{
  "best_metric": 0.6366316676139832,
  "best_model_checkpoint": "./exper3_mesum5/checkpoint-2800",
  "epoch": 8.0,
  "global_step": 3440,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "learning_rate": 0.0001994186046511628,
      "loss": 4.954,
      "step": 10
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0001988372093023256,
      "loss": 4.8641,
      "step": 20
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00019825581395348837,
      "loss": 4.6647,
      "step": 30
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00019767441860465116,
      "loss": 4.5255,
      "step": 40
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00019709302325581396,
      "loss": 4.4509,
      "step": 50
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00019651162790697676,
      "loss": 4.253,
      "step": 60
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00019593023255813952,
      "loss": 4.2104,
      "step": 70
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00019534883720930232,
      "loss": 4.0778,
      "step": 80
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00019476744186046511,
      "loss": 3.9537,
      "step": 90
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0001941860465116279,
      "loss": 3.895,
      "step": 100
    },
    {
      "epoch": 0.23,
      "eval_accuracy": 0.19349112426035503,
      "eval_loss": 3.8276302814483643,
      "eval_runtime": 19.0606,
      "eval_samples_per_second": 88.665,
      "eval_steps_per_second": 11.122,
      "step": 100
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0001936046511627907,
      "loss": 3.825,
      "step": 110
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0001930232558139535,
      "loss": 3.6952,
      "step": 120
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0001924418604651163,
      "loss": 3.4767,
      "step": 130
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0001918604651162791,
      "loss": 3.5417,
      "step": 140
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0001912790697674419,
      "loss": 3.4797,
      "step": 150
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00019069767441860466,
      "loss": 3.3749,
      "step": 160
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00019011627906976745,
      "loss": 3.4024,
      "step": 170
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00018953488372093025,
      "loss": 3.2742,
      "step": 180
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00018895348837209304,
      "loss": 3.3841,
      "step": 190
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00018837209302325584,
      "loss": 3.1174,
      "step": 200
    },
    {
      "epoch": 0.47,
      "eval_accuracy": 0.3106508875739645,
      "eval_loss": 3.1216797828674316,
      "eval_runtime": 21.0487,
      "eval_samples_per_second": 80.29,
      "eval_steps_per_second": 10.072,
      "step": 200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0001877906976744186,
      "loss": 3.012,
      "step": 210
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0001872093023255814,
      "loss": 3.0221,
      "step": 220
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0001866279069767442,
      "loss": 2.9971,
      "step": 230
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000186046511627907,
      "loss": 3.0478,
      "step": 240
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00018546511627906976,
      "loss": 2.9971,
      "step": 250
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00018488372093023256,
      "loss": 2.7655,
      "step": 260
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018430232558139535,
      "loss": 2.703,
      "step": 270
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00018372093023255815,
      "loss": 2.8271,
      "step": 280
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00018313953488372094,
      "loss": 2.6716,
      "step": 290
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001825581395348837,
      "loss": 2.6,
      "step": 300
    },
    {
      "epoch": 0.7,
      "eval_accuracy": 0.42071005917159765,
      "eval_loss": 2.5399255752563477,
      "eval_runtime": 20.8035,
      "eval_samples_per_second": 81.236,
      "eval_steps_per_second": 10.191,
      "step": 300
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001819767441860465,
      "loss": 2.5875,
      "step": 310
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001813953488372093,
      "loss": 2.6057,
      "step": 320
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00018081395348837212,
      "loss": 2.5459,
      "step": 330
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001802325581395349,
      "loss": 2.4955,
      "step": 340
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001796511627906977,
      "loss": 2.3718,
      "step": 350
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00017906976744186048,
      "loss": 2.3314,
      "step": 360
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00017848837209302328,
      "loss": 2.3855,
      "step": 370
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.00017790697674418605,
      "loss": 2.313,
      "step": 380
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.00017732558139534884,
      "loss": 2.1767,
      "step": 390
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.00017674418604651164,
      "loss": 2.256,
      "step": 400
    },
    {
      "epoch": 0.93,
      "eval_accuracy": 0.5159763313609468,
      "eval_loss": 2.176730155944824,
      "eval_runtime": 21.1447,
      "eval_samples_per_second": 79.925,
      "eval_steps_per_second": 10.026,
      "step": 400
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.00017616279069767443,
      "loss": 2.2881,
      "step": 410
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.00017558139534883723,
      "loss": 2.0996,
      "step": 420
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.000175,
      "loss": 2.0865,
      "step": 430
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0001744186046511628,
      "loss": 1.84,
      "step": 440
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0001738372093023256,
      "loss": 1.7042,
      "step": 450
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00017325581395348838,
      "loss": 1.6687,
      "step": 460
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00017267441860465118,
      "loss": 1.8394,
      "step": 470
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.00017209302325581395,
      "loss": 1.6962,
      "step": 480
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.00017151162790697674,
      "loss": 1.7329,
      "step": 490
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.00017093023255813954,
      "loss": 1.5441,
      "step": 500
    },
    {
      "epoch": 1.16,
      "eval_accuracy": 0.585207100591716,
      "eval_loss": 1.8085863590240479,
      "eval_runtime": 21.5882,
      "eval_samples_per_second": 78.284,
      "eval_steps_per_second": 9.82,
      "step": 500
    },
    {
      "epoch": 1.19,
      "learning_rate": 0.00017034883720930233,
      "loss": 1.6099,
      "step": 510
    },
    {
      "epoch": 1.21,
      "learning_rate": 0.0001697674418604651,
      "loss": 1.5833,
      "step": 520
    },
    {
      "epoch": 1.23,
      "learning_rate": 0.0001691860465116279,
      "loss": 1.5696,
      "step": 530
    },
    {
      "epoch": 1.26,
      "learning_rate": 0.00016860465116279072,
      "loss": 1.3757,
      "step": 540
    },
    {
      "epoch": 1.28,
      "learning_rate": 0.00016802325581395352,
      "loss": 1.4537,
      "step": 550
    },
    {
      "epoch": 1.3,
      "learning_rate": 0.00016744186046511629,
      "loss": 1.445,
      "step": 560
    },
    {
      "epoch": 1.33,
      "learning_rate": 0.00016686046511627908,
      "loss": 1.3975,
      "step": 570
    },
    {
      "epoch": 1.35,
      "learning_rate": 0.00016627906976744188,
      "loss": 1.4086,
      "step": 580
    },
    {
      "epoch": 1.37,
      "learning_rate": 0.00016569767441860467,
      "loss": 1.3367,
      "step": 590
    },
    {
      "epoch": 1.4,
      "learning_rate": 0.00016511627906976747,
      "loss": 1.3834,
      "step": 600
    },
    {
      "epoch": 1.4,
      "eval_accuracy": 0.6325443786982249,
      "eval_loss": 1.556496500968933,
      "eval_runtime": 21.486,
      "eval_samples_per_second": 78.656,
      "eval_steps_per_second": 9.867,
      "step": 600
    },
    {
      "epoch": 1.42,
      "learning_rate": 0.00016453488372093024,
      "loss": 1.2953,
      "step": 610
    },
    {
      "epoch": 1.44,
      "learning_rate": 0.00016395348837209303,
      "loss": 1.2843,
      "step": 620
    },
    {
      "epoch": 1.47,
      "learning_rate": 0.00016337209302325583,
      "loss": 1.1906,
      "step": 630
    },
    {
      "epoch": 1.49,
      "learning_rate": 0.00016279069767441862,
      "loss": 1.3458,
      "step": 640
    },
    {
      "epoch": 1.51,
      "learning_rate": 0.0001622093023255814,
      "loss": 1.1714,
      "step": 650
    },
    {
      "epoch": 1.53,
      "learning_rate": 0.00016162790697674419,
      "loss": 1.191,
      "step": 660
    },
    {
      "epoch": 1.56,
      "learning_rate": 0.00016104651162790698,
      "loss": 1.159,
      "step": 670
    },
    {
      "epoch": 1.58,
      "learning_rate": 0.00016046511627906978,
      "loss": 1.2594,
      "step": 680
    },
    {
      "epoch": 1.6,
      "learning_rate": 0.00015988372093023257,
      "loss": 1.1533,
      "step": 690
    },
    {
      "epoch": 1.63,
      "learning_rate": 0.00015930232558139534,
      "loss": 1.1995,
      "step": 700
    },
    {
      "epoch": 1.63,
      "eval_accuracy": 0.6763313609467455,
      "eval_loss": 1.3339420557022095,
      "eval_runtime": 19.734,
      "eval_samples_per_second": 85.639,
      "eval_steps_per_second": 10.743,
      "step": 700
    },
    {
      "epoch": 1.65,
      "learning_rate": 0.00015872093023255814,
      "loss": 1.0989,
      "step": 710
    },
    {
      "epoch": 1.67,
      "learning_rate": 0.00015813953488372093,
      "loss": 1.2864,
      "step": 720
    },
    {
      "epoch": 1.7,
      "learning_rate": 0.00015755813953488373,
      "loss": 1.2124,
      "step": 730
    },
    {
      "epoch": 1.72,
      "learning_rate": 0.00015697674418604652,
      "loss": 1.1752,
      "step": 740
    },
    {
      "epoch": 1.74,
      "learning_rate": 0.0001563953488372093,
      "loss": 1.2127,
      "step": 750
    },
    {
      "epoch": 1.77,
      "learning_rate": 0.0001558139534883721,
      "loss": 1.1634,
      "step": 760
    },
    {
      "epoch": 1.79,
      "learning_rate": 0.0001552325581395349,
      "loss": 1.0915,
      "step": 770
    },
    {
      "epoch": 1.81,
      "learning_rate": 0.00015465116279069768,
      "loss": 1.103,
      "step": 780
    },
    {
      "epoch": 1.84,
      "learning_rate": 0.00015406976744186047,
      "loss": 1.0951,
      "step": 790
    },
    {
      "epoch": 1.86,
      "learning_rate": 0.00015348837209302327,
      "loss": 1.0845,
      "step": 800
    },
    {
      "epoch": 1.86,
      "eval_accuracy": 0.6532544378698225,
      "eval_loss": 1.3299002647399902,
      "eval_runtime": 19.6334,
      "eval_samples_per_second": 86.078,
      "eval_steps_per_second": 10.798,
      "step": 800
    },
    {
      "epoch": 1.88,
      "learning_rate": 0.00015290697674418606,
      "loss": 1.03,
      "step": 810
    },
    {
      "epoch": 1.91,
      "learning_rate": 0.00015232558139534886,
      "loss": 0.9288,
      "step": 820
    },
    {
      "epoch": 1.93,
      "learning_rate": 0.00015174418604651163,
      "loss": 0.9963,
      "step": 830
    },
    {
      "epoch": 1.95,
      "learning_rate": 0.00015116279069767442,
      "loss": 1.0363,
      "step": 840
    },
    {
      "epoch": 1.98,
      "learning_rate": 0.00015058139534883722,
      "loss": 0.9671,
      "step": 850
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00015000000000000001,
      "loss": 0.8207,
      "step": 860
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0001494186046511628,
      "loss": 0.6051,
      "step": 870
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00014883720930232558,
      "loss": 0.6698,
      "step": 880
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00014825581395348837,
      "loss": 0.6969,
      "step": 890
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00014767441860465117,
      "loss": 0.6472,
      "step": 900
    },
    {
      "epoch": 2.09,
      "eval_accuracy": 0.7218934911242604,
      "eval_loss": 1.0679467916488647,
      "eval_runtime": 19.5979,
      "eval_samples_per_second": 86.234,
      "eval_steps_per_second": 10.818,
      "step": 900
    },
    {
      "epoch": 2.12,
      "learning_rate": 0.00014709302325581396,
      "loss": 0.6366,
      "step": 910
    },
    {
      "epoch": 2.14,
      "learning_rate": 0.00014651162790697673,
      "loss": 0.6353,
      "step": 920
    },
    {
      "epoch": 2.16,
      "learning_rate": 0.00014593023255813953,
      "loss": 0.4765,
      "step": 930
    },
    {
      "epoch": 2.19,
      "learning_rate": 0.00014534883720930232,
      "loss": 0.453,
      "step": 940
    },
    {
      "epoch": 2.21,
      "learning_rate": 0.00014476744186046512,
      "loss": 0.5234,
      "step": 950
    },
    {
      "epoch": 2.23,
      "learning_rate": 0.00014418604651162791,
      "loss": 0.5019,
      "step": 960
    },
    {
      "epoch": 2.26,
      "learning_rate": 0.0001436046511627907,
      "loss": 0.6719,
      "step": 970
    },
    {
      "epoch": 2.28,
      "learning_rate": 0.0001430232558139535,
      "loss": 0.5294,
      "step": 980
    },
    {
      "epoch": 2.3,
      "learning_rate": 0.0001424418604651163,
      "loss": 0.6135,
      "step": 990
    },
    {
      "epoch": 2.33,
      "learning_rate": 0.0001418604651162791,
      "loss": 0.5948,
      "step": 1000
    },
    {
      "epoch": 2.33,
      "eval_accuracy": 0.7124260355029586,
      "eval_loss": 1.0286362171173096,
      "eval_runtime": 19.4904,
      "eval_samples_per_second": 86.709,
      "eval_steps_per_second": 10.877,
      "step": 1000
    },
    {
      "epoch": 2.35,
      "learning_rate": 0.00014127906976744186,
      "loss": 0.6138,
      "step": 1010
    },
    {
      "epoch": 2.37,
      "learning_rate": 0.00014069767441860466,
      "loss": 0.6543,
      "step": 1020
    },
    {
      "epoch": 2.4,
      "learning_rate": 0.00014011627906976746,
      "loss": 0.5534,
      "step": 1030
    },
    {
      "epoch": 2.42,
      "learning_rate": 0.00013953488372093025,
      "loss": 0.6408,
      "step": 1040
    },
    {
      "epoch": 2.44,
      "learning_rate": 0.00013895348837209302,
      "loss": 0.4687,
      "step": 1050
    },
    {
      "epoch": 2.47,
      "learning_rate": 0.00013837209302325582,
      "loss": 0.4635,
      "step": 1060
    },
    {
      "epoch": 2.49,
      "learning_rate": 0.0001377906976744186,
      "loss": 0.4466,
      "step": 1070
    },
    {
      "epoch": 2.51,
      "learning_rate": 0.0001372093023255814,
      "loss": 0.4255,
      "step": 1080
    },
    {
      "epoch": 2.53,
      "learning_rate": 0.0001366279069767442,
      "loss": 0.5848,
      "step": 1090
    },
    {
      "epoch": 2.56,
      "learning_rate": 0.00013604651162790697,
      "loss": 0.5565,
      "step": 1100
    },
    {
      "epoch": 2.56,
      "eval_accuracy": 0.7284023668639054,
      "eval_loss": 0.9595437049865723,
      "eval_runtime": 19.5222,
      "eval_samples_per_second": 86.568,
      "eval_steps_per_second": 10.859,
      "step": 1100
    },
    {
      "epoch": 2.58,
      "learning_rate": 0.00013546511627906977,
      "loss": 0.4625,
      "step": 1110
    },
    {
      "epoch": 2.6,
      "learning_rate": 0.00013488372093023256,
      "loss": 0.7103,
      "step": 1120
    },
    {
      "epoch": 2.63,
      "learning_rate": 0.00013430232558139536,
      "loss": 0.5923,
      "step": 1130
    },
    {
      "epoch": 2.65,
      "learning_rate": 0.00013372093023255815,
      "loss": 0.4913,
      "step": 1140
    },
    {
      "epoch": 2.67,
      "learning_rate": 0.00013313953488372092,
      "loss": 0.4915,
      "step": 1150
    },
    {
      "epoch": 2.7,
      "learning_rate": 0.00013255813953488372,
      "loss": 0.401,
      "step": 1160
    },
    {
      "epoch": 2.72,
      "learning_rate": 0.0001319767441860465,
      "loss": 0.4169,
      "step": 1170
    },
    {
      "epoch": 2.74,
      "learning_rate": 0.0001313953488372093,
      "loss": 0.52,
      "step": 1180
    },
    {
      "epoch": 2.77,
      "learning_rate": 0.0001308139534883721,
      "loss": 0.4018,
      "step": 1190
    },
    {
      "epoch": 2.79,
      "learning_rate": 0.0001302325581395349,
      "loss": 0.4879,
      "step": 1200
    },
    {
      "epoch": 2.79,
      "eval_accuracy": 0.7420118343195267,
      "eval_loss": 0.8915188312530518,
      "eval_runtime": 19.4697,
      "eval_samples_per_second": 86.801,
      "eval_steps_per_second": 10.889,
      "step": 1200
    },
    {
      "epoch": 2.81,
      "learning_rate": 0.0001296511627906977,
      "loss": 0.527,
      "step": 1210
    },
    {
      "epoch": 2.84,
      "learning_rate": 0.0001290697674418605,
      "loss": 0.4114,
      "step": 1220
    },
    {
      "epoch": 2.86,
      "learning_rate": 0.00012848837209302326,
      "loss": 0.5728,
      "step": 1230
    },
    {
      "epoch": 2.88,
      "learning_rate": 0.00012790697674418605,
      "loss": 0.347,
      "step": 1240
    },
    {
      "epoch": 2.91,
      "learning_rate": 0.00012732558139534885,
      "loss": 0.3652,
      "step": 1250
    },
    {
      "epoch": 2.93,
      "learning_rate": 0.00012674418604651164,
      "loss": 0.5574,
      "step": 1260
    },
    {
      "epoch": 2.95,
      "learning_rate": 0.00012616279069767444,
      "loss": 0.4363,
      "step": 1270
    },
    {
      "epoch": 2.98,
      "learning_rate": 0.0001255813953488372,
      "loss": 0.4769,
      "step": 1280
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.000125,
      "loss": 0.5541,
      "step": 1290
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0001244186046511628,
      "loss": 0.2816,
      "step": 1300
    },
    {
      "epoch": 3.02,
      "eval_accuracy": 0.7763313609467456,
      "eval_loss": 0.8158556818962097,
      "eval_runtime": 19.4244,
      "eval_samples_per_second": 87.004,
      "eval_steps_per_second": 10.914,
      "step": 1300
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0001238372093023256,
      "loss": 0.2724,
      "step": 1310
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00012325581395348836,
      "loss": 0.234,
      "step": 1320
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00012267441860465116,
      "loss": 0.2116,
      "step": 1330
    },
    {
      "epoch": 3.12,
      "learning_rate": 0.00012209302325581395,
      "loss": 0.2968,
      "step": 1340
    },
    {
      "epoch": 3.14,
      "learning_rate": 0.00012151162790697675,
      "loss": 0.2539,
      "step": 1350
    },
    {
      "epoch": 3.16,
      "learning_rate": 0.00012093023255813953,
      "loss": 0.2837,
      "step": 1360
    },
    {
      "epoch": 3.19,
      "learning_rate": 0.00012034883720930233,
      "loss": 0.2571,
      "step": 1370
    },
    {
      "epoch": 3.21,
      "learning_rate": 0.00011976744186046511,
      "loss": 0.2502,
      "step": 1380
    },
    {
      "epoch": 3.23,
      "learning_rate": 0.0001191860465116279,
      "loss": 0.2629,
      "step": 1390
    },
    {
      "epoch": 3.26,
      "learning_rate": 0.00011860465116279071,
      "loss": 0.2412,
      "step": 1400
    },
    {
      "epoch": 3.26,
      "eval_accuracy": 0.7911242603550296,
      "eval_loss": 0.776643693447113,
      "eval_runtime": 19.7719,
      "eval_samples_per_second": 85.475,
      "eval_steps_per_second": 10.722,
      "step": 1400
    },
    {
      "epoch": 3.28,
      "learning_rate": 0.00011802325581395351,
      "loss": 0.2721,
      "step": 1410
    },
    {
      "epoch": 3.3,
      "learning_rate": 0.00011744186046511629,
      "loss": 0.1876,
      "step": 1420
    },
    {
      "epoch": 3.33,
      "learning_rate": 0.00011686046511627909,
      "loss": 0.2417,
      "step": 1430
    },
    {
      "epoch": 3.35,
      "learning_rate": 0.00011627906976744187,
      "loss": 0.2029,
      "step": 1440
    },
    {
      "epoch": 3.37,
      "learning_rate": 0.00011569767441860466,
      "loss": 0.1991,
      "step": 1450
    },
    {
      "epoch": 3.4,
      "learning_rate": 0.00011511627906976746,
      "loss": 0.2763,
      "step": 1460
    },
    {
      "epoch": 3.42,
      "learning_rate": 0.00011453488372093024,
      "loss": 0.1539,
      "step": 1470
    },
    {
      "epoch": 3.44,
      "learning_rate": 0.00011395348837209304,
      "loss": 0.2287,
      "step": 1480
    },
    {
      "epoch": 3.47,
      "learning_rate": 0.00011337209302325582,
      "loss": 0.2572,
      "step": 1490
    },
    {
      "epoch": 3.49,
      "learning_rate": 0.00011279069767441861,
      "loss": 0.2015,
      "step": 1500
    },
    {
      "epoch": 3.49,
      "eval_accuracy": 0.7828402366863906,
      "eval_loss": 0.784956157207489,
      "eval_runtime": 19.4284,
      "eval_samples_per_second": 86.986,
      "eval_steps_per_second": 10.912,
      "step": 1500
    },
    {
      "epoch": 3.51,
      "learning_rate": 0.0001122093023255814,
      "loss": 0.2263,
      "step": 1510
    },
    {
      "epoch": 3.53,
      "learning_rate": 0.00011162790697674419,
      "loss": 0.2843,
      "step": 1520
    },
    {
      "epoch": 3.56,
      "learning_rate": 0.00011104651162790699,
      "loss": 0.2371,
      "step": 1530
    },
    {
      "epoch": 3.58,
      "learning_rate": 0.00011046511627906977,
      "loss": 0.1717,
      "step": 1540
    },
    {
      "epoch": 3.6,
      "learning_rate": 0.00010988372093023256,
      "loss": 0.311,
      "step": 1550
    },
    {
      "epoch": 3.63,
      "learning_rate": 0.00010930232558139534,
      "loss": 0.2567,
      "step": 1560
    },
    {
      "epoch": 3.65,
      "learning_rate": 0.00010872093023255814,
      "loss": 0.1739,
      "step": 1570
    },
    {
      "epoch": 3.67,
      "learning_rate": 0.00010813953488372092,
      "loss": 0.2185,
      "step": 1580
    },
    {
      "epoch": 3.7,
      "learning_rate": 0.00010755813953488372,
      "loss": 0.2158,
      "step": 1590
    },
    {
      "epoch": 3.72,
      "learning_rate": 0.00010697674418604651,
      "loss": 0.274,
      "step": 1600
    },
    {
      "epoch": 3.72,
      "eval_accuracy": 0.7934911242603551,
      "eval_loss": 0.7361425757408142,
      "eval_runtime": 19.2623,
      "eval_samples_per_second": 87.736,
      "eval_steps_per_second": 11.006,
      "step": 1600
    },
    {
      "epoch": 3.74,
      "learning_rate": 0.0001063953488372093,
      "loss": 0.1164,
      "step": 1610
    },
    {
      "epoch": 3.77,
      "learning_rate": 0.0001058139534883721,
      "loss": 0.1515,
      "step": 1620
    },
    {
      "epoch": 3.79,
      "learning_rate": 0.0001052325581395349,
      "loss": 0.2399,
      "step": 1630
    },
    {
      "epoch": 3.81,
      "learning_rate": 0.00010465116279069768,
      "loss": 0.2516,
      "step": 1640
    },
    {
      "epoch": 3.84,
      "learning_rate": 0.00010406976744186048,
      "loss": 0.2054,
      "step": 1650
    },
    {
      "epoch": 3.86,
      "learning_rate": 0.00010348837209302327,
      "loss": 0.1758,
      "step": 1660
    },
    {
      "epoch": 3.88,
      "learning_rate": 0.00010290697674418605,
      "loss": 0.1903,
      "step": 1670
    },
    {
      "epoch": 3.91,
      "learning_rate": 0.00010232558139534885,
      "loss": 0.1922,
      "step": 1680
    },
    {
      "epoch": 3.93,
      "learning_rate": 0.00010174418604651163,
      "loss": 0.2019,
      "step": 1690
    },
    {
      "epoch": 3.95,
      "learning_rate": 0.00010116279069767443,
      "loss": 0.1244,
      "step": 1700
    },
    {
      "epoch": 3.95,
      "eval_accuracy": 0.7911242603550296,
      "eval_loss": 0.7299075126647949,
      "eval_runtime": 19.238,
      "eval_samples_per_second": 87.847,
      "eval_steps_per_second": 11.02,
      "step": 1700
    },
    {
      "epoch": 3.98,
      "learning_rate": 0.00010058139534883721,
      "loss": 0.2361,
      "step": 1710
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0001,
      "loss": 0.1389,
      "step": 1720
    },
    {
      "epoch": 4.02,
      "learning_rate": 9.94186046511628e-05,
      "loss": 0.0844,
      "step": 1730
    },
    {
      "epoch": 4.05,
      "learning_rate": 9.883720930232558e-05,
      "loss": 0.1463,
      "step": 1740
    },
    {
      "epoch": 4.07,
      "learning_rate": 9.825581395348838e-05,
      "loss": 0.09,
      "step": 1750
    },
    {
      "epoch": 4.09,
      "learning_rate": 9.767441860465116e-05,
      "loss": 0.1106,
      "step": 1760
    },
    {
      "epoch": 4.12,
      "learning_rate": 9.709302325581396e-05,
      "loss": 0.1387,
      "step": 1770
    },
    {
      "epoch": 4.14,
      "learning_rate": 9.651162790697675e-05,
      "loss": 0.1231,
      "step": 1780
    },
    {
      "epoch": 4.16,
      "learning_rate": 9.593023255813955e-05,
      "loss": 0.0734,
      "step": 1790
    },
    {
      "epoch": 4.19,
      "learning_rate": 9.534883720930233e-05,
      "loss": 0.0794,
      "step": 1800
    },
    {
      "epoch": 4.19,
      "eval_accuracy": 0.7846153846153846,
      "eval_loss": 0.7440704107284546,
      "eval_runtime": 19.1417,
      "eval_samples_per_second": 88.289,
      "eval_steps_per_second": 11.075,
      "step": 1800
    },
    {
      "epoch": 4.21,
      "learning_rate": 9.476744186046512e-05,
      "loss": 0.0885,
      "step": 1810
    },
    {
      "epoch": 4.23,
      "learning_rate": 9.418604651162792e-05,
      "loss": 0.0781,
      "step": 1820
    },
    {
      "epoch": 4.26,
      "learning_rate": 9.36046511627907e-05,
      "loss": 0.0842,
      "step": 1830
    },
    {
      "epoch": 4.28,
      "learning_rate": 9.30232558139535e-05,
      "loss": 0.0957,
      "step": 1840
    },
    {
      "epoch": 4.3,
      "learning_rate": 9.244186046511628e-05,
      "loss": 0.0561,
      "step": 1850
    },
    {
      "epoch": 4.33,
      "learning_rate": 9.186046511627907e-05,
      "loss": 0.0905,
      "step": 1860
    },
    {
      "epoch": 4.35,
      "learning_rate": 9.127906976744186e-05,
      "loss": 0.075,
      "step": 1870
    },
    {
      "epoch": 4.37,
      "learning_rate": 9.069767441860465e-05,
      "loss": 0.1165,
      "step": 1880
    },
    {
      "epoch": 4.4,
      "learning_rate": 9.011627906976745e-05,
      "loss": 0.0809,
      "step": 1890
    },
    {
      "epoch": 4.42,
      "learning_rate": 8.953488372093024e-05,
      "loss": 0.0915,
      "step": 1900
    },
    {
      "epoch": 4.42,
      "eval_accuracy": 0.7940828402366864,
      "eval_loss": 0.7614301443099976,
      "eval_runtime": 19.2799,
      "eval_samples_per_second": 87.656,
      "eval_steps_per_second": 10.996,
      "step": 1900
    },
    {
      "epoch": 4.44,
      "learning_rate": 8.895348837209302e-05,
      "loss": 0.1187,
      "step": 1910
    },
    {
      "epoch": 4.47,
      "learning_rate": 8.837209302325582e-05,
      "loss": 0.1082,
      "step": 1920
    },
    {
      "epoch": 4.49,
      "learning_rate": 8.779069767441861e-05,
      "loss": 0.0698,
      "step": 1930
    },
    {
      "epoch": 4.51,
      "learning_rate": 8.72093023255814e-05,
      "loss": 0.1073,
      "step": 1940
    },
    {
      "epoch": 4.53,
      "learning_rate": 8.662790697674419e-05,
      "loss": 0.0638,
      "step": 1950
    },
    {
      "epoch": 4.56,
      "learning_rate": 8.604651162790697e-05,
      "loss": 0.157,
      "step": 1960
    },
    {
      "epoch": 4.58,
      "learning_rate": 8.546511627906977e-05,
      "loss": 0.0641,
      "step": 1970
    },
    {
      "epoch": 4.6,
      "learning_rate": 8.488372093023255e-05,
      "loss": 0.0536,
      "step": 1980
    },
    {
      "epoch": 4.63,
      "learning_rate": 8.430232558139536e-05,
      "loss": 0.0921,
      "step": 1990
    },
    {
      "epoch": 4.65,
      "learning_rate": 8.372093023255814e-05,
      "loss": 0.0817,
      "step": 2000
    },
    {
      "epoch": 4.65,
      "eval_accuracy": 0.8011834319526627,
      "eval_loss": 0.7310301065444946,
      "eval_runtime": 19.2052,
      "eval_samples_per_second": 87.997,
      "eval_steps_per_second": 11.039,
      "step": 2000
    },
    {
      "epoch": 4.67,
      "learning_rate": 8.313953488372094e-05,
      "loss": 0.0876,
      "step": 2010
    },
    {
      "epoch": 4.7,
      "learning_rate": 8.255813953488373e-05,
      "loss": 0.0959,
      "step": 2020
    },
    {
      "epoch": 4.72,
      "learning_rate": 8.197674418604652e-05,
      "loss": 0.0945,
      "step": 2030
    },
    {
      "epoch": 4.74,
      "learning_rate": 8.139534883720931e-05,
      "loss": 0.0375,
      "step": 2040
    },
    {
      "epoch": 4.77,
      "learning_rate": 8.081395348837209e-05,
      "loss": 0.0877,
      "step": 2050
    },
    {
      "epoch": 4.79,
      "learning_rate": 8.023255813953489e-05,
      "loss": 0.053,
      "step": 2060
    },
    {
      "epoch": 4.81,
      "learning_rate": 7.965116279069767e-05,
      "loss": 0.1233,
      "step": 2070
    },
    {
      "epoch": 4.84,
      "learning_rate": 7.906976744186047e-05,
      "loss": 0.1089,
      "step": 2080
    },
    {
      "epoch": 4.86,
      "learning_rate": 7.848837209302326e-05,
      "loss": 0.1326,
      "step": 2090
    },
    {
      "epoch": 4.88,
      "learning_rate": 7.790697674418606e-05,
      "loss": 0.0561,
      "step": 2100
    },
    {
      "epoch": 4.88,
      "eval_accuracy": 0.806508875739645,
      "eval_loss": 0.722186267375946,
      "eval_runtime": 19.2931,
      "eval_samples_per_second": 87.596,
      "eval_steps_per_second": 10.988,
      "step": 2100
    },
    {
      "epoch": 4.91,
      "learning_rate": 7.732558139534884e-05,
      "loss": 0.1245,
      "step": 2110
    },
    {
      "epoch": 4.93,
      "learning_rate": 7.674418604651163e-05,
      "loss": 0.0414,
      "step": 2120
    },
    {
      "epoch": 4.95,
      "learning_rate": 7.616279069767443e-05,
      "loss": 0.0901,
      "step": 2130
    },
    {
      "epoch": 4.98,
      "learning_rate": 7.558139534883721e-05,
      "loss": 0.0751,
      "step": 2140
    },
    {
      "epoch": 5.0,
      "learning_rate": 7.500000000000001e-05,
      "loss": 0.1456,
      "step": 2150
    },
    {
      "epoch": 5.02,
      "learning_rate": 7.441860465116279e-05,
      "loss": 0.0245,
      "step": 2160
    },
    {
      "epoch": 5.05,
      "learning_rate": 7.383720930232558e-05,
      "loss": 0.0377,
      "step": 2170
    },
    {
      "epoch": 5.07,
      "learning_rate": 7.325581395348837e-05,
      "loss": 0.0435,
      "step": 2180
    },
    {
      "epoch": 5.09,
      "learning_rate": 7.267441860465116e-05,
      "loss": 0.0639,
      "step": 2190
    },
    {
      "epoch": 5.12,
      "learning_rate": 7.209302325581396e-05,
      "loss": 0.0165,
      "step": 2200
    },
    {
      "epoch": 5.12,
      "eval_accuracy": 0.8059171597633136,
      "eval_loss": 0.7515397667884827,
      "eval_runtime": 19.4193,
      "eval_samples_per_second": 87.027,
      "eval_steps_per_second": 10.917,
      "step": 2200
    },
    {
      "epoch": 5.14,
      "learning_rate": 7.151162790697675e-05,
      "loss": 0.0338,
      "step": 2210
    },
    {
      "epoch": 5.16,
      "learning_rate": 7.093023255813955e-05,
      "loss": 0.0586,
      "step": 2220
    },
    {
      "epoch": 5.19,
      "learning_rate": 7.034883720930233e-05,
      "loss": 0.0787,
      "step": 2230
    },
    {
      "epoch": 5.21,
      "learning_rate": 6.976744186046513e-05,
      "loss": 0.023,
      "step": 2240
    },
    {
      "epoch": 5.23,
      "learning_rate": 6.918604651162791e-05,
      "loss": 0.0681,
      "step": 2250
    },
    {
      "epoch": 5.26,
      "learning_rate": 6.86046511627907e-05,
      "loss": 0.0569,
      "step": 2260
    },
    {
      "epoch": 5.28,
      "learning_rate": 6.802325581395348e-05,
      "loss": 0.0206,
      "step": 2270
    },
    {
      "epoch": 5.3,
      "learning_rate": 6.744186046511628e-05,
      "loss": 0.0369,
      "step": 2280
    },
    {
      "epoch": 5.33,
      "learning_rate": 6.686046511627908e-05,
      "loss": 0.0526,
      "step": 2290
    },
    {
      "epoch": 5.35,
      "learning_rate": 6.627906976744186e-05,
      "loss": 0.0168,
      "step": 2300
    },
    {
      "epoch": 5.35,
      "eval_accuracy": 0.821301775147929,
      "eval_loss": 0.6687235832214355,
      "eval_runtime": 19.3189,
      "eval_samples_per_second": 87.479,
      "eval_steps_per_second": 10.974,
      "step": 2300
    },
    {
      "epoch": 5.37,
      "learning_rate": 6.569767441860465e-05,
      "loss": 0.0185,
      "step": 2310
    },
    {
      "epoch": 5.4,
      "learning_rate": 6.511627906976745e-05,
      "loss": 0.018,
      "step": 2320
    },
    {
      "epoch": 5.42,
      "learning_rate": 6.453488372093024e-05,
      "loss": 0.0503,
      "step": 2330
    },
    {
      "epoch": 5.44,
      "learning_rate": 6.395348837209303e-05,
      "loss": 0.0145,
      "step": 2340
    },
    {
      "epoch": 5.47,
      "learning_rate": 6.337209302325582e-05,
      "loss": 0.0139,
      "step": 2350
    },
    {
      "epoch": 5.49,
      "learning_rate": 6.27906976744186e-05,
      "loss": 0.0134,
      "step": 2360
    },
    {
      "epoch": 5.51,
      "learning_rate": 6.22093023255814e-05,
      "loss": 0.0168,
      "step": 2370
    },
    {
      "epoch": 5.53,
      "learning_rate": 6.162790697674418e-05,
      "loss": 0.0234,
      "step": 2380
    },
    {
      "epoch": 5.56,
      "learning_rate": 6.104651162790698e-05,
      "loss": 0.046,
      "step": 2390
    },
    {
      "epoch": 5.58,
      "learning_rate": 6.0465116279069765e-05,
      "loss": 0.0212,
      "step": 2400
    },
    {
      "epoch": 5.58,
      "eval_accuracy": 0.8248520710059172,
      "eval_loss": 0.6671048402786255,
      "eval_runtime": 19.6081,
      "eval_samples_per_second": 86.189,
      "eval_steps_per_second": 10.812,
      "step": 2400
    },
    {
      "epoch": 5.6,
      "learning_rate": 5.9883720930232554e-05,
      "loss": 0.0182,
      "step": 2410
    },
    {
      "epoch": 5.63,
      "learning_rate": 5.9302325581395356e-05,
      "loss": 0.0101,
      "step": 2420
    },
    {
      "epoch": 5.65,
      "learning_rate": 5.8720930232558145e-05,
      "loss": 0.0131,
      "step": 2430
    },
    {
      "epoch": 5.67,
      "learning_rate": 5.8139534883720933e-05,
      "loss": 0.0592,
      "step": 2440
    },
    {
      "epoch": 5.7,
      "learning_rate": 5.755813953488373e-05,
      "loss": 0.0317,
      "step": 2450
    },
    {
      "epoch": 5.72,
      "learning_rate": 5.697674418604652e-05,
      "loss": 0.0136,
      "step": 2460
    },
    {
      "epoch": 5.74,
      "learning_rate": 5.6395348837209306e-05,
      "loss": 0.024,
      "step": 2470
    },
    {
      "epoch": 5.77,
      "learning_rate": 5.5813953488372095e-05,
      "loss": 0.0233,
      "step": 2480
    },
    {
      "epoch": 5.79,
      "learning_rate": 5.5232558139534884e-05,
      "loss": 0.034,
      "step": 2490
    },
    {
      "epoch": 5.81,
      "learning_rate": 5.465116279069767e-05,
      "loss": 0.0389,
      "step": 2500
    },
    {
      "epoch": 5.81,
      "eval_accuracy": 0.827810650887574,
      "eval_loss": 0.6893125176429749,
      "eval_runtime": 19.5009,
      "eval_samples_per_second": 86.663,
      "eval_steps_per_second": 10.871,
      "step": 2500
    },
    {
      "epoch": 5.84,
      "learning_rate": 5.406976744186046e-05,
      "loss": 0.0227,
      "step": 2510
    },
    {
      "epoch": 5.86,
      "learning_rate": 5.348837209302326e-05,
      "loss": 0.0186,
      "step": 2520
    },
    {
      "epoch": 5.88,
      "learning_rate": 5.290697674418605e-05,
      "loss": 0.022,
      "step": 2530
    },
    {
      "epoch": 5.91,
      "learning_rate": 5.232558139534884e-05,
      "loss": 0.0373,
      "step": 2540
    },
    {
      "epoch": 5.93,
      "learning_rate": 5.1744186046511636e-05,
      "loss": 0.0144,
      "step": 2550
    },
    {
      "epoch": 5.95,
      "learning_rate": 5.1162790697674425e-05,
      "loss": 0.0112,
      "step": 2560
    },
    {
      "epoch": 5.98,
      "learning_rate": 5.0581395348837214e-05,
      "loss": 0.0331,
      "step": 2570
    },
    {
      "epoch": 6.0,
      "learning_rate": 5e-05,
      "loss": 0.0139,
      "step": 2580
    },
    {
      "epoch": 6.02,
      "learning_rate": 4.941860465116279e-05,
      "loss": 0.0142,
      "step": 2590
    },
    {
      "epoch": 6.05,
      "learning_rate": 4.883720930232558e-05,
      "loss": 0.0087,
      "step": 2600
    },
    {
      "epoch": 6.05,
      "eval_accuracy": 0.8260355029585799,
      "eval_loss": 0.6839348077774048,
      "eval_runtime": 19.5465,
      "eval_samples_per_second": 86.461,
      "eval_steps_per_second": 10.846,
      "step": 2600
    },
    {
      "epoch": 6.07,
      "learning_rate": 4.8255813953488375e-05,
      "loss": 0.0089,
      "step": 2610
    },
    {
      "epoch": 6.09,
      "learning_rate": 4.7674418604651164e-05,
      "loss": 0.0168,
      "step": 2620
    },
    {
      "epoch": 6.12,
      "learning_rate": 4.709302325581396e-05,
      "loss": 0.009,
      "step": 2630
    },
    {
      "epoch": 6.14,
      "learning_rate": 4.651162790697675e-05,
      "loss": 0.0079,
      "step": 2640
    },
    {
      "epoch": 6.16,
      "learning_rate": 4.593023255813954e-05,
      "loss": 0.0092,
      "step": 2650
    },
    {
      "epoch": 6.19,
      "learning_rate": 4.5348837209302326e-05,
      "loss": 0.01,
      "step": 2660
    },
    {
      "epoch": 6.21,
      "learning_rate": 4.476744186046512e-05,
      "loss": 0.0134,
      "step": 2670
    },
    {
      "epoch": 6.23,
      "learning_rate": 4.418604651162791e-05,
      "loss": 0.0265,
      "step": 2680
    },
    {
      "epoch": 6.26,
      "learning_rate": 4.36046511627907e-05,
      "loss": 0.0079,
      "step": 2690
    },
    {
      "epoch": 6.28,
      "learning_rate": 4.302325581395349e-05,
      "loss": 0.0087,
      "step": 2700
    },
    {
      "epoch": 6.28,
      "eval_accuracy": 0.8319526627218935,
      "eval_loss": 0.6412006616592407,
      "eval_runtime": 19.4572,
      "eval_samples_per_second": 86.857,
      "eval_steps_per_second": 10.896,
      "step": 2700
    },
    {
      "epoch": 6.3,
      "learning_rate": 4.2441860465116276e-05,
      "loss": 0.0242,
      "step": 2710
    },
    {
      "epoch": 6.33,
      "learning_rate": 4.186046511627907e-05,
      "loss": 0.0087,
      "step": 2720
    },
    {
      "epoch": 6.35,
      "learning_rate": 4.127906976744187e-05,
      "loss": 0.0097,
      "step": 2730
    },
    {
      "epoch": 6.37,
      "learning_rate": 4.0697674418604655e-05,
      "loss": 0.0073,
      "step": 2740
    },
    {
      "epoch": 6.4,
      "learning_rate": 4.0116279069767444e-05,
      "loss": 0.0077,
      "step": 2750
    },
    {
      "epoch": 6.42,
      "learning_rate": 3.953488372093023e-05,
      "loss": 0.0115,
      "step": 2760
    },
    {
      "epoch": 6.44,
      "learning_rate": 3.895348837209303e-05,
      "loss": 0.0076,
      "step": 2770
    },
    {
      "epoch": 6.47,
      "learning_rate": 3.837209302325582e-05,
      "loss": 0.0092,
      "step": 2780
    },
    {
      "epoch": 6.49,
      "learning_rate": 3.7790697674418606e-05,
      "loss": 0.0071,
      "step": 2790
    },
    {
      "epoch": 6.51,
      "learning_rate": 3.7209302325581394e-05,
      "loss": 0.0077,
      "step": 2800
    },
    {
      "epoch": 6.51,
      "eval_accuracy": 0.8366863905325443,
      "eval_loss": 0.6366316676139832,
      "eval_runtime": 19.6422,
      "eval_samples_per_second": 86.039,
      "eval_steps_per_second": 10.793,
      "step": 2800
    },
    {
      "epoch": 6.53,
      "learning_rate": 3.662790697674418e-05,
      "loss": 0.0081,
      "step": 2810
    },
    {
      "epoch": 6.56,
      "learning_rate": 3.604651162790698e-05,
      "loss": 0.0077,
      "step": 2820
    },
    {
      "epoch": 6.58,
      "learning_rate": 3.5465116279069774e-05,
      "loss": 0.0068,
      "step": 2830
    },
    {
      "epoch": 6.6,
      "learning_rate": 3.488372093023256e-05,
      "loss": 0.0069,
      "step": 2840
    },
    {
      "epoch": 6.63,
      "learning_rate": 3.430232558139535e-05,
      "loss": 0.0078,
      "step": 2850
    },
    {
      "epoch": 6.65,
      "learning_rate": 3.372093023255814e-05,
      "loss": 0.0069,
      "step": 2860
    },
    {
      "epoch": 6.67,
      "learning_rate": 3.313953488372093e-05,
      "loss": 0.0075,
      "step": 2870
    },
    {
      "epoch": 6.7,
      "learning_rate": 3.2558139534883724e-05,
      "loss": 0.0088,
      "step": 2880
    },
    {
      "epoch": 6.72,
      "learning_rate": 3.197674418604651e-05,
      "loss": 0.007,
      "step": 2890
    },
    {
      "epoch": 6.74,
      "learning_rate": 3.13953488372093e-05,
      "loss": 0.0065,
      "step": 2900
    },
    {
      "epoch": 6.74,
      "eval_accuracy": 0.8272189349112427,
      "eval_loss": 0.6696515679359436,
      "eval_runtime": 19.4791,
      "eval_samples_per_second": 86.76,
      "eval_steps_per_second": 10.883,
      "step": 2900
    },
    {
      "epoch": 6.77,
      "learning_rate": 3.081395348837209e-05,
      "loss": 0.0276,
      "step": 2910
    },
    {
      "epoch": 6.79,
      "learning_rate": 3.0232558139534883e-05,
      "loss": 0.0064,
      "step": 2920
    },
    {
      "epoch": 6.81,
      "learning_rate": 2.9651162790697678e-05,
      "loss": 0.0064,
      "step": 2930
    },
    {
      "epoch": 6.84,
      "learning_rate": 2.9069767441860467e-05,
      "loss": 0.0271,
      "step": 2940
    },
    {
      "epoch": 6.86,
      "learning_rate": 2.848837209302326e-05,
      "loss": 0.0062,
      "step": 2950
    },
    {
      "epoch": 6.88,
      "learning_rate": 2.7906976744186048e-05,
      "loss": 0.0073,
      "step": 2960
    },
    {
      "epoch": 6.91,
      "learning_rate": 2.7325581395348836e-05,
      "loss": 0.0074,
      "step": 2970
    },
    {
      "epoch": 6.93,
      "learning_rate": 2.674418604651163e-05,
      "loss": 0.0068,
      "step": 2980
    },
    {
      "epoch": 6.95,
      "learning_rate": 2.616279069767442e-05,
      "loss": 0.007,
      "step": 2990
    },
    {
      "epoch": 6.98,
      "learning_rate": 2.5581395348837212e-05,
      "loss": 0.0061,
      "step": 3000
    },
    {
      "epoch": 6.98,
      "eval_accuracy": 0.8349112426035503,
      "eval_loss": 0.6509989500045776,
      "eval_runtime": 19.534,
      "eval_samples_per_second": 86.516,
      "eval_steps_per_second": 10.853,
      "step": 3000
    },
    {
      "epoch": 7.0,
      "learning_rate": 2.5e-05,
      "loss": 0.0411,
      "step": 3010
    },
    {
      "epoch": 7.02,
      "learning_rate": 2.441860465116279e-05,
      "loss": 0.006,
      "step": 3020
    },
    {
      "epoch": 7.05,
      "learning_rate": 2.3837209302325582e-05,
      "loss": 0.0066,
      "step": 3030
    },
    {
      "epoch": 7.07,
      "learning_rate": 2.3255813953488374e-05,
      "loss": 0.0058,
      "step": 3040
    },
    {
      "epoch": 7.09,
      "learning_rate": 2.2674418604651163e-05,
      "loss": 0.0064,
      "step": 3050
    },
    {
      "epoch": 7.12,
      "learning_rate": 2.2093023255813955e-05,
      "loss": 0.006,
      "step": 3060
    },
    {
      "epoch": 7.14,
      "learning_rate": 2.1511627906976744e-05,
      "loss": 0.0064,
      "step": 3070
    },
    {
      "epoch": 7.16,
      "learning_rate": 2.0930232558139536e-05,
      "loss": 0.0063,
      "step": 3080
    },
    {
      "epoch": 7.19,
      "learning_rate": 2.0348837209302328e-05,
      "loss": 0.0243,
      "step": 3090
    },
    {
      "epoch": 7.21,
      "learning_rate": 1.9767441860465116e-05,
      "loss": 0.0185,
      "step": 3100
    },
    {
      "epoch": 7.21,
      "eval_accuracy": 0.8366863905325443,
      "eval_loss": 0.6451619267463684,
      "eval_runtime": 19.7485,
      "eval_samples_per_second": 85.576,
      "eval_steps_per_second": 10.735,
      "step": 3100
    },
    {
      "epoch": 7.23,
      "learning_rate": 1.918604651162791e-05,
      "loss": 0.0111,
      "step": 3110
    },
    {
      "epoch": 7.26,
      "learning_rate": 1.8604651162790697e-05,
      "loss": 0.0063,
      "step": 3120
    },
    {
      "epoch": 7.28,
      "learning_rate": 1.802325581395349e-05,
      "loss": 0.007,
      "step": 3130
    },
    {
      "epoch": 7.3,
      "learning_rate": 1.744186046511628e-05,
      "loss": 0.006,
      "step": 3140
    },
    {
      "epoch": 7.33,
      "learning_rate": 1.686046511627907e-05,
      "loss": 0.0064,
      "step": 3150
    },
    {
      "epoch": 7.35,
      "learning_rate": 1.6279069767441862e-05,
      "loss": 0.006,
      "step": 3160
    },
    {
      "epoch": 7.37,
      "learning_rate": 1.569767441860465e-05,
      "loss": 0.0059,
      "step": 3170
    },
    {
      "epoch": 7.4,
      "learning_rate": 1.5116279069767441e-05,
      "loss": 0.0053,
      "step": 3180
    },
    {
      "epoch": 7.42,
      "learning_rate": 1.4534883720930233e-05,
      "loss": 0.006,
      "step": 3190
    },
    {
      "epoch": 7.44,
      "learning_rate": 1.3953488372093024e-05,
      "loss": 0.0059,
      "step": 3200
    },
    {
      "epoch": 7.44,
      "eval_accuracy": 0.8378698224852071,
      "eval_loss": 0.6426283717155457,
      "eval_runtime": 19.368,
      "eval_samples_per_second": 87.257,
      "eval_steps_per_second": 10.946,
      "step": 3200
    },
    {
      "epoch": 7.47,
      "learning_rate": 1.3372093023255814e-05,
      "loss": 0.0059,
      "step": 3210
    },
    {
      "epoch": 7.49,
      "learning_rate": 1.2790697674418606e-05,
      "loss": 0.0057,
      "step": 3220
    },
    {
      "epoch": 7.51,
      "learning_rate": 1.2209302325581395e-05,
      "loss": 0.006,
      "step": 3230
    },
    {
      "epoch": 7.53,
      "learning_rate": 1.1627906976744187e-05,
      "loss": 0.0053,
      "step": 3240
    },
    {
      "epoch": 7.56,
      "learning_rate": 1.1046511627906977e-05,
      "loss": 0.0053,
      "step": 3250
    },
    {
      "epoch": 7.58,
      "learning_rate": 1.0465116279069768e-05,
      "loss": 0.0062,
      "step": 3260
    },
    {
      "epoch": 7.6,
      "learning_rate": 9.883720930232558e-06,
      "loss": 0.0061,
      "step": 3270
    },
    {
      "epoch": 7.63,
      "learning_rate": 9.302325581395349e-06,
      "loss": 0.0061,
      "step": 3280
    },
    {
      "epoch": 7.65,
      "learning_rate": 8.72093023255814e-06,
      "loss": 0.0053,
      "step": 3290
    },
    {
      "epoch": 7.67,
      "learning_rate": 8.139534883720931e-06,
      "loss": 0.0062,
      "step": 3300
    },
    {
      "epoch": 7.67,
      "eval_accuracy": 0.8378698224852071,
      "eval_loss": 0.6398439407348633,
      "eval_runtime": 20.4188,
      "eval_samples_per_second": 82.767,
      "eval_steps_per_second": 10.383,
      "step": 3300
    },
    {
      "epoch": 7.7,
      "learning_rate": 7.558139534883721e-06,
      "loss": 0.0056,
      "step": 3310
    },
    {
      "epoch": 7.72,
      "learning_rate": 6.976744186046512e-06,
      "loss": 0.0058,
      "step": 3320
    },
    {
      "epoch": 7.74,
      "learning_rate": 6.395348837209303e-06,
      "loss": 0.0058,
      "step": 3330
    },
    {
      "epoch": 7.77,
      "learning_rate": 5.8139534883720935e-06,
      "loss": 0.0058,
      "step": 3340
    },
    {
      "epoch": 7.79,
      "learning_rate": 5.232558139534884e-06,
      "loss": 0.0062,
      "step": 3350
    },
    {
      "epoch": 7.81,
      "learning_rate": 4.651162790697674e-06,
      "loss": 0.0055,
      "step": 3360
    },
    {
      "epoch": 7.84,
      "learning_rate": 4.0697674418604655e-06,
      "loss": 0.0053,
      "step": 3370
    },
    {
      "epoch": 7.86,
      "learning_rate": 3.488372093023256e-06,
      "loss": 0.0057,
      "step": 3380
    },
    {
      "epoch": 7.88,
      "learning_rate": 2.9069767441860468e-06,
      "loss": 0.006,
      "step": 3390
    },
    {
      "epoch": 7.91,
      "learning_rate": 2.325581395348837e-06,
      "loss": 0.0315,
      "step": 3400
    },
    {
      "epoch": 7.91,
      "eval_accuracy": 0.8384615384615385,
      "eval_loss": 0.6396650075912476,
      "eval_runtime": 19.5469,
      "eval_samples_per_second": 86.459,
      "eval_steps_per_second": 10.846,
      "step": 3400
    },
    {
      "epoch": 7.93,
      "learning_rate": 1.744186046511628e-06,
      "loss": 0.0074,
      "step": 3410
    },
    {
      "epoch": 7.95,
      "learning_rate": 1.1627906976744186e-06,
      "loss": 0.0057,
      "step": 3420
    },
    {
      "epoch": 7.98,
      "learning_rate": 5.813953488372093e-07,
      "loss": 0.0058,
      "step": 3430
    },
    {
      "epoch": 8.0,
      "learning_rate": 0.0,
      "loss": 0.0057,
      "step": 3440
    },
    {
      "epoch": 8.0,
      "step": 3440,
      "total_flos": 4.2707785173722726e+18,
      "train_loss": 0.6754590564342432,
      "train_runtime": 2008.5618,
      "train_samples_per_second": 27.403,
      "train_steps_per_second": 1.713
    }
  ],
  "max_steps": 3440,
  "num_train_epochs": 8,
  "total_flos": 4.2707785173722726e+18,
  "trial_name": null,
  "trial_params": null
}