{
  "best_metric": 0.6994230151176453,
  "best_model_checkpoint": "./kd_results/facebook/deit-small-distilled-patch16-224_alpha0.5_temp3.0/checkpoint-1620",
  "epoch": 20.0,
  "eval_steps": 500,
  "global_step": 1800,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.2777777777777778,
      "grad_norm": 4.104116916656494,
      "learning_rate": 6.944444444444445e-06,
      "loss": 1.3245,
      "step": 25
    },
    {
      "epoch": 0.5555555555555556,
      "grad_norm": 2.875807523727417,
      "learning_rate": 1.388888888888889e-05,
      "loss": 1.2378,
      "step": 50
    },
    {
      "epoch": 0.8333333333333334,
      "grad_norm": 3.105369806289673,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 1.2256,
      "step": 75
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.2707509881422925,
      "eval_loss": 1.420397400856018,
      "eval_runtime": 29.6146,
      "eval_samples_per_second": 34.172,
      "eval_steps_per_second": 0.54,
      "step": 90
    },
    {
      "epoch": 1.1111111111111112,
      "grad_norm": 3.8822920322418213,
      "learning_rate": 2.777777777777778e-05,
      "loss": 1.1963,
      "step": 100
    },
    {
      "epoch": 1.3888888888888888,
      "grad_norm": 3.240900754928589,
      "learning_rate": 3.472222222222222e-05,
      "loss": 1.1576,
      "step": 125
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 3.8485963344573975,
      "learning_rate": 4.166666666666667e-05,
      "loss": 1.1301,
      "step": 150
    },
    {
      "epoch": 1.9444444444444444,
      "grad_norm": 4.880220890045166,
      "learning_rate": 4.8611111111111115e-05,
      "loss": 1.0697,
      "step": 175
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.43774703557312256,
      "eval_loss": 1.2418382167816162,
      "eval_runtime": 30.2043,
      "eval_samples_per_second": 33.505,
      "eval_steps_per_second": 0.53,
      "step": 180
    },
    {
      "epoch": 2.2222222222222223,
      "grad_norm": 8.320867538452148,
      "learning_rate": 4.938271604938271e-05,
      "loss": 0.9319,
      "step": 200
    },
    {
      "epoch": 2.5,
      "grad_norm": 6.680625915527344,
      "learning_rate": 4.8611111111111115e-05,
      "loss": 0.8352,
      "step": 225
    },
    {
      "epoch": 2.7777777777777777,
      "grad_norm": 4.338289260864258,
      "learning_rate": 4.783950617283951e-05,
      "loss": 0.745,
      "step": 250
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6343873517786561,
      "eval_loss": 0.8559542298316956,
      "eval_runtime": 29.7805,
      "eval_samples_per_second": 33.982,
      "eval_steps_per_second": 0.537,
      "step": 270
    },
    {
      "epoch": 3.0555555555555554,
      "grad_norm": 4.460063934326172,
      "learning_rate": 4.70679012345679e-05,
      "loss": 0.673,
      "step": 275
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 3.742156982421875,
      "learning_rate": 4.62962962962963e-05,
      "loss": 0.5085,
      "step": 300
    },
    {
      "epoch": 3.611111111111111,
      "grad_norm": 3.4500081539154053,
      "learning_rate": 4.5524691358024696e-05,
      "loss": 0.4901,
      "step": 325
    },
    {
      "epoch": 3.888888888888889,
      "grad_norm": 3.695544958114624,
      "learning_rate": 4.4753086419753084e-05,
      "loss": 0.4822,
      "step": 350
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7055335968379447,
      "eval_loss": 0.7596805691719055,
      "eval_runtime": 29.2784,
      "eval_samples_per_second": 34.565,
      "eval_steps_per_second": 0.546,
      "step": 360
    },
    {
      "epoch": 4.166666666666667,
      "grad_norm": 3.700939655303955,
      "learning_rate": 4.3981481481481486e-05,
      "loss": 0.3773,
      "step": 375
    },
    {
      "epoch": 4.444444444444445,
      "grad_norm": 4.073795795440674,
      "learning_rate": 4.3209876543209875e-05,
      "loss": 0.2994,
      "step": 400
    },
    {
      "epoch": 4.722222222222222,
      "grad_norm": 3.6610753536224365,
      "learning_rate": 4.243827160493827e-05,
      "loss": 0.2977,
      "step": 425
    },
    {
      "epoch": 5.0,
      "grad_norm": 4.114968776702881,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.296,
      "step": 450
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7035573122529645,
      "eval_loss": 0.7598134875297546,
      "eval_runtime": 29.5479,
      "eval_samples_per_second": 34.25,
      "eval_steps_per_second": 0.541,
      "step": 450
    },
    {
      "epoch": 5.277777777777778,
      "grad_norm": 2.291775941848755,
      "learning_rate": 4.089506172839506e-05,
      "loss": 0.2227,
      "step": 475
    },
    {
      "epoch": 5.555555555555555,
      "grad_norm": 3.2807505130767822,
      "learning_rate": 4.012345679012346e-05,
      "loss": 0.2138,
      "step": 500
    },
    {
      "epoch": 5.833333333333333,
      "grad_norm": 2.3632428646087646,
      "learning_rate": 3.935185185185186e-05,
      "loss": 0.2083,
      "step": 525
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6867588932806324,
      "eval_loss": 0.7894434332847595,
      "eval_runtime": 29.3598,
      "eval_samples_per_second": 34.469,
      "eval_steps_per_second": 0.545,
      "step": 540
    },
    {
      "epoch": 6.111111111111111,
      "grad_norm": 2.4907870292663574,
      "learning_rate": 3.8580246913580246e-05,
      "loss": 0.2149,
      "step": 550
    },
    {
      "epoch": 6.388888888888889,
      "grad_norm": 3.0467641353607178,
      "learning_rate": 3.780864197530865e-05,
      "loss": 0.1969,
      "step": 575
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 1.9196618795394897,
      "learning_rate": 3.7037037037037037e-05,
      "loss": 0.1929,
      "step": 600
    },
    {
      "epoch": 6.944444444444445,
      "grad_norm": 1.7986650466918945,
      "learning_rate": 3.626543209876543e-05,
      "loss": 0.1911,
      "step": 625
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.7134387351778656,
      "eval_loss": 0.7575205564498901,
      "eval_runtime": 29.6909,
      "eval_samples_per_second": 34.084,
      "eval_steps_per_second": 0.539,
      "step": 630
    },
    {
      "epoch": 7.222222222222222,
      "grad_norm": 1.0618268251419067,
      "learning_rate": 3.5493827160493834e-05,
      "loss": 0.1792,
      "step": 650
    },
    {
      "epoch": 7.5,
      "grad_norm": 1.4524767398834229,
      "learning_rate": 3.472222222222222e-05,
      "loss": 0.1837,
      "step": 675
    },
    {
      "epoch": 7.777777777777778,
      "grad_norm": 1.0937868356704712,
      "learning_rate": 3.395061728395062e-05,
      "loss": 0.1783,
      "step": 700
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.7302371541501976,
      "eval_loss": 0.7299193143844604,
      "eval_runtime": 29.5938,
      "eval_samples_per_second": 34.196,
      "eval_steps_per_second": 0.541,
      "step": 720
    },
    {
      "epoch": 8.055555555555555,
      "grad_norm": 1.8283872604370117,
      "learning_rate": 3.317901234567901e-05,
      "loss": 0.1841,
      "step": 725
    },
    {
      "epoch": 8.333333333333334,
      "grad_norm": 1.2189834117889404,
      "learning_rate": 3.240740740740741e-05,
      "loss": 0.1708,
      "step": 750
    },
    {
      "epoch": 8.61111111111111,
      "grad_norm": 1.5531915426254272,
      "learning_rate": 3.16358024691358e-05,
      "loss": 0.1736,
      "step": 775
    },
    {
      "epoch": 8.88888888888889,
      "grad_norm": 1.221135139465332,
      "learning_rate": 3.08641975308642e-05,
      "loss": 0.1743,
      "step": 800
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.7183794466403162,
      "eval_loss": 0.7375351786613464,
      "eval_runtime": 29.5919,
      "eval_samples_per_second": 34.199,
      "eval_steps_per_second": 0.541,
      "step": 810
    },
    {
      "epoch": 9.166666666666666,
      "grad_norm": 1.11898672580719,
      "learning_rate": 3.0092592592592593e-05,
      "loss": 0.1633,
      "step": 825
    },
    {
      "epoch": 9.444444444444445,
      "grad_norm": 1.2036685943603516,
      "learning_rate": 2.9320987654320992e-05,
      "loss": 0.1658,
      "step": 850
    },
    {
      "epoch": 9.722222222222221,
      "grad_norm": 1.015676736831665,
      "learning_rate": 2.8549382716049384e-05,
      "loss": 0.1648,
      "step": 875
    },
    {
      "epoch": 10.0,
      "grad_norm": 1.6817418336868286,
      "learning_rate": 2.777777777777778e-05,
      "loss": 0.1645,
      "step": 900
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.7272727272727273,
      "eval_loss": 0.7215421795845032,
      "eval_runtime": 29.8084,
      "eval_samples_per_second": 33.95,
      "eval_steps_per_second": 0.537,
      "step": 900
    },
    {
      "epoch": 10.277777777777779,
      "grad_norm": 0.8497186303138733,
      "learning_rate": 2.700617283950617e-05,
      "loss": 0.161,
      "step": 925
    },
    {
      "epoch": 10.555555555555555,
      "grad_norm": 0.9341030120849609,
      "learning_rate": 2.623456790123457e-05,
      "loss": 0.1614,
      "step": 950
    },
    {
      "epoch": 10.833333333333334,
      "grad_norm": 1.0269076824188232,
      "learning_rate": 2.5462962962962965e-05,
      "loss": 0.1548,
      "step": 975
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.741106719367589,
      "eval_loss": 0.7100472450256348,
      "eval_runtime": 30.2305,
      "eval_samples_per_second": 33.476,
      "eval_steps_per_second": 0.529,
      "step": 990
    },
    {
      "epoch": 11.11111111111111,
      "grad_norm": 0.7800092697143555,
      "learning_rate": 2.4691358024691357e-05,
      "loss": 0.1509,
      "step": 1000
    },
    {
      "epoch": 11.38888888888889,
      "grad_norm": 0.8492220044136047,
      "learning_rate": 2.3919753086419755e-05,
      "loss": 0.15,
      "step": 1025
    },
    {
      "epoch": 11.666666666666666,
      "grad_norm": 0.7080733180046082,
      "learning_rate": 2.314814814814815e-05,
      "loss": 0.1572,
      "step": 1050
    },
    {
      "epoch": 11.944444444444445,
      "grad_norm": 0.579961895942688,
      "learning_rate": 2.2376543209876542e-05,
      "loss": 0.1544,
      "step": 1075
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.7351778656126482,
      "eval_loss": 0.7043272852897644,
      "eval_runtime": 29.3995,
      "eval_samples_per_second": 34.422,
      "eval_steps_per_second": 0.544,
      "step": 1080
    },
    {
      "epoch": 12.222222222222221,
      "grad_norm": 0.7947866916656494,
      "learning_rate": 2.1604938271604937e-05,
      "loss": 0.1481,
      "step": 1100
    },
    {
      "epoch": 12.5,
      "grad_norm": 0.7545654773712158,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 0.1545,
      "step": 1125
    },
    {
      "epoch": 12.777777777777779,
      "grad_norm": 0.6851534247398376,
      "learning_rate": 2.006172839506173e-05,
      "loss": 0.1461,
      "step": 1150
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.7351778656126482,
      "eval_loss": 0.709749698638916,
      "eval_runtime": 29.6843,
      "eval_samples_per_second": 34.092,
      "eval_steps_per_second": 0.539,
      "step": 1170
    },
    {
      "epoch": 13.055555555555555,
      "grad_norm": 0.6691556572914124,
      "learning_rate": 1.9290123456790123e-05,
      "loss": 0.1434,
      "step": 1175
    },
    {
      "epoch": 13.333333333333334,
      "grad_norm": 0.7591906189918518,
      "learning_rate": 1.8518518518518518e-05,
      "loss": 0.1493,
      "step": 1200
    },
    {
      "epoch": 13.61111111111111,
      "grad_norm": 0.7788909673690796,
      "learning_rate": 1.7746913580246917e-05,
      "loss": 0.1472,
      "step": 1225
    },
    {
      "epoch": 13.88888888888889,
      "grad_norm": 0.5040611624717712,
      "learning_rate": 1.697530864197531e-05,
      "loss": 0.1439,
      "step": 1250
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.7430830039525692,
      "eval_loss": 0.7007103562355042,
      "eval_runtime": 29.6092,
      "eval_samples_per_second": 34.179,
      "eval_steps_per_second": 0.54,
      "step": 1260
    },
    {
      "epoch": 14.166666666666666,
      "grad_norm": 0.6750937700271606,
      "learning_rate": 1.6203703703703704e-05,
      "loss": 0.1432,
      "step": 1275
    },
    {
      "epoch": 14.444444444444445,
      "grad_norm": 0.4908043146133423,
      "learning_rate": 1.54320987654321e-05,
      "loss": 0.1412,
      "step": 1300
    },
    {
      "epoch": 14.722222222222221,
      "grad_norm": 0.5028871297836304,
      "learning_rate": 1.4660493827160496e-05,
      "loss": 0.1403,
      "step": 1325
    },
    {
      "epoch": 15.0,
      "grad_norm": 0.579842746257782,
      "learning_rate": 1.388888888888889e-05,
      "loss": 0.1428,
      "step": 1350
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.7401185770750988,
      "eval_loss": 0.7012726068496704,
      "eval_runtime": 29.6572,
      "eval_samples_per_second": 34.123,
      "eval_steps_per_second": 0.539,
      "step": 1350
    },
    {
      "epoch": 15.277777777777779,
      "grad_norm": 0.372830331325531,
      "learning_rate": 1.3117283950617285e-05,
      "loss": 0.142,
      "step": 1375
    },
    {
      "epoch": 15.555555555555555,
      "grad_norm": 0.30693933367729187,
      "learning_rate": 1.2345679012345678e-05,
      "loss": 0.1399,
      "step": 1400
    },
    {
      "epoch": 15.833333333333334,
      "grad_norm": 0.4288826286792755,
      "learning_rate": 1.1574074074074075e-05,
      "loss": 0.1409,
      "step": 1425
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.7401185770750988,
      "eval_loss": 0.7025883197784424,
      "eval_runtime": 29.677,
      "eval_samples_per_second": 34.1,
      "eval_steps_per_second": 0.539,
      "step": 1440
    },
    {
      "epoch": 16.11111111111111,
      "grad_norm": 0.3597729504108429,
      "learning_rate": 1.0802469135802469e-05,
      "loss": 0.1359,
      "step": 1450
    },
    {
      "epoch": 16.38888888888889,
      "grad_norm": 0.36627763509750366,
      "learning_rate": 1.0030864197530866e-05,
      "loss": 0.1388,
      "step": 1475
    },
    {
      "epoch": 16.666666666666668,
      "grad_norm": 0.40269795060157776,
      "learning_rate": 9.259259259259259e-06,
      "loss": 0.1417,
      "step": 1500
    },
    {
      "epoch": 16.944444444444443,
      "grad_norm": 0.42748990654945374,
      "learning_rate": 8.487654320987654e-06,
      "loss": 0.1402,
      "step": 1525
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.7430830039525692,
      "eval_loss": 0.704167902469635,
      "eval_runtime": 29.7414,
      "eval_samples_per_second": 34.027,
      "eval_steps_per_second": 0.538,
      "step": 1530
    },
    {
      "epoch": 17.22222222222222,
      "grad_norm": 0.40921568870544434,
      "learning_rate": 7.71604938271605e-06,
      "loss": 0.1402,
      "step": 1550
    },
    {
      "epoch": 17.5,
      "grad_norm": 0.37299108505249023,
      "learning_rate": 6.944444444444445e-06,
      "loss": 0.1379,
      "step": 1575
    },
    {
      "epoch": 17.77777777777778,
      "grad_norm": 0.39989030361175537,
      "learning_rate": 6.172839506172839e-06,
      "loss": 0.1325,
      "step": 1600
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.7351778656126482,
      "eval_loss": 0.6994230151176453,
      "eval_runtime": 29.536,
      "eval_samples_per_second": 34.263,
      "eval_steps_per_second": 0.542,
      "step": 1620
    },
    {
      "epoch": 18.055555555555557,
      "grad_norm": 0.4737277925014496,
      "learning_rate": 5.401234567901234e-06,
      "loss": 0.1447,
      "step": 1625
    },
    {
      "epoch": 18.333333333333332,
      "grad_norm": 0.2812809944152832,
      "learning_rate": 4.6296296296296296e-06,
      "loss": 0.1418,
      "step": 1650
    },
    {
      "epoch": 18.61111111111111,
      "grad_norm": 0.3522038161754608,
      "learning_rate": 3.858024691358025e-06,
      "loss": 0.1337,
      "step": 1675
    },
    {
      "epoch": 18.88888888888889,
      "grad_norm": 0.34057098627090454,
      "learning_rate": 3.0864197530864196e-06,
      "loss": 0.1364,
      "step": 1700
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.7460474308300395,
      "eval_loss": 0.7015388607978821,
      "eval_runtime": 29.1817,
      "eval_samples_per_second": 34.679,
      "eval_steps_per_second": 0.548,
      "step": 1710
    },
    {
      "epoch": 19.166666666666668,
      "grad_norm": 0.4172286093235016,
      "learning_rate": 2.3148148148148148e-06,
      "loss": 0.1344,
      "step": 1725
    },
    {
      "epoch": 19.444444444444443,
      "grad_norm": 0.5193507671356201,
      "learning_rate": 1.5432098765432098e-06,
      "loss": 0.1415,
      "step": 1750
    },
    {
      "epoch": 19.72222222222222,
      "grad_norm": 0.3823259174823761,
      "learning_rate": 7.716049382716049e-07,
      "loss": 0.1379,
      "step": 1775
    },
    {
      "epoch": 20.0,
      "grad_norm": 0.4077320694923401,
      "learning_rate": 0.0,
      "loss": 0.1329,
      "step": 1800
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.7460474308300395,
      "eval_loss": 0.7007161378860474,
      "eval_runtime": 29.7968,
      "eval_samples_per_second": 33.963,
      "eval_steps_per_second": 0.537,
      "step": 1800
    },
    {
      "epoch": 20.0,
      "step": 1800,
      "total_flos": 0.0,
      "train_loss": 0.31672255052460563,
      "train_runtime": 10514.5803,
      "train_samples_per_second": 10.899,
      "train_steps_per_second": 0.171
    }
  ],
  "logging_steps": 25,
  "max_steps": 1800,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}