step": 2630
    },
    {
      "epoch": 2.74,
      "learning_rate": 6.292834890965732e-05,
      "loss": 0.135,
      "step": 2640
    },
    {
      "epoch": 2.75,
      "learning_rate": 6.24091381100727e-05,
      "loss": 0.1582,
      "step": 2650
    },
    {
      "epoch": 2.76,
      "learning_rate": 6.188992731048806e-05,
      "loss": 0.0668,
      "step": 2660
    },
    {
      "epoch": 2.77,
      "learning_rate": 6.137071651090343e-05,
      "loss": 0.1317,
      "step": 2670
    },
    {
      "epoch": 2.78,
      "learning_rate": 6.08515057113188e-05,
      "loss": 0.0756,
      "step": 2680
    },
    {
      "epoch": 2.79,
      "learning_rate": 6.0332294911734164e-05,
      "loss": 0.0565,
      "step": 2690
    },
    {
      "epoch": 2.8,
      "learning_rate": 5.981308411214953e-05,
      "loss": 0.0097,
      "step": 2700
    },
    {
      "epoch": 2.8,
      "eval_accuracy": 0.9265262838963794,
      "eval_loss": 0.22841490805149078,
      "eval_runtime": 104.4263,
      "eval_samples_per_second": 63.212,
      "eval_steps_per_second": 7.91,
      "step": 2700
    },
    {
      "epoch": 2.81,
      "learning_rate": 5.9293873312564906e-05,
      "loss": 0.0621,
      "step": 2710
    },
    {
      "epoch": 2.82,
      "learning_rate": 5.877466251298027e-05,
      "loss": 0.0684,
      "step": 2720
    },
    {
      "epoch": 2.83,
      "learning_rate": 5.8255451713395635e-05,
      "loss": 0.0408,
      "step": 2730
    },
    {
      "epoch": 2.85,
      "learning_rate": 5.773624091381101e-05,
      "loss": 0.1156,
      "step": 2740
    },
    {
      "epoch": 2.86,
      "learning_rate": 5.721703011422638e-05,
      "loss": 0.0228,
      "step": 2750
    },
    {
      "epoch": 2.87,
      "learning_rate": 5.669781931464174e-05,
      "loss": 0.1417,
      "step": 2760
    },
    {
      "epoch": 2.88,
      "learning_rate": 5.617860851505712e-05,
      "loss": 0.0443,
      "step": 2770
    },
    {
      "epoch": 2.89,
      "learning_rate": 5.5659397715472484e-05,
      "loss": 0.0903,
      "step": 2780
    },
    {
      "epoch": 2.9,
      "learning_rate": 5.514018691588785e-05,
      "loss": 0.0395,
      "step": 2790
    },
    {
      "epoch": 2.91,
      "learning_rate": 5.4620976116303226e-05,
      "loss": 0.0775,
      "step": 2800
    },
    {
      "epoch": 2.91,
      "eval_accuracy": 0.9321314952279958,
      "eval_loss": 0.22424499690532684,
      "eval_runtime": 104.7535,
      "eval_samples_per_second": 63.015,
      "eval_steps_per_second": 7.885,
      "step": 2800
    },
    {
      "epoch": 2.92,
      "learning_rate": 5.410176531671859e-05,
      "loss": 0.0281,
      "step": 2810
    },
    {
      "epoch": 2.93,
      "learning_rate": 5.3582554517133955e-05,
      "loss": 0.0756,
      "step": 2820
    },
    {
      "epoch": 2.94,
      "learning_rate": 5.306334371754933e-05,
      "loss": 0.083,
      "step": 2830
    },
    {
      "epoch": 2.95,
      "learning_rate": 5.25441329179647e-05,
      "loss": 0.0804,
      "step": 2840
    },
    {
      "epoch": 2.96,
      "learning_rate": 5.202492211838006e-05,
      "loss": 0.068,
      "step": 2850
    },
    {
      "epoch": 2.97,
      "learning_rate": 5.150571131879543e-05,
      "loss": 0.0549,
      "step": 2860
    },
    {
      "epoch": 2.98,
      "learning_rate": 5.09865005192108e-05,
      "loss": 0.1113,
      "step": 2870
    },
    {
      "epoch": 2.99,
      "learning_rate": 5.046728971962617e-05,
      "loss": 0.1297,
      "step": 2880
    },
    {
      "epoch": 3.0,
      "learning_rate": 4.994807892004154e-05,
      "loss": 0.0397,
      "step": 2890
    },
    {
      "epoch": 3.01,
      "learning_rate": 4.9428868120456904e-05,
      "loss": 0.0634,
      "step": 2900
    },
    {
      "epoch": 3.01,
      "eval_accuracy": 0.9286471746705045,
      "eval_loss": 0.23135913908481598,
      "eval_runtime": 104.8164,
      "eval_samples_per_second": 62.977,
      "eval_steps_per_second": 7.88,
      "step": 2900
    },
    {
      "epoch": 3.02,
      "learning_rate": 4.8909657320872275e-05,
      "loss": 0.02,
      "step": 2910
    },
    {
      "epoch": 3.03,
      "learning_rate": 4.8390446521287646e-05,
      "loss": 0.0324,
      "step": 2920
    },
    {
      "epoch": 3.04,
      "learning_rate": 4.787123572170301e-05,
      "loss": 0.0064,
      "step": 2930
    },
    {
      "epoch": 3.05,
      "learning_rate": 4.735202492211838e-05,
      "loss": 0.0063,
      "step": 2940
    },
    {
      "epoch": 3.06,
      "learning_rate": 4.683281412253375e-05,
      "loss": 0.0372,
      "step": 2950
    },
    {
      "epoch": 3.07,
      "learning_rate": 4.631360332294912e-05,
      "loss": 0.0679,
      "step": 2960
    },
    {
      "epoch": 3.08,
      "learning_rate": 4.579439252336449e-05,
      "loss": 0.0517,
      "step": 2970
    },
    {
      "epoch": 3.09,
      "learning_rate": 4.527518172377986e-05,
      "loss": 0.097,
      "step": 2980
    },
    {
      "epoch": 3.1,
      "learning_rate": 4.4755970924195224e-05,
      "loss": 0.1,
      "step": 2990
    },
    {
      "epoch": 3.12,
      "learning_rate": 4.4236760124610595e-05,
      "loss": 0.0109,
      "step": 3000
    },
    {
      "epoch": 3.12,
      "eval_accuracy": 0.9337979094076655,
      "eval_loss": 0.22030013799667358,
      "eval_runtime": 105.6713,
      "eval_samples_per_second": 62.467,
      "eval_steps_per_second": 7.817,
      "step": 3000
    },
    {
      "epoch": 3.13,
      "learning_rate": 4.3717549325025966e-05,
      "loss": 0.0143,
      "step": 3010
    },
    {
      "epoch": 3.14,
      "learning_rate": 4.319833852544133e-05,
      "loss": 0.0638,
      "step": 3020
    },
    {
      "epoch": 3.15,
      "learning_rate": 4.26791277258567e-05,
      "loss": 0.0213,
      "step": 3030
    },
    {
      "epoch": 3.16,
      "learning_rate": 4.2159916926272066e-05,
      "loss": 0.0153,
      "step": 3040
    },
    {
      "epoch": 3.17,
      "learning_rate": 4.164070612668744e-05,
      "loss": 0.0052,
      "step": 3050
    },
    {
      "epoch": 3.18,
      "learning_rate": 4.11214953271028e-05,
      "loss": 0.0406,
      "step": 3060
    },
    {
      "epoch": 3.19,
      "learning_rate": 4.060228452751817e-05,
      "loss": 0.0109,
      "step": 3070
    },
    {
      "epoch": 3.2,
      "learning_rate": 4.0083073727933544e-05,
      "loss": 0.0058,
      "step": 3080
    },
    {
      "epoch": 3.21,
      "learning_rate": 3.956386292834891e-05,
      "loss": 0.005,
      "step": 3090
    },
    {
      "epoch": 3.22,
      "learning_rate": 3.904465212876428e-05,
      "loss": 0.0039,
      "step": 3100
    },
    {
      "epoch": 3.22,
      "eval_accuracy": 0.9357673079836388,
      "eval_loss": 0.2575285732746124,
      "eval_runtime": 105.4895,
      "eval_samples_per_second": 62.575,
      "eval_steps_per_second": 7.83,
      "step": 3100
    },
    {
      "epoch": 3.23,
      "learning_rate": 3.852544132917965e-05,
      "loss": 0.0284,
      "step": 3110
    },
    {
      "epoch": 3.24,
      "learning_rate": 3.8006230529595015e-05,
      "loss": 0.0042,
      "step": 3120
    },
    {
      "epoch": 3.25,
      "learning_rate": 3.7487019730010386e-05,
      "loss": 0.0407,
      "step": 3130
    },
    {
      "epoch": 3.26,
      "learning_rate": 3.696780893042576e-05,
      "loss": 0.009,
      "step": 3140
    },
    {
      "epoch": 3.27,
      "learning_rate": 3.644859813084112e-05,
      "loss": 0.0342,
      "step": 3150
    },
    {
      "epoch": 3.28,
      "learning_rate": 3.592938733125649e-05,
      "loss": 0.056,
      "step": 3160
    },
    {
      "epoch": 3.29,
      "learning_rate": 3.5410176531671864e-05,
      "loss": 0.0384,
      "step": 3170
    },
    {
      "epoch": 3.3,
      "learning_rate": 3.489096573208723e-05,
      "loss": 0.0319,
      "step": 3180
    },
    {
      "epoch": 3.31,
      "learning_rate": 3.43717549325026e-05,
      "loss": 0.0619,
      "step": 3190
    },
    {
      "epoch": 3.32,
      "learning_rate": 3.385254413291797e-05,
      "loss": 0.0139,
      "step": 3200
    },
    {
      "epoch": 3.32,
      "eval_accuracy": 0.935615815785487,
      "eval_loss": 0.25701212882995605,
      "eval_runtime": 105.4027,
      "eval_samples_per_second": 62.626,
      "eval_steps_per_second": 7.837,
      "step": 3200
    },
    {
      "epoch": 3.33,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 0.0035,
      "step": 3210
    },
    {
      "epoch": 3.34,
      "learning_rate": 3.2814122533748706e-05,
      "loss": 0.0232,
      "step": 3220
    },
    {
      "epoch": 3.35,
      "learning_rate": 3.229491173416408e-05,
      "loss": 0.0078,
      "step": 3230
    },
    {
      "epoch": 3.36,
      "learning_rate": 3.177570093457944e-05,
      "loss": 0.041,
      "step": 3240
    },
    {
      "epoch": 3.37,
      "learning_rate": 3.1256490134994806e-05,
      "loss": 0.0408,
      "step": 3250
    },
    {
      "epoch": 3.39,
      "learning_rate": 3.073727933541018e-05,
      "loss": 0.0372,
      "step": 3260
    },
    {
      "epoch": 3.4,
      "learning_rate": 3.0218068535825545e-05,
      "loss": 0.0639,
      "step": 3270
    },
    {
      "epoch": 3.41,
      "learning_rate": 2.9698857736240916e-05,
      "loss": 0.0205,
      "step": 3280
    },
    {
      "epoch": 3.42,
      "learning_rate": 2.9179646936656284e-05,
      "loss": 0.0579,
      "step": 3290
    },
    {
      "epoch": 3.43,
      "learning_rate": 2.866043613707165e-05,
      "loss": 0.0358,
      "step": 3300
    },
    {
      "epoch": 3.43,
      "eval_accuracy": 0.9334949250113619,
      "eval_loss": 0.2629750072956085,
      "eval_runtime": 105.1034,
      "eval_samples_per_second": 62.805,
      "eval_steps_per_second": 7.859,
      "step": 3300
    },
    {
      "epoch": 3.44,
      "learning_rate": 2.814122533748702e-05,
      "loss": 0.0598,
      "step": 3310
    },
    {
      "epoch": 3.45,
      "learning_rate": 2.762201453790239e-05,
      "loss": 0.0089,
      "step": 3320
    },
    {
      "epoch": 3.46,
      "learning_rate": 2.7102803738317755e-05,
      "loss": 0.0032,
      "step": 3330
    },
    {
      "epoch": 3.47,
      "learning_rate": 2.6583592938733126e-05,
      "loss": 0.0039,
      "step": 3340
    },
    {
      "epoch": 3.48,
      "learning_rate": 2.6064382139148497e-05,
      "loss": 0.0053,
      "step": 3350
    },
    {
      "epoch": 3.49,
      "learning_rate": 2.554517133956386e-05,
      "loss": 0.0093,
      "step": 3360
    },
    {
      "epoch": 3.5,
      "learning_rate": 2.5025960539979233e-05,
      "loss": 0.0657,
      "step": 3370
    },
    {
      "epoch": 3.51,
      "learning_rate": 2.4506749740394604e-05,
      "loss": 0.0027,
      "step": 3380
    },
    {
      "epoch": 3.52,
      "learning_rate": 2.398753894080997e-05,
      "loss": 0.0775,
      "step": 3390
    },
    {
      "epoch": 3.53,
      "learning_rate": 2.3468328141225336e-05,
      "loss": 0.0347,
      "step": 3400
    },
    {
      "epoch": 3.53,
      "eval_accuracy": 0.9357673079836388,
      "eval_loss": 0.2633425295352936,
      "eval_runtime": 105.2554,
      "eval_samples_per_second": 62.714,
      "eval_steps_per_second": 7.848,
      "step": 3400
    },
    {
      "epoch": 3.54,
      "learning_rate": 2.2949117341640707e-05,
      "loss": 0.0288,
      "step": 3410
    },
    {
      "epoch": 3.55,
      "learning_rate": 2.2429906542056075e-05,
      "loss": 0.039,
      "step": 3420
    },
    {
      "epoch": 3.56,
      "learning_rate": 2.1910695742471443e-05,
      "loss": 0.0565,
      "step": 3430
    },
    {
      "epoch": 3.57,
      "learning_rate": 2.1391484942886814e-05,
      "loss": 0.0412,
      "step": 3440
    },
    {
      "epoch": 3.58,
      "learning_rate": 2.087227414330218e-05,
      "loss": 0.0391,
      "step": 3450
    },
    {
      "epoch": 3.59,
      "learning_rate": 2.0353063343717553e-05,
      "loss": 0.0036,
      "step": 3460
    },
    {
      "epoch": 3.6,
      "learning_rate": 1.983385254413292e-05,
      "loss": 0.0045,
      "step": 3470
    },
    {
      "epoch": 3.61,
      "learning_rate": 1.9314641744548288e-05,
      "loss": 0.0679,
      "step": 3480
    },
    {
      "epoch": 3.62,
      "learning_rate": 1.8795430944963656e-05,
      "loss": 0.0169,
      "step": 3490
    },
    {
      "epoch": 3.63,
      "learning_rate": 1.8276220145379024e-05,
      "loss": 0.0408,
      "step": 3500
    },
    {
      "epoch": 3.63,
      "eval_accuracy": 0.9334949250113619,
      "eval_loss": 0.2590745985507965,
      "eval_runtime": 105.6016,
      "eval_samples_per_second": 62.509,
      "eval_steps_per_second": 7.822,
      "step": 3500
    },
    {
      "epoch": 3.64,
      "learning_rate": 1.775700934579439e-05,
      "loss": 0.0217,
      "step": 3510
    },
    {
      "epoch": 3.66,
      "learning_rate": 1.7237798546209763e-05,
      "loss": 0.0186,
      "step": 3520
    },
    {
      "epoch": 3.67,
      "learning_rate": 1.671858774662513e-05,
      "loss": 0.006,
      "step": 3530
    },
    {
      "epoch": 3.68,
      "learning_rate": 1.6199376947040498e-05,
      "loss": 0.0216,
      "step": 3540
    },
    {
      "epoch": 3.69,
      "learning_rate": 1.568016614745587e-05,
      "loss": 0.0108,
      "step": 3550
    },
    {
      "epoch": 3.7,
      "learning_rate": 1.5160955347871237e-05,
      "loss": 0.0033,
      "step": 3560
    },
    {
      "epoch": 3.71,
      "learning_rate": 1.4641744548286603e-05,
      "loss": 0.0038,
      "step": 3570
    },
    {
      "epoch": 3.72,
      "learning_rate": 1.4122533748701974e-05,
      "loss": 0.0039,
      "step": 3580
    },
    {
      "epoch": 3.73,
      "learning_rate": 1.3603322949117342e-05,
      "loss": 0.0311,
      "step": 3590
    },
    {
      "epoch": 3.74,
      "learning_rate": 1.308411214953271e-05,
      "loss": 0.041,
      "step": 3600
    },
    {
      "epoch": 3.74,
      "eval_accuracy": 0.9366762611725497,
      "eval_loss": 0.2612759470939636,
      "eval_runtime": 105.1174,
      "eval_samples_per_second": 62.796,
      "eval_steps_per_second": 7.858,
      "step": 3600
    },
    {
      "epoch": 3.75,
      "learning_rate": 1.256490134994808e-05,
      "loss": 0.0035,
      "step": 3610
    },
    {
      "epoch": 3.76,
      "learning_rate": 1.2045690550363447e-05,
      "loss": 0.0501,
      "step": 3620
    },
    {
      "epoch": 3.77,
      "learning_rate": 1.1526479750778816e-05,
      "loss": 0.0398,
      "step": 3630
    },
    {
      "epoch": 3.78,
      "learning_rate": 1.1007268951194186e-05,
      "loss": 0.0042,
      "step": 3640
    },
    {
      "epoch": 3.79,
      "learning_rate": 1.0488058151609554e-05,
      "loss": 0.0044,
      "step": 3650
    },
    {
      "epoch": 3.8,
      "learning_rate": 9.968847352024923e-06,
      "loss": 0.087,
      "step": 3660
    },
    {
      "epoch": 3.81,
      "learning_rate": 9.44963655244029e-06,
      "loss": 0.005,
      "step": 3670
    },
    {
      "epoch": 3.82,
      "learning_rate": 8.93042575285566e-06,
      "loss": 0.0404,
      "step": 3680
    },
    {
      "epoch": 3.83,
      "learning_rate": 8.411214953271028e-06,
      "loss": 0.0401,
      "step": 3690
    },
    {
      "epoch": 3.84,
      "learning_rate": 7.892004153686398e-06,
      "loss": 0.004,
      "step": 3700
    },
    {
      "epoch": 3.84,
      "eval_accuracy": 0.9369792455688533,
      "eval_loss": 0.25872641801834106,
      "eval_runtime": 106.8562,
      "eval_samples_per_second": 61.775,
      "eval_steps_per_second": 7.73,
      "step": 3700
    },
    {
      "epoch": 3.85,
      "learning_rate": 7.372793354101766e-06,
      "loss": 0.0411,
      "step": 3710
    },
    {
      "epoch": 3.86,
      "learning_rate": 6.853582554517134e-06,
      "loss": 0.0689,
      "step": 3720
    },
    {
      "epoch": 3.87,
      "learning_rate": 6.3343717549325025e-06,
      "loss": 0.0428,
      "step": 3730
    },
    {
      "epoch": 3.88,
      "learning_rate": 5.815160955347872e-06,
      "loss": 0.0464,
      "step": 3740
    },
    {
      "epoch": 3.89,
      "learning_rate": 5.29595015576324e-06,
      "loss": 0.024,
      "step": 3750
    },
    {
      "epoch": 3.9,
      "learning_rate": 4.776739356178609e-06,
      "loss": 0.0732,
      "step": 3760
    },
    {
      "epoch": 3.91,
      "learning_rate": 4.257528556593978e-06,
      "loss": 0.0386,
      "step": 3770
    },
    {
      "epoch": 3.93,
      "learning_rate": 3.7383177570093455e-06,
      "loss": 0.0532,
      "step": 3780
    },
    {
      "epoch": 3.94,
      "learning_rate": 3.2191069574247146e-06,
      "loss": 0.0413,
      "step": 3790
    },
    {
      "epoch": 3.95,
      "learning_rate": 2.699896157840083e-06,
      "loss": 0.0389,
      "step": 3800
    },
    {
      "epoch": 3.95,
      "eval_accuracy": 0.9372822299651568,
      "eval_loss": 0.25349465012550354,
      "eval_runtime": 110.0172,
      "eval_samples_per_second": 60.0,
      "eval_steps_per_second": 7.508,
      "step": 3800
    },
    {
      "epoch": 3.96,
      "learning_rate": 2.1806853582554518e-06,
      "loss": 0.0383,
      "step": 3810
    },
    {
      "epoch": 3.97,
      "learning_rate": 1.6614745586708204e-06,
      "loss": 0.0169,
      "step": 3820
    },
    {
      "epoch": 3.98,
      "learning_rate": 1.142263759086189e-06,
      "loss": 0.0072,
      "step": 3830
    },
    {
      "epoch": 3.99,
      "learning_rate": 6.230529595015577e-07,
      "loss": 0.0054,
      "step": 3840
    },
    {
      "epoch": 4.0,
      "learning_rate": 1.0384215991692627e-07,
      "loss": 0.0044,
      "step": 3850
    },
    {
      "epoch": 4.0,
      "step": 3852,
      "total_flos": 4.774126496175784e+18,
      "train_loss": 0.1544622457234444,
      "train_runtime": 5879.4453,
      "train_samples_per_second": 10.479,
      "train_steps_per_second": 0.655
    }
  ],
  "logging_steps": 10,
  "max_steps": 3852,
  "num_train_epochs": 4,
  "save_steps": 100,
  "total_flos": 4.774126496175784e+18,
  "trial_name": null,
  "trial_params": null
}