{
  "best_metric": 0.3908935785293579,
  "best_model_checkpoint": "./exper_batch_32_e4/checkpoint-1200",
  "epoch": 4.0,
  "global_step": 1280,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.03,
      "learning_rate": 0.00019843750000000002,
      "loss": 4.9283,
      "step": 10
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000196875,
      "loss": 4.7121,
      "step": 20
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0001953125,
      "loss": 4.4878,
      "step": 30
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00019375000000000002,
      "loss": 4.3163,
      "step": 40
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0001921875,
      "loss": 4.1709,
      "step": 50
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000190625,
      "loss": 4.0397,
      "step": 60
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00018906250000000002,
      "loss": 3.8221,
      "step": 70
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0001875,
      "loss": 3.7326,
      "step": 80
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0001859375,
      "loss": 3.541,
      "step": 90
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000184375,
      "loss": 3.4295,
      "step": 100
    },
    {
      "epoch": 0.31,
      "eval_accuracy": 0.28373168851195063,
      "eval_loss": 3.4026525020599365,
      "eval_runtime": 43.01,
      "eval_samples_per_second": 60.312,
      "eval_steps_per_second": 7.556,
      "step": 100
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0001828125,
      "loss": 3.2807,
      "step": 110
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00018125000000000001,
      "loss": 3.2503,
      "step": 120
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0001796875,
      "loss": 3.0753,
      "step": 130
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000178125,
      "loss": 3.0426,
      "step": 140
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00017656250000000002,
      "loss": 2.8475,
      "step": 150
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000175,
      "loss": 2.8298,
      "step": 160
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0001734375,
      "loss": 2.7558,
      "step": 170
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00017187500000000002,
      "loss": 2.5849,
      "step": 180
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0001703125,
      "loss": 2.5335,
      "step": 190
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00016875,
      "loss": 2.5035,
      "step": 200
    },
    {
      "epoch": 0.62,
      "eval_accuracy": 0.5246723207401697,
      "eval_loss": 2.433885097503662,
      "eval_runtime": 43.0144,
      "eval_samples_per_second": 60.305,
      "eval_steps_per_second": 7.556,
      "step": 200
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016718750000000002,
      "loss": 2.4416,
      "step": 210
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000165625,
      "loss": 2.2758,
      "step": 220
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001640625,
      "loss": 2.2845,
      "step": 230
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00016250000000000002,
      "loss": 2.2053,
      "step": 240
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001609375,
      "loss": 2.165,
      "step": 250
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000159375,
      "loss": 1.9497,
      "step": 260
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00015781250000000002,
      "loss": 1.9581,
      "step": 270
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.00015625,
      "loss": 1.8753,
      "step": 280
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0001546875,
      "loss": 1.7976,
      "step": 290
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.000153125,
      "loss": 1.6542,
      "step": 300
    },
    {
      "epoch": 0.94,
      "eval_accuracy": 0.6387818041634541,
      "eval_loss": 1.7689646482467651,
      "eval_runtime": 42.9608,
      "eval_samples_per_second": 60.381,
      "eval_steps_per_second": 7.565,
      "step": 300
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0001515625,
      "loss": 1.7968,
      "step": 310
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00015000000000000001,
      "loss": 1.84,
      "step": 320
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0001484375,
      "loss": 1.5215,
      "step": 330
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000146875,
      "loss": 1.3148,
      "step": 340
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00014531250000000002,
      "loss": 1.3781,
      "step": 350
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.00014375,
      "loss": 1.3547,
      "step": 360
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0001421875,
      "loss": 1.2466,
      "step": 370
    },
    {
      "epoch": 1.19,
      "learning_rate": 0.00014062500000000002,
      "loss": 1.1866,
      "step": 380
    },
    {
      "epoch": 1.22,
      "learning_rate": 0.0001390625,
      "loss": 1.187,
      "step": 390
    },
    {
      "epoch": 1.25,
      "learning_rate": 0.0001375,
      "loss": 1.1589,
      "step": 400
    },
    {
      "epoch": 1.25,
      "eval_accuracy": 0.7459521973785659,
      "eval_loss": 1.3106426000595093,
      "eval_runtime": 43.0294,
      "eval_samples_per_second": 60.284,
      "eval_steps_per_second": 7.553,
      "step": 400
    },
    {
      "epoch": 1.28,
      "learning_rate": 0.00013593750000000002,
      "loss": 1.0901,
      "step": 410
    },
    {
      "epoch": 1.31,
      "learning_rate": 0.000134375,
      "loss": 1.0998,
      "step": 420
    },
    {
      "epoch": 1.34,
      "learning_rate": 0.0001328125,
      "loss": 1.028,
      "step": 430
    },
    {
      "epoch": 1.38,
      "learning_rate": 0.00013125000000000002,
      "loss": 1.1139,
      "step": 440
    },
    {
      "epoch": 1.41,
      "learning_rate": 0.0001296875,
      "loss": 0.9936,
      "step": 450
    },
    {
      "epoch": 1.44,
      "learning_rate": 0.000128125,
      "loss": 0.9392,
      "step": 460
    },
    {
      "epoch": 1.47,
      "learning_rate": 0.0001265625,
      "loss": 0.8785,
      "step": 470
    },
    {
      "epoch": 1.5,
      "learning_rate": 0.000125,
      "loss": 0.8851,
      "step": 480
    },
    {
      "epoch": 1.53,
      "learning_rate": 0.0001234375,
      "loss": 0.8302,
      "step": 490
    },
    {
      "epoch": 1.56,
      "learning_rate": 0.00012187500000000001,
      "loss": 0.9363,
      "step": 500
    },
    {
      "epoch": 1.56,
      "eval_accuracy": 0.7802621434078643,
      "eval_loss": 0.997726321220398,
      "eval_runtime": 42.8785,
      "eval_samples_per_second": 60.496,
      "eval_steps_per_second": 7.58,
      "step": 500
    },
    {
      "epoch": 1.59,
      "learning_rate": 0.0001203125,
      "loss": 0.826,
      "step": 510
    },
    {
      "epoch": 1.62,
      "learning_rate": 0.00011875,
      "loss": 0.8231,
      "step": 520
    },
    {
      "epoch": 1.66,
      "learning_rate": 0.00011718750000000001,
      "loss": 0.8144,
      "step": 530
    },
    {
      "epoch": 1.69,
      "learning_rate": 0.000115625,
      "loss": 0.8248,
      "step": 540
    },
    {
      "epoch": 1.72,
      "learning_rate": 0.0001140625,
      "loss": 0.8278,
      "step": 550
    },
    {
      "epoch": 1.75,
      "learning_rate": 0.00011250000000000001,
      "loss": 0.8477,
      "step": 560
    },
    {
      "epoch": 1.78,
      "learning_rate": 0.0001109375,
      "loss": 0.714,
      "step": 570
    },
    {
      "epoch": 1.81,
      "learning_rate": 0.000109375,
      "loss": 0.7819,
      "step": 580
    },
    {
      "epoch": 1.84,
      "learning_rate": 0.00010781250000000001,
      "loss": 0.7466,
      "step": 590
    },
    {
      "epoch": 1.88,
      "learning_rate": 0.00010625000000000001,
      "loss": 0.6946,
      "step": 600
    },
    {
      "epoch": 1.88,
      "eval_accuracy": 0.8207401696222051,
      "eval_loss": 0.8138005137443542,
      "eval_runtime": 42.9052,
      "eval_samples_per_second": 60.459,
      "eval_steps_per_second": 7.575,
      "step": 600
    },
    {
      "epoch": 1.91,
      "learning_rate": 0.0001046875,
      "loss": 0.699,
      "step": 610
    },
    {
      "epoch": 1.94,
      "learning_rate": 0.000103125,
      "loss": 0.5999,
      "step": 620
    },
    {
      "epoch": 1.97,
      "learning_rate": 0.00010156250000000001,
      "loss": 0.5976,
      "step": 630
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0001,
      "loss": 0.5776,
      "step": 640
    },
    {
      "epoch": 2.03,
      "learning_rate": 9.84375e-05,
      "loss": 0.4293,
      "step": 650
    },
    {
      "epoch": 2.06,
      "learning_rate": 9.687500000000001e-05,
      "loss": 0.4565,
      "step": 660
    },
    {
      "epoch": 2.09,
      "learning_rate": 9.53125e-05,
      "loss": 0.4259,
      "step": 670
    },
    {
      "epoch": 2.12,
      "learning_rate": 9.375e-05,
      "loss": 0.3683,
      "step": 680
    },
    {
      "epoch": 2.16,
      "learning_rate": 9.21875e-05,
      "loss": 0.4605,
      "step": 690
    },
    {
      "epoch": 2.19,
      "learning_rate": 9.062500000000001e-05,
      "loss": 0.3488,
      "step": 700
    },
    {
      "epoch": 2.19,
      "eval_accuracy": 0.848882035466461,
      "eval_loss": 0.6592601537704468,
      "eval_runtime": 42.672,
      "eval_samples_per_second": 60.789,
      "eval_steps_per_second": 7.616,
      "step": 700
    },
    {
      "epoch": 2.22,
      "learning_rate": 8.90625e-05,
      "loss": 0.4594,
      "step": 710
    },
    {
      "epoch": 2.25,
      "learning_rate": 8.75e-05,
      "loss": 0.3706,
      "step": 720
    },
    {
      "epoch": 2.28,
      "learning_rate": 8.593750000000001e-05,
      "loss": 0.3855,
      "step": 730
    },
    {
      "epoch": 2.31,
      "learning_rate": 8.4375e-05,
      "loss": 0.4005,
      "step": 740
    },
    {
      "epoch": 2.34,
      "learning_rate": 8.28125e-05,
      "loss": 0.3648,
      "step": 750
    },
    {
      "epoch": 2.38,
      "learning_rate": 8.125000000000001e-05,
      "loss": 0.3759,
      "step": 760
    },
    {
      "epoch": 2.41,
      "learning_rate": 7.96875e-05,
      "loss": 0.3622,
      "step": 770
    },
    {
      "epoch": 2.44,
      "learning_rate": 7.8125e-05,
      "loss": 0.3209,
      "step": 780
    },
    {
      "epoch": 2.47,
      "learning_rate": 7.65625e-05,
      "loss": 0.2794,
      "step": 790
    },
    {
      "epoch": 2.5,
      "learning_rate": 7.500000000000001e-05,
      "loss": 0.2935,
      "step": 800
    },
    {
      "epoch": 2.5,
      "eval_accuracy": 0.8662297609868929,
      "eval_loss": 0.5725119113922119,
      "eval_runtime": 42.6907,
      "eval_samples_per_second": 60.763,
      "eval_steps_per_second": 7.613,
      "step": 800
    },
    {
      "epoch": 2.53,
      "learning_rate": 7.34375e-05,
      "loss": 0.3629,
      "step": 810
    },
    {
      "epoch": 2.56,
      "learning_rate": 7.1875e-05,
      "loss": 0.3006,
      "step": 820
    },
    {
      "epoch": 2.59,
      "learning_rate": 7.031250000000001e-05,
      "loss": 0.2876,
      "step": 830
    },
    {
      "epoch": 2.62,
      "learning_rate": 6.875e-05,
      "loss": 0.2612,
      "step": 840
    },
    {
      "epoch": 2.66,
      "learning_rate": 6.71875e-05,
      "loss": 0.3033,
      "step": 850
    },
    {
      "epoch": 2.69,
      "learning_rate": 6.562500000000001e-05,
      "loss": 0.2857,
      "step": 860
    },
    {
      "epoch": 2.72,
      "learning_rate": 6.40625e-05,
      "loss": 0.2877,
      "step": 870
    },
    {
      "epoch": 2.75,
      "learning_rate": 6.25e-05,
      "loss": 0.2429,
      "step": 880
    },
    {
      "epoch": 2.78,
      "learning_rate": 6.0937500000000004e-05,
      "loss": 0.2133,
      "step": 890
    },
    {
      "epoch": 2.81,
      "learning_rate": 5.9375e-05,
      "loss": 0.2557,
      "step": 900
    },
    {
      "epoch": 2.81,
      "eval_accuracy": 0.8855050115651504,
      "eval_loss": 0.5088278651237488,
      "eval_runtime": 43.0972,
      "eval_samples_per_second": 60.189,
      "eval_steps_per_second": 7.541,
      "step": 900
    },
    {
      "epoch": 2.84,
      "learning_rate": 5.78125e-05,
      "loss": 0.3011,
      "step": 910
    },
    {
      "epoch": 2.88,
      "learning_rate": 5.6250000000000005e-05,
      "loss": 0.2739,
      "step": 920
    },
    {
      "epoch": 2.91,
      "learning_rate": 5.46875e-05,
      "loss": 0.215,
      "step": 930
    },
    {
      "epoch": 2.94,
      "learning_rate": 5.3125000000000004e-05,
      "loss": 0.2399,
      "step": 940
    },
    {
      "epoch": 2.97,
      "learning_rate": 5.15625e-05,
      "loss": 0.2356,
      "step": 950
    },
    {
      "epoch": 3.0,
      "learning_rate": 5e-05,
      "loss": 0.2304,
      "step": 960
    },
    {
      "epoch": 3.03,
      "learning_rate": 4.8437500000000005e-05,
      "loss": 0.21,
      "step": 970
    },
    {
      "epoch": 3.06,
      "learning_rate": 4.6875e-05,
      "loss": 0.1801,
      "step": 980
    },
    {
      "epoch": 3.09,
      "learning_rate": 4.5312500000000004e-05,
      "loss": 0.12,
      "step": 990
    },
    {
      "epoch": 3.12,
      "learning_rate": 4.375e-05,
      "loss": 0.1509,
      "step": 1000
    },
    {
      "epoch": 3.12,
      "eval_accuracy": 0.8970701619121049,
      "eval_loss": 0.4571980834007263,
      "eval_runtime": 42.5823,
      "eval_samples_per_second": 60.917,
      "eval_steps_per_second": 7.632,
      "step": 1000
    },
    {
      "epoch": 3.16,
      "learning_rate": 4.21875e-05,
      "loss": 0.1448,
      "step": 1010
    },
    {
      "epoch": 3.19,
      "learning_rate": 4.0625000000000005e-05,
      "loss": 0.1272,
      "step": 1020
    },
    {
      "epoch": 3.22,
      "learning_rate": 3.90625e-05,
      "loss": 0.1215,
      "step": 1030
    },
    {
      "epoch": 3.25,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.1236,
      "step": 1040
    },
    {
      "epoch": 3.28,
      "learning_rate": 3.59375e-05,
      "loss": 0.1364,
      "step": 1050
    },
    {
      "epoch": 3.31,
      "learning_rate": 3.4375e-05,
      "loss": 0.1074,
      "step": 1060
    },
    {
      "epoch": 3.34,
      "learning_rate": 3.2812500000000005e-05,
      "loss": 0.1383,
      "step": 1070
    },
    {
      "epoch": 3.38,
      "learning_rate": 3.125e-05,
      "loss": 0.1304,
      "step": 1080
    },
    {
      "epoch": 3.41,
      "learning_rate": 2.96875e-05,
      "loss": 0.1431,
      "step": 1090
    },
    {
      "epoch": 3.44,
      "learning_rate": 2.8125000000000003e-05,
      "loss": 0.1367,
      "step": 1100
    },
    {
      "epoch": 3.44,
      "eval_accuracy": 0.9090208172706246,
      "eval_loss": 0.4128749370574951,
      "eval_runtime": 42.4657,
      "eval_samples_per_second": 61.085,
      "eval_steps_per_second": 7.653,
      "step": 1100
    },
    {
      "epoch": 3.47,
      "learning_rate": 2.6562500000000002e-05,
      "loss": 0.1202,
      "step": 1110
    },
    {
      "epoch": 3.5,
      "learning_rate": 2.5e-05,
      "loss": 0.1309,
      "step": 1120
    },
    {
      "epoch": 3.53,
      "learning_rate": 2.34375e-05,
      "loss": 0.1081,
      "step": 1130
    },
    {
      "epoch": 3.56,
      "learning_rate": 2.1875e-05,
      "loss": 0.144,
      "step": 1140
    },
    {
      "epoch": 3.59,
      "learning_rate": 2.0312500000000002e-05,
      "loss": 0.1104,
      "step": 1150
    },
    {
      "epoch": 3.62,
      "learning_rate": 1.8750000000000002e-05,
      "loss": 0.0947,
      "step": 1160
    },
    {
      "epoch": 3.66,
      "learning_rate": 1.71875e-05,
      "loss": 0.1008,
      "step": 1170
    },
    {
      "epoch": 3.69,
      "learning_rate": 1.5625e-05,
      "loss": 0.1177,
      "step": 1180
    },
    {
      "epoch": 3.72,
      "learning_rate": 1.4062500000000001e-05,
      "loss": 0.1347,
      "step": 1190
    },
    {
      "epoch": 3.75,
      "learning_rate": 1.25e-05,
      "loss": 0.1078,
      "step": 1200
    },
    {
      "epoch": 3.75,
      "eval_accuracy": 0.9067077872012336,
      "eval_loss": 0.3908935785293579,
      "eval_runtime": 42.3296,
      "eval_samples_per_second": 61.281,
      "eval_steps_per_second": 7.678,
      "step": 1200
    },
    {
      "epoch": 3.78,
      "learning_rate": 1.09375e-05,
      "loss": 0.1007,
      "step": 1210
    },
    {
      "epoch": 3.81,
      "learning_rate": 9.375000000000001e-06,
      "loss": 0.1205,
      "step": 1220
    },
    {
      "epoch": 3.84,
      "learning_rate": 7.8125e-06,
      "loss": 0.1124,
      "step": 1230
    },
    {
      "epoch": 3.88,
      "learning_rate": 6.25e-06,
      "loss": 0.1064,
      "step": 1240
    },
    {
      "epoch": 3.91,
      "learning_rate": 4.6875000000000004e-06,
      "loss": 0.1117,
      "step": 1250
    },
    {
      "epoch": 3.94,
      "learning_rate": 3.125e-06,
      "loss": 0.1043,
      "step": 1260
    },
    {
      "epoch": 3.97,
      "learning_rate": 1.5625e-06,
      "loss": 0.1033,
      "step": 1270
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0,
      "loss": 0.1101,
      "step": 1280
    },
    {
      "epoch": 4.0,
      "step": 1280,
      "total_flos": 3.1689424899978854e+18,
      "train_loss": 1.0863585265353322,
      "train_runtime": 1526.0713,
      "train_samples_per_second": 26.762,
      "train_steps_per_second": 0.839
    }
  ],
  "max_steps": 1280,
  "num_train_epochs": 4,
  "total_flos": 3.1689424899978854e+18,
  "trial_name": null,
  "trial_params": null
}