{
  "best_metric": 0.5012136101722717,
  "best_model_checkpoint": "./kd_results/microsoft/beit-base-patch16-384_alpha0.7_temp5.0/checkpoint-1440",
  "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": 5.184907913208008,
      "learning_rate": 6.944444444444445e-06,
      "loss": 1.1858,
      "step": 25
    },
    {
      "epoch": 0.5555555555555556,
      "grad_norm": 5.849033355712891,
      "learning_rate": 1.388888888888889e-05,
      "loss": 1.1366,
      "step": 50
    },
    {
      "epoch": 0.8333333333333334,
      "grad_norm": 5.295121669769287,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 1.098,
      "step": 75
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.30632411067193677,
      "eval_loss": 1.3922169208526611,
      "eval_runtime": 38.7879,
      "eval_samples_per_second": 26.091,
      "eval_steps_per_second": 0.412,
      "step": 90
    },
    {
      "epoch": 1.1111111111111112,
      "grad_norm": 7.918435573577881,
      "learning_rate": 2.777777777777778e-05,
      "loss": 1.0812,
      "step": 100
    },
    {
      "epoch": 1.3888888888888888,
      "grad_norm": 7.588355541229248,
      "learning_rate": 3.472222222222222e-05,
      "loss": 0.9704,
      "step": 125
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 8.122336387634277,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.9292,
      "step": 150
    },
    {
      "epoch": 1.9444444444444444,
      "grad_norm": 5.83294677734375,
      "learning_rate": 4.8611111111111115e-05,
      "loss": 0.8736,
      "step": 175
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6136363636363636,
      "eval_loss": 0.9673107266426086,
      "eval_runtime": 38.6388,
      "eval_samples_per_second": 26.191,
      "eval_steps_per_second": 0.414,
      "step": 180
    },
    {
      "epoch": 2.2222222222222223,
      "grad_norm": 5.593318462371826,
      "learning_rate": 4.938271604938271e-05,
      "loss": 0.6777,
      "step": 200
    },
    {
      "epoch": 2.5,
      "grad_norm": 3.0493323802948,
      "learning_rate": 4.8611111111111115e-05,
      "loss": 0.6155,
      "step": 225
    },
    {
      "epoch": 2.7777777777777777,
      "grad_norm": 6.693108081817627,
      "learning_rate": 4.783950617283951e-05,
      "loss": 0.5416,
      "step": 250
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7598814229249012,
      "eval_loss": 0.6812384724617004,
      "eval_runtime": 38.0797,
      "eval_samples_per_second": 26.576,
      "eval_steps_per_second": 0.42,
      "step": 270
    },
    {
      "epoch": 3.0555555555555554,
      "grad_norm": 3.002751588821411,
      "learning_rate": 4.70679012345679e-05,
      "loss": 0.4754,
      "step": 275
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 4.859847068786621,
      "learning_rate": 4.62962962962963e-05,
      "loss": 0.4004,
      "step": 300
    },
    {
      "epoch": 3.611111111111111,
      "grad_norm": 5.904755115509033,
      "learning_rate": 4.5524691358024696e-05,
      "loss": 0.3873,
      "step": 325
    },
    {
      "epoch": 3.888888888888889,
      "grad_norm": 3.493598461151123,
      "learning_rate": 4.4753086419753084e-05,
      "loss": 0.3877,
      "step": 350
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7984189723320159,
      "eval_loss": 0.5915946960449219,
      "eval_runtime": 38.2782,
      "eval_samples_per_second": 26.438,
      "eval_steps_per_second": 0.418,
      "step": 360
    },
    {
      "epoch": 4.166666666666667,
      "grad_norm": 2.4440574645996094,
      "learning_rate": 4.3981481481481486e-05,
      "loss": 0.3319,
      "step": 375
    },
    {
      "epoch": 4.444444444444445,
      "grad_norm": 4.575101375579834,
      "learning_rate": 4.3209876543209875e-05,
      "loss": 0.2766,
      "step": 400
    },
    {
      "epoch": 4.722222222222222,
      "grad_norm": 4.615421772003174,
      "learning_rate": 4.243827160493827e-05,
      "loss": 0.3046,
      "step": 425
    },
    {
      "epoch": 5.0,
      "grad_norm": 5.199356555938721,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.2839,
      "step": 450
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8063241106719368,
      "eval_loss": 0.5837989449501038,
      "eval_runtime": 37.2653,
      "eval_samples_per_second": 27.157,
      "eval_steps_per_second": 0.429,
      "step": 450
    },
    {
      "epoch": 5.277777777777778,
      "grad_norm": 3.1010215282440186,
      "learning_rate": 4.089506172839506e-05,
      "loss": 0.237,
      "step": 475
    },
    {
      "epoch": 5.555555555555555,
      "grad_norm": 2.9845540523529053,
      "learning_rate": 4.012345679012346e-05,
      "loss": 0.2372,
      "step": 500
    },
    {
      "epoch": 5.833333333333333,
      "grad_norm": 2.53747820854187,
      "learning_rate": 3.935185185185186e-05,
      "loss": 0.2368,
      "step": 525
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8181818181818182,
      "eval_loss": 0.549915611743927,
      "eval_runtime": 37.6349,
      "eval_samples_per_second": 26.89,
      "eval_steps_per_second": 0.425,
      "step": 540
    },
    {
      "epoch": 6.111111111111111,
      "grad_norm": 2.4945929050445557,
      "learning_rate": 3.8580246913580246e-05,
      "loss": 0.2242,
      "step": 550
    },
    {
      "epoch": 6.388888888888889,
      "grad_norm": 3.1071951389312744,
      "learning_rate": 3.780864197530865e-05,
      "loss": 0.2067,
      "step": 575
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 2.733633518218994,
      "learning_rate": 3.7037037037037037e-05,
      "loss": 0.2155,
      "step": 600
    },
    {
      "epoch": 6.944444444444445,
      "grad_norm": 4.4593987464904785,
      "learning_rate": 3.626543209876543e-05,
      "loss": 0.2149,
      "step": 625
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.7984189723320159,
      "eval_loss": 0.601948618888855,
      "eval_runtime": 39.5799,
      "eval_samples_per_second": 25.569,
      "eval_steps_per_second": 0.404,
      "step": 630
    },
    {
      "epoch": 7.222222222222222,
      "grad_norm": 3.548896074295044,
      "learning_rate": 3.5493827160493834e-05,
      "loss": 0.1999,
      "step": 650
    },
    {
      "epoch": 7.5,
      "grad_norm": 1.9559125900268555,
      "learning_rate": 3.472222222222222e-05,
      "loss": 0.1843,
      "step": 675
    },
    {
      "epoch": 7.777777777777778,
      "grad_norm": 2.4996557235717773,
      "learning_rate": 3.395061728395062e-05,
      "loss": 0.1883,
      "step": 700
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8053359683794467,
      "eval_loss": 0.5796774625778198,
      "eval_runtime": 38.3082,
      "eval_samples_per_second": 26.417,
      "eval_steps_per_second": 0.418,
      "step": 720
    },
    {
      "epoch": 8.055555555555555,
      "grad_norm": 2.0347297191619873,
      "learning_rate": 3.317901234567901e-05,
      "loss": 0.1838,
      "step": 725
    },
    {
      "epoch": 8.333333333333334,
      "grad_norm": 1.6269855499267578,
      "learning_rate": 3.240740740740741e-05,
      "loss": 0.1786,
      "step": 750
    },
    {
      "epoch": 8.61111111111111,
      "grad_norm": 1.7134425640106201,
      "learning_rate": 3.16358024691358e-05,
      "loss": 0.1739,
      "step": 775
    },
    {
      "epoch": 8.88888888888889,
      "grad_norm": 2.123713254928589,
      "learning_rate": 3.08641975308642e-05,
      "loss": 0.1724,
      "step": 800
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.825098814229249,
      "eval_loss": 0.5546149015426636,
      "eval_runtime": 40.3516,
      "eval_samples_per_second": 25.08,
      "eval_steps_per_second": 0.397,
      "step": 810
    },
    {
      "epoch": 9.166666666666666,
      "grad_norm": 1.7199671268463135,
      "learning_rate": 3.0092592592592593e-05,
      "loss": 0.1581,
      "step": 825
    },
    {
      "epoch": 9.444444444444445,
      "grad_norm": 1.6997764110565186,
      "learning_rate": 2.9320987654320992e-05,
      "loss": 0.1658,
      "step": 850
    },
    {
      "epoch": 9.722222222222221,
      "grad_norm": 1.5390872955322266,
      "learning_rate": 2.8549382716049384e-05,
      "loss": 0.1639,
      "step": 875
    },
    {
      "epoch": 10.0,
      "grad_norm": 1.5496571063995361,
      "learning_rate": 2.777777777777778e-05,
      "loss": 0.1587,
      "step": 900
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.8231225296442688,
      "eval_loss": 0.5657897591590881,
      "eval_runtime": 39.7046,
      "eval_samples_per_second": 25.488,
      "eval_steps_per_second": 0.403,
      "step": 900
    },
    {
      "epoch": 10.277777777777779,
      "grad_norm": 1.098414659500122,
      "learning_rate": 2.700617283950617e-05,
      "loss": 0.155,
      "step": 925
    },
    {
      "epoch": 10.555555555555555,
      "grad_norm": 0.9889047145843506,
      "learning_rate": 2.623456790123457e-05,
      "loss": 0.1584,
      "step": 950
    },
    {
      "epoch": 10.833333333333334,
      "grad_norm": 2.0119264125823975,
      "learning_rate": 2.5462962962962965e-05,
      "loss": 0.1592,
      "step": 975
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.8231225296442688,
      "eval_loss": 0.516467809677124,
      "eval_runtime": 38.3777,
      "eval_samples_per_second": 26.369,
      "eval_steps_per_second": 0.417,
      "step": 990
    },
    {
      "epoch": 11.11111111111111,
      "grad_norm": 2.5024330615997314,
      "learning_rate": 2.4691358024691357e-05,
      "loss": 0.1541,
      "step": 1000
    },
    {
      "epoch": 11.38888888888889,
      "grad_norm": 1.625465989112854,
      "learning_rate": 2.3919753086419755e-05,
      "loss": 0.1485,
      "step": 1025
    },
    {
      "epoch": 11.666666666666666,
      "grad_norm": 1.1128909587860107,
      "learning_rate": 2.314814814814815e-05,
      "loss": 0.1401,
      "step": 1050
    },
    {
      "epoch": 11.944444444444445,
      "grad_norm": 1.7594417333602905,
      "learning_rate": 2.2376543209876542e-05,
      "loss": 0.1455,
      "step": 1075
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.8300395256916996,
      "eval_loss": 0.5143478512763977,
      "eval_runtime": 39.6911,
      "eval_samples_per_second": 25.497,
      "eval_steps_per_second": 0.403,
      "step": 1080
    },
    {
      "epoch": 12.222222222222221,
      "grad_norm": 1.4825284481048584,
      "learning_rate": 2.1604938271604937e-05,
      "loss": 0.1392,
      "step": 1100
    },
    {
      "epoch": 12.5,
      "grad_norm": 1.117546796798706,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 0.1425,
      "step": 1125
    },
    {
      "epoch": 12.777777777777779,
      "grad_norm": 1.0373642444610596,
      "learning_rate": 2.006172839506173e-05,
      "loss": 0.1438,
      "step": 1150
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.8330039525691699,
      "eval_loss": 0.5300087332725525,
      "eval_runtime": 38.4193,
      "eval_samples_per_second": 26.341,
      "eval_steps_per_second": 0.416,
      "step": 1170
    },
    {
      "epoch": 13.055555555555555,
      "grad_norm": 1.4194252490997314,
      "learning_rate": 1.9290123456790123e-05,
      "loss": 0.132,
      "step": 1175
    },
    {
      "epoch": 13.333333333333334,
      "grad_norm": 1.4579600095748901,
      "learning_rate": 1.8518518518518518e-05,
      "loss": 0.1353,
      "step": 1200
    },
    {
      "epoch": 13.61111111111111,
      "grad_norm": 1.5753164291381836,
      "learning_rate": 1.7746913580246917e-05,
      "loss": 0.1379,
      "step": 1225
    },
    {
      "epoch": 13.88888888888889,
      "grad_norm": 1.0776703357696533,
      "learning_rate": 1.697530864197531e-05,
      "loss": 0.1303,
      "step": 1250
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.8270750988142292,
      "eval_loss": 0.5375888347625732,
      "eval_runtime": 39.1838,
      "eval_samples_per_second": 25.827,
      "eval_steps_per_second": 0.408,
      "step": 1260
    },
    {
      "epoch": 14.166666666666666,
      "grad_norm": 1.2338591814041138,
      "learning_rate": 1.6203703703703704e-05,
      "loss": 0.1285,
      "step": 1275
    },
    {
      "epoch": 14.444444444444445,
      "grad_norm": 1.0960110425949097,
      "learning_rate": 1.54320987654321e-05,
      "loss": 0.1312,
      "step": 1300
    },
    {
      "epoch": 14.722222222222221,
      "grad_norm": 1.6634916067123413,
      "learning_rate": 1.4660493827160496e-05,
      "loss": 0.1337,
      "step": 1325
    },
    {
      "epoch": 15.0,
      "grad_norm": 1.1385433673858643,
      "learning_rate": 1.388888888888889e-05,
      "loss": 0.1306,
      "step": 1350
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.8280632411067194,
      "eval_loss": 0.5235399007797241,
      "eval_runtime": 40.6053,
      "eval_samples_per_second": 24.923,
      "eval_steps_per_second": 0.394,
      "step": 1350
    },
    {
      "epoch": 15.277777777777779,
      "grad_norm": 0.9714440703392029,
      "learning_rate": 1.3117283950617285e-05,
      "loss": 0.1249,
      "step": 1375
    },
    {
      "epoch": 15.555555555555555,
      "grad_norm": 0.6641523838043213,
      "learning_rate": 1.2345679012345678e-05,
      "loss": 0.1234,
      "step": 1400
    },
    {
      "epoch": 15.833333333333334,
      "grad_norm": 0.7077529430389404,
      "learning_rate": 1.1574074074074075e-05,
      "loss": 0.1274,
      "step": 1425
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.8399209486166008,
      "eval_loss": 0.5012136101722717,
      "eval_runtime": 39.893,
      "eval_samples_per_second": 25.368,
      "eval_steps_per_second": 0.401,
      "step": 1440
    },
    {
      "epoch": 16.11111111111111,
      "grad_norm": 0.6755304336547852,
      "learning_rate": 1.0802469135802469e-05,
      "loss": 0.1278,
      "step": 1450
    },
    {
      "epoch": 16.38888888888889,
      "grad_norm": 0.9831761121749878,
      "learning_rate": 1.0030864197530866e-05,
      "loss": 0.1242,
      "step": 1475
    },
    {
      "epoch": 16.666666666666668,
      "grad_norm": 0.6216269135475159,
      "learning_rate": 9.259259259259259e-06,
      "loss": 0.1194,
      "step": 1500
    },
    {
      "epoch": 16.944444444444443,
      "grad_norm": 0.6754944324493408,
      "learning_rate": 8.487654320987654e-06,
      "loss": 0.121,
      "step": 1525
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.8369565217391305,
      "eval_loss": 0.5206666588783264,
      "eval_runtime": 38.6688,
      "eval_samples_per_second": 26.171,
      "eval_steps_per_second": 0.414,
      "step": 1530
    },
    {
      "epoch": 17.22222222222222,
      "grad_norm": 1.1738343238830566,
      "learning_rate": 7.71604938271605e-06,
      "loss": 0.1257,
      "step": 1550
    },
    {
      "epoch": 17.5,
      "grad_norm": 0.6815004348754883,
      "learning_rate": 6.944444444444445e-06,
      "loss": 0.1134,
      "step": 1575
    },
    {
      "epoch": 17.77777777777778,
      "grad_norm": 0.6849784255027771,
      "learning_rate": 6.172839506172839e-06,
      "loss": 0.1197,
      "step": 1600
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.8369565217391305,
      "eval_loss": 0.5180361270904541,
      "eval_runtime": 40.2512,
      "eval_samples_per_second": 25.142,
      "eval_steps_per_second": 0.398,
      "step": 1620
    },
    {
      "epoch": 18.055555555555557,
      "grad_norm": 0.7509155869483948,
      "learning_rate": 5.401234567901234e-06,
      "loss": 0.1252,
      "step": 1625
    },
    {
      "epoch": 18.333333333333332,
      "grad_norm": 0.6722740530967712,
      "learning_rate": 4.6296296296296296e-06,
      "loss": 0.1158,
      "step": 1650
    },
    {
      "epoch": 18.61111111111111,
      "grad_norm": 0.5532152652740479,
      "learning_rate": 3.858024691358025e-06,
      "loss": 0.119,
      "step": 1675
    },
    {
      "epoch": 18.88888888888889,
      "grad_norm": 0.6120570302009583,
      "learning_rate": 3.0864197530864196e-06,
      "loss": 0.1189,
      "step": 1700
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.8409090909090909,
      "eval_loss": 0.5124243497848511,
      "eval_runtime": 39.1972,
      "eval_samples_per_second": 25.818,
      "eval_steps_per_second": 0.408,
      "step": 1710
    },
    {
      "epoch": 19.166666666666668,
      "grad_norm": 0.6396480202674866,
      "learning_rate": 2.3148148148148148e-06,
      "loss": 0.1204,
      "step": 1725
    },
    {
      "epoch": 19.444444444444443,
      "grad_norm": 0.5467541217803955,
      "learning_rate": 1.5432098765432098e-06,
      "loss": 0.1175,
      "step": 1750
    },
    {
      "epoch": 19.72222222222222,
      "grad_norm": 0.7086827754974365,
      "learning_rate": 7.716049382716049e-07,
      "loss": 0.1155,
      "step": 1775
    },
    {
      "epoch": 20.0,
      "grad_norm": 1.0691556930541992,
      "learning_rate": 0.0,
      "loss": 0.1157,
      "step": 1800
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.841897233201581,
      "eval_loss": 0.5116459131240845,
      "eval_runtime": 38.7759,
      "eval_samples_per_second": 26.099,
      "eval_steps_per_second": 0.413,
      "step": 1800
    },
    {
      "epoch": 20.0,
      "step": 1800,
      "total_flos": 0.0,
      "train_loss": 0.28024863150384693,
      "train_runtime": 12279.1623,
      "train_samples_per_second": 9.333,
      "train_steps_per_second": 0.147
    }
  ],
  "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
}