{
  "best_metric": 0.5912562012672424,
  "best_model_checkpoint": "./kd_results/microsoft/swin-tiny-patch4-window7-224_alpha0.5_temp5.0/checkpoint-1350",
  "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": 4.644997596740723,
      "learning_rate": 6.944444444444445e-06,
      "loss": 1.2487,
      "step": 25
    },
    {
      "epoch": 0.5555555555555556,
      "grad_norm": 3.5007596015930176,
      "learning_rate": 1.388888888888889e-05,
      "loss": 1.2211,
      "step": 50
    },
    {
      "epoch": 0.8333333333333334,
      "grad_norm": 5.315507888793945,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 1.2082,
      "step": 75
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.26679841897233203,
      "eval_loss": 1.4079054594039917,
      "eval_runtime": 30.1961,
      "eval_samples_per_second": 33.514,
      "eval_steps_per_second": 0.53,
      "step": 90
    },
    {
      "epoch": 1.1111111111111112,
      "grad_norm": 5.468733310699463,
      "learning_rate": 2.777777777777778e-05,
      "loss": 1.19,
      "step": 100
    },
    {
      "epoch": 1.3888888888888888,
      "grad_norm": 4.748199462890625,
      "learning_rate": 3.472222222222222e-05,
      "loss": 1.1606,
      "step": 125
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 4.719135761260986,
      "learning_rate": 4.166666666666667e-05,
      "loss": 1.1654,
      "step": 150
    },
    {
      "epoch": 1.9444444444444444,
      "grad_norm": 7.299838066101074,
      "learning_rate": 4.8611111111111115e-05,
      "loss": 1.0779,
      "step": 175
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.43774703557312256,
      "eval_loss": 1.2161993980407715,
      "eval_runtime": 30.3579,
      "eval_samples_per_second": 33.336,
      "eval_steps_per_second": 0.527,
      "step": 180
    },
    {
      "epoch": 2.2222222222222223,
      "grad_norm": 9.123008728027344,
      "learning_rate": 4.938271604938271e-05,
      "loss": 0.9602,
      "step": 200
    },
    {
      "epoch": 2.5,
      "grad_norm": 4.930814266204834,
      "learning_rate": 4.8611111111111115e-05,
      "loss": 0.8791,
      "step": 225
    },
    {
      "epoch": 2.7777777777777777,
      "grad_norm": 10.713147163391113,
      "learning_rate": 4.783950617283951e-05,
      "loss": 0.8302,
      "step": 250
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.616600790513834,
      "eval_loss": 0.9250781536102295,
      "eval_runtime": 34.1996,
      "eval_samples_per_second": 29.591,
      "eval_steps_per_second": 0.468,
      "step": 270
    },
    {
      "epoch": 3.0555555555555554,
      "grad_norm": 7.605921745300293,
      "learning_rate": 4.70679012345679e-05,
      "loss": 0.7859,
      "step": 275
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 5.485124588012695,
      "learning_rate": 4.62962962962963e-05,
      "loss": 0.637,
      "step": 300
    },
    {
      "epoch": 3.611111111111111,
      "grad_norm": 7.356892108917236,
      "learning_rate": 4.5524691358024696e-05,
      "loss": 0.6399,
      "step": 325
    },
    {
      "epoch": 3.888888888888889,
      "grad_norm": 5.9463958740234375,
      "learning_rate": 4.4753086419753084e-05,
      "loss": 0.628,
      "step": 350
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7144268774703557,
      "eval_loss": 0.7441917657852173,
      "eval_runtime": 31.3566,
      "eval_samples_per_second": 32.274,
      "eval_steps_per_second": 0.51,
      "step": 360
    },
    {
      "epoch": 4.166666666666667,
      "grad_norm": 6.7027788162231445,
      "learning_rate": 4.3981481481481486e-05,
      "loss": 0.5377,
      "step": 375
    },
    {
      "epoch": 4.444444444444445,
      "grad_norm": 9.396843910217285,
      "learning_rate": 4.3209876543209875e-05,
      "loss": 0.4929,
      "step": 400
    },
    {
      "epoch": 4.722222222222222,
      "grad_norm": 7.425771713256836,
      "learning_rate": 4.243827160493827e-05,
      "loss": 0.4842,
      "step": 425
    },
    {
      "epoch": 5.0,
      "grad_norm": 7.368725776672363,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.4817,
      "step": 450
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7539525691699605,
      "eval_loss": 0.6760750412940979,
      "eval_runtime": 31.4867,
      "eval_samples_per_second": 32.141,
      "eval_steps_per_second": 0.508,
      "step": 450
    },
    {
      "epoch": 5.277777777777778,
      "grad_norm": 3.3067617416381836,
      "learning_rate": 4.089506172839506e-05,
      "loss": 0.3911,
      "step": 475
    },
    {
      "epoch": 5.555555555555555,
      "grad_norm": 4.621086120605469,
      "learning_rate": 4.012345679012346e-05,
      "loss": 0.3781,
      "step": 500
    },
    {
      "epoch": 5.833333333333333,
      "grad_norm": 6.426018714904785,
      "learning_rate": 3.935185185185186e-05,
      "loss": 0.378,
      "step": 525
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7559288537549407,
      "eval_loss": 0.7097513675689697,
      "eval_runtime": 31.3257,
      "eval_samples_per_second": 32.306,
      "eval_steps_per_second": 0.511,
      "step": 540
    },
    {
      "epoch": 6.111111111111111,
      "grad_norm": 5.1462297439575195,
      "learning_rate": 3.8580246913580246e-05,
      "loss": 0.3532,
      "step": 550
    },
    {
      "epoch": 6.388888888888889,
      "grad_norm": 5.476616859436035,
      "learning_rate": 3.780864197530865e-05,
      "loss": 0.3103,
      "step": 575
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 5.179195404052734,
      "learning_rate": 3.7037037037037037e-05,
      "loss": 0.3165,
      "step": 600
    },
    {
      "epoch": 6.944444444444445,
      "grad_norm": 4.515374660491943,
      "learning_rate": 3.626543209876543e-05,
      "loss": 0.3129,
      "step": 625
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.7737154150197628,
      "eval_loss": 0.6407632231712341,
      "eval_runtime": 31.4456,
      "eval_samples_per_second": 32.183,
      "eval_steps_per_second": 0.509,
      "step": 630
    },
    {
      "epoch": 7.222222222222222,
      "grad_norm": 3.401458740234375,
      "learning_rate": 3.5493827160493834e-05,
      "loss": 0.2719,
      "step": 650
    },
    {
      "epoch": 7.5,
      "grad_norm": 6.166269779205322,
      "learning_rate": 3.472222222222222e-05,
      "loss": 0.2642,
      "step": 675
    },
    {
      "epoch": 7.777777777777778,
      "grad_norm": 3.761406183242798,
      "learning_rate": 3.395061728395062e-05,
      "loss": 0.2627,
      "step": 700
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.7806324110671937,
      "eval_loss": 0.6368152499198914,
      "eval_runtime": 30.8953,
      "eval_samples_per_second": 32.756,
      "eval_steps_per_second": 0.518,
      "step": 720
    },
    {
      "epoch": 8.055555555555555,
      "grad_norm": 2.4448399543762207,
      "learning_rate": 3.317901234567901e-05,
      "loss": 0.2608,
      "step": 725
    },
    {
      "epoch": 8.333333333333334,
      "grad_norm": 3.3776795864105225,
      "learning_rate": 3.240740740740741e-05,
      "loss": 0.2297,
      "step": 750
    },
    {
      "epoch": 8.61111111111111,
      "grad_norm": 4.182880878448486,
      "learning_rate": 3.16358024691358e-05,
      "loss": 0.2377,
      "step": 775
    },
    {
      "epoch": 8.88888888888889,
      "grad_norm": 4.512351989746094,
      "learning_rate": 3.08641975308642e-05,
      "loss": 0.2458,
      "step": 800
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.7697628458498024,
      "eval_loss": 0.6722161173820496,
      "eval_runtime": 30.4764,
      "eval_samples_per_second": 33.206,
      "eval_steps_per_second": 0.525,
      "step": 810
    },
    {
      "epoch": 9.166666666666666,
      "grad_norm": 5.714383602142334,
      "learning_rate": 3.0092592592592593e-05,
      "loss": 0.2258,
      "step": 825
    },
    {
      "epoch": 9.444444444444445,
      "grad_norm": 3.8343403339385986,
      "learning_rate": 2.9320987654320992e-05,
      "loss": 0.2154,
      "step": 850
    },
    {
      "epoch": 9.722222222222221,
      "grad_norm": 3.6408908367156982,
      "learning_rate": 2.8549382716049384e-05,
      "loss": 0.2107,
      "step": 875
    },
    {
      "epoch": 10.0,
      "grad_norm": 5.553225040435791,
      "learning_rate": 2.777777777777778e-05,
      "loss": 0.2303,
      "step": 900
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.7806324110671937,
      "eval_loss": 0.6332230567932129,
      "eval_runtime": 29.4298,
      "eval_samples_per_second": 34.387,
      "eval_steps_per_second": 0.544,
      "step": 900
    },
    {
      "epoch": 10.277777777777779,
      "grad_norm": 2.42313289642334,
      "learning_rate": 2.700617283950617e-05,
      "loss": 0.2038,
      "step": 925
    },
    {
      "epoch": 10.555555555555555,
      "grad_norm": 3.482445478439331,
      "learning_rate": 2.623456790123457e-05,
      "loss": 0.2067,
      "step": 950
    },
    {
      "epoch": 10.833333333333334,
      "grad_norm": 4.0919084548950195,
      "learning_rate": 2.5462962962962965e-05,
      "loss": 0.2018,
      "step": 975
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.7885375494071146,
      "eval_loss": 0.6202763915061951,
      "eval_runtime": 30.7248,
      "eval_samples_per_second": 32.938,
      "eval_steps_per_second": 0.521,
      "step": 990
    },
    {
      "epoch": 11.11111111111111,
      "grad_norm": 2.223787546157837,
      "learning_rate": 2.4691358024691357e-05,
      "loss": 0.1936,
      "step": 1000
    },
    {
      "epoch": 11.38888888888889,
      "grad_norm": 2.527418375015259,
      "learning_rate": 2.3919753086419755e-05,
      "loss": 0.1905,
      "step": 1025
    },
    {
      "epoch": 11.666666666666666,
      "grad_norm": 1.6127002239227295,
      "learning_rate": 2.314814814814815e-05,
      "loss": 0.1959,
      "step": 1050
    },
    {
      "epoch": 11.944444444444445,
      "grad_norm": 3.127549409866333,
      "learning_rate": 2.2376543209876542e-05,
      "loss": 0.1942,
      "step": 1075
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.7885375494071146,
      "eval_loss": 0.6080687046051025,
      "eval_runtime": 30.2917,
      "eval_samples_per_second": 33.408,
      "eval_steps_per_second": 0.528,
      "step": 1080
    },
    {
      "epoch": 12.222222222222221,
      "grad_norm": 3.3274600505828857,
      "learning_rate": 2.1604938271604937e-05,
      "loss": 0.1888,
      "step": 1100
    },
    {
      "epoch": 12.5,
      "grad_norm": 2.4724836349487305,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 0.1923,
      "step": 1125
    },
    {
      "epoch": 12.777777777777779,
      "grad_norm": 2.305563449859619,
      "learning_rate": 2.006172839506173e-05,
      "loss": 0.1831,
      "step": 1150
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.7905138339920948,
      "eval_loss": 0.6040802597999573,
      "eval_runtime": 30.7894,
      "eval_samples_per_second": 32.868,
      "eval_steps_per_second": 0.52,
      "step": 1170
    },
    {
      "epoch": 13.055555555555555,
      "grad_norm": 2.3536362648010254,
      "learning_rate": 1.9290123456790123e-05,
      "loss": 0.1816,
      "step": 1175
    },
    {
      "epoch": 13.333333333333334,
      "grad_norm": 2.2629213333129883,
      "learning_rate": 1.8518518518518518e-05,
      "loss": 0.1887,
      "step": 1200
    },
    {
      "epoch": 13.61111111111111,
      "grad_norm": 3.821051836013794,
      "learning_rate": 1.7746913580246917e-05,
      "loss": 0.1851,
      "step": 1225
    },
    {
      "epoch": 13.88888888888889,
      "grad_norm": 2.3054168224334717,
      "learning_rate": 1.697530864197531e-05,
      "loss": 0.1827,
      "step": 1250
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.7895256916996047,
      "eval_loss": 0.6042537093162537,
      "eval_runtime": 30.3988,
      "eval_samples_per_second": 33.291,
      "eval_steps_per_second": 0.526,
      "step": 1260
    },
    {
      "epoch": 14.166666666666666,
      "grad_norm": 2.693080425262451,
      "learning_rate": 1.6203703703703704e-05,
      "loss": 0.1846,
      "step": 1275
    },
    {
      "epoch": 14.444444444444445,
      "grad_norm": 3.4559383392333984,
      "learning_rate": 1.54320987654321e-05,
      "loss": 0.18,
      "step": 1300
    },
    {
      "epoch": 14.722222222222221,
      "grad_norm": 4.556414604187012,
      "learning_rate": 1.4660493827160496e-05,
      "loss": 0.1724,
      "step": 1325
    },
    {
      "epoch": 15.0,
      "grad_norm": 3.5349297523498535,
      "learning_rate": 1.388888888888889e-05,
      "loss": 0.1825,
      "step": 1350
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.7974308300395256,
      "eval_loss": 0.5912562012672424,
      "eval_runtime": 30.1636,
      "eval_samples_per_second": 33.55,
      "eval_steps_per_second": 0.53,
      "step": 1350
    },
    {
      "epoch": 15.277777777777779,
      "grad_norm": 2.2571911811828613,
      "learning_rate": 1.3117283950617285e-05,
      "loss": 0.1735,
      "step": 1375
    },
    {
      "epoch": 15.555555555555555,
      "grad_norm": 1.961810827255249,
      "learning_rate": 1.2345679012345678e-05,
      "loss": 0.1722,
      "step": 1400
    },
    {
      "epoch": 15.833333333333334,
      "grad_norm": 1.5210778713226318,
      "learning_rate": 1.1574074074074075e-05,
      "loss": 0.1727,
      "step": 1425
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.7974308300395256,
      "eval_loss": 0.6063701510429382,
      "eval_runtime": 29.3895,
      "eval_samples_per_second": 34.434,
      "eval_steps_per_second": 0.544,
      "step": 1440
    },
    {
      "epoch": 16.11111111111111,
      "grad_norm": 3.2144644260406494,
      "learning_rate": 1.0802469135802469e-05,
      "loss": 0.1681,
      "step": 1450
    },
    {
      "epoch": 16.38888888888889,
      "grad_norm": 2.0110483169555664,
      "learning_rate": 1.0030864197530866e-05,
      "loss": 0.169,
      "step": 1475
    },
    {
      "epoch": 16.666666666666668,
      "grad_norm": 1.479304552078247,
      "learning_rate": 9.259259259259259e-06,
      "loss": 0.1679,
      "step": 1500
    },
    {
      "epoch": 16.944444444444443,
      "grad_norm": 1.2866755723953247,
      "learning_rate": 8.487654320987654e-06,
      "loss": 0.1701,
      "step": 1525
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.8033596837944664,
      "eval_loss": 0.6057149767875671,
      "eval_runtime": 29.9373,
      "eval_samples_per_second": 33.804,
      "eval_steps_per_second": 0.534,
      "step": 1530
    },
    {
      "epoch": 17.22222222222222,
      "grad_norm": 1.6458144187927246,
      "learning_rate": 7.71604938271605e-06,
      "loss": 0.1707,
      "step": 1550
    },
    {
      "epoch": 17.5,
      "grad_norm": 1.190675973892212,
      "learning_rate": 6.944444444444445e-06,
      "loss": 0.1626,
      "step": 1575
    },
    {
      "epoch": 17.77777777777778,
      "grad_norm": 1.7997467517852783,
      "learning_rate": 6.172839506172839e-06,
      "loss": 0.1595,
      "step": 1600
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.7924901185770751,
      "eval_loss": 0.5968884229660034,
      "eval_runtime": 29.8415,
      "eval_samples_per_second": 33.913,
      "eval_steps_per_second": 0.536,
      "step": 1620
    },
    {
      "epoch": 18.055555555555557,
      "grad_norm": 1.313234567642212,
      "learning_rate": 5.401234567901234e-06,
      "loss": 0.1663,
      "step": 1625
    },
    {
      "epoch": 18.333333333333332,
      "grad_norm": 1.4124459028244019,
      "learning_rate": 4.6296296296296296e-06,
      "loss": 0.1677,
      "step": 1650
    },
    {
      "epoch": 18.61111111111111,
      "grad_norm": 2.6002743244171143,
      "learning_rate": 3.858024691358025e-06,
      "loss": 0.1605,
      "step": 1675
    },
    {
      "epoch": 18.88888888888889,
      "grad_norm": 1.4194871187210083,
      "learning_rate": 3.0864197530864196e-06,
      "loss": 0.1616,
      "step": 1700
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.7964426877470355,
      "eval_loss": 0.5923498272895813,
      "eval_runtime": 30.1541,
      "eval_samples_per_second": 33.561,
      "eval_steps_per_second": 0.531,
      "step": 1710
    },
    {
      "epoch": 19.166666666666668,
      "grad_norm": 1.4580141305923462,
      "learning_rate": 2.3148148148148148e-06,
      "loss": 0.1572,
      "step": 1725
    },
    {
      "epoch": 19.444444444444443,
      "grad_norm": 1.037297248840332,
      "learning_rate": 1.5432098765432098e-06,
      "loss": 0.1635,
      "step": 1750
    },
    {
      "epoch": 19.72222222222222,
      "grad_norm": 1.164696216583252,
      "learning_rate": 7.716049382716049e-07,
      "loss": 0.1619,
      "step": 1775
    },
    {
      "epoch": 20.0,
      "grad_norm": 2.1098556518554688,
      "learning_rate": 0.0,
      "loss": 0.1594,
      "step": 1800
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.7964426877470355,
      "eval_loss": 0.5932034850120544,
      "eval_runtime": 30.4907,
      "eval_samples_per_second": 33.19,
      "eval_steps_per_second": 0.525,
      "step": 1800
    },
    {
      "epoch": 20.0,
      "step": 1800,
      "total_flos": 0.0,
      "train_loss": 0.37662137985229494,
      "train_runtime": 10637.343,
      "train_samples_per_second": 10.773,
      "train_steps_per_second": 0.169
    }
  ],
  "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
}