{
  "best_metric": 0.8025210499763489,
  "best_model_checkpoint": "./kd_results/facebook/deit-tiny-distilled-patch16-224_alpha0.5_temp5.0/checkpoint-1800",
  "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": 3.6476805210113525,
      "learning_rate": 6.944444444444445e-06,
      "loss": 1.3499,
      "step": 25
    },
    {
      "epoch": 0.5555555555555556,
      "grad_norm": 2.6681578159332275,
      "learning_rate": 1.388888888888889e-05,
      "loss": 1.2143,
      "step": 50
    },
    {
      "epoch": 0.8333333333333334,
      "grad_norm": 3.053227186203003,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 1.207,
      "step": 75
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.28952569169960474,
      "eval_loss": 1.4124675989151,
      "eval_runtime": 29.0528,
      "eval_samples_per_second": 34.833,
      "eval_steps_per_second": 0.551,
      "step": 90
    },
    {
      "epoch": 1.1111111111111112,
      "grad_norm": 2.9482927322387695,
      "learning_rate": 2.777777777777778e-05,
      "loss": 1.181,
      "step": 100
    },
    {
      "epoch": 1.3888888888888888,
      "grad_norm": 3.6009232997894287,
      "learning_rate": 3.472222222222222e-05,
      "loss": 1.147,
      "step": 125
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 3.493739604949951,
      "learning_rate": 4.166666666666667e-05,
      "loss": 1.1264,
      "step": 150
    },
    {
      "epoch": 1.9444444444444444,
      "grad_norm": 4.056971073150635,
      "learning_rate": 4.8611111111111115e-05,
      "loss": 1.0715,
      "step": 175
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.41205533596837945,
      "eval_loss": 1.2813795804977417,
      "eval_runtime": 29.1904,
      "eval_samples_per_second": 34.669,
      "eval_steps_per_second": 0.548,
      "step": 180
    },
    {
      "epoch": 2.2222222222222223,
      "grad_norm": 8.939580917358398,
      "learning_rate": 4.938271604938271e-05,
      "loss": 1.0009,
      "step": 200
    },
    {
      "epoch": 2.5,
      "grad_norm": 5.8859758377075195,
      "learning_rate": 4.8611111111111115e-05,
      "loss": 0.9391,
      "step": 225
    },
    {
      "epoch": 2.7777777777777777,
      "grad_norm": 4.805037021636963,
      "learning_rate": 4.783950617283951e-05,
      "loss": 0.9394,
      "step": 250
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5049407114624506,
      "eval_loss": 1.1279006004333496,
      "eval_runtime": 28.909,
      "eval_samples_per_second": 35.006,
      "eval_steps_per_second": 0.553,
      "step": 270
    },
    {
      "epoch": 3.0555555555555554,
      "grad_norm": 6.43830680847168,
      "learning_rate": 4.70679012345679e-05,
      "loss": 0.8696,
      "step": 275
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 5.744662761688232,
      "learning_rate": 4.62962962962963e-05,
      "loss": 0.7338,
      "step": 300
    },
    {
      "epoch": 3.611111111111111,
      "grad_norm": 5.218618392944336,
      "learning_rate": 4.5524691358024696e-05,
      "loss": 0.7265,
      "step": 325
    },
    {
      "epoch": 3.888888888888889,
      "grad_norm": 6.001340866088867,
      "learning_rate": 4.4753086419753084e-05,
      "loss": 0.6911,
      "step": 350
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6215415019762845,
      "eval_loss": 0.93367999792099,
      "eval_runtime": 28.6221,
      "eval_samples_per_second": 35.357,
      "eval_steps_per_second": 0.559,
      "step": 360
    },
    {
      "epoch": 4.166666666666667,
      "grad_norm": 6.301717758178711,
      "learning_rate": 4.3981481481481486e-05,
      "loss": 0.5784,
      "step": 375
    },
    {
      "epoch": 4.444444444444445,
      "grad_norm": 8.525504112243652,
      "learning_rate": 4.3209876543209875e-05,
      "loss": 0.5074,
      "step": 400
    },
    {
      "epoch": 4.722222222222222,
      "grad_norm": 7.204672336578369,
      "learning_rate": 4.243827160493827e-05,
      "loss": 0.5119,
      "step": 425
    },
    {
      "epoch": 5.0,
      "grad_norm": 8.81457233428955,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.4932,
      "step": 450
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6551383399209486,
      "eval_loss": 0.8380208611488342,
      "eval_runtime": 28.8464,
      "eval_samples_per_second": 35.082,
      "eval_steps_per_second": 0.555,
      "step": 450
    },
    {
      "epoch": 5.277777777777778,
      "grad_norm": 7.421513557434082,
      "learning_rate": 4.089506172839506e-05,
      "loss": 0.3899,
      "step": 475
    },
    {
      "epoch": 5.555555555555555,
      "grad_norm": 4.949956893920898,
      "learning_rate": 4.012345679012346e-05,
      "loss": 0.3503,
      "step": 500
    },
    {
      "epoch": 5.833333333333333,
      "grad_norm": 5.17600154876709,
      "learning_rate": 3.935185185185186e-05,
      "loss": 0.3481,
      "step": 525
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6432806324110671,
      "eval_loss": 0.8895840048789978,
      "eval_runtime": 29.2381,
      "eval_samples_per_second": 34.612,
      "eval_steps_per_second": 0.547,
      "step": 540
    },
    {
      "epoch": 6.111111111111111,
      "grad_norm": 4.453238487243652,
      "learning_rate": 3.8580246913580246e-05,
      "loss": 0.338,
      "step": 550
    },
    {
      "epoch": 6.388888888888889,
      "grad_norm": 5.4958343505859375,
      "learning_rate": 3.780864197530865e-05,
      "loss": 0.2561,
      "step": 575
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 8.34775161743164,
      "learning_rate": 3.7037037037037037e-05,
      "loss": 0.2657,
      "step": 600
    },
    {
      "epoch": 6.944444444444445,
      "grad_norm": 5.20697546005249,
      "learning_rate": 3.626543209876543e-05,
      "loss": 0.2781,
      "step": 625
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6620553359683794,
      "eval_loss": 0.8509350419044495,
      "eval_runtime": 28.8664,
      "eval_samples_per_second": 35.058,
      "eval_steps_per_second": 0.554,
      "step": 630
    },
    {
      "epoch": 7.222222222222222,
      "grad_norm": 3.5376052856445312,
      "learning_rate": 3.5493827160493834e-05,
      "loss": 0.2287,
      "step": 650
    },
    {
      "epoch": 7.5,
      "grad_norm": 3.5099503993988037,
      "learning_rate": 3.472222222222222e-05,
      "loss": 0.2179,
      "step": 675
    },
    {
      "epoch": 7.777777777777778,
      "grad_norm": 3.958256483078003,
      "learning_rate": 3.395061728395062e-05,
      "loss": 0.2196,
      "step": 700
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.674901185770751,
      "eval_loss": 0.8347586989402771,
      "eval_runtime": 28.7104,
      "eval_samples_per_second": 35.249,
      "eval_steps_per_second": 0.557,
      "step": 720
    },
    {
      "epoch": 8.055555555555555,
      "grad_norm": 2.8401408195495605,
      "learning_rate": 3.317901234567901e-05,
      "loss": 0.2187,
      "step": 725
    },
    {
      "epoch": 8.333333333333334,
      "grad_norm": 3.6627988815307617,
      "learning_rate": 3.240740740740741e-05,
      "loss": 0.1922,
      "step": 750
    },
    {
      "epoch": 8.61111111111111,
      "grad_norm": 3.0782864093780518,
      "learning_rate": 3.16358024691358e-05,
      "loss": 0.1996,
      "step": 775
    },
    {
      "epoch": 8.88888888888889,
      "grad_norm": 3.3091213703155518,
      "learning_rate": 3.08641975308642e-05,
      "loss": 0.2002,
      "step": 800
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.6640316205533597,
      "eval_loss": 0.8324930667877197,
      "eval_runtime": 29.273,
      "eval_samples_per_second": 34.571,
      "eval_steps_per_second": 0.547,
      "step": 810
    },
    {
      "epoch": 9.166666666666666,
      "grad_norm": 2.4105794429779053,
      "learning_rate": 3.0092592592592593e-05,
      "loss": 0.1874,
      "step": 825
    },
    {
      "epoch": 9.444444444444445,
      "grad_norm": 2.552306652069092,
      "learning_rate": 2.9320987654320992e-05,
      "loss": 0.1842,
      "step": 850
    },
    {
      "epoch": 9.722222222222221,
      "grad_norm": 3.465869426727295,
      "learning_rate": 2.8549382716049384e-05,
      "loss": 0.1821,
      "step": 875
    },
    {
      "epoch": 10.0,
      "grad_norm": 3.8747434616088867,
      "learning_rate": 2.777777777777778e-05,
      "loss": 0.1821,
      "step": 900
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.6699604743083004,
      "eval_loss": 0.8579540252685547,
      "eval_runtime": 29.2708,
      "eval_samples_per_second": 34.574,
      "eval_steps_per_second": 0.547,
      "step": 900
    },
    {
      "epoch": 10.277777777777779,
      "grad_norm": 2.2930424213409424,
      "learning_rate": 2.700617283950617e-05,
      "loss": 0.1782,
      "step": 925
    },
    {
      "epoch": 10.555555555555555,
      "grad_norm": 2.286329984664917,
      "learning_rate": 2.623456790123457e-05,
      "loss": 0.1787,
      "step": 950
    },
    {
      "epoch": 10.833333333333334,
      "grad_norm": 2.499483108520508,
      "learning_rate": 2.5462962962962965e-05,
      "loss": 0.1722,
      "step": 975
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.6818181818181818,
      "eval_loss": 0.8279414772987366,
      "eval_runtime": 28.7784,
      "eval_samples_per_second": 35.165,
      "eval_steps_per_second": 0.556,
      "step": 990
    },
    {
      "epoch": 11.11111111111111,
      "grad_norm": 1.292887806892395,
      "learning_rate": 2.4691358024691357e-05,
      "loss": 0.1675,
      "step": 1000
    },
    {
      "epoch": 11.38888888888889,
      "grad_norm": 1.8185718059539795,
      "learning_rate": 2.3919753086419755e-05,
      "loss": 0.1642,
      "step": 1025
    },
    {
      "epoch": 11.666666666666666,
      "grad_norm": 1.7777647972106934,
      "learning_rate": 2.314814814814815e-05,
      "loss": 0.175,
      "step": 1050
    },
    {
      "epoch": 11.944444444444445,
      "grad_norm": 1.606765627861023,
      "learning_rate": 2.2376543209876542e-05,
      "loss": 0.1726,
      "step": 1075
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.6897233201581028,
      "eval_loss": 0.8298330307006836,
      "eval_runtime": 28.7529,
      "eval_samples_per_second": 35.196,
      "eval_steps_per_second": 0.556,
      "step": 1080
    },
    {
      "epoch": 12.222222222222221,
      "grad_norm": 1.7732797861099243,
      "learning_rate": 2.1604938271604937e-05,
      "loss": 0.1644,
      "step": 1100
    },
    {
      "epoch": 12.5,
      "grad_norm": 2.459123373031616,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 0.1734,
      "step": 1125
    },
    {
      "epoch": 12.777777777777779,
      "grad_norm": 1.4211158752441406,
      "learning_rate": 2.006172839506173e-05,
      "loss": 0.1583,
      "step": 1150
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.6837944664031621,
      "eval_loss": 0.8128145933151245,
      "eval_runtime": 28.7373,
      "eval_samples_per_second": 35.216,
      "eval_steps_per_second": 0.557,
      "step": 1170
    },
    {
      "epoch": 13.055555555555555,
      "grad_norm": 2.5340707302093506,
      "learning_rate": 1.9290123456790123e-05,
      "loss": 0.159,
      "step": 1175
    },
    {
      "epoch": 13.333333333333334,
      "grad_norm": 1.599650502204895,
      "learning_rate": 1.8518518518518518e-05,
      "loss": 0.1619,
      "step": 1200
    },
    {
      "epoch": 13.61111111111111,
      "grad_norm": 2.159294366836548,
      "learning_rate": 1.7746913580246917e-05,
      "loss": 0.1616,
      "step": 1225
    },
    {
      "epoch": 13.88888888888889,
      "grad_norm": 2.541760206222534,
      "learning_rate": 1.697530864197531e-05,
      "loss": 0.1596,
      "step": 1250
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.6758893280632411,
      "eval_loss": 0.8156259655952454,
      "eval_runtime": 29.1036,
      "eval_samples_per_second": 34.772,
      "eval_steps_per_second": 0.55,
      "step": 1260
    },
    {
      "epoch": 14.166666666666666,
      "grad_norm": 1.4837608337402344,
      "learning_rate": 1.6203703703703704e-05,
      "loss": 0.1541,
      "step": 1275
    },
    {
      "epoch": 14.444444444444445,
      "grad_norm": 1.8819512128829956,
      "learning_rate": 1.54320987654321e-05,
      "loss": 0.152,
      "step": 1300
    },
    {
      "epoch": 14.722222222222221,
      "grad_norm": 1.5394985675811768,
      "learning_rate": 1.4660493827160496e-05,
      "loss": 0.1506,
      "step": 1325
    },
    {
      "epoch": 15.0,
      "grad_norm": 1.7922059297561646,
      "learning_rate": 1.388888888888889e-05,
      "loss": 0.1551,
      "step": 1350
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.6867588932806324,
      "eval_loss": 0.8225850462913513,
      "eval_runtime": 28.2463,
      "eval_samples_per_second": 35.828,
      "eval_steps_per_second": 0.566,
      "step": 1350
    },
    {
      "epoch": 15.277777777777779,
      "grad_norm": 1.3428395986557007,
      "learning_rate": 1.3117283950617285e-05,
      "loss": 0.1527,
      "step": 1375
    },
    {
      "epoch": 15.555555555555555,
      "grad_norm": 1.2559345960617065,
      "learning_rate": 1.2345679012345678e-05,
      "loss": 0.1512,
      "step": 1400
    },
    {
      "epoch": 15.833333333333334,
      "grad_norm": 1.0937647819519043,
      "learning_rate": 1.1574074074074075e-05,
      "loss": 0.1518,
      "step": 1425
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.6857707509881423,
      "eval_loss": 0.8178162574768066,
      "eval_runtime": 28.862,
      "eval_samples_per_second": 35.063,
      "eval_steps_per_second": 0.554,
      "step": 1440
    },
    {
      "epoch": 16.11111111111111,
      "grad_norm": 0.9090328812599182,
      "learning_rate": 1.0802469135802469e-05,
      "loss": 0.1453,
      "step": 1450
    },
    {
      "epoch": 16.38888888888889,
      "grad_norm": 0.8728322982788086,
      "learning_rate": 1.0030864197530866e-05,
      "loss": 0.148,
      "step": 1475
    },
    {
      "epoch": 16.666666666666668,
      "grad_norm": 0.7685718536376953,
      "learning_rate": 9.259259259259259e-06,
      "loss": 0.1493,
      "step": 1500
    },
    {
      "epoch": 16.944444444444443,
      "grad_norm": 1.0115023851394653,
      "learning_rate": 8.487654320987654e-06,
      "loss": 0.1485,
      "step": 1525
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.6837944664031621,
      "eval_loss": 0.8044672608375549,
      "eval_runtime": 29.1269,
      "eval_samples_per_second": 34.744,
      "eval_steps_per_second": 0.549,
      "step": 1530
    },
    {
      "epoch": 17.22222222222222,
      "grad_norm": 0.8916856646537781,
      "learning_rate": 7.71604938271605e-06,
      "loss": 0.1474,
      "step": 1550
    },
    {
      "epoch": 17.5,
      "grad_norm": 0.7983067035675049,
      "learning_rate": 6.944444444444445e-06,
      "loss": 0.1442,
      "step": 1575
    },
    {
      "epoch": 17.77777777777778,
      "grad_norm": 0.9365713596343994,
      "learning_rate": 6.172839506172839e-06,
      "loss": 0.1404,
      "step": 1600
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.6768774703557312,
      "eval_loss": 0.8060990571975708,
      "eval_runtime": 28.3597,
      "eval_samples_per_second": 35.684,
      "eval_steps_per_second": 0.564,
      "step": 1620
    },
    {
      "epoch": 18.055555555555557,
      "grad_norm": 1.0071580410003662,
      "learning_rate": 5.401234567901234e-06,
      "loss": 0.1501,
      "step": 1625
    },
    {
      "epoch": 18.333333333333332,
      "grad_norm": 0.9413686394691467,
      "learning_rate": 4.6296296296296296e-06,
      "loss": 0.1481,
      "step": 1650
    },
    {
      "epoch": 18.61111111111111,
      "grad_norm": 0.9161611795425415,
      "learning_rate": 3.858024691358025e-06,
      "loss": 0.1393,
      "step": 1675
    },
    {
      "epoch": 18.88888888888889,
      "grad_norm": 0.8299447894096375,
      "learning_rate": 3.0864197530864196e-06,
      "loss": 0.1415,
      "step": 1700
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.6877470355731226,
      "eval_loss": 0.8038429021835327,
      "eval_runtime": 29.0965,
      "eval_samples_per_second": 34.781,
      "eval_steps_per_second": 0.55,
      "step": 1710
    },
    {
      "epoch": 19.166666666666668,
      "grad_norm": 0.9107728600502014,
      "learning_rate": 2.3148148148148148e-06,
      "loss": 0.1393,
      "step": 1725
    },
    {
      "epoch": 19.444444444444443,
      "grad_norm": 0.8105676770210266,
      "learning_rate": 1.5432098765432098e-06,
      "loss": 0.1453,
      "step": 1750
    },
    {
      "epoch": 19.72222222222222,
      "grad_norm": 0.7728349566459656,
      "learning_rate": 7.716049382716049e-07,
      "loss": 0.1428,
      "step": 1775
    },
    {
      "epoch": 20.0,
      "grad_norm": 1.2281630039215088,
      "learning_rate": 0.0,
      "loss": 0.137,
      "step": 1800
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.6847826086956522,
      "eval_loss": 0.8025210499763489,
      "eval_runtime": 28.8277,
      "eval_samples_per_second": 35.105,
      "eval_steps_per_second": 0.555,
      "step": 1800
    },
    {
      "epoch": 20.0,
      "step": 1800,
      "total_flos": 0.0,
      "train_loss": 0.365383341577318,
      "train_runtime": 10382.0898,
      "train_samples_per_second": 11.038,
      "train_steps_per_second": 0.173
    }
  ],
  "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
}