{
  "best_metric": 0.964,
  "best_model_checkpoint": "./Swim/checkpoint-2016",
  "epoch": 20.0,
  "eval_steps": 500,
  "global_step": 2240,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.09,
      "learning_rate": 7.142857142857143e-07,
      "loss": 4.6953,
      "step": 10
    },
    {
      "epoch": 0.18,
      "learning_rate": 1.6071428571428574e-06,
      "loss": 4.6947,
      "step": 20
    },
    {
      "epoch": 0.27,
      "learning_rate": 2.5e-06,
      "loss": 4.6787,
      "step": 30
    },
    {
      "epoch": 0.36,
      "learning_rate": 3.3928571428571435e-06,
      "loss": 4.6788,
      "step": 40
    },
    {
      "epoch": 0.45,
      "learning_rate": 4.2857142857142855e-06,
      "loss": 4.6518,
      "step": 50
    },
    {
      "epoch": 0.54,
      "learning_rate": 5.1785714285714296e-06,
      "loss": 4.5876,
      "step": 60
    },
    {
      "epoch": 0.62,
      "learning_rate": 6.071428571428571e-06,
      "loss": 4.5313,
      "step": 70
    },
    {
      "epoch": 0.71,
      "learning_rate": 6.964285714285714e-06,
      "loss": 4.5096,
      "step": 80
    },
    {
      "epoch": 0.8,
      "learning_rate": 7.857142857142858e-06,
      "loss": 4.4597,
      "step": 90
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.750000000000001e-06,
      "loss": 4.3871,
      "step": 100
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.642857142857144e-06,
      "loss": 4.2666,
      "step": 110
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.168,
      "eval_loss": 4.156820297241211,
      "eval_runtime": 35.382,
      "eval_samples_per_second": 56.526,
      "eval_steps_per_second": 7.066,
      "step": 112
    },
    {
      "epoch": 1.07,
      "learning_rate": 1.0535714285714287e-05,
      "loss": 4.1574,
      "step": 120
    },
    {
      "epoch": 1.16,
      "learning_rate": 1.1428571428571429e-05,
      "loss": 3.9255,
      "step": 130
    },
    {
      "epoch": 1.25,
      "learning_rate": 1.2321428571428572e-05,
      "loss": 3.6748,
      "step": 140
    },
    {
      "epoch": 1.34,
      "learning_rate": 1.3125e-05,
      "loss": 3.4105,
      "step": 150
    },
    {
      "epoch": 1.43,
      "learning_rate": 1.4017857142857144e-05,
      "loss": 3.0024,
      "step": 160
    },
    {
      "epoch": 1.52,
      "learning_rate": 1.4910714285714287e-05,
      "loss": 2.7239,
      "step": 170
    },
    {
      "epoch": 1.61,
      "learning_rate": 1.580357142857143e-05,
      "loss": 2.3972,
      "step": 180
    },
    {
      "epoch": 1.7,
      "learning_rate": 1.6696428571428574e-05,
      "loss": 2.0513,
      "step": 190
    },
    {
      "epoch": 1.79,
      "learning_rate": 1.7589285714285717e-05,
      "loss": 1.7811,
      "step": 200
    },
    {
      "epoch": 1.88,
      "learning_rate": 1.848214285714286e-05,
      "loss": 1.4509,
      "step": 210
    },
    {
      "epoch": 1.96,
      "learning_rate": 1.9375e-05,
      "loss": 1.178,
      "step": 220
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8465,
      "eval_loss": 0.8606186509132385,
      "eval_runtime": 37.8478,
      "eval_samples_per_second": 52.843,
      "eval_steps_per_second": 6.605,
      "step": 224
    },
    {
      "epoch": 2.05,
      "learning_rate": 1.9970238095238096e-05,
      "loss": 1.0238,
      "step": 230
    },
    {
      "epoch": 2.14,
      "learning_rate": 1.9871031746031747e-05,
      "loss": 0.8224,
      "step": 240
    },
    {
      "epoch": 2.23,
      "learning_rate": 1.9781746031746036e-05,
      "loss": 0.744,
      "step": 250
    },
    {
      "epoch": 2.32,
      "learning_rate": 1.9682539682539684e-05,
      "loss": 0.6564,
      "step": 260
    },
    {
      "epoch": 2.41,
      "learning_rate": 1.9583333333333333e-05,
      "loss": 0.6285,
      "step": 270
    },
    {
      "epoch": 2.5,
      "learning_rate": 1.9484126984126985e-05,
      "loss": 0.5989,
      "step": 280
    },
    {
      "epoch": 2.59,
      "learning_rate": 1.9384920634920637e-05,
      "loss": 0.4642,
      "step": 290
    },
    {
      "epoch": 2.68,
      "learning_rate": 1.928571428571429e-05,
      "loss": 0.448,
      "step": 300
    },
    {
      "epoch": 2.77,
      "learning_rate": 1.918650793650794e-05,
      "loss": 0.4743,
      "step": 310
    },
    {
      "epoch": 2.86,
      "learning_rate": 1.908730158730159e-05,
      "loss": 0.4283,
      "step": 320
    },
    {
      "epoch": 2.95,
      "learning_rate": 1.8988095238095237e-05,
      "loss": 0.3841,
      "step": 330
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9125,
      "eval_loss": 0.326899915933609,
      "eval_runtime": 37.3802,
      "eval_samples_per_second": 53.504,
      "eval_steps_per_second": 6.688,
      "step": 336
    },
    {
      "epoch": 3.04,
      "learning_rate": 1.888888888888889e-05,
      "loss": 0.4184,
      "step": 340
    },
    {
      "epoch": 3.12,
      "learning_rate": 1.878968253968254e-05,
      "loss": 0.2988,
      "step": 350
    },
    {
      "epoch": 3.21,
      "learning_rate": 1.8690476190476193e-05,
      "loss": 0.2439,
      "step": 360
    },
    {
      "epoch": 3.3,
      "learning_rate": 1.8591269841269845e-05,
      "loss": 0.2364,
      "step": 370
    },
    {
      "epoch": 3.39,
      "learning_rate": 1.8492063492063493e-05,
      "loss": 0.2264,
      "step": 380
    },
    {
      "epoch": 3.48,
      "learning_rate": 1.8392857142857142e-05,
      "loss": 0.2282,
      "step": 390
    },
    {
      "epoch": 3.57,
      "learning_rate": 1.8293650793650794e-05,
      "loss": 0.2825,
      "step": 400
    },
    {
      "epoch": 3.66,
      "learning_rate": 1.8194444444444445e-05,
      "loss": 0.2303,
      "step": 410
    },
    {
      "epoch": 3.75,
      "learning_rate": 1.8095238095238097e-05,
      "loss": 0.237,
      "step": 420
    },
    {
      "epoch": 3.84,
      "learning_rate": 1.799603174603175e-05,
      "loss": 0.216,
      "step": 430
    },
    {
      "epoch": 3.93,
      "learning_rate": 1.7896825396825398e-05,
      "loss": 0.2023,
      "step": 440
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.945,
      "eval_loss": 0.21039776504039764,
      "eval_runtime": 35.9017,
      "eval_samples_per_second": 55.708,
      "eval_steps_per_second": 6.963,
      "step": 448
    },
    {
      "epoch": 4.02,
      "learning_rate": 1.779761904761905e-05,
      "loss": 0.1821,
      "step": 450
    },
    {
      "epoch": 4.11,
      "learning_rate": 1.7698412698412698e-05,
      "loss": 0.1563,
      "step": 460
    },
    {
      "epoch": 4.2,
      "learning_rate": 1.759920634920635e-05,
      "loss": 0.1472,
      "step": 470
    },
    {
      "epoch": 4.29,
      "learning_rate": 1.7500000000000002e-05,
      "loss": 0.1544,
      "step": 480
    },
    {
      "epoch": 4.38,
      "learning_rate": 1.7400793650793654e-05,
      "loss": 0.1244,
      "step": 490
    },
    {
      "epoch": 4.46,
      "learning_rate": 1.7301587301587302e-05,
      "loss": 0.1229,
      "step": 500
    },
    {
      "epoch": 4.55,
      "learning_rate": 1.7202380952380954e-05,
      "loss": 0.1651,
      "step": 510
    },
    {
      "epoch": 4.64,
      "learning_rate": 1.7103174603174603e-05,
      "loss": 0.1111,
      "step": 520
    },
    {
      "epoch": 4.73,
      "learning_rate": 1.7003968253968254e-05,
      "loss": 0.1201,
      "step": 530
    },
    {
      "epoch": 4.82,
      "learning_rate": 1.6904761904761906e-05,
      "loss": 0.1292,
      "step": 540
    },
    {
      "epoch": 4.91,
      "learning_rate": 1.6805555555555558e-05,
      "loss": 0.1134,
      "step": 550
    },
    {
      "epoch": 5.0,
      "learning_rate": 1.6706349206349207e-05,
      "loss": 0.1551,
      "step": 560
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9495,
      "eval_loss": 0.18565970659255981,
      "eval_runtime": 37.3803,
      "eval_samples_per_second": 53.504,
      "eval_steps_per_second": 6.688,
      "step": 560
    },
    {
      "epoch": 5.09,
      "learning_rate": 1.660714285714286e-05,
      "loss": 0.09,
      "step": 570
    },
    {
      "epoch": 5.18,
      "learning_rate": 1.6507936507936507e-05,
      "loss": 0.0771,
      "step": 580
    },
    {
      "epoch": 5.27,
      "learning_rate": 1.640873015873016e-05,
      "loss": 0.0916,
      "step": 590
    },
    {
      "epoch": 5.36,
      "learning_rate": 1.630952380952381e-05,
      "loss": 0.0962,
      "step": 600
    },
    {
      "epoch": 5.45,
      "learning_rate": 1.6210317460317463e-05,
      "loss": 0.0885,
      "step": 610
    },
    {
      "epoch": 5.54,
      "learning_rate": 1.6111111111111115e-05,
      "loss": 0.0604,
      "step": 620
    },
    {
      "epoch": 5.62,
      "learning_rate": 1.6011904761904763e-05,
      "loss": 0.0801,
      "step": 630
    },
    {
      "epoch": 5.71,
      "learning_rate": 1.591269841269841e-05,
      "loss": 0.0798,
      "step": 640
    },
    {
      "epoch": 5.8,
      "learning_rate": 1.5813492063492063e-05,
      "loss": 0.0775,
      "step": 650
    },
    {
      "epoch": 5.89,
      "learning_rate": 1.5714285714285715e-05,
      "loss": 0.0715,
      "step": 660
    },
    {
      "epoch": 5.98,
      "learning_rate": 1.5615079365079367e-05,
      "loss": 0.0791,
      "step": 670
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9525,
      "eval_loss": 0.16961055994033813,
      "eval_runtime": 35.7579,
      "eval_samples_per_second": 55.932,
      "eval_steps_per_second": 6.991,
      "step": 672
    },
    {
      "epoch": 6.07,
      "learning_rate": 1.551587301587302e-05,
      "loss": 0.0438,
      "step": 680
    },
    {
      "epoch": 6.16,
      "learning_rate": 1.5416666666666668e-05,
      "loss": 0.0512,
      "step": 690
    },
    {
      "epoch": 6.25,
      "learning_rate": 1.531746031746032e-05,
      "loss": 0.0415,
      "step": 700
    },
    {
      "epoch": 6.34,
      "learning_rate": 1.5218253968253968e-05,
      "loss": 0.0554,
      "step": 710
    },
    {
      "epoch": 6.43,
      "learning_rate": 1.511904761904762e-05,
      "loss": 0.041,
      "step": 720
    },
    {
      "epoch": 6.52,
      "learning_rate": 1.5019841269841272e-05,
      "loss": 0.0498,
      "step": 730
    },
    {
      "epoch": 6.61,
      "learning_rate": 1.4920634920634922e-05,
      "loss": 0.0448,
      "step": 740
    },
    {
      "epoch": 6.7,
      "learning_rate": 1.4821428571428574e-05,
      "loss": 0.0479,
      "step": 750
    },
    {
      "epoch": 6.79,
      "learning_rate": 1.4722222222222224e-05,
      "loss": 0.0409,
      "step": 760
    },
    {
      "epoch": 6.88,
      "learning_rate": 1.4623015873015874e-05,
      "loss": 0.0573,
      "step": 770
    },
    {
      "epoch": 6.96,
      "learning_rate": 1.4523809523809524e-05,
      "loss": 0.0416,
      "step": 780
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9565,
      "eval_loss": 0.15801626443862915,
      "eval_runtime": 35.8258,
      "eval_samples_per_second": 55.826,
      "eval_steps_per_second": 6.978,
      "step": 784
    },
    {
      "epoch": 7.05,
      "learning_rate": 1.4424603174603176e-05,
      "loss": 0.034,
      "step": 790
    },
    {
      "epoch": 7.14,
      "learning_rate": 1.4325396825396826e-05,
      "loss": 0.043,
      "step": 800
    },
    {
      "epoch": 7.23,
      "learning_rate": 1.4226190476190478e-05,
      "loss": 0.0351,
      "step": 810
    },
    {
      "epoch": 7.32,
      "learning_rate": 1.4126984126984128e-05,
      "loss": 0.0254,
      "step": 820
    },
    {
      "epoch": 7.41,
      "learning_rate": 1.4027777777777779e-05,
      "loss": 0.0228,
      "step": 830
    },
    {
      "epoch": 7.5,
      "learning_rate": 1.3928571428571429e-05,
      "loss": 0.0277,
      "step": 840
    },
    {
      "epoch": 7.59,
      "learning_rate": 1.382936507936508e-05,
      "loss": 0.0269,
      "step": 850
    },
    {
      "epoch": 7.68,
      "learning_rate": 1.3730158730158731e-05,
      "loss": 0.026,
      "step": 860
    },
    {
      "epoch": 7.77,
      "learning_rate": 1.3630952380952383e-05,
      "loss": 0.0426,
      "step": 870
    },
    {
      "epoch": 7.86,
      "learning_rate": 1.3531746031746033e-05,
      "loss": 0.0289,
      "step": 880
    },
    {
      "epoch": 7.95,
      "learning_rate": 1.3432539682539683e-05,
      "loss": 0.0297,
      "step": 890
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9535,
      "eval_loss": 0.17498256266117096,
      "eval_runtime": 35.8718,
      "eval_samples_per_second": 55.754,
      "eval_steps_per_second": 6.969,
      "step": 896
    },
    {
      "epoch": 8.04,
      "learning_rate": 1.3333333333333333e-05,
      "loss": 0.0221,
      "step": 900
    },
    {
      "epoch": 8.12,
      "learning_rate": 1.3234126984126985e-05,
      "loss": 0.018,
      "step": 910
    },
    {
      "epoch": 8.21,
      "learning_rate": 1.3134920634920635e-05,
      "loss": 0.0214,
      "step": 920
    },
    {
      "epoch": 8.3,
      "learning_rate": 1.3035714285714287e-05,
      "loss": 0.0197,
      "step": 930
    },
    {
      "epoch": 8.39,
      "learning_rate": 1.2936507936507937e-05,
      "loss": 0.0208,
      "step": 940
    },
    {
      "epoch": 8.48,
      "learning_rate": 1.283730158730159e-05,
      "loss": 0.022,
      "step": 950
    },
    {
      "epoch": 8.57,
      "learning_rate": 1.2738095238095238e-05,
      "loss": 0.0181,
      "step": 960
    },
    {
      "epoch": 8.66,
      "learning_rate": 1.263888888888889e-05,
      "loss": 0.021,
      "step": 970
    },
    {
      "epoch": 8.75,
      "learning_rate": 1.253968253968254e-05,
      "loss": 0.0151,
      "step": 980
    },
    {
      "epoch": 8.84,
      "learning_rate": 1.2440476190476192e-05,
      "loss": 0.0156,
      "step": 990
    },
    {
      "epoch": 8.93,
      "learning_rate": 1.2341269841269844e-05,
      "loss": 0.0134,
      "step": 1000
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9555,
      "eval_loss": 0.16116014122962952,
      "eval_runtime": 35.6775,
      "eval_samples_per_second": 56.058,
      "eval_steps_per_second": 7.007,
      "step": 1008
    },
    {
      "epoch": 9.02,
      "learning_rate": 1.2242063492063494e-05,
      "loss": 0.0134,
      "step": 1010
    },
    {
      "epoch": 9.11,
      "learning_rate": 1.2142857142857142e-05,
      "loss": 0.0194,
      "step": 1020
    },
    {
      "epoch": 9.2,
      "learning_rate": 1.2043650793650794e-05,
      "loss": 0.0136,
      "step": 1030
    },
    {
      "epoch": 9.29,
      "learning_rate": 1.1944444444444444e-05,
      "loss": 0.0126,
      "step": 1040
    },
    {
      "epoch": 9.38,
      "learning_rate": 1.1845238095238096e-05,
      "loss": 0.0084,
      "step": 1050
    },
    {
      "epoch": 9.46,
      "learning_rate": 1.1746031746031748e-05,
      "loss": 0.0179,
      "step": 1060
    },
    {
      "epoch": 9.55,
      "learning_rate": 1.1646825396825398e-05,
      "loss": 0.0164,
      "step": 1070
    },
    {
      "epoch": 9.64,
      "learning_rate": 1.1547619047619047e-05,
      "loss": 0.0106,
      "step": 1080
    },
    {
      "epoch": 9.73,
      "learning_rate": 1.1448412698412699e-05,
      "loss": 0.0154,
      "step": 1090
    },
    {
      "epoch": 9.82,
      "learning_rate": 1.134920634920635e-05,
      "loss": 0.0116,
      "step": 1100
    },
    {
      "epoch": 9.91,
      "learning_rate": 1.125e-05,
      "loss": 0.0195,
      "step": 1110
    },
    {
      "epoch": 10.0,
      "learning_rate": 1.1150793650793653e-05,
      "loss": 0.0107,
      "step": 1120
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9545,
      "eval_loss": 0.15605691075325012,
      "eval_runtime": 35.901,
      "eval_samples_per_second": 55.709,
      "eval_steps_per_second": 6.964,
      "step": 1120
    },
    {
      "epoch": 10.09,
      "learning_rate": 1.1051587301587303e-05,
      "loss": 0.0146,
      "step": 1130
    },
    {
      "epoch": 10.18,
      "learning_rate": 1.0952380952380955e-05,
      "loss": 0.0077,
      "step": 1140
    },
    {
      "epoch": 10.27,
      "learning_rate": 1.0853174603174603e-05,
      "loss": 0.0119,
      "step": 1150
    },
    {
      "epoch": 10.36,
      "learning_rate": 1.0753968253968255e-05,
      "loss": 0.0043,
      "step": 1160
    },
    {
      "epoch": 10.45,
      "learning_rate": 1.0654761904761905e-05,
      "loss": 0.0061,
      "step": 1170
    },
    {
      "epoch": 10.54,
      "learning_rate": 1.0555555555555557e-05,
      "loss": 0.0075,
      "step": 1180
    },
    {
      "epoch": 10.62,
      "learning_rate": 1.0456349206349207e-05,
      "loss": 0.0066,
      "step": 1190
    },
    {
      "epoch": 10.71,
      "learning_rate": 1.0357142857142859e-05,
      "loss": 0.0067,
      "step": 1200
    },
    {
      "epoch": 10.8,
      "learning_rate": 1.0257936507936508e-05,
      "loss": 0.0115,
      "step": 1210
    },
    {
      "epoch": 10.89,
      "learning_rate": 1.015873015873016e-05,
      "loss": 0.0065,
      "step": 1220
    },
    {
      "epoch": 10.98,
      "learning_rate": 1.005952380952381e-05,
      "loss": 0.0092,
      "step": 1230
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.962,
      "eval_loss": 0.15236468613147736,
      "eval_runtime": 36.7344,
      "eval_samples_per_second": 54.445,
      "eval_steps_per_second": 6.806,
      "step": 1232
    },
    {
      "epoch": 11.07,
      "learning_rate": 9.960317460317462e-06,
      "loss": 0.0058,
      "step": 1240
    },
    {
      "epoch": 11.16,
      "learning_rate": 9.861111111111112e-06,
      "loss": 0.0088,
      "step": 1250
    },
    {
      "epoch": 11.25,
      "learning_rate": 9.761904761904762e-06,
      "loss": 0.0078,
      "step": 1260
    },
    {
      "epoch": 11.34,
      "learning_rate": 9.662698412698414e-06,
      "loss": 0.0059,
      "step": 1270
    },
    {
      "epoch": 11.43,
      "learning_rate": 9.563492063492064e-06,
      "loss": 0.0041,
      "step": 1280
    },
    {
      "epoch": 11.52,
      "learning_rate": 9.464285714285714e-06,
      "loss": 0.0063,
      "step": 1290
    },
    {
      "epoch": 11.61,
      "learning_rate": 9.365079365079366e-06,
      "loss": 0.0078,
      "step": 1300
    },
    {
      "epoch": 11.7,
      "learning_rate": 9.265873015873016e-06,
      "loss": 0.0065,
      "step": 1310
    },
    {
      "epoch": 11.79,
      "learning_rate": 9.166666666666666e-06,
      "loss": 0.0101,
      "step": 1320
    },
    {
      "epoch": 11.88,
      "learning_rate": 9.067460317460318e-06,
      "loss": 0.0113,
      "step": 1330
    },
    {
      "epoch": 11.96,
      "learning_rate": 8.968253968253968e-06,
      "loss": 0.004,
      "step": 1340
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9585,
      "eval_loss": 0.1630759984254837,
      "eval_runtime": 35.3041,
      "eval_samples_per_second": 56.651,
      "eval_steps_per_second": 7.081,
      "step": 1344
    },
    {
      "epoch": 12.05,
      "learning_rate": 8.869047619047619e-06,
      "loss": 0.0075,
      "step": 1350
    },
    {
      "epoch": 12.14,
      "learning_rate": 8.76984126984127e-06,
      "loss": 0.0054,
      "step": 1360
    },
    {
      "epoch": 12.23,
      "learning_rate": 8.670634920634922e-06,
      "loss": 0.0056,
      "step": 1370
    },
    {
      "epoch": 12.32,
      "learning_rate": 8.571428571428571e-06,
      "loss": 0.0049,
      "step": 1380
    },
    {
      "epoch": 12.41,
      "learning_rate": 8.472222222222223e-06,
      "loss": 0.0071,
      "step": 1390
    },
    {
      "epoch": 12.5,
      "learning_rate": 8.373015873015875e-06,
      "loss": 0.0076,
      "step": 1400
    },
    {
      "epoch": 12.59,
      "learning_rate": 8.273809523809523e-06,
      "loss": 0.0038,
      "step": 1410
    },
    {
      "epoch": 12.68,
      "learning_rate": 8.174603174603175e-06,
      "loss": 0.0083,
      "step": 1420
    },
    {
      "epoch": 12.77,
      "learning_rate": 8.075396825396827e-06,
      "loss": 0.0043,
      "step": 1430
    },
    {
      "epoch": 12.86,
      "learning_rate": 7.976190476190477e-06,
      "loss": 0.0041,
      "step": 1440
    },
    {
      "epoch": 12.95,
      "learning_rate": 7.876984126984127e-06,
      "loss": 0.0052,
      "step": 1450
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.963,
      "eval_loss": 0.14644142985343933,
      "eval_runtime": 35.1388,
      "eval_samples_per_second": 56.917,
      "eval_steps_per_second": 7.115,
      "step": 1456
    },
    {
      "epoch": 13.04,
      "learning_rate": 7.77777777777778e-06,
      "loss": 0.004,
      "step": 1460
    },
    {
      "epoch": 13.12,
      "learning_rate": 7.67857142857143e-06,
      "loss": 0.0041,
      "step": 1470
    },
    {
      "epoch": 13.21,
      "learning_rate": 7.5793650793650795e-06,
      "loss": 0.0051,
      "step": 1480
    },
    {
      "epoch": 13.3,
      "learning_rate": 7.4801587301587306e-06,
      "loss": 0.0025,
      "step": 1490
    },
    {
      "epoch": 13.39,
      "learning_rate": 7.380952380952382e-06,
      "loss": 0.0035,
      "step": 1500
    },
    {
      "epoch": 13.48,
      "learning_rate": 7.281746031746032e-06,
      "loss": 0.0049,
      "step": 1510
    },
    {
      "epoch": 13.57,
      "learning_rate": 7.182539682539683e-06,
      "loss": 0.0037,
      "step": 1520
    },
    {
      "epoch": 13.66,
      "learning_rate": 7.083333333333335e-06,
      "loss": 0.0045,
      "step": 1530
    },
    {
      "epoch": 13.75,
      "learning_rate": 6.984126984126984e-06,
      "loss": 0.0039,
      "step": 1540
    },
    {
      "epoch": 13.84,
      "learning_rate": 6.884920634920635e-06,
      "loss": 0.0055,
      "step": 1550
    },
    {
      "epoch": 13.93,
      "learning_rate": 6.785714285714287e-06,
      "loss": 0.0119,
      "step": 1560
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.9625,
      "eval_loss": 0.15529316663742065,
      "eval_runtime": 35.172,
      "eval_samples_per_second": 56.863,
      "eval_steps_per_second": 7.108,
      "step": 1568
    },
    {
      "epoch": 14.02,
      "learning_rate": 6.686507936507936e-06,
      "loss": 0.0039,
      "step": 1570
    },
    {
      "epoch": 14.11,
      "learning_rate": 6.587301587301588e-06,
      "loss": 0.0054,
      "step": 1580
    },
    {
      "epoch": 14.2,
      "learning_rate": 6.488095238095239e-06,
      "loss": 0.0046,
      "step": 1590
    },
    {
      "epoch": 14.29,
      "learning_rate": 6.3888888888888885e-06,
      "loss": 0.0052,
      "step": 1600
    },
    {
      "epoch": 14.38,
      "learning_rate": 6.28968253968254e-06,
      "loss": 0.0064,
      "step": 1610
    },
    {
      "epoch": 14.46,
      "learning_rate": 6.1904761904761914e-06,
      "loss": 0.0079,
      "step": 1620
    },
    {
      "epoch": 14.55,
      "learning_rate": 6.091269841269842e-06,
      "loss": 0.0043,
      "step": 1630
    },
    {
      "epoch": 14.64,
      "learning_rate": 5.992063492063493e-06,
      "loss": 0.0054,
      "step": 1640
    },
    {
      "epoch": 14.73,
      "learning_rate": 5.892857142857144e-06,
      "loss": 0.0041,
      "step": 1650
    },
    {
      "epoch": 14.82,
      "learning_rate": 5.793650793650795e-06,
      "loss": 0.004,
      "step": 1660
    },
    {
      "epoch": 14.91,
      "learning_rate": 5.694444444444445e-06,
      "loss": 0.0072,
      "step": 1670
    },
    {
      "epoch": 15.0,
      "learning_rate": 5.595238095238096e-06,
      "loss": 0.007,
      "step": 1680
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.962,
      "eval_loss": 0.16472585499286652,
      "eval_runtime": 37.4061,
      "eval_samples_per_second": 53.467,
      "eval_steps_per_second": 6.683,
      "step": 1680
    },
    {
      "epoch": 15.09,
      "learning_rate": 5.496031746031747e-06,
      "loss": 0.0037,
      "step": 1690
    },
    {
      "epoch": 15.18,
      "learning_rate": 5.396825396825397e-06,
      "loss": 0.0081,
      "step": 1700
    },
    {
      "epoch": 15.27,
      "learning_rate": 5.297619047619048e-06,
      "loss": 0.0031,
      "step": 1710
    },
    {
      "epoch": 15.36,
      "learning_rate": 5.198412698412699e-06,
      "loss": 0.0055,
      "step": 1720
    },
    {
      "epoch": 15.45,
      "learning_rate": 5.099206349206349e-06,
      "loss": 0.0036,
      "step": 1730
    },
    {
      "epoch": 15.54,
      "learning_rate": 5e-06,
      "loss": 0.0027,
      "step": 1740
    },
    {
      "epoch": 15.62,
      "learning_rate": 4.900793650793651e-06,
      "loss": 0.0053,
      "step": 1750
    },
    {
      "epoch": 15.71,
      "learning_rate": 4.8015873015873025e-06,
      "loss": 0.0046,
      "step": 1760
    },
    {
      "epoch": 15.8,
      "learning_rate": 4.702380952380953e-06,
      "loss": 0.0059,
      "step": 1770
    },
    {
      "epoch": 15.89,
      "learning_rate": 4.603174603174604e-06,
      "loss": 0.0052,
      "step": 1780
    },
    {
      "epoch": 15.98,
      "learning_rate": 4.503968253968255e-06,
      "loss": 0.0042,
      "step": 1790
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.9625,
      "eval_loss": 0.15602682530879974,
      "eval_runtime": 35.8384,
      "eval_samples_per_second": 55.806,
      "eval_steps_per_second": 6.976,
      "step": 1792
    },
    {
      "epoch": 16.07,
      "learning_rate": 4.404761904761905e-06,
      "loss": 0.0014,
      "step": 1800
    },
    {
      "epoch": 16.16,
      "learning_rate": 4.305555555555556e-06,
      "loss": 0.0039,
      "step": 1810
    },
    {
      "epoch": 16.25,
      "learning_rate": 4.206349206349207e-06,
      "loss": 0.0038,
      "step": 1820
    },
    {
      "epoch": 16.34,
      "learning_rate": 4.107142857142857e-06,
      "loss": 0.0027,
      "step": 1830
    },
    {
      "epoch": 16.43,
      "learning_rate": 4.007936507936508e-06,
      "loss": 0.0018,
      "step": 1840
    },
    {
      "epoch": 16.52,
      "learning_rate": 3.908730158730159e-06,
      "loss": 0.0024,
      "step": 1850
    },
    {
      "epoch": 16.61,
      "learning_rate": 3.80952380952381e-06,
      "loss": 0.0064,
      "step": 1860
    },
    {
      "epoch": 16.7,
      "learning_rate": 3.710317460317461e-06,
      "loss": 0.0019,
      "step": 1870
    },
    {
      "epoch": 16.79,
      "learning_rate": 3.6111111111111115e-06,
      "loss": 0.0033,
      "step": 1880
    },
    {
      "epoch": 16.88,
      "learning_rate": 3.511904761904762e-06,
      "loss": 0.0021,
      "step": 1890
    },
    {
      "epoch": 16.96,
      "learning_rate": 3.412698412698413e-06,
      "loss": 0.0027,
      "step": 1900
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.9635,
      "eval_loss": 0.15760424733161926,
      "eval_runtime": 35.2983,
      "eval_samples_per_second": 56.66,
      "eval_steps_per_second": 7.082,
      "step": 1904
    },
    {
      "epoch": 17.05,
      "learning_rate": 3.3134920634920637e-06,
      "loss": 0.0043,
      "step": 1910
    },
    {
      "epoch": 17.14,
      "learning_rate": 3.2142857142857147e-06,
      "loss": 0.0021,
      "step": 1920
    },
    {
      "epoch": 17.23,
      "learning_rate": 3.1150793650793653e-06,
      "loss": 0.003,
      "step": 1930
    },
    {
      "epoch": 17.32,
      "learning_rate": 3.015873015873016e-06,
      "loss": 0.0055,
      "step": 1940
    },
    {
      "epoch": 17.41,
      "learning_rate": 2.916666666666667e-06,
      "loss": 0.0098,
      "step": 1950
    },
    {
      "epoch": 17.5,
      "learning_rate": 2.8174603174603176e-06,
      "loss": 0.0023,
      "step": 1960
    },
    {
      "epoch": 17.59,
      "learning_rate": 2.718253968253968e-06,
      "loss": 0.0076,
      "step": 1970
    },
    {
      "epoch": 17.68,
      "learning_rate": 2.6190476190476192e-06,
      "loss": 0.0092,
      "step": 1980
    },
    {
      "epoch": 17.77,
      "learning_rate": 2.51984126984127e-06,
      "loss": 0.0024,
      "step": 1990
    },
    {
      "epoch": 17.86,
      "learning_rate": 2.420634920634921e-06,
      "loss": 0.0066,
      "step": 2000
    },
    {
      "epoch": 17.95,
      "learning_rate": 2.321428571428572e-06,
      "loss": 0.0019,
      "step": 2010
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.964,
      "eval_loss": 0.15231870114803314,
      "eval_runtime": 35.8219,
      "eval_samples_per_second": 55.832,
      "eval_steps_per_second": 6.979,
      "step": 2016
    },
    {
      "epoch": 18.04,
      "learning_rate": 2.222222222222222e-06,
      "loss": 0.0015,
      "step": 2020
    },
    {
      "epoch": 18.12,
      "learning_rate": 2.123015873015873e-06,
      "loss": 0.0035,
      "step": 2030
    },
    {
      "epoch": 18.21,
      "learning_rate": 2.023809523809524e-06,
      "loss": 0.0032,
      "step": 2040
    },
    {
      "epoch": 18.3,
      "learning_rate": 1.9246031746031747e-06,
      "loss": 0.003,
      "step": 2050
    },
    {
      "epoch": 18.39,
      "learning_rate": 1.8253968253968254e-06,
      "loss": 0.0027,
      "step": 2060
    },
    {
      "epoch": 18.48,
      "learning_rate": 1.7261904761904764e-06,
      "loss": 0.0028,
      "step": 2070
    },
    {
      "epoch": 18.57,
      "learning_rate": 1.6269841269841272e-06,
      "loss": 0.0031,
      "step": 2080
    },
    {
      "epoch": 18.66,
      "learning_rate": 1.527777777777778e-06,
      "loss": 0.0018,
      "step": 2090
    },
    {
      "epoch": 18.75,
      "learning_rate": 1.4285714285714286e-06,
      "loss": 0.002,
      "step": 2100
    },
    {
      "epoch": 18.84,
      "learning_rate": 1.3293650793650794e-06,
      "loss": 0.0025,
      "step": 2110
    },
    {
      "epoch": 18.93,
      "learning_rate": 1.2301587301587303e-06,
      "loss": 0.0018,
      "step": 2120
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.963,
      "eval_loss": 0.15332922339439392,
      "eval_runtime": 36.9815,
      "eval_samples_per_second": 54.081,
      "eval_steps_per_second": 6.76,
      "step": 2128
    },
    {
      "epoch": 19.02,
      "learning_rate": 1.130952380952381e-06,
      "loss": 0.0031,
      "step": 2130
    },
    {
      "epoch": 19.11,
      "learning_rate": 1.0317460317460317e-06,
      "loss": 0.0036,
      "step": 2140
    },
    {
      "epoch": 19.2,
      "learning_rate": 9.325396825396826e-07,
      "loss": 0.0031,
      "step": 2150
    },
    {
      "epoch": 19.29,
      "learning_rate": 8.333333333333333e-07,
      "loss": 0.0049,
      "step": 2160
    },
    {
      "epoch": 19.38,
      "learning_rate": 7.341269841269843e-07,
      "loss": 0.0039,
      "step": 2170
    },
    {
      "epoch": 19.46,
      "learning_rate": 6.34920634920635e-07,
      "loss": 0.0043,
      "step": 2180
    },
    {
      "epoch": 19.55,
      "learning_rate": 5.357142857142857e-07,
      "loss": 0.0038,
      "step": 2190
    },
    {
      "epoch": 19.64,
      "learning_rate": 4.365079365079365e-07,
      "loss": 0.0023,
      "step": 2200
    },
    {
      "epoch": 19.73,
      "learning_rate": 3.373015873015873e-07,
      "loss": 0.0036,
      "step": 2210
    },
    {
      "epoch": 19.82,
      "learning_rate": 2.3809523809523811e-07,
      "loss": 0.0025,
      "step": 2220
    },
    {
      "epoch": 19.91,
      "learning_rate": 1.3888888888888888e-07,
      "loss": 0.0037,
      "step": 2230
    },
    {
      "epoch": 20.0,
      "learning_rate": 3.9682539682539686e-08,
      "loss": 0.0023,
      "step": 2240
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.9635,
      "eval_loss": 0.15336257219314575,
      "eval_runtime": 37.728,
      "eval_samples_per_second": 53.011,
      "eval_steps_per_second": 6.626,
      "step": 2240
    },
    {
      "epoch": 20.0,
      "step": 2240,
      "total_flos": 1.254980461658112e+19,
      "train_loss": 0.41855473115200376,
      "train_runtime": 5002.9699,
      "train_samples_per_second": 31.981,
      "train_steps_per_second": 0.448
    }
  ],
  "logging_steps": 10,
  "max_steps": 2240,
  "num_train_epochs": 20,
  "save_steps": 500,
  "total_flos": 1.254980461658112e+19,
  "trial_name": null,
  "trial_params": null
}