{
  "best_metric": 0.5081688165664673,
  "best_model_checkpoint": "./kd_results/microsoft/beit-base-patch16-384_alpha0.7_temp3.0/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": 5.21443510055542,
      "learning_rate": 6.944444444444445e-06,
      "loss": 1.2049,
      "step": 25
    },
    {
      "epoch": 0.5555555555555556,
      "grad_norm": 5.94625997543335,
      "learning_rate": 1.388888888888889e-05,
      "loss": 1.1563,
      "step": 50
    },
    {
      "epoch": 0.8333333333333334,
      "grad_norm": 4.836014747619629,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 1.1161,
      "step": 75
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.30335968379446643,
      "eval_loss": 1.395824670791626,
      "eval_runtime": 39.5759,
      "eval_samples_per_second": 25.571,
      "eval_steps_per_second": 0.404,
      "step": 90
    },
    {
      "epoch": 1.1111111111111112,
      "grad_norm": 7.636369228363037,
      "learning_rate": 2.777777777777778e-05,
      "loss": 1.1054,
      "step": 100
    },
    {
      "epoch": 1.3888888888888888,
      "grad_norm": 7.937506198883057,
      "learning_rate": 3.472222222222222e-05,
      "loss": 0.9892,
      "step": 125
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 9.13624095916748,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.9409,
      "step": 150
    },
    {
      "epoch": 1.9444444444444444,
      "grad_norm": 5.304960250854492,
      "learning_rate": 4.8611111111111115e-05,
      "loss": 0.8866,
      "step": 175
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5830039525691699,
      "eval_loss": 0.9934173822402954,
      "eval_runtime": 38.8493,
      "eval_samples_per_second": 26.049,
      "eval_steps_per_second": 0.412,
      "step": 180
    },
    {
      "epoch": 2.2222222222222223,
      "grad_norm": 7.893396854400635,
      "learning_rate": 4.938271604938271e-05,
      "loss": 0.6869,
      "step": 200
    },
    {
      "epoch": 2.5,
      "grad_norm": 5.390753746032715,
      "learning_rate": 4.8611111111111115e-05,
      "loss": 0.6352,
      "step": 225
    },
    {
      "epoch": 2.7777777777777777,
      "grad_norm": 5.973294258117676,
      "learning_rate": 4.783950617283951e-05,
      "loss": 0.5473,
      "step": 250
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7450592885375494,
      "eval_loss": 0.6915313601493835,
      "eval_runtime": 40.0437,
      "eval_samples_per_second": 25.272,
      "eval_steps_per_second": 0.4,
      "step": 270
    },
    {
      "epoch": 3.0555555555555554,
      "grad_norm": 4.769130706787109,
      "learning_rate": 4.70679012345679e-05,
      "loss": 0.4833,
      "step": 275
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 4.838928699493408,
      "learning_rate": 4.62962962962963e-05,
      "loss": 0.4083,
      "step": 300
    },
    {
      "epoch": 3.611111111111111,
      "grad_norm": 5.378454208374023,
      "learning_rate": 4.5524691358024696e-05,
      "loss": 0.372,
      "step": 325
    },
    {
      "epoch": 3.888888888888889,
      "grad_norm": 3.364759922027588,
      "learning_rate": 4.4753086419753084e-05,
      "loss": 0.4068,
      "step": 350
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7835968379446641,
      "eval_loss": 0.6147316098213196,
      "eval_runtime": 39.2106,
      "eval_samples_per_second": 25.809,
      "eval_steps_per_second": 0.408,
      "step": 360
    },
    {
      "epoch": 4.166666666666667,
      "grad_norm": 2.3518576622009277,
      "learning_rate": 4.3981481481481486e-05,
      "loss": 0.3311,
      "step": 375
    },
    {
      "epoch": 4.444444444444445,
      "grad_norm": 4.567741870880127,
      "learning_rate": 4.3209876543209875e-05,
      "loss": 0.2842,
      "step": 400
    },
    {
      "epoch": 4.722222222222222,
      "grad_norm": 5.316351413726807,
      "learning_rate": 4.243827160493827e-05,
      "loss": 0.3013,
      "step": 425
    },
    {
      "epoch": 5.0,
      "grad_norm": 4.875609874725342,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.2803,
      "step": 450
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8132411067193676,
      "eval_loss": 0.5612610578536987,
      "eval_runtime": 39.6839,
      "eval_samples_per_second": 25.502,
      "eval_steps_per_second": 0.403,
      "step": 450
    },
    {
      "epoch": 5.277777777777778,
      "grad_norm": 3.3461155891418457,
      "learning_rate": 4.089506172839506e-05,
      "loss": 0.2348,
      "step": 475
    },
    {
      "epoch": 5.555555555555555,
      "grad_norm": 4.080212593078613,
      "learning_rate": 4.012345679012346e-05,
      "loss": 0.2367,
      "step": 500
    },
    {
      "epoch": 5.833333333333333,
      "grad_norm": 3.0743508338928223,
      "learning_rate": 3.935185185185186e-05,
      "loss": 0.232,
      "step": 525
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8073122529644269,
      "eval_loss": 0.5583400130271912,
      "eval_runtime": 40.5296,
      "eval_samples_per_second": 24.969,
      "eval_steps_per_second": 0.395,
      "step": 540
    },
    {
      "epoch": 6.111111111111111,
      "grad_norm": 2.7358102798461914,
      "learning_rate": 3.8580246913580246e-05,
      "loss": 0.222,
      "step": 550
    },
    {
      "epoch": 6.388888888888889,
      "grad_norm": 2.1522977352142334,
      "learning_rate": 3.780864197530865e-05,
      "loss": 0.2081,
      "step": 575
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 5.203805446624756,
      "learning_rate": 3.7037037037037037e-05,
      "loss": 0.2082,
      "step": 600
    },
    {
      "epoch": 6.944444444444445,
      "grad_norm": 4.83523416519165,
      "learning_rate": 3.626543209876543e-05,
      "loss": 0.2132,
      "step": 625
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8102766798418972,
      "eval_loss": 0.5762213468551636,
      "eval_runtime": 41.0695,
      "eval_samples_per_second": 24.641,
      "eval_steps_per_second": 0.39,
      "step": 630
    },
    {
      "epoch": 7.222222222222222,
      "grad_norm": 4.739799976348877,
      "learning_rate": 3.5493827160493834e-05,
      "loss": 0.1997,
      "step": 650
    },
    {
      "epoch": 7.5,
      "grad_norm": 2.2983713150024414,
      "learning_rate": 3.472222222222222e-05,
      "loss": 0.1857,
      "step": 675
    },
    {
      "epoch": 7.777777777777778,
      "grad_norm": 1.8098427057266235,
      "learning_rate": 3.395061728395062e-05,
      "loss": 0.1911,
      "step": 700
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8191699604743083,
      "eval_loss": 0.5656912922859192,
      "eval_runtime": 40.3498,
      "eval_samples_per_second": 25.081,
      "eval_steps_per_second": 0.397,
      "step": 720
    },
    {
      "epoch": 8.055555555555555,
      "grad_norm": 1.349625587463379,
      "learning_rate": 3.317901234567901e-05,
      "loss": 0.1768,
      "step": 725
    },
    {
      "epoch": 8.333333333333334,
      "grad_norm": 1.4469423294067383,
      "learning_rate": 3.240740740740741e-05,
      "loss": 0.1764,
      "step": 750
    },
    {
      "epoch": 8.61111111111111,
      "grad_norm": 2.6558518409729004,
      "learning_rate": 3.16358024691358e-05,
      "loss": 0.1717,
      "step": 775
    },
    {
      "epoch": 8.88888888888889,
      "grad_norm": 1.6594892740249634,
      "learning_rate": 3.08641975308642e-05,
      "loss": 0.1705,
      "step": 800
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8102766798418972,
      "eval_loss": 0.5529190301895142,
      "eval_runtime": 40.4087,
      "eval_samples_per_second": 25.044,
      "eval_steps_per_second": 0.396,
      "step": 810
    },
    {
      "epoch": 9.166666666666666,
      "grad_norm": 2.369534492492676,
      "learning_rate": 3.0092592592592593e-05,
      "loss": 0.1593,
      "step": 825
    },
    {
      "epoch": 9.444444444444445,
      "grad_norm": 1.01450514793396,
      "learning_rate": 2.9320987654320992e-05,
      "loss": 0.1631,
      "step": 850
    },
    {
      "epoch": 9.722222222222221,
      "grad_norm": 1.973649263381958,
      "learning_rate": 2.8549382716049384e-05,
      "loss": 0.1633,
      "step": 875
    },
    {
      "epoch": 10.0,
      "grad_norm": 2.4213805198669434,
      "learning_rate": 2.777777777777778e-05,
      "loss": 0.1603,
      "step": 900
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.817193675889328,
      "eval_loss": 0.5434157252311707,
      "eval_runtime": 38.8087,
      "eval_samples_per_second": 26.077,
      "eval_steps_per_second": 0.412,
      "step": 900
    },
    {
      "epoch": 10.277777777777779,
      "grad_norm": 1.6117361783981323,
      "learning_rate": 2.700617283950617e-05,
      "loss": 0.154,
      "step": 925
    },
    {
      "epoch": 10.555555555555555,
      "grad_norm": 1.486670732498169,
      "learning_rate": 2.623456790123457e-05,
      "loss": 0.1562,
      "step": 950
    },
    {
      "epoch": 10.833333333333334,
      "grad_norm": 1.4686287641525269,
      "learning_rate": 2.5462962962962965e-05,
      "loss": 0.1553,
      "step": 975
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.8221343873517787,
      "eval_loss": 0.5208997130393982,
      "eval_runtime": 39.5335,
      "eval_samples_per_second": 25.599,
      "eval_steps_per_second": 0.405,
      "step": 990
    },
    {
      "epoch": 11.11111111111111,
      "grad_norm": 1.855955719947815,
      "learning_rate": 2.4691358024691357e-05,
      "loss": 0.1539,
      "step": 1000
    },
    {
      "epoch": 11.38888888888889,
      "grad_norm": 1.5377458333969116,
      "learning_rate": 2.3919753086419755e-05,
      "loss": 0.1521,
      "step": 1025
    },
    {
      "epoch": 11.666666666666666,
      "grad_norm": 1.1357789039611816,
      "learning_rate": 2.314814814814815e-05,
      "loss": 0.143,
      "step": 1050
    },
    {
      "epoch": 11.944444444444445,
      "grad_norm": 1.9014394283294678,
      "learning_rate": 2.2376543209876542e-05,
      "loss": 0.1475,
      "step": 1075
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.8181818181818182,
      "eval_loss": 0.5408589243888855,
      "eval_runtime": 39.6591,
      "eval_samples_per_second": 25.517,
      "eval_steps_per_second": 0.403,
      "step": 1080
    },
    {
      "epoch": 12.222222222222221,
      "grad_norm": 1.2245171070098877,
      "learning_rate": 2.1604938271604937e-05,
      "loss": 0.1406,
      "step": 1100
    },
    {
      "epoch": 12.5,
      "grad_norm": 1.3014460802078247,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 0.1434,
      "step": 1125
    },
    {
      "epoch": 12.777777777777779,
      "grad_norm": 1.600350022315979,
      "learning_rate": 2.006172839506173e-05,
      "loss": 0.1446,
      "step": 1150
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.83399209486166,
      "eval_loss": 0.5162519812583923,
      "eval_runtime": 39.4145,
      "eval_samples_per_second": 25.676,
      "eval_steps_per_second": 0.406,
      "step": 1170
    },
    {
      "epoch": 13.055555555555555,
      "grad_norm": 1.1929854154586792,
      "learning_rate": 1.9290123456790123e-05,
      "loss": 0.1316,
      "step": 1175
    },
    {
      "epoch": 13.333333333333334,
      "grad_norm": 1.6338070631027222,
      "learning_rate": 1.8518518518518518e-05,
      "loss": 0.1387,
      "step": 1200
    },
    {
      "epoch": 13.61111111111111,
      "grad_norm": 1.1349387168884277,
      "learning_rate": 1.7746913580246917e-05,
      "loss": 0.1411,
      "step": 1225
    },
    {
      "epoch": 13.88888888888889,
      "grad_norm": 0.7745327949523926,
      "learning_rate": 1.697530864197531e-05,
      "loss": 0.1316,
      "step": 1250
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.825098814229249,
      "eval_loss": 0.5329201221466064,
      "eval_runtime": 38.7271,
      "eval_samples_per_second": 26.132,
      "eval_steps_per_second": 0.413,
      "step": 1260
    },
    {
      "epoch": 14.166666666666666,
      "grad_norm": 1.1600468158721924,
      "learning_rate": 1.6203703703703704e-05,
      "loss": 0.1286,
      "step": 1275
    },
    {
      "epoch": 14.444444444444445,
      "grad_norm": 1.093441128730774,
      "learning_rate": 1.54320987654321e-05,
      "loss": 0.1323,
      "step": 1300
    },
    {
      "epoch": 14.722222222222221,
      "grad_norm": 1.7087829113006592,
      "learning_rate": 1.4660493827160496e-05,
      "loss": 0.1345,
      "step": 1325
    },
    {
      "epoch": 15.0,
      "grad_norm": 1.0991994142532349,
      "learning_rate": 1.388888888888889e-05,
      "loss": 0.1317,
      "step": 1350
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.8389328063241107,
      "eval_loss": 0.5183126926422119,
      "eval_runtime": 38.8714,
      "eval_samples_per_second": 26.035,
      "eval_steps_per_second": 0.412,
      "step": 1350
    },
    {
      "epoch": 15.277777777777779,
      "grad_norm": 1.002341628074646,
      "learning_rate": 1.3117283950617285e-05,
      "loss": 0.1257,
      "step": 1375
    },
    {
      "epoch": 15.555555555555555,
      "grad_norm": 0.891163170337677,
      "learning_rate": 1.2345679012345678e-05,
      "loss": 0.1229,
      "step": 1400
    },
    {
      "epoch": 15.833333333333334,
      "grad_norm": 0.743229329586029,
      "learning_rate": 1.1574074074074075e-05,
      "loss": 0.1284,
      "step": 1425
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.8409090909090909,
      "eval_loss": 0.5081688165664673,
      "eval_runtime": 37.962,
      "eval_samples_per_second": 26.658,
      "eval_steps_per_second": 0.421,
      "step": 1440
    },
    {
      "epoch": 16.11111111111111,
      "grad_norm": 0.6938183903694153,
      "learning_rate": 1.0802469135802469e-05,
      "loss": 0.1287,
      "step": 1450
    },
    {
      "epoch": 16.38888888888889,
      "grad_norm": 0.9573596715927124,
      "learning_rate": 1.0030864197530866e-05,
      "loss": 0.1253,
      "step": 1475
    },
    {
      "epoch": 16.666666666666668,
      "grad_norm": 0.7567666172981262,
      "learning_rate": 9.259259259259259e-06,
      "loss": 0.1211,
      "step": 1500
    },
    {
      "epoch": 16.944444444444443,
      "grad_norm": 0.7587071657180786,
      "learning_rate": 8.487654320987654e-06,
      "loss": 0.1221,
      "step": 1525
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.8389328063241107,
      "eval_loss": 0.5179046988487244,
      "eval_runtime": 39.6955,
      "eval_samples_per_second": 25.494,
      "eval_steps_per_second": 0.403,
      "step": 1530
    },
    {
      "epoch": 17.22222222222222,
      "grad_norm": 1.9301629066467285,
      "learning_rate": 7.71604938271605e-06,
      "loss": 0.1272,
      "step": 1550
    },
    {
      "epoch": 17.5,
      "grad_norm": 0.6017625331878662,
      "learning_rate": 6.944444444444445e-06,
      "loss": 0.1151,
      "step": 1575
    },
    {
      "epoch": 17.77777777777778,
      "grad_norm": 0.6997950077056885,
      "learning_rate": 6.172839506172839e-06,
      "loss": 0.1204,
      "step": 1600
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.8369565217391305,
      "eval_loss": 0.5164082050323486,
      "eval_runtime": 39.4407,
      "eval_samples_per_second": 25.659,
      "eval_steps_per_second": 0.406,
      "step": 1620
    },
    {
      "epoch": 18.055555555555557,
      "grad_norm": 0.7749896049499512,
      "learning_rate": 5.401234567901234e-06,
      "loss": 0.1264,
      "step": 1625
    },
    {
      "epoch": 18.333333333333332,
      "grad_norm": 0.4899977445602417,
      "learning_rate": 4.6296296296296296e-06,
      "loss": 0.1173,
      "step": 1650
    },
    {
      "epoch": 18.61111111111111,
      "grad_norm": 0.5992828607559204,
      "learning_rate": 3.858024691358025e-06,
      "loss": 0.1201,
      "step": 1675
    },
    {
      "epoch": 18.88888888888889,
      "grad_norm": 0.7319512367248535,
      "learning_rate": 3.0864197530864196e-06,
      "loss": 0.1192,
      "step": 1700
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.8438735177865613,
      "eval_loss": 0.5116754174232483,
      "eval_runtime": 38.6717,
      "eval_samples_per_second": 26.169,
      "eval_steps_per_second": 0.414,
      "step": 1710
    },
    {
      "epoch": 19.166666666666668,
      "grad_norm": 0.6426663994789124,
      "learning_rate": 2.3148148148148148e-06,
      "loss": 0.1216,
      "step": 1725
    },
    {
      "epoch": 19.444444444444443,
      "grad_norm": 0.6498861312866211,
      "learning_rate": 1.5432098765432098e-06,
      "loss": 0.1182,
      "step": 1750
    },
    {
      "epoch": 19.72222222222222,
      "grad_norm": 0.7802283763885498,
      "learning_rate": 7.716049382716049e-07,
      "loss": 0.1164,
      "step": 1775
    },
    {
      "epoch": 20.0,
      "grad_norm": 0.6459742188453674,
      "learning_rate": 0.0,
      "loss": 0.1167,
      "step": 1800
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.8428853754940712,
      "eval_loss": 0.512320339679718,
      "eval_runtime": 40.1616,
      "eval_samples_per_second": 25.198,
      "eval_steps_per_second": 0.398,
      "step": 1800
    },
    {
      "epoch": 20.0,
      "step": 1800,
      "total_flos": 0.0,
      "train_loss": 0.28280486318800185,
      "train_runtime": 12369.536,
      "train_samples_per_second": 9.265,
      "train_steps_per_second": 0.146
    }
  ],
  "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
}