{
  "best_metric": 0.9130434782608695,
  "best_model_checkpoint": "BEiT-DMAE-DA\\checkpoint-379",
  "epoch": 38.26086956521739,
  "eval_steps": 500,
  "global_step": 440,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.87,
      "learning_rate": 4.886363636363637e-05,
      "loss": 1.3475,
      "step": 10
    },
    {
      "epoch": 0.96,
      "eval_accuracy": 0.32608695652173914,
      "eval_loss": 1.327863335609436,
      "eval_runtime": 0.6895,
      "eval_samples_per_second": 66.715,
      "eval_steps_per_second": 4.351,
      "step": 11
    },
    {
      "epoch": 1.74,
      "learning_rate": 4.772727272727273e-05,
      "loss": 1.1875,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.34782608695652173,
      "eval_loss": 1.13204824924469,
      "eval_runtime": 0.7412,
      "eval_samples_per_second": 62.06,
      "eval_steps_per_second": 4.047,
      "step": 23
    },
    {
      "epoch": 2.61,
      "learning_rate": 4.659090909090909e-05,
      "loss": 0.9998,
      "step": 30
    },
    {
      "epoch": 2.96,
      "eval_accuracy": 0.5434782608695652,
      "eval_loss": 0.9957121014595032,
      "eval_runtime": 0.6867,
      "eval_samples_per_second": 66.99,
      "eval_steps_per_second": 4.369,
      "step": 34
    },
    {
      "epoch": 3.48,
      "learning_rate": 4.545454545454546e-05,
      "loss": 0.8836,
      "step": 40
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5869565217391305,
      "eval_loss": 0.8435541987419128,
      "eval_runtime": 0.7072,
      "eval_samples_per_second": 65.049,
      "eval_steps_per_second": 4.242,
      "step": 46
    },
    {
      "epoch": 4.35,
      "learning_rate": 4.431818181818182e-05,
      "loss": 0.7593,
      "step": 50
    },
    {
      "epoch": 4.96,
      "eval_accuracy": 0.6304347826086957,
      "eval_loss": 0.7904003262519836,
      "eval_runtime": 0.7214,
      "eval_samples_per_second": 63.763,
      "eval_steps_per_second": 4.158,
      "step": 57
    },
    {
      "epoch": 5.22,
      "learning_rate": 4.318181818181819e-05,
      "loss": 0.6939,
      "step": 60
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8260869565217391,
      "eval_loss": 0.6656226515769958,
      "eval_runtime": 0.6452,
      "eval_samples_per_second": 71.3,
      "eval_steps_per_second": 4.65,
      "step": 69
    },
    {
      "epoch": 6.09,
      "learning_rate": 4.204545454545455e-05,
      "loss": 0.5942,
      "step": 70
    },
    {
      "epoch": 6.96,
      "learning_rate": 4.0909090909090915e-05,
      "loss": 0.4924,
      "step": 80
    },
    {
      "epoch": 6.96,
      "eval_accuracy": 0.6739130434782609,
      "eval_loss": 0.6724327206611633,
      "eval_runtime": 0.6712,
      "eval_samples_per_second": 68.537,
      "eval_steps_per_second": 4.47,
      "step": 80
    },
    {
      "epoch": 7.83,
      "learning_rate": 3.9772727272727275e-05,
      "loss": 0.4444,
      "step": 90
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.782608695652174,
      "eval_loss": 0.5950984954833984,
      "eval_runtime": 0.6562,
      "eval_samples_per_second": 70.105,
      "eval_steps_per_second": 4.572,
      "step": 92
    },
    {
      "epoch": 8.7,
      "learning_rate": 3.8636363636363636e-05,
      "loss": 0.337,
      "step": 100
    },
    {
      "epoch": 8.96,
      "eval_accuracy": 0.8260869565217391,
      "eval_loss": 0.5222101211547852,
      "eval_runtime": 0.6647,
      "eval_samples_per_second": 69.208,
      "eval_steps_per_second": 4.514,
      "step": 103
    },
    {
      "epoch": 9.57,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.3213,
      "step": 110
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.8043478260869565,
      "eval_loss": 0.6813698410987854,
      "eval_runtime": 0.6601,
      "eval_samples_per_second": 69.681,
      "eval_steps_per_second": 4.544,
      "step": 115
    },
    {
      "epoch": 10.43,
      "learning_rate": 3.6363636363636364e-05,
      "loss": 0.2689,
      "step": 120
    },
    {
      "epoch": 10.96,
      "eval_accuracy": 0.782608695652174,
      "eval_loss": 0.5912956595420837,
      "eval_runtime": 0.6652,
      "eval_samples_per_second": 69.156,
      "eval_steps_per_second": 4.51,
      "step": 126
    },
    {
      "epoch": 11.3,
      "learning_rate": 3.522727272727273e-05,
      "loss": 0.2538,
      "step": 130
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.782608695652174,
      "eval_loss": 0.6227802634239197,
      "eval_runtime": 0.6802,
      "eval_samples_per_second": 67.63,
      "eval_steps_per_second": 4.411,
      "step": 138
    },
    {
      "epoch": 12.17,
      "learning_rate": 3.409090909090909e-05,
      "loss": 0.2032,
      "step": 140
    },
    {
      "epoch": 12.96,
      "eval_accuracy": 0.7608695652173914,
      "eval_loss": 0.6992256045341492,
      "eval_runtime": 0.6597,
      "eval_samples_per_second": 69.734,
      "eval_steps_per_second": 4.548,
      "step": 149
    },
    {
      "epoch": 13.04,
      "learning_rate": 3.295454545454545e-05,
      "loss": 0.2439,
      "step": 150
    },
    {
      "epoch": 13.91,
      "learning_rate": 3.181818181818182e-05,
      "loss": 0.2152,
      "step": 160
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.7608695652173914,
      "eval_loss": 0.7730478048324585,
      "eval_runtime": 0.7082,
      "eval_samples_per_second": 64.956,
      "eval_steps_per_second": 4.236,
      "step": 161
    },
    {
      "epoch": 14.78,
      "learning_rate": 3.068181818181818e-05,
      "loss": 0.1713,
      "step": 170
    },
    {
      "epoch": 14.96,
      "eval_accuracy": 0.7608695652173914,
      "eval_loss": 0.776236891746521,
      "eval_runtime": 0.6617,
      "eval_samples_per_second": 69.522,
      "eval_steps_per_second": 4.534,
      "step": 172
    },
    {
      "epoch": 15.65,
      "learning_rate": 2.954545454545455e-05,
      "loss": 0.2042,
      "step": 180
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.717391304347826,
      "eval_loss": 0.7652217745780945,
      "eval_runtime": 0.6431,
      "eval_samples_per_second": 71.524,
      "eval_steps_per_second": 4.665,
      "step": 184
    },
    {
      "epoch": 16.52,
      "learning_rate": 2.8409090909090912e-05,
      "loss": 0.1668,
      "step": 190
    },
    {
      "epoch": 16.96,
      "eval_accuracy": 0.8478260869565217,
      "eval_loss": 0.55118328332901,
      "eval_runtime": 0.6664,
      "eval_samples_per_second": 69.031,
      "eval_steps_per_second": 4.502,
      "step": 195
    },
    {
      "epoch": 17.39,
      "learning_rate": 2.7272727272727273e-05,
      "loss": 0.1743,
      "step": 200
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.782608695652174,
      "eval_loss": 0.7310671806335449,
      "eval_runtime": 0.6702,
      "eval_samples_per_second": 68.639,
      "eval_steps_per_second": 4.476,
      "step": 207
    },
    {
      "epoch": 18.26,
      "learning_rate": 2.6136363636363637e-05,
      "loss": 0.1226,
      "step": 210
    },
    {
      "epoch": 18.96,
      "eval_accuracy": 0.8043478260869565,
      "eval_loss": 0.711499810218811,
      "eval_runtime": 1.0547,
      "eval_samples_per_second": 43.612,
      "eval_steps_per_second": 2.844,
      "step": 218
    },
    {
      "epoch": 19.13,
      "learning_rate": 2.5e-05,
      "loss": 0.1628,
      "step": 220
    },
    {
      "epoch": 20.0,
      "learning_rate": 2.3863636363636365e-05,
      "loss": 0.1537,
      "step": 230
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.7608695652173914,
      "eval_loss": 0.6799980998039246,
      "eval_runtime": 0.7192,
      "eval_samples_per_second": 63.962,
      "eval_steps_per_second": 4.171,
      "step": 230
    },
    {
      "epoch": 20.87,
      "learning_rate": 2.272727272727273e-05,
      "loss": 0.1311,
      "step": 240
    },
    {
      "epoch": 20.96,
      "eval_accuracy": 0.8478260869565217,
      "eval_loss": 0.5863925218582153,
      "eval_runtime": 0.7092,
      "eval_samples_per_second": 64.866,
      "eval_steps_per_second": 4.23,
      "step": 241
    },
    {
      "epoch": 21.74,
      "learning_rate": 2.1590909090909093e-05,
      "loss": 0.1335,
      "step": 250
    },
    {
      "epoch": 22.0,
      "eval_accuracy": 0.8260869565217391,
      "eval_loss": 0.6345816254615784,
      "eval_runtime": 0.6942,
      "eval_samples_per_second": 66.266,
      "eval_steps_per_second": 4.322,
      "step": 253
    },
    {
      "epoch": 22.61,
      "learning_rate": 2.0454545454545457e-05,
      "loss": 0.0981,
      "step": 260
    },
    {
      "epoch": 22.96,
      "eval_accuracy": 0.8043478260869565,
      "eval_loss": 0.6541118621826172,
      "eval_runtime": 0.7142,
      "eval_samples_per_second": 64.41,
      "eval_steps_per_second": 4.201,
      "step": 264
    },
    {
      "epoch": 23.48,
      "learning_rate": 1.9318181818181818e-05,
      "loss": 0.1248,
      "step": 270
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.8260869565217391,
      "eval_loss": 0.7017006874084473,
      "eval_runtime": 0.7302,
      "eval_samples_per_second": 62.999,
      "eval_steps_per_second": 4.109,
      "step": 276
    },
    {
      "epoch": 24.35,
      "learning_rate": 1.8181818181818182e-05,
      "loss": 0.1183,
      "step": 280
    },
    {
      "epoch": 24.96,
      "eval_accuracy": 0.8260869565217391,
      "eval_loss": 0.69643634557724,
      "eval_runtime": 0.6507,
      "eval_samples_per_second": 70.698,
      "eval_steps_per_second": 4.611,
      "step": 287
    },
    {
      "epoch": 25.22,
      "learning_rate": 1.7045454545454546e-05,
      "loss": 0.0946,
      "step": 290
    },
    {
      "epoch": 26.0,
      "eval_accuracy": 0.8260869565217391,
      "eval_loss": 0.6449636220932007,
      "eval_runtime": 0.6636,
      "eval_samples_per_second": 69.314,
      "eval_steps_per_second": 4.52,
      "step": 299
    },
    {
      "epoch": 26.09,
      "learning_rate": 1.590909090909091e-05,
      "loss": 0.1312,
      "step": 300
    },
    {
      "epoch": 26.96,
      "learning_rate": 1.4772727272727274e-05,
      "loss": 0.0957,
      "step": 310
    },
    {
      "epoch": 26.96,
      "eval_accuracy": 0.8043478260869565,
      "eval_loss": 0.7056821584701538,
      "eval_runtime": 0.6997,
      "eval_samples_per_second": 65.746,
      "eval_steps_per_second": 4.288,
      "step": 310
    },
    {
      "epoch": 27.83,
      "learning_rate": 1.3636363636363637e-05,
      "loss": 0.1692,
      "step": 320
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.8043478260869565,
      "eval_loss": 0.6635299324989319,
      "eval_runtime": 0.6605,
      "eval_samples_per_second": 69.644,
      "eval_steps_per_second": 4.542,
      "step": 322
    },
    {
      "epoch": 28.7,
      "learning_rate": 1.25e-05,
      "loss": 0.0967,
      "step": 330
    },
    {
      "epoch": 28.96,
      "eval_accuracy": 0.8695652173913043,
      "eval_loss": 0.5040035843849182,
      "eval_runtime": 0.6991,
      "eval_samples_per_second": 65.8,
      "eval_steps_per_second": 4.291,
      "step": 333
    },
    {
      "epoch": 29.57,
      "learning_rate": 1.1363636363636365e-05,
      "loss": 0.094,
      "step": 340
    },
    {
      "epoch": 30.0,
      "eval_accuracy": 0.8913043478260869,
      "eval_loss": 0.5587611198425293,
      "eval_runtime": 0.6827,
      "eval_samples_per_second": 67.384,
      "eval_steps_per_second": 4.395,
      "step": 345
    },
    {
      "epoch": 30.43,
      "learning_rate": 1.0227272727272729e-05,
      "loss": 0.0843,
      "step": 350
    },
    {
      "epoch": 30.96,
      "eval_accuracy": 0.8695652173913043,
      "eval_loss": 0.539750874042511,
      "eval_runtime": 0.6677,
      "eval_samples_per_second": 68.898,
      "eval_steps_per_second": 4.493,
      "step": 356
    },
    {
      "epoch": 31.3,
      "learning_rate": 9.090909090909091e-06,
      "loss": 0.0851,
      "step": 360
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.8478260869565217,
      "eval_loss": 0.5806054472923279,
      "eval_runtime": 0.6481,
      "eval_samples_per_second": 70.972,
      "eval_steps_per_second": 4.629,
      "step": 368
    },
    {
      "epoch": 32.17,
      "learning_rate": 7.954545454545455e-06,
      "loss": 0.0955,
      "step": 370
    },
    {
      "epoch": 32.96,
      "eval_accuracy": 0.9130434782608695,
      "eval_loss": 0.48163074254989624,
      "eval_runtime": 0.6626,
      "eval_samples_per_second": 69.42,
      "eval_steps_per_second": 4.527,
      "step": 379
    },
    {
      "epoch": 33.04,
      "learning_rate": 6.818181818181818e-06,
      "loss": 0.1194,
      "step": 380
    },
    {
      "epoch": 33.91,
      "learning_rate": 5.681818181818182e-06,
      "loss": 0.1157,
      "step": 390
    },
    {
      "epoch": 34.0,
      "eval_accuracy": 0.8695652173913043,
      "eval_loss": 0.5289022922515869,
      "eval_runtime": 0.6557,
      "eval_samples_per_second": 70.158,
      "eval_steps_per_second": 4.576,
      "step": 391
    },
    {
      "epoch": 34.78,
      "learning_rate": 4.5454545454545455e-06,
      "loss": 0.072,
      "step": 400
    },
    {
      "epoch": 34.96,
      "eval_accuracy": 0.8913043478260869,
      "eval_loss": 0.5657246112823486,
      "eval_runtime": 0.6672,
      "eval_samples_per_second": 68.949,
      "eval_steps_per_second": 4.497,
      "step": 402
    },
    {
      "epoch": 35.65,
      "learning_rate": 3.409090909090909e-06,
      "loss": 0.091,
      "step": 410
    },
    {
      "epoch": 36.0,
      "eval_accuracy": 0.8478260869565217,
      "eval_loss": 0.5566320419311523,
      "eval_runtime": 0.6752,
      "eval_samples_per_second": 68.133,
      "eval_steps_per_second": 4.443,
      "step": 414
    },
    {
      "epoch": 36.52,
      "learning_rate": 2.2727272727272728e-06,
      "loss": 0.0891,
      "step": 420
    },
    {
      "epoch": 36.96,
      "eval_accuracy": 0.8478260869565217,
      "eval_loss": 0.5729131102561951,
      "eval_runtime": 0.6672,
      "eval_samples_per_second": 68.948,
      "eval_steps_per_second": 4.497,
      "step": 425
    },
    {
      "epoch": 37.39,
      "learning_rate": 1.1363636363636364e-06,
      "loss": 0.0732,
      "step": 430
    },
    {
      "epoch": 38.0,
      "eval_accuracy": 0.8260869565217391,
      "eval_loss": 0.5915136933326721,
      "eval_runtime": 0.6622,
      "eval_samples_per_second": 69.469,
      "eval_steps_per_second": 4.531,
      "step": 437
    },
    {
      "epoch": 38.26,
      "learning_rate": 0.0,
      "loss": 0.0647,
      "step": 440
    },
    {
      "epoch": 38.26,
      "eval_accuracy": 0.8260869565217391,
      "eval_loss": 0.5901583433151245,
      "eval_runtime": 0.7022,
      "eval_samples_per_second": 65.512,
      "eval_steps_per_second": 4.273,
      "step": 440
    },
    {
      "epoch": 38.26,
      "step": 440,
      "total_flos": 2.1400720000140902e+18,
      "train_loss": 0.2858855720270764,
      "train_runtime": 727.0673,
      "train_samples_per_second": 39.721,
      "train_steps_per_second": 0.605
    }
  ],
  "logging_steps": 10,
  "max_steps": 440,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 40,
  "save_steps": 500,
  "total_flos": 2.1400720000140902e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}