{
  "best_metric": 0.8166666666666667,
  "best_model_checkpoint": "beit-base-patch16-224-dmae-va-U5-42/checkpoint-224",
  "epoch": 37.935483870967744,
  "eval_steps": 500,
  "global_step": 294,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.9,
      "eval_accuracy": 0.4666666666666667,
      "eval_loss": 1.347098708152771,
      "eval_runtime": 1.5986,
      "eval_samples_per_second": 37.533,
      "eval_steps_per_second": 1.251,
      "step": 7
    },
    {
      "epoch": 1.29,
      "grad_norm": 10.756444931030273,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 1.6023,
      "step": 10
    },
    {
      "epoch": 1.94,
      "eval_accuracy": 0.5833333333333334,
      "eval_loss": 1.0873388051986694,
      "eval_runtime": 1.6091,
      "eval_samples_per_second": 37.288,
      "eval_steps_per_second": 1.243,
      "step": 15
    },
    {
      "epoch": 2.58,
      "grad_norm": 9.189943313598633,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 1.1509,
      "step": 20
    },
    {
      "epoch": 2.97,
      "eval_accuracy": 0.5833333333333334,
      "eval_loss": 0.9947898983955383,
      "eval_runtime": 2.0991,
      "eval_samples_per_second": 28.584,
      "eval_steps_per_second": 0.953,
      "step": 23
    },
    {
      "epoch": 3.87,
      "grad_norm": 11.76810073852539,
      "learning_rate": 5e-05,
      "loss": 0.826,
      "step": 30
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7166666666666667,
      "eval_loss": 0.7244278192520142,
      "eval_runtime": 1.6205,
      "eval_samples_per_second": 37.025,
      "eval_steps_per_second": 1.234,
      "step": 31
    },
    {
      "epoch": 4.9,
      "eval_accuracy": 0.7333333333333333,
      "eval_loss": 0.5740881562232971,
      "eval_runtime": 2.1053,
      "eval_samples_per_second": 28.5,
      "eval_steps_per_second": 0.95,
      "step": 38
    },
    {
      "epoch": 5.16,
      "grad_norm": 10.483253479003906,
      "learning_rate": 4.810606060606061e-05,
      "loss": 0.5551,
      "step": 40
    },
    {
      "epoch": 5.94,
      "eval_accuracy": 0.75,
      "eval_loss": 0.6568958759307861,
      "eval_runtime": 2.0199,
      "eval_samples_per_second": 29.704,
      "eval_steps_per_second": 0.99,
      "step": 46
    },
    {
      "epoch": 6.45,
      "grad_norm": 6.183708667755127,
      "learning_rate": 4.621212121212121e-05,
      "loss": 0.3649,
      "step": 50
    },
    {
      "epoch": 6.97,
      "eval_accuracy": 0.7166666666666667,
      "eval_loss": 0.6321703195571899,
      "eval_runtime": 1.6185,
      "eval_samples_per_second": 37.07,
      "eval_steps_per_second": 1.236,
      "step": 54
    },
    {
      "epoch": 7.74,
      "grad_norm": 8.211569786071777,
      "learning_rate": 4.431818181818182e-05,
      "loss": 0.2592,
      "step": 60
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.7333333333333333,
      "eval_loss": 0.6993545889854431,
      "eval_runtime": 1.7914,
      "eval_samples_per_second": 33.494,
      "eval_steps_per_second": 1.116,
      "step": 62
    },
    {
      "epoch": 8.9,
      "eval_accuracy": 0.7333333333333333,
      "eval_loss": 0.6589908599853516,
      "eval_runtime": 1.6239,
      "eval_samples_per_second": 36.949,
      "eval_steps_per_second": 1.232,
      "step": 69
    },
    {
      "epoch": 9.03,
      "grad_norm": 6.782934665679932,
      "learning_rate": 4.242424242424243e-05,
      "loss": 0.1958,
      "step": 70
    },
    {
      "epoch": 9.94,
      "eval_accuracy": 0.7666666666666667,
      "eval_loss": 0.6845510601997375,
      "eval_runtime": 1.8061,
      "eval_samples_per_second": 33.22,
      "eval_steps_per_second": 1.107,
      "step": 77
    },
    {
      "epoch": 10.32,
      "grad_norm": 6.670173645019531,
      "learning_rate": 4.053030303030303e-05,
      "loss": 0.1664,
      "step": 80
    },
    {
      "epoch": 10.97,
      "eval_accuracy": 0.7666666666666667,
      "eval_loss": 0.7165997624397278,
      "eval_runtime": 1.6803,
      "eval_samples_per_second": 35.708,
      "eval_steps_per_second": 1.19,
      "step": 85
    },
    {
      "epoch": 11.61,
      "grad_norm": 7.649813175201416,
      "learning_rate": 3.8636363636363636e-05,
      "loss": 0.1571,
      "step": 90
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.7833333333333333,
      "eval_loss": 0.7842047214508057,
      "eval_runtime": 1.6667,
      "eval_samples_per_second": 35.999,
      "eval_steps_per_second": 1.2,
      "step": 93
    },
    {
      "epoch": 12.9,
      "grad_norm": 5.415360927581787,
      "learning_rate": 3.6742424242424246e-05,
      "loss": 0.1174,
      "step": 100
    },
    {
      "epoch": 12.9,
      "eval_accuracy": 0.8,
      "eval_loss": 0.8464832901954651,
      "eval_runtime": 1.8073,
      "eval_samples_per_second": 33.198,
      "eval_steps_per_second": 1.107,
      "step": 100
    },
    {
      "epoch": 13.94,
      "eval_accuracy": 0.7666666666666667,
      "eval_loss": 0.9116414189338684,
      "eval_runtime": 1.6421,
      "eval_samples_per_second": 36.539,
      "eval_steps_per_second": 1.218,
      "step": 108
    },
    {
      "epoch": 14.19,
      "grad_norm": 3.7090325355529785,
      "learning_rate": 3.484848484848485e-05,
      "loss": 0.0956,
      "step": 110
    },
    {
      "epoch": 14.97,
      "eval_accuracy": 0.75,
      "eval_loss": 0.9740654230117798,
      "eval_runtime": 1.6076,
      "eval_samples_per_second": 37.322,
      "eval_steps_per_second": 1.244,
      "step": 116
    },
    {
      "epoch": 15.48,
      "grad_norm": 7.993117809295654,
      "learning_rate": 3.295454545454545e-05,
      "loss": 0.1252,
      "step": 120
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.8,
      "eval_loss": 0.7759976983070374,
      "eval_runtime": 1.6595,
      "eval_samples_per_second": 36.154,
      "eval_steps_per_second": 1.205,
      "step": 124
    },
    {
      "epoch": 16.77,
      "grad_norm": 3.9146523475646973,
      "learning_rate": 3.106060606060606e-05,
      "loss": 0.0933,
      "step": 130
    },
    {
      "epoch": 16.9,
      "eval_accuracy": 0.7833333333333333,
      "eval_loss": 0.9424102902412415,
      "eval_runtime": 2.1274,
      "eval_samples_per_second": 28.203,
      "eval_steps_per_second": 0.94,
      "step": 131
    },
    {
      "epoch": 17.94,
      "eval_accuracy": 0.7333333333333333,
      "eval_loss": 1.0444563627243042,
      "eval_runtime": 1.6154,
      "eval_samples_per_second": 37.143,
      "eval_steps_per_second": 1.238,
      "step": 139
    },
    {
      "epoch": 18.06,
      "grad_norm": 7.39138126373291,
      "learning_rate": 2.916666666666667e-05,
      "loss": 0.1455,
      "step": 140
    },
    {
      "epoch": 18.97,
      "eval_accuracy": 0.7333333333333333,
      "eval_loss": 0.8524726629257202,
      "eval_runtime": 1.681,
      "eval_samples_per_second": 35.693,
      "eval_steps_per_second": 1.19,
      "step": 147
    },
    {
      "epoch": 19.35,
      "grad_norm": 3.7276432514190674,
      "learning_rate": 2.7272727272727273e-05,
      "loss": 0.1034,
      "step": 150
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.7666666666666667,
      "eval_loss": 0.822151243686676,
      "eval_runtime": 2.0162,
      "eval_samples_per_second": 29.759,
      "eval_steps_per_second": 0.992,
      "step": 155
    },
    {
      "epoch": 20.65,
      "grad_norm": 4.372833728790283,
      "learning_rate": 2.537878787878788e-05,
      "loss": 0.0855,
      "step": 160
    },
    {
      "epoch": 20.9,
      "eval_accuracy": 0.7833333333333333,
      "eval_loss": 0.8990534543991089,
      "eval_runtime": 2.1376,
      "eval_samples_per_second": 28.069,
      "eval_steps_per_second": 0.936,
      "step": 162
    },
    {
      "epoch": 21.94,
      "grad_norm": 6.204973220825195,
      "learning_rate": 2.3484848484848487e-05,
      "loss": 0.0985,
      "step": 170
    },
    {
      "epoch": 21.94,
      "eval_accuracy": 0.8,
      "eval_loss": 0.8954617977142334,
      "eval_runtime": 1.7159,
      "eval_samples_per_second": 34.968,
      "eval_steps_per_second": 1.166,
      "step": 170
    },
    {
      "epoch": 22.97,
      "eval_accuracy": 0.7666666666666667,
      "eval_loss": 0.9603295922279358,
      "eval_runtime": 1.628,
      "eval_samples_per_second": 36.854,
      "eval_steps_per_second": 1.228,
      "step": 178
    },
    {
      "epoch": 23.23,
      "grad_norm": 2.5669796466827393,
      "learning_rate": 2.1590909090909093e-05,
      "loss": 0.087,
      "step": 180
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.7833333333333333,
      "eval_loss": 0.9932332634925842,
      "eval_runtime": 1.6549,
      "eval_samples_per_second": 36.255,
      "eval_steps_per_second": 1.209,
      "step": 186
    },
    {
      "epoch": 24.52,
      "grad_norm": 4.33466911315918,
      "learning_rate": 1.9696969696969697e-05,
      "loss": 0.0832,
      "step": 190
    },
    {
      "epoch": 24.9,
      "eval_accuracy": 0.7833333333333333,
      "eval_loss": 1.0099666118621826,
      "eval_runtime": 1.6219,
      "eval_samples_per_second": 36.994,
      "eval_steps_per_second": 1.233,
      "step": 193
    },
    {
      "epoch": 25.81,
      "grad_norm": 2.6737489700317383,
      "learning_rate": 1.7803030303030303e-05,
      "loss": 0.0632,
      "step": 200
    },
    {
      "epoch": 25.94,
      "eval_accuracy": 0.7666666666666667,
      "eval_loss": 0.9393168091773987,
      "eval_runtime": 1.62,
      "eval_samples_per_second": 37.037,
      "eval_steps_per_second": 1.235,
      "step": 201
    },
    {
      "epoch": 26.97,
      "eval_accuracy": 0.7833333333333333,
      "eval_loss": 0.9061955213546753,
      "eval_runtime": 2.0226,
      "eval_samples_per_second": 29.665,
      "eval_steps_per_second": 0.989,
      "step": 209
    },
    {
      "epoch": 27.1,
      "grad_norm": 6.029117107391357,
      "learning_rate": 1.590909090909091e-05,
      "loss": 0.0778,
      "step": 210
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.8,
      "eval_loss": 0.9339290857315063,
      "eval_runtime": 2.1654,
      "eval_samples_per_second": 27.708,
      "eval_steps_per_second": 0.924,
      "step": 217
    },
    {
      "epoch": 28.39,
      "grad_norm": 3.781505584716797,
      "learning_rate": 1.4015151515151515e-05,
      "loss": 0.0627,
      "step": 220
    },
    {
      "epoch": 28.9,
      "eval_accuracy": 0.8166666666666667,
      "eval_loss": 1.003859043121338,
      "eval_runtime": 1.6395,
      "eval_samples_per_second": 36.596,
      "eval_steps_per_second": 1.22,
      "step": 224
    },
    {
      "epoch": 29.68,
      "grad_norm": 8.317666053771973,
      "learning_rate": 1.2121212121212122e-05,
      "loss": 0.0837,
      "step": 230
    },
    {
      "epoch": 29.94,
      "eval_accuracy": 0.7333333333333333,
      "eval_loss": 1.0636054277420044,
      "eval_runtime": 2.1368,
      "eval_samples_per_second": 28.079,
      "eval_steps_per_second": 0.936,
      "step": 232
    },
    {
      "epoch": 30.97,
      "grad_norm": 4.580255508422852,
      "learning_rate": 1.0227272727272729e-05,
      "loss": 0.0595,
      "step": 240
    },
    {
      "epoch": 30.97,
      "eval_accuracy": 0.75,
      "eval_loss": 1.0424180030822754,
      "eval_runtime": 1.9745,
      "eval_samples_per_second": 30.387,
      "eval_steps_per_second": 1.013,
      "step": 240
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.8,
      "eval_loss": 1.0514304637908936,
      "eval_runtime": 1.6234,
      "eval_samples_per_second": 36.96,
      "eval_steps_per_second": 1.232,
      "step": 248
    },
    {
      "epoch": 32.26,
      "grad_norm": 3.7219626903533936,
      "learning_rate": 8.333333333333334e-06,
      "loss": 0.0706,
      "step": 250
    },
    {
      "epoch": 32.9,
      "eval_accuracy": 0.7833333333333333,
      "eval_loss": 1.0638715028762817,
      "eval_runtime": 1.6117,
      "eval_samples_per_second": 37.228,
      "eval_steps_per_second": 1.241,
      "step": 255
    },
    {
      "epoch": 33.55,
      "grad_norm": 2.273451566696167,
      "learning_rate": 6.43939393939394e-06,
      "loss": 0.0565,
      "step": 260
    },
    {
      "epoch": 33.94,
      "eval_accuracy": 0.7666666666666667,
      "eval_loss": 1.0494216680526733,
      "eval_runtime": 1.9391,
      "eval_samples_per_second": 30.942,
      "eval_steps_per_second": 1.031,
      "step": 263
    },
    {
      "epoch": 34.84,
      "grad_norm": 2.9928078651428223,
      "learning_rate": 4.5454545454545455e-06,
      "loss": 0.0515,
      "step": 270
    },
    {
      "epoch": 34.97,
      "eval_accuracy": 0.7666666666666667,
      "eval_loss": 1.0627790689468384,
      "eval_runtime": 1.5942,
      "eval_samples_per_second": 37.636,
      "eval_steps_per_second": 1.255,
      "step": 271
    },
    {
      "epoch": 36.0,
      "eval_accuracy": 0.7666666666666667,
      "eval_loss": 1.1089335680007935,
      "eval_runtime": 1.74,
      "eval_samples_per_second": 34.484,
      "eval_steps_per_second": 1.149,
      "step": 279
    },
    {
      "epoch": 36.13,
      "grad_norm": 6.05413818359375,
      "learning_rate": 2.651515151515152e-06,
      "loss": 0.0614,
      "step": 280
    },
    {
      "epoch": 36.9,
      "eval_accuracy": 0.8,
      "eval_loss": 1.086135983467102,
      "eval_runtime": 1.5851,
      "eval_samples_per_second": 37.853,
      "eval_steps_per_second": 1.262,
      "step": 286
    },
    {
      "epoch": 37.42,
      "grad_norm": 3.7136027812957764,
      "learning_rate": 7.575757575757576e-07,
      "loss": 0.0496,
      "step": 290
    },
    {
      "epoch": 37.94,
      "eval_accuracy": 0.8,
      "eval_loss": 1.0713495016098022,
      "eval_runtime": 1.6733,
      "eval_samples_per_second": 35.857,
      "eval_steps_per_second": 1.195,
      "step": 294
    },
    {
      "epoch": 37.94,
      "step": 294,
      "total_flos": 2.8633958865108173e+18,
      "train_loss": 0.23703576421656578,
      "train_runtime": 1686.3425,
      "train_samples_per_second": 24.258,
      "train_steps_per_second": 0.174
    }
  ],
  "logging_steps": 10,
  "max_steps": 294,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 42,
  "save_steps": 500,
  "total_flos": 2.8633958865108173e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}