{
  "best_metric": 0.8431372549019608,
  "best_model_checkpoint": "vit-base-patch16-224-U8-40d\\checkpoint-200",
  "epoch": 40.0,
  "eval_steps": 500,
  "global_step": 800,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5,
      "learning_rate": 1.5e-05,
      "loss": 1.3819,
      "step": 10
    },
    {
      "epoch": 1.0,
      "learning_rate": 3e-05,
      "loss": 1.3419,
      "step": 20
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.47058823529411764,
      "eval_loss": 1.299819827079773,
      "eval_runtime": 0.7952,
      "eval_samples_per_second": 64.133,
      "eval_steps_per_second": 2.515,
      "step": 20
    },
    {
      "epoch": 1.5,
      "learning_rate": 4.5e-05,
      "loss": 1.263,
      "step": 30
    },
    {
      "epoch": 2.0,
      "learning_rate": 6e-05,
      "loss": 1.1313,
      "step": 40
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5686274509803921,
      "eval_loss": 1.0832325220108032,
      "eval_runtime": 0.8195,
      "eval_samples_per_second": 62.236,
      "eval_steps_per_second": 2.441,
      "step": 40
    },
    {
      "epoch": 2.5,
      "learning_rate": 5.921052631578947e-05,
      "loss": 0.978,
      "step": 50
    },
    {
      "epoch": 3.0,
      "learning_rate": 5.842105263157895e-05,
      "loss": 0.7969,
      "step": 60
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6666666666666666,
      "eval_loss": 0.809360682964325,
      "eval_runtime": 0.8396,
      "eval_samples_per_second": 60.742,
      "eval_steps_per_second": 2.382,
      "step": 60
    },
    {
      "epoch": 3.5,
      "learning_rate": 5.7631578947368423e-05,
      "loss": 0.6378,
      "step": 70
    },
    {
      "epoch": 4.0,
      "learning_rate": 5.684210526315789e-05,
      "loss": 0.5063,
      "step": 80
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7843137254901961,
      "eval_loss": 0.6573488116264343,
      "eval_runtime": 0.7944,
      "eval_samples_per_second": 64.199,
      "eval_steps_per_second": 2.518,
      "step": 80
    },
    {
      "epoch": 4.5,
      "learning_rate": 5.605263157894737e-05,
      "loss": 0.3989,
      "step": 90
    },
    {
      "epoch": 5.0,
      "learning_rate": 5.5263157894736845e-05,
      "loss": 0.3367,
      "step": 100
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7647058823529411,
      "eval_loss": 0.6389498114585876,
      "eval_runtime": 0.7895,
      "eval_samples_per_second": 64.594,
      "eval_steps_per_second": 2.533,
      "step": 100
    },
    {
      "epoch": 5.5,
      "learning_rate": 5.447368421052632e-05,
      "loss": 0.2707,
      "step": 110
    },
    {
      "epoch": 6.0,
      "learning_rate": 5.368421052631579e-05,
      "loss": 0.242,
      "step": 120
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7450980392156863,
      "eval_loss": 0.6878873705863953,
      "eval_runtime": 0.8009,
      "eval_samples_per_second": 63.676,
      "eval_steps_per_second": 2.497,
      "step": 120
    },
    {
      "epoch": 6.5,
      "learning_rate": 5.289473684210526e-05,
      "loss": 0.2263,
      "step": 130
    },
    {
      "epoch": 7.0,
      "learning_rate": 5.210526315789474e-05,
      "loss": 0.1881,
      "step": 140
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.7058823529411765,
      "eval_loss": 0.7939884066581726,
      "eval_runtime": 0.8019,
      "eval_samples_per_second": 63.596,
      "eval_steps_per_second": 2.494,
      "step": 140
    },
    {
      "epoch": 7.5,
      "learning_rate": 5.131578947368421e-05,
      "loss": 0.2096,
      "step": 150
    },
    {
      "epoch": 8.0,
      "learning_rate": 5.052631578947368e-05,
      "loss": 0.1561,
      "step": 160
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.7647058823529411,
      "eval_loss": 0.8029699325561523,
      "eval_runtime": 0.7855,
      "eval_samples_per_second": 64.927,
      "eval_steps_per_second": 2.546,
      "step": 160
    },
    {
      "epoch": 8.5,
      "learning_rate": 4.973684210526316e-05,
      "loss": 0.1468,
      "step": 170
    },
    {
      "epoch": 9.0,
      "learning_rate": 4.8947368421052635e-05,
      "loss": 0.1557,
      "step": 180
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8235294117647058,
      "eval_loss": 0.7004449367523193,
      "eval_runtime": 0.797,
      "eval_samples_per_second": 63.992,
      "eval_steps_per_second": 2.509,
      "step": 180
    },
    {
      "epoch": 9.5,
      "learning_rate": 4.815789473684211e-05,
      "loss": 0.1385,
      "step": 190
    },
    {
      "epoch": 10.0,
      "learning_rate": 4.736842105263158e-05,
      "loss": 0.1154,
      "step": 200
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.8431372549019608,
      "eval_loss": 0.649506688117981,
      "eval_runtime": 0.7901,
      "eval_samples_per_second": 64.552,
      "eval_steps_per_second": 2.531,
      "step": 200
    },
    {
      "epoch": 10.5,
      "learning_rate": 4.657894736842105e-05,
      "loss": 0.1235,
      "step": 210
    },
    {
      "epoch": 11.0,
      "learning_rate": 4.5789473684210527e-05,
      "loss": 0.1469,
      "step": 220
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.7058823529411765,
      "eval_loss": 1.1387523412704468,
      "eval_runtime": 0.7986,
      "eval_samples_per_second": 63.858,
      "eval_steps_per_second": 2.504,
      "step": 220
    },
    {
      "epoch": 11.5,
      "learning_rate": 4.5e-05,
      "loss": 0.1223,
      "step": 230
    },
    {
      "epoch": 12.0,
      "learning_rate": 4.421052631578947e-05,
      "loss": 0.0898,
      "step": 240
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.7647058823529411,
      "eval_loss": 0.7966563105583191,
      "eval_runtime": 0.8161,
      "eval_samples_per_second": 62.495,
      "eval_steps_per_second": 2.451,
      "step": 240
    },
    {
      "epoch": 12.5,
      "learning_rate": 4.342105263157895e-05,
      "loss": 0.1042,
      "step": 250
    },
    {
      "epoch": 13.0,
      "learning_rate": 4.2631578947368425e-05,
      "loss": 0.0719,
      "step": 260
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.803921568627451,
      "eval_loss": 0.8934146165847778,
      "eval_runtime": 0.7865,
      "eval_samples_per_second": 64.845,
      "eval_steps_per_second": 2.543,
      "step": 260
    },
    {
      "epoch": 13.5,
      "learning_rate": 4.1842105263157894e-05,
      "loss": 0.0927,
      "step": 270
    },
    {
      "epoch": 14.0,
      "learning_rate": 4.105263157894737e-05,
      "loss": 0.0739,
      "step": 280
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.7647058823529411,
      "eval_loss": 0.8476159572601318,
      "eval_runtime": 0.8141,
      "eval_samples_per_second": 62.643,
      "eval_steps_per_second": 2.457,
      "step": 280
    },
    {
      "epoch": 14.5,
      "learning_rate": 4.026315789473684e-05,
      "loss": 0.088,
      "step": 290
    },
    {
      "epoch": 15.0,
      "learning_rate": 3.9473684210526316e-05,
      "loss": 0.0823,
      "step": 300
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.7647058823529411,
      "eval_loss": 0.9692044854164124,
      "eval_runtime": 0.82,
      "eval_samples_per_second": 62.196,
      "eval_steps_per_second": 2.439,
      "step": 300
    },
    {
      "epoch": 15.5,
      "learning_rate": 3.868421052631579e-05,
      "loss": 0.0714,
      "step": 310
    },
    {
      "epoch": 16.0,
      "learning_rate": 3.789473684210526e-05,
      "loss": 0.0828,
      "step": 320
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.7843137254901961,
      "eval_loss": 0.9384645819664001,
      "eval_runtime": 0.8043,
      "eval_samples_per_second": 63.407,
      "eval_steps_per_second": 2.487,
      "step": 320
    },
    {
      "epoch": 16.5,
      "learning_rate": 3.710526315789474e-05,
      "loss": 0.0762,
      "step": 330
    },
    {
      "epoch": 17.0,
      "learning_rate": 3.6315789473684214e-05,
      "loss": 0.0761,
      "step": 340
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.7254901960784313,
      "eval_loss": 1.1684223413467407,
      "eval_runtime": 0.8011,
      "eval_samples_per_second": 63.661,
      "eval_steps_per_second": 2.496,
      "step": 340
    },
    {
      "epoch": 17.5,
      "learning_rate": 3.5526315789473684e-05,
      "loss": 0.0925,
      "step": 350
    },
    {
      "epoch": 18.0,
      "learning_rate": 3.473684210526316e-05,
      "loss": 0.0597,
      "step": 360
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.7647058823529411,
      "eval_loss": 0.9413917660713196,
      "eval_runtime": 0.8568,
      "eval_samples_per_second": 59.521,
      "eval_steps_per_second": 2.334,
      "step": 360
    },
    {
      "epoch": 18.5,
      "learning_rate": 3.394736842105263e-05,
      "loss": 0.0806,
      "step": 370
    },
    {
      "epoch": 19.0,
      "learning_rate": 3.3157894736842106e-05,
      "loss": 0.0727,
      "step": 380
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.7058823529411765,
      "eval_loss": 1.020107388496399,
      "eval_runtime": 0.8388,
      "eval_samples_per_second": 60.801,
      "eval_steps_per_second": 2.384,
      "step": 380
    },
    {
      "epoch": 19.5,
      "learning_rate": 3.236842105263158e-05,
      "loss": 0.0789,
      "step": 390
    },
    {
      "epoch": 20.0,
      "learning_rate": 3.157894736842105e-05,
      "loss": 0.0507,
      "step": 400
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.803921568627451,
      "eval_loss": 0.8562985062599182,
      "eval_runtime": 0.7915,
      "eval_samples_per_second": 64.435,
      "eval_steps_per_second": 2.527,
      "step": 400
    },
    {
      "epoch": 20.5,
      "learning_rate": 3.078947368421053e-05,
      "loss": 0.0557,
      "step": 410
    },
    {
      "epoch": 21.0,
      "learning_rate": 3e-05,
      "loss": 0.0587,
      "step": 420
    },
    {
      "epoch": 21.0,
      "eval_accuracy": 0.7843137254901961,
      "eval_loss": 0.8476255536079407,
      "eval_runtime": 0.7773,
      "eval_samples_per_second": 65.611,
      "eval_steps_per_second": 2.573,
      "step": 420
    },
    {
      "epoch": 21.5,
      "learning_rate": 2.9210526315789474e-05,
      "loss": 0.0615,
      "step": 430
    },
    {
      "epoch": 22.0,
      "learning_rate": 2.8421052631578946e-05,
      "loss": 0.0608,
      "step": 440
    },
    {
      "epoch": 22.0,
      "eval_accuracy": 0.803921568627451,
      "eval_loss": 0.9399316310882568,
      "eval_runtime": 0.8324,
      "eval_samples_per_second": 61.27,
      "eval_steps_per_second": 2.403,
      "step": 440
    },
    {
      "epoch": 22.5,
      "learning_rate": 2.7631578947368423e-05,
      "loss": 0.0505,
      "step": 450
    },
    {
      "epoch": 23.0,
      "learning_rate": 2.6842105263157896e-05,
      "loss": 0.055,
      "step": 460
    },
    {
      "epoch": 23.0,
      "eval_accuracy": 0.7450980392156863,
      "eval_loss": 0.8819794654846191,
      "eval_runtime": 0.7867,
      "eval_samples_per_second": 64.824,
      "eval_steps_per_second": 2.542,
      "step": 460
    },
    {
      "epoch": 23.5,
      "learning_rate": 2.605263157894737e-05,
      "loss": 0.0682,
      "step": 470
    },
    {
      "epoch": 24.0,
      "learning_rate": 2.526315789473684e-05,
      "loss": 0.0619,
      "step": 480
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.7647058823529411,
      "eval_loss": 1.0459517240524292,
      "eval_runtime": 0.7938,
      "eval_samples_per_second": 64.25,
      "eval_steps_per_second": 2.52,
      "step": 480
    },
    {
      "epoch": 24.5,
      "learning_rate": 2.4473684210526318e-05,
      "loss": 0.0448,
      "step": 490
    },
    {
      "epoch": 25.0,
      "learning_rate": 2.368421052631579e-05,
      "loss": 0.0615,
      "step": 500
    },
    {
      "epoch": 25.0,
      "eval_accuracy": 0.8235294117647058,
      "eval_loss": 0.9392306804656982,
      "eval_runtime": 0.7883,
      "eval_samples_per_second": 64.697,
      "eval_steps_per_second": 2.537,
      "step": 500
    },
    {
      "epoch": 25.5,
      "learning_rate": 2.2894736842105263e-05,
      "loss": 0.0488,
      "step": 510
    },
    {
      "epoch": 26.0,
      "learning_rate": 2.2105263157894736e-05,
      "loss": 0.0455,
      "step": 520
    },
    {
      "epoch": 26.0,
      "eval_accuracy": 0.8235294117647058,
      "eval_loss": 0.9267483353614807,
      "eval_runtime": 0.7948,
      "eval_samples_per_second": 64.17,
      "eval_steps_per_second": 2.516,
      "step": 520
    },
    {
      "epoch": 26.5,
      "learning_rate": 2.1315789473684212e-05,
      "loss": 0.0493,
      "step": 530
    },
    {
      "epoch": 27.0,
      "learning_rate": 2.0526315789473685e-05,
      "loss": 0.0567,
      "step": 540
    },
    {
      "epoch": 27.0,
      "eval_accuracy": 0.7843137254901961,
      "eval_loss": 0.9784489870071411,
      "eval_runtime": 0.8082,
      "eval_samples_per_second": 63.101,
      "eval_steps_per_second": 2.475,
      "step": 540
    },
    {
      "epoch": 27.5,
      "learning_rate": 1.9736842105263158e-05,
      "loss": 0.0467,
      "step": 550
    },
    {
      "epoch": 28.0,
      "learning_rate": 1.894736842105263e-05,
      "loss": 0.032,
      "step": 560
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.7647058823529411,
      "eval_loss": 1.1540778875350952,
      "eval_runtime": 0.7989,
      "eval_samples_per_second": 63.834,
      "eval_steps_per_second": 2.503,
      "step": 560
    },
    {
      "epoch": 28.5,
      "learning_rate": 1.8157894736842107e-05,
      "loss": 0.0242,
      "step": 570
    },
    {
      "epoch": 29.0,
      "learning_rate": 1.736842105263158e-05,
      "loss": 0.0276,
      "step": 580
    },
    {
      "epoch": 29.0,
      "eval_accuracy": 0.7843137254901961,
      "eval_loss": 0.8864995837211609,
      "eval_runtime": 0.8109,
      "eval_samples_per_second": 62.894,
      "eval_steps_per_second": 2.466,
      "step": 580
    },
    {
      "epoch": 29.5,
      "learning_rate": 1.6578947368421053e-05,
      "loss": 0.058,
      "step": 590
    },
    {
      "epoch": 30.0,
      "learning_rate": 1.5789473684210526e-05,
      "loss": 0.0368,
      "step": 600
    },
    {
      "epoch": 30.0,
      "eval_accuracy": 0.803921568627451,
      "eval_loss": 1.0847781896591187,
      "eval_runtime": 0.7819,
      "eval_samples_per_second": 65.229,
      "eval_steps_per_second": 2.558,
      "step": 600
    },
    {
      "epoch": 30.5,
      "learning_rate": 1.5e-05,
      "loss": 0.0479,
      "step": 610
    },
    {
      "epoch": 31.0,
      "learning_rate": 1.4210526315789473e-05,
      "loss": 0.0342,
      "step": 620
    },
    {
      "epoch": 31.0,
      "eval_accuracy": 0.803921568627451,
      "eval_loss": 0.9638255834579468,
      "eval_runtime": 0.7841,
      "eval_samples_per_second": 65.044,
      "eval_steps_per_second": 2.551,
      "step": 620
    },
    {
      "epoch": 31.5,
      "learning_rate": 1.3421052631578948e-05,
      "loss": 0.0352,
      "step": 630
    },
    {
      "epoch": 32.0,
      "learning_rate": 1.263157894736842e-05,
      "loss": 0.037,
      "step": 640
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.803921568627451,
      "eval_loss": 0.961588978767395,
      "eval_runtime": 0.7905,
      "eval_samples_per_second": 64.516,
      "eval_steps_per_second": 2.53,
      "step": 640
    },
    {
      "epoch": 32.5,
      "learning_rate": 1.1842105263157895e-05,
      "loss": 0.0295,
      "step": 650
    },
    {
      "epoch": 33.0,
      "learning_rate": 1.1052631578947368e-05,
      "loss": 0.0371,
      "step": 660
    },
    {
      "epoch": 33.0,
      "eval_accuracy": 0.803921568627451,
      "eval_loss": 1.0072985887527466,
      "eval_runtime": 0.7709,
      "eval_samples_per_second": 66.152,
      "eval_steps_per_second": 2.594,
      "step": 660
    },
    {
      "epoch": 33.5,
      "learning_rate": 1.0263157894736843e-05,
      "loss": 0.0299,
      "step": 670
    },
    {
      "epoch": 34.0,
      "learning_rate": 9.473684210526315e-06,
      "loss": 0.0371,
      "step": 680
    },
    {
      "epoch": 34.0,
      "eval_accuracy": 0.803921568627451,
      "eval_loss": 1.0493559837341309,
      "eval_runtime": 0.7834,
      "eval_samples_per_second": 65.102,
      "eval_steps_per_second": 2.553,
      "step": 680
    },
    {
      "epoch": 34.5,
      "learning_rate": 8.68421052631579e-06,
      "loss": 0.0217,
      "step": 690
    },
    {
      "epoch": 35.0,
      "learning_rate": 7.894736842105263e-06,
      "loss": 0.0359,
      "step": 700
    },
    {
      "epoch": 35.0,
      "eval_accuracy": 0.7843137254901961,
      "eval_loss": 1.1287018060684204,
      "eval_runtime": 0.7799,
      "eval_samples_per_second": 65.39,
      "eval_steps_per_second": 2.564,
      "step": 700
    },
    {
      "epoch": 35.5,
      "learning_rate": 7.105263157894737e-06,
      "loss": 0.0483,
      "step": 710
    },
    {
      "epoch": 36.0,
      "learning_rate": 6.31578947368421e-06,
      "loss": 0.0255,
      "step": 720
    },
    {
      "epoch": 36.0,
      "eval_accuracy": 0.7647058823529411,
      "eval_loss": 1.1830930709838867,
      "eval_runtime": 0.7736,
      "eval_samples_per_second": 65.928,
      "eval_steps_per_second": 2.585,
      "step": 720
    },
    {
      "epoch": 36.5,
      "learning_rate": 5.526315789473684e-06,
      "loss": 0.026,
      "step": 730
    },
    {
      "epoch": 37.0,
      "learning_rate": 4.736842105263158e-06,
      "loss": 0.0269,
      "step": 740
    },
    {
      "epoch": 37.0,
      "eval_accuracy": 0.7843137254901961,
      "eval_loss": 1.1609560251235962,
      "eval_runtime": 0.8022,
      "eval_samples_per_second": 63.578,
      "eval_steps_per_second": 2.493,
      "step": 740
    },
    {
      "epoch": 37.5,
      "learning_rate": 3.9473684210526315e-06,
      "loss": 0.023,
      "step": 750
    },
    {
      "epoch": 38.0,
      "learning_rate": 3.157894736842105e-06,
      "loss": 0.0292,
      "step": 760
    },
    {
      "epoch": 38.0,
      "eval_accuracy": 0.7843137254901961,
      "eval_loss": 1.1842255592346191,
      "eval_runtime": 0.815,
      "eval_samples_per_second": 62.574,
      "eval_steps_per_second": 2.454,
      "step": 760
    },
    {
      "epoch": 38.5,
      "learning_rate": 2.368421052631579e-06,
      "loss": 0.0328,
      "step": 770
    },
    {
      "epoch": 39.0,
      "learning_rate": 1.5789473684210526e-06,
      "loss": 0.0161,
      "step": 780
    },
    {
      "epoch": 39.0,
      "eval_accuracy": 0.803921568627451,
      "eval_loss": 1.109218716621399,
      "eval_runtime": 0.7881,
      "eval_samples_per_second": 64.712,
      "eval_steps_per_second": 2.538,
      "step": 780
    },
    {
      "epoch": 39.5,
      "learning_rate": 7.894736842105263e-07,
      "loss": 0.0197,
      "step": 790
    },
    {
      "epoch": 40.0,
      "learning_rate": 0.0,
      "loss": 0.0333,
      "step": 800
    },
    {
      "epoch": 40.0,
      "eval_accuracy": 0.803921568627451,
      "eval_loss": 1.1185595989227295,
      "eval_runtime": 0.7848,
      "eval_samples_per_second": 64.984,
      "eval_steps_per_second": 2.548,
      "step": 800
    },
    {
      "epoch": 40.0,
      "step": 800,
      "total_flos": 7.873327274596762e+18,
      "train_loss": 0.17486795043572784,
      "train_runtime": 1567.8548,
      "train_samples_per_second": 64.802,
      "train_steps_per_second": 0.51
    }
  ],
  "logging_steps": 10,
  "max_steps": 800,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 40,
  "save_steps": 500,
  "total_flos": 7.873327274596762e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}