{
  "best_metric": 0.35198959708213806,
  "best_model_checkpoint": "./exper_batch_32_e8/checkpoint-2500",
  "epoch": 8.0,
  "global_step": 2560,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.03,
      "learning_rate": 0.00019921875000000001,
      "loss": 4.9133,
      "step": 10
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00019843750000000002,
      "loss": 4.6764,
      "step": 20
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00019765625,
      "loss": 4.467,
      "step": 30
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000196875,
      "loss": 4.2923,
      "step": 40
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00019609375,
      "loss": 4.1396,
      "step": 50
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0001953125,
      "loss": 3.9854,
      "step": 60
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00019453125000000002,
      "loss": 3.7673,
      "step": 70
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00019375000000000002,
      "loss": 3.6646,
      "step": 80
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00019296875,
      "loss": 3.5048,
      "step": 90
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0001921875,
      "loss": 3.3787,
      "step": 100
    },
    {
      "epoch": 0.31,
      "eval_accuracy": 0.35659213569776405,
      "eval_loss": 3.309976100921631,
      "eval_runtime": 43.5696,
      "eval_samples_per_second": 59.537,
      "eval_steps_per_second": 7.459,
      "step": 100
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00019140625,
      "loss": 3.2397,
      "step": 110
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000190625,
      "loss": 3.1479,
      "step": 120
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00018984375000000002,
      "loss": 3.0083,
      "step": 130
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00018906250000000002,
      "loss": 2.9685,
      "step": 140
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00018828125,
      "loss": 2.7569,
      "step": 150
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0001875,
      "loss": 2.7539,
      "step": 160
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00018671875,
      "loss": 2.6816,
      "step": 170
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0001859375,
      "loss": 2.4719,
      "step": 180
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00018515625000000002,
      "loss": 2.4216,
      "step": 190
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000184375,
      "loss": 2.3975,
      "step": 200
    },
    {
      "epoch": 0.62,
      "eval_accuracy": 0.5717039321511179,
      "eval_loss": 2.3196070194244385,
      "eval_runtime": 43.0702,
      "eval_samples_per_second": 60.227,
      "eval_steps_per_second": 7.546,
      "step": 200
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00018359375,
      "loss": 2.3173,
      "step": 210
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001828125,
      "loss": 2.1784,
      "step": 220
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00018203125,
      "loss": 2.1889,
      "step": 230
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00018125000000000001,
      "loss": 2.077,
      "step": 240
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00018046875000000002,
      "loss": 2.0334,
      "step": 250
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001796875,
      "loss": 1.8344,
      "step": 260
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00017890625,
      "loss": 1.8512,
      "step": 270
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.000178125,
      "loss": 1.7556,
      "step": 280
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.00017734375,
      "loss": 1.6702,
      "step": 290
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.00017656250000000002,
      "loss": 1.5578,
      "step": 300
    },
    {
      "epoch": 0.94,
      "eval_accuracy": 0.646106399383192,
      "eval_loss": 1.676370620727539,
      "eval_runtime": 42.4496,
      "eval_samples_per_second": 61.108,
      "eval_steps_per_second": 7.656,
      "step": 300
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.00017578125000000002,
      "loss": 1.687,
      "step": 310
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.000175,
      "loss": 1.6243,
      "step": 320
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00017421875,
      "loss": 1.398,
      "step": 330
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0001734375,
      "loss": 1.188,
      "step": 340
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00017265625,
      "loss": 1.2636,
      "step": 350
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.00017187500000000002,
      "loss": 1.2485,
      "step": 360
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.00017109375,
      "loss": 1.1226,
      "step": 370
    },
    {
      "epoch": 1.19,
      "learning_rate": 0.0001703125,
      "loss": 1.0828,
      "step": 380
    },
    {
      "epoch": 1.22,
      "learning_rate": 0.00016953125,
      "loss": 1.0664,
      "step": 390
    },
    {
      "epoch": 1.25,
      "learning_rate": 0.00016875,
      "loss": 1.0291,
      "step": 400
    },
    {
      "epoch": 1.25,
      "eval_accuracy": 0.7463377023901311,
      "eval_loss": 1.1713167428970337,
      "eval_runtime": 43.3111,
      "eval_samples_per_second": 59.892,
      "eval_steps_per_second": 7.504,
      "step": 400
    },
    {
      "epoch": 1.28,
      "learning_rate": 0.00016796875000000001,
      "loss": 0.9675,
      "step": 410
    },
    {
      "epoch": 1.31,
      "learning_rate": 0.00016718750000000002,
      "loss": 1.015,
      "step": 420
    },
    {
      "epoch": 1.34,
      "learning_rate": 0.00016640625,
      "loss": 0.9497,
      "step": 430
    },
    {
      "epoch": 1.38,
      "learning_rate": 0.000165625,
      "loss": 1.0306,
      "step": 440
    },
    {
      "epoch": 1.41,
      "learning_rate": 0.00016484375,
      "loss": 0.8879,
      "step": 450
    },
    {
      "epoch": 1.44,
      "learning_rate": 0.0001640625,
      "loss": 0.8771,
      "step": 460
    },
    {
      "epoch": 1.47,
      "learning_rate": 0.00016328125000000001,
      "loss": 0.7988,
      "step": 470
    },
    {
      "epoch": 1.5,
      "learning_rate": 0.00016250000000000002,
      "loss": 0.8302,
      "step": 480
    },
    {
      "epoch": 1.53,
      "learning_rate": 0.00016171875,
      "loss": 0.7392,
      "step": 490
    },
    {
      "epoch": 1.56,
      "learning_rate": 0.0001609375,
      "loss": 0.8185,
      "step": 500
    },
    {
      "epoch": 1.56,
      "eval_accuracy": 0.7952968388589051,
      "eval_loss": 0.9292389154434204,
      "eval_runtime": 43.1031,
      "eval_samples_per_second": 60.181,
      "eval_steps_per_second": 7.54,
      "step": 500
    },
    {
      "epoch": 1.59,
      "learning_rate": 0.00016015625,
      "loss": 0.7336,
      "step": 510
    },
    {
      "epoch": 1.62,
      "learning_rate": 0.000159375,
      "loss": 0.753,
      "step": 520
    },
    {
      "epoch": 1.66,
      "learning_rate": 0.00015859375000000002,
      "loss": 0.753,
      "step": 530
    },
    {
      "epoch": 1.69,
      "learning_rate": 0.00015781250000000002,
      "loss": 0.7302,
      "step": 540
    },
    {
      "epoch": 1.72,
      "learning_rate": 0.00015703125,
      "loss": 0.7229,
      "step": 550
    },
    {
      "epoch": 1.75,
      "learning_rate": 0.00015625,
      "loss": 0.7445,
      "step": 560
    },
    {
      "epoch": 1.78,
      "learning_rate": 0.00015546875,
      "loss": 0.684,
      "step": 570
    },
    {
      "epoch": 1.81,
      "learning_rate": 0.0001546875,
      "loss": 0.7035,
      "step": 580
    },
    {
      "epoch": 1.84,
      "learning_rate": 0.00015390625000000002,
      "loss": 0.717,
      "step": 590
    },
    {
      "epoch": 1.88,
      "learning_rate": 0.000153125,
      "loss": 0.6181,
      "step": 600
    },
    {
      "epoch": 1.88,
      "eval_accuracy": 0.8168851195065536,
      "eval_loss": 0.7732186317443848,
      "eval_runtime": 42.7674,
      "eval_samples_per_second": 60.654,
      "eval_steps_per_second": 7.599,
      "step": 600
    },
    {
      "epoch": 1.91,
      "learning_rate": 0.00015234375,
      "loss": 0.6409,
      "step": 610
    },
    {
      "epoch": 1.94,
      "learning_rate": 0.0001515625,
      "loss": 0.5785,
      "step": 620
    },
    {
      "epoch": 1.97,
      "learning_rate": 0.00015078125,
      "loss": 0.5554,
      "step": 630
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00015000000000000001,
      "loss": 0.5513,
      "step": 640
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00014921875000000002,
      "loss": 0.3915,
      "step": 650
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0001484375,
      "loss": 0.41,
      "step": 660
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00014765625,
      "loss": 0.413,
      "step": 670
    },
    {
      "epoch": 2.12,
      "learning_rate": 0.000146875,
      "loss": 0.4791,
      "step": 680
    },
    {
      "epoch": 2.16,
      "learning_rate": 0.00014609375,
      "loss": 0.4581,
      "step": 690
    },
    {
      "epoch": 2.19,
      "learning_rate": 0.00014531250000000002,
      "loss": 0.3873,
      "step": 700
    },
    {
      "epoch": 2.19,
      "eval_accuracy": 0.8276792598303778,
      "eval_loss": 0.6876775622367859,
      "eval_runtime": 42.6013,
      "eval_samples_per_second": 60.89,
      "eval_steps_per_second": 7.629,
      "step": 700
    },
    {
      "epoch": 2.22,
      "learning_rate": 0.00014453125000000002,
      "loss": 0.4062,
      "step": 710
    },
    {
      "epoch": 2.25,
      "learning_rate": 0.00014375,
      "loss": 0.3818,
      "step": 720
    },
    {
      "epoch": 2.28,
      "learning_rate": 0.00014296875,
      "loss": 0.3834,
      "step": 730
    },
    {
      "epoch": 2.31,
      "learning_rate": 0.0001421875,
      "loss": 0.3881,
      "step": 740
    },
    {
      "epoch": 2.34,
      "learning_rate": 0.00014140625,
      "loss": 0.4153,
      "step": 750
    },
    {
      "epoch": 2.38,
      "learning_rate": 0.00014062500000000002,
      "loss": 0.3353,
      "step": 760
    },
    {
      "epoch": 2.41,
      "learning_rate": 0.00013984375,
      "loss": 0.3368,
      "step": 770
    },
    {
      "epoch": 2.44,
      "learning_rate": 0.0001390625,
      "loss": 0.3217,
      "step": 780
    },
    {
      "epoch": 2.47,
      "learning_rate": 0.00013828125,
      "loss": 0.2734,
      "step": 790
    },
    {
      "epoch": 2.5,
      "learning_rate": 0.0001375,
      "loss": 0.2979,
      "step": 800
    },
    {
      "epoch": 2.5,
      "eval_accuracy": 0.8404009252120278,
      "eval_loss": 0.625044584274292,
      "eval_runtime": 43.4522,
      "eval_samples_per_second": 59.698,
      "eval_steps_per_second": 7.479,
      "step": 800
    },
    {
      "epoch": 2.53,
      "learning_rate": 0.00013671875,
      "loss": 0.4192,
      "step": 810
    },
    {
      "epoch": 2.56,
      "learning_rate": 0.00013593750000000002,
      "loss": 0.3249,
      "step": 820
    },
    {
      "epoch": 2.59,
      "learning_rate": 0.00013515625,
      "loss": 0.329,
      "step": 830
    },
    {
      "epoch": 2.62,
      "learning_rate": 0.000134375,
      "loss": 0.2907,
      "step": 840
    },
    {
      "epoch": 2.66,
      "learning_rate": 0.00013359375,
      "loss": 0.3428,
      "step": 850
    },
    {
      "epoch": 2.69,
      "learning_rate": 0.0001328125,
      "loss": 0.3506,
      "step": 860
    },
    {
      "epoch": 2.72,
      "learning_rate": 0.00013203125000000001,
      "loss": 0.311,
      "step": 870
    },
    {
      "epoch": 2.75,
      "learning_rate": 0.00013125000000000002,
      "loss": 0.2628,
      "step": 880
    },
    {
      "epoch": 2.78,
      "learning_rate": 0.00013046875,
      "loss": 0.2148,
      "step": 890
    },
    {
      "epoch": 2.81,
      "learning_rate": 0.0001296875,
      "loss": 0.2967,
      "step": 900
    },
    {
      "epoch": 2.81,
      "eval_accuracy": 0.8365458750963762,
      "eval_loss": 0.6150881052017212,
      "eval_runtime": 42.6639,
      "eval_samples_per_second": 60.801,
      "eval_steps_per_second": 7.618,
      "step": 900
    },
    {
      "epoch": 2.84,
      "learning_rate": 0.00012890625,
      "loss": 0.3521,
      "step": 910
    },
    {
      "epoch": 2.88,
      "learning_rate": 0.000128125,
      "loss": 0.3112,
      "step": 920
    },
    {
      "epoch": 2.91,
      "learning_rate": 0.00012734375000000002,
      "loss": 0.2827,
      "step": 930
    },
    {
      "epoch": 2.94,
      "learning_rate": 0.0001265625,
      "loss": 0.2714,
      "step": 940
    },
    {
      "epoch": 2.97,
      "learning_rate": 0.00012578125,
      "loss": 0.2865,
      "step": 950
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.000125,
      "loss": 0.2594,
      "step": 960
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00012421875,
      "loss": 0.2627,
      "step": 970
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0001234375,
      "loss": 0.1978,
      "step": 980
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00012265625000000002,
      "loss": 0.1228,
      "step": 990
    },
    {
      "epoch": 3.12,
      "learning_rate": 0.00012187500000000001,
      "loss": 0.1874,
      "step": 1000
    },
    {
      "epoch": 3.12,
      "eval_accuracy": 0.8608326908249807,
      "eval_loss": 0.5401089787483215,
      "eval_runtime": 43.4743,
      "eval_samples_per_second": 59.668,
      "eval_steps_per_second": 7.476,
      "step": 1000
    },
    {
      "epoch": 3.16,
      "learning_rate": 0.00012109375,
      "loss": 0.1935,
      "step": 1010
    },
    {
      "epoch": 3.19,
      "learning_rate": 0.0001203125,
      "loss": 0.1469,
      "step": 1020
    },
    {
      "epoch": 3.22,
      "learning_rate": 0.00011953125000000001,
      "loss": 0.1394,
      "step": 1030
    },
    {
      "epoch": 3.25,
      "learning_rate": 0.00011875,
      "loss": 0.1353,
      "step": 1040
    },
    {
      "epoch": 3.28,
      "learning_rate": 0.00011796875,
      "loss": 0.1193,
      "step": 1050
    },
    {
      "epoch": 3.31,
      "learning_rate": 0.00011718750000000001,
      "loss": 0.1413,
      "step": 1060
    },
    {
      "epoch": 3.34,
      "learning_rate": 0.00011640625,
      "loss": 0.1398,
      "step": 1070
    },
    {
      "epoch": 3.38,
      "learning_rate": 0.000115625,
      "loss": 0.1682,
      "step": 1080
    },
    {
      "epoch": 3.41,
      "learning_rate": 0.00011484375000000001,
      "loss": 0.1991,
      "step": 1090
    },
    {
      "epoch": 3.44,
      "learning_rate": 0.0001140625,
      "loss": 0.2232,
      "step": 1100
    },
    {
      "epoch": 3.44,
      "eval_accuracy": 0.8712413261372398,
      "eval_loss": 0.5031955242156982,
      "eval_runtime": 43.2709,
      "eval_samples_per_second": 59.948,
      "eval_steps_per_second": 7.511,
      "step": 1100
    },
    {
      "epoch": 3.47,
      "learning_rate": 0.00011328125,
      "loss": 0.1051,
      "step": 1110
    },
    {
      "epoch": 3.5,
      "learning_rate": 0.00011250000000000001,
      "loss": 0.1798,
      "step": 1120
    },
    {
      "epoch": 3.53,
      "learning_rate": 0.00011171875,
      "loss": 0.1333,
      "step": 1130
    },
    {
      "epoch": 3.56,
      "learning_rate": 0.0001109375,
      "loss": 0.1734,
      "step": 1140
    },
    {
      "epoch": 3.59,
      "learning_rate": 0.00011015625000000001,
      "loss": 0.1398,
      "step": 1150
    },
    {
      "epoch": 3.62,
      "learning_rate": 0.000109375,
      "loss": 0.1554,
      "step": 1160
    },
    {
      "epoch": 3.66,
      "learning_rate": 0.00010859375000000001,
      "loss": 0.1485,
      "step": 1170
    },
    {
      "epoch": 3.69,
      "learning_rate": 0.00010781250000000001,
      "loss": 0.1506,
      "step": 1180
    },
    {
      "epoch": 3.72,
      "learning_rate": 0.00010703125,
      "loss": 0.1467,
      "step": 1190
    },
    {
      "epoch": 3.75,
      "learning_rate": 0.00010625000000000001,
      "loss": 0.1109,
      "step": 1200
    },
    {
      "epoch": 3.75,
      "eval_accuracy": 0.8774094063222821,
      "eval_loss": 0.4635218679904938,
      "eval_runtime": 42.8981,
      "eval_samples_per_second": 60.469,
      "eval_steps_per_second": 7.576,
      "step": 1200
    },
    {
      "epoch": 3.78,
      "learning_rate": 0.00010546875,
      "loss": 0.1112,
      "step": 1210
    },
    {
      "epoch": 3.81,
      "learning_rate": 0.0001046875,
      "loss": 0.1135,
      "step": 1220
    },
    {
      "epoch": 3.84,
      "learning_rate": 0.00010390625000000001,
      "loss": 0.1418,
      "step": 1230
    },
    {
      "epoch": 3.88,
      "learning_rate": 0.000103125,
      "loss": 0.1416,
      "step": 1240
    },
    {
      "epoch": 3.91,
      "learning_rate": 0.00010234375,
      "loss": 0.1266,
      "step": 1250
    },
    {
      "epoch": 3.94,
      "learning_rate": 0.00010156250000000001,
      "loss": 0.1228,
      "step": 1260
    },
    {
      "epoch": 3.97,
      "learning_rate": 0.00010078125,
      "loss": 0.1496,
      "step": 1270
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0001,
      "loss": 0.1067,
      "step": 1280
    },
    {
      "epoch": 4.03,
      "learning_rate": 9.921875000000001e-05,
      "loss": 0.1101,
      "step": 1290
    },
    {
      "epoch": 4.06,
      "learning_rate": 9.84375e-05,
      "loss": 0.0539,
      "step": 1300
    },
    {
      "epoch": 4.06,
      "eval_accuracy": 0.8843484965304549,
      "eval_loss": 0.44948554039001465,
      "eval_runtime": 42.6302,
      "eval_samples_per_second": 60.849,
      "eval_steps_per_second": 7.624,
      "step": 1300
    },
    {
      "epoch": 4.09,
      "learning_rate": 9.765625e-05,
      "loss": 0.0912,
      "step": 1310
    },
    {
      "epoch": 4.12,
      "learning_rate": 9.687500000000001e-05,
      "loss": 0.0828,
      "step": 1320
    },
    {
      "epoch": 4.16,
      "learning_rate": 9.609375e-05,
      "loss": 0.0896,
      "step": 1330
    },
    {
      "epoch": 4.19,
      "learning_rate": 9.53125e-05,
      "loss": 0.0528,
      "step": 1340
    },
    {
      "epoch": 4.22,
      "learning_rate": 9.453125000000001e-05,
      "loss": 0.0768,
      "step": 1350
    },
    {
      "epoch": 4.25,
      "learning_rate": 9.375e-05,
      "loss": 0.0762,
      "step": 1360
    },
    {
      "epoch": 4.28,
      "learning_rate": 9.296875e-05,
      "loss": 0.052,
      "step": 1370
    },
    {
      "epoch": 4.31,
      "learning_rate": 9.21875e-05,
      "loss": 0.0577,
      "step": 1380
    },
    {
      "epoch": 4.34,
      "learning_rate": 9.140625e-05,
      "loss": 0.0849,
      "step": 1390
    },
    {
      "epoch": 4.38,
      "learning_rate": 9.062500000000001e-05,
      "loss": 0.0668,
      "step": 1400
    },
    {
      "epoch": 4.38,
      "eval_accuracy": 0.8951426368542791,
      "eval_loss": 0.42733636498451233,
      "eval_runtime": 42.8966,
      "eval_samples_per_second": 60.471,
      "eval_steps_per_second": 7.576,
      "step": 1400
    },
    {
      "epoch": 4.41,
      "learning_rate": 8.984375e-05,
      "loss": 0.1054,
      "step": 1410
    },
    {
      "epoch": 4.44,
      "learning_rate": 8.90625e-05,
      "loss": 0.0728,
      "step": 1420
    },
    {
      "epoch": 4.47,
      "learning_rate": 8.828125000000001e-05,
      "loss": 0.072,
      "step": 1430
    },
    {
      "epoch": 4.5,
      "learning_rate": 8.75e-05,
      "loss": 0.0641,
      "step": 1440
    },
    {
      "epoch": 4.53,
      "learning_rate": 8.671875e-05,
      "loss": 0.0908,
      "step": 1450
    },
    {
      "epoch": 4.56,
      "learning_rate": 8.593750000000001e-05,
      "loss": 0.1079,
      "step": 1460
    },
    {
      "epoch": 4.59,
      "learning_rate": 8.515625e-05,
      "loss": 0.0623,
      "step": 1470
    },
    {
      "epoch": 4.62,
      "learning_rate": 8.4375e-05,
      "loss": 0.0922,
      "step": 1480
    },
    {
      "epoch": 4.66,
      "learning_rate": 8.359375000000001e-05,
      "loss": 0.0853,
      "step": 1490
    },
    {
      "epoch": 4.69,
      "learning_rate": 8.28125e-05,
      "loss": 0.0567,
      "step": 1500
    },
    {
      "epoch": 4.69,
      "eval_accuracy": 0.8866615265998458,
      "eval_loss": 0.44270050525665283,
      "eval_runtime": 42.9006,
      "eval_samples_per_second": 60.465,
      "eval_steps_per_second": 7.576,
      "step": 1500
    },
    {
      "epoch": 4.72,
      "learning_rate": 8.203125e-05,
      "loss": 0.0499,
      "step": 1510
    },
    {
      "epoch": 4.75,
      "learning_rate": 8.125000000000001e-05,
      "loss": 0.0497,
      "step": 1520
    },
    {
      "epoch": 4.78,
      "learning_rate": 8.046875e-05,
      "loss": 0.0325,
      "step": 1530
    },
    {
      "epoch": 4.81,
      "learning_rate": 7.96875e-05,
      "loss": 0.0338,
      "step": 1540
    },
    {
      "epoch": 4.84,
      "learning_rate": 7.890625000000001e-05,
      "loss": 0.0281,
      "step": 1550
    },
    {
      "epoch": 4.88,
      "learning_rate": 7.8125e-05,
      "loss": 0.0597,
      "step": 1560
    },
    {
      "epoch": 4.91,
      "learning_rate": 7.734375e-05,
      "loss": 0.0904,
      "step": 1570
    },
    {
      "epoch": 4.94,
      "learning_rate": 7.65625e-05,
      "loss": 0.0521,
      "step": 1580
    },
    {
      "epoch": 4.97,
      "learning_rate": 7.578125e-05,
      "loss": 0.0496,
      "step": 1590
    },
    {
      "epoch": 5.0,
      "learning_rate": 7.500000000000001e-05,
      "loss": 0.0285,
      "step": 1600
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8955281418658443,
      "eval_loss": 0.40922868251800537,
      "eval_runtime": 42.4136,
      "eval_samples_per_second": 61.16,
      "eval_steps_per_second": 7.663,
      "step": 1600
    },
    {
      "epoch": 5.03,
      "learning_rate": 7.421875e-05,
      "loss": 0.0629,
      "step": 1610
    },
    {
      "epoch": 5.06,
      "learning_rate": 7.34375e-05,
      "loss": 0.0391,
      "step": 1620
    },
    {
      "epoch": 5.09,
      "learning_rate": 7.265625000000001e-05,
      "loss": 0.057,
      "step": 1630
    },
    {
      "epoch": 5.12,
      "learning_rate": 7.1875e-05,
      "loss": 0.0413,
      "step": 1640
    },
    {
      "epoch": 5.16,
      "learning_rate": 7.109375e-05,
      "loss": 0.0342,
      "step": 1650
    },
    {
      "epoch": 5.19,
      "learning_rate": 7.031250000000001e-05,
      "loss": 0.0272,
      "step": 1660
    },
    {
      "epoch": 5.22,
      "learning_rate": 6.953125e-05,
      "loss": 0.02,
      "step": 1670
    },
    {
      "epoch": 5.25,
      "learning_rate": 6.875e-05,
      "loss": 0.0237,
      "step": 1680
    },
    {
      "epoch": 5.28,
      "learning_rate": 6.796875000000001e-05,
      "loss": 0.0232,
      "step": 1690
    },
    {
      "epoch": 5.31,
      "learning_rate": 6.71875e-05,
      "loss": 0.0473,
      "step": 1700
    },
    {
      "epoch": 5.31,
      "eval_accuracy": 0.9070932922127988,
      "eval_loss": 0.37195590138435364,
      "eval_runtime": 43.0497,
      "eval_samples_per_second": 60.256,
      "eval_steps_per_second": 7.549,
      "step": 1700
    },
    {
      "epoch": 5.34,
      "learning_rate": 6.640625e-05,
      "loss": 0.0569,
      "step": 1710
    },
    {
      "epoch": 5.38,
      "learning_rate": 6.562500000000001e-05,
      "loss": 0.0424,
      "step": 1720
    },
    {
      "epoch": 5.41,
      "learning_rate": 6.484375e-05,
      "loss": 0.0457,
      "step": 1730
    },
    {
      "epoch": 5.44,
      "learning_rate": 6.40625e-05,
      "loss": 0.0217,
      "step": 1740
    },
    {
      "epoch": 5.47,
      "learning_rate": 6.328125e-05,
      "loss": 0.0141,
      "step": 1750
    },
    {
      "epoch": 5.5,
      "learning_rate": 6.25e-05,
      "loss": 0.0153,
      "step": 1760
    },
    {
      "epoch": 5.53,
      "learning_rate": 6.171875e-05,
      "loss": 0.0341,
      "step": 1770
    },
    {
      "epoch": 5.56,
      "learning_rate": 6.0937500000000004e-05,
      "loss": 0.0159,
      "step": 1780
    },
    {
      "epoch": 5.59,
      "learning_rate": 6.015625e-05,
      "loss": 0.0249,
      "step": 1790
    },
    {
      "epoch": 5.62,
      "learning_rate": 5.9375e-05,
      "loss": 0.0225,
      "step": 1800
    },
    {
      "epoch": 5.62,
      "eval_accuracy": 0.9063222821896685,
      "eval_loss": 0.3691478669643402,
      "eval_runtime": 43.044,
      "eval_samples_per_second": 60.264,
      "eval_steps_per_second": 7.55,
      "step": 1800
    },
    {
      "epoch": 5.66,
      "learning_rate": 5.8593750000000005e-05,
      "loss": 0.0151,
      "step": 1810
    },
    {
      "epoch": 5.69,
      "learning_rate": 5.78125e-05,
      "loss": 0.0157,
      "step": 1820
    },
    {
      "epoch": 5.72,
      "learning_rate": 5.703125e-05,
      "loss": 0.0184,
      "step": 1830
    },
    {
      "epoch": 5.75,
      "learning_rate": 5.6250000000000005e-05,
      "loss": 0.0123,
      "step": 1840
    },
    {
      "epoch": 5.78,
      "learning_rate": 5.546875e-05,
      "loss": 0.0141,
      "step": 1850
    },
    {
      "epoch": 5.81,
      "learning_rate": 5.46875e-05,
      "loss": 0.0155,
      "step": 1860
    },
    {
      "epoch": 5.84,
      "learning_rate": 5.3906250000000006e-05,
      "loss": 0.0474,
      "step": 1870
    },
    {
      "epoch": 5.88,
      "learning_rate": 5.3125000000000004e-05,
      "loss": 0.0138,
      "step": 1880
    },
    {
      "epoch": 5.91,
      "learning_rate": 5.234375e-05,
      "loss": 0.0305,
      "step": 1890
    },
    {
      "epoch": 5.94,
      "learning_rate": 5.15625e-05,
      "loss": 0.0196,
      "step": 1900
    },
    {
      "epoch": 5.94,
      "eval_accuracy": 0.9047802621434079,
      "eval_loss": 0.3775177597999573,
      "eval_runtime": 43.4223,
      "eval_samples_per_second": 59.739,
      "eval_steps_per_second": 7.485,
      "step": 1900
    },
    {
      "epoch": 5.97,
      "learning_rate": 5.0781250000000004e-05,
      "loss": 0.0178,
      "step": 1910
    },
    {
      "epoch": 6.0,
      "learning_rate": 5e-05,
      "loss": 0.0161,
      "step": 1920
    },
    {
      "epoch": 6.03,
      "learning_rate": 4.921875e-05,
      "loss": 0.012,
      "step": 1930
    },
    {
      "epoch": 6.06,
      "learning_rate": 4.8437500000000005e-05,
      "loss": 0.0101,
      "step": 1940
    },
    {
      "epoch": 6.09,
      "learning_rate": 4.765625e-05,
      "loss": 0.0213,
      "step": 1950
    },
    {
      "epoch": 6.12,
      "learning_rate": 4.6875e-05,
      "loss": 0.0256,
      "step": 1960
    },
    {
      "epoch": 6.16,
      "learning_rate": 4.609375e-05,
      "loss": 0.0155,
      "step": 1970
    },
    {
      "epoch": 6.19,
      "learning_rate": 4.5312500000000004e-05,
      "loss": 0.0141,
      "step": 1980
    },
    {
      "epoch": 6.22,
      "learning_rate": 4.453125e-05,
      "loss": 0.0109,
      "step": 1990
    },
    {
      "epoch": 6.25,
      "learning_rate": 4.375e-05,
      "loss": 0.0173,
      "step": 2000
    },
    {
      "epoch": 6.25,
      "eval_accuracy": 0.9040092521202776,
      "eval_loss": 0.36413344740867615,
      "eval_runtime": 42.5316,
      "eval_samples_per_second": 60.99,
      "eval_steps_per_second": 7.641,
      "step": 2000
    },
    {
      "epoch": 6.28,
      "learning_rate": 4.2968750000000004e-05,
      "loss": 0.0093,
      "step": 2010
    },
    {
      "epoch": 6.31,
      "learning_rate": 4.21875e-05,
      "loss": 0.0101,
      "step": 2020
    },
    {
      "epoch": 6.34,
      "learning_rate": 4.140625e-05,
      "loss": 0.0118,
      "step": 2030
    },
    {
      "epoch": 6.38,
      "learning_rate": 4.0625000000000005e-05,
      "loss": 0.01,
      "step": 2040
    },
    {
      "epoch": 6.41,
      "learning_rate": 3.984375e-05,
      "loss": 0.0089,
      "step": 2050
    },
    {
      "epoch": 6.44,
      "learning_rate": 3.90625e-05,
      "loss": 0.01,
      "step": 2060
    },
    {
      "epoch": 6.47,
      "learning_rate": 3.828125e-05,
      "loss": 0.0098,
      "step": 2070
    },
    {
      "epoch": 6.5,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.0084,
      "step": 2080
    },
    {
      "epoch": 6.53,
      "learning_rate": 3.671875e-05,
      "loss": 0.0099,
      "step": 2090
    },
    {
      "epoch": 6.56,
      "learning_rate": 3.59375e-05,
      "loss": 0.0092,
      "step": 2100
    },
    {
      "epoch": 6.56,
      "eval_accuracy": 0.9090208172706246,
      "eval_loss": 0.3551329970359802,
      "eval_runtime": 42.9682,
      "eval_samples_per_second": 60.37,
      "eval_steps_per_second": 7.564,
      "step": 2100
    },
    {
      "epoch": 6.59,
      "learning_rate": 3.5156250000000004e-05,
      "loss": 0.0098,
      "step": 2110
    },
    {
      "epoch": 6.62,
      "learning_rate": 3.4375e-05,
      "loss": 0.0217,
      "step": 2120
    },
    {
      "epoch": 6.66,
      "learning_rate": 3.359375e-05,
      "loss": 0.0084,
      "step": 2130
    },
    {
      "epoch": 6.69,
      "learning_rate": 3.2812500000000005e-05,
      "loss": 0.009,
      "step": 2140
    },
    {
      "epoch": 6.72,
      "learning_rate": 3.203125e-05,
      "loss": 0.0135,
      "step": 2150
    },
    {
      "epoch": 6.75,
      "learning_rate": 3.125e-05,
      "loss": 0.0085,
      "step": 2160
    },
    {
      "epoch": 6.78,
      "learning_rate": 3.0468750000000002e-05,
      "loss": 0.0089,
      "step": 2170
    },
    {
      "epoch": 6.81,
      "learning_rate": 2.96875e-05,
      "loss": 0.0087,
      "step": 2180
    },
    {
      "epoch": 6.84,
      "learning_rate": 2.890625e-05,
      "loss": 0.0093,
      "step": 2190
    },
    {
      "epoch": 6.88,
      "learning_rate": 2.8125000000000003e-05,
      "loss": 0.008,
      "step": 2200
    },
    {
      "epoch": 6.88,
      "eval_accuracy": 0.9124903623747108,
      "eval_loss": 0.3591383993625641,
      "eval_runtime": 42.7423,
      "eval_samples_per_second": 60.689,
      "eval_steps_per_second": 7.604,
      "step": 2200
    },
    {
      "epoch": 6.91,
      "learning_rate": 2.734375e-05,
      "loss": 0.009,
      "step": 2210
    },
    {
      "epoch": 6.94,
      "learning_rate": 2.6562500000000002e-05,
      "loss": 0.0086,
      "step": 2220
    },
    {
      "epoch": 6.97,
      "learning_rate": 2.578125e-05,
      "loss": 0.0161,
      "step": 2230
    },
    {
      "epoch": 7.0,
      "learning_rate": 2.5e-05,
      "loss": 0.0222,
      "step": 2240
    },
    {
      "epoch": 7.03,
      "learning_rate": 2.4218750000000003e-05,
      "loss": 0.0092,
      "step": 2250
    },
    {
      "epoch": 7.06,
      "learning_rate": 2.34375e-05,
      "loss": 0.0096,
      "step": 2260
    },
    {
      "epoch": 7.09,
      "learning_rate": 2.2656250000000002e-05,
      "loss": 0.0083,
      "step": 2270
    },
    {
      "epoch": 7.12,
      "learning_rate": 2.1875e-05,
      "loss": 0.0076,
      "step": 2280
    },
    {
      "epoch": 7.16,
      "learning_rate": 2.109375e-05,
      "loss": 0.0082,
      "step": 2290
    },
    {
      "epoch": 7.19,
      "learning_rate": 2.0312500000000002e-05,
      "loss": 0.0072,
      "step": 2300
    },
    {
      "epoch": 7.19,
      "eval_accuracy": 0.9121048573631457,
      "eval_loss": 0.3541559875011444,
      "eval_runtime": 42.8947,
      "eval_samples_per_second": 60.474,
      "eval_steps_per_second": 7.577,
      "step": 2300
    },
    {
      "epoch": 7.22,
      "learning_rate": 1.953125e-05,
      "loss": 0.0075,
      "step": 2310
    },
    {
      "epoch": 7.25,
      "learning_rate": 1.8750000000000002e-05,
      "loss": 0.0077,
      "step": 2320
    },
    {
      "epoch": 7.28,
      "learning_rate": 1.796875e-05,
      "loss": 0.0077,
      "step": 2330
    },
    {
      "epoch": 7.31,
      "learning_rate": 1.71875e-05,
      "loss": 0.008,
      "step": 2340
    },
    {
      "epoch": 7.34,
      "learning_rate": 1.6406250000000002e-05,
      "loss": 0.0156,
      "step": 2350
    },
    {
      "epoch": 7.38,
      "learning_rate": 1.5625e-05,
      "loss": 0.0079,
      "step": 2360
    },
    {
      "epoch": 7.41,
      "learning_rate": 1.484375e-05,
      "loss": 0.0082,
      "step": 2370
    },
    {
      "epoch": 7.44,
      "learning_rate": 1.4062500000000001e-05,
      "loss": 0.007,
      "step": 2380
    },
    {
      "epoch": 7.47,
      "learning_rate": 1.3281250000000001e-05,
      "loss": 0.0073,
      "step": 2390
    },
    {
      "epoch": 7.5,
      "learning_rate": 1.25e-05,
      "loss": 0.007,
      "step": 2400
    },
    {
      "epoch": 7.5,
      "eval_accuracy": 0.9105628373168851,
      "eval_loss": 0.3532123565673828,
      "eval_runtime": 43.0876,
      "eval_samples_per_second": 60.203,
      "eval_steps_per_second": 7.543,
      "step": 2400
    },
    {
      "epoch": 7.53,
      "learning_rate": 1.171875e-05,
      "loss": 0.0079,
      "step": 2410
    },
    {
      "epoch": 7.56,
      "learning_rate": 1.09375e-05,
      "loss": 0.0092,
      "step": 2420
    },
    {
      "epoch": 7.59,
      "learning_rate": 1.0156250000000001e-05,
      "loss": 0.0073,
      "step": 2430
    },
    {
      "epoch": 7.62,
      "learning_rate": 9.375000000000001e-06,
      "loss": 0.0077,
      "step": 2440
    },
    {
      "epoch": 7.66,
      "learning_rate": 8.59375e-06,
      "loss": 0.0077,
      "step": 2450
    },
    {
      "epoch": 7.69,
      "learning_rate": 7.8125e-06,
      "loss": 0.007,
      "step": 2460
    },
    {
      "epoch": 7.72,
      "learning_rate": 7.031250000000001e-06,
      "loss": 0.0082,
      "step": 2470
    },
    {
      "epoch": 7.75,
      "learning_rate": 6.25e-06,
      "loss": 0.0073,
      "step": 2480
    },
    {
      "epoch": 7.78,
      "learning_rate": 5.46875e-06,
      "loss": 0.007,
      "step": 2490
    },
    {
      "epoch": 7.81,
      "learning_rate": 4.6875000000000004e-06,
      "loss": 0.007,
      "step": 2500
    },
    {
      "epoch": 7.81,
      "eval_accuracy": 0.9113338473400154,
      "eval_loss": 0.35198959708213806,
      "eval_runtime": 42.6761,
      "eval_samples_per_second": 60.783,
      "eval_steps_per_second": 7.615,
      "step": 2500
    },
    {
      "epoch": 7.84,
      "learning_rate": 3.90625e-06,
      "loss": 0.0069,
      "step": 2510
    },
    {
      "epoch": 7.88,
      "learning_rate": 3.125e-06,
      "loss": 0.0077,
      "step": 2520
    },
    {
      "epoch": 7.91,
      "learning_rate": 2.3437500000000002e-06,
      "loss": 0.0065,
      "step": 2530
    },
    {
      "epoch": 7.94,
      "learning_rate": 1.5625e-06,
      "loss": 0.007,
      "step": 2540
    },
    {
      "epoch": 7.97,
      "learning_rate": 7.8125e-07,
      "loss": 0.0082,
      "step": 2550
    },
    {
      "epoch": 8.0,
      "learning_rate": 0.0,
      "loss": 0.0089,
      "step": 2560
    },
    {
      "epoch": 8.0,
      "step": 2560,
      "total_flos": 6.337884979995771e+18,
      "train_loss": 0.5420162634982262,
      "train_runtime": 3330.2524,
      "train_samples_per_second": 24.527,
      "train_steps_per_second": 0.769
    }
  ],
  "max_steps": 2560,
  "num_train_epochs": 8,
  "total_flos": 6.337884979995771e+18,
  "trial_name": null,
  "trial_params": null
}