{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 15.0,
  "global_step": 64050,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 9.825795424664739e-05,
      "loss": 7.6544,
      "step": 4270
    },
    {
      "epoch": 1.0,
      "eval_cer": 0.21212121212121213,
      "eval_loss": 1.1680512428283691,
      "eval_new_wer": 0.18095238095238095,
      "eval_old_wer": 0.4,
      "eval_runtime": 7.6679,
      "eval_samples_per_second": 27.387,
      "eval_steps_per_second": 3.521,
      "step": 4270
    },
    {
      "epoch": 2.0,
      "learning_rate": 9.124046805153827e-05,
      "loss": 0.867,
      "step": 8540
    },
    {
      "epoch": 2.0,
      "eval_cer": 0.14545454545454545,
      "eval_loss": 0.9439918398857117,
      "eval_new_wer": 0.10952380952380952,
      "eval_old_wer": 0.28095238095238095,
      "eval_runtime": 7.5859,
      "eval_samples_per_second": 27.683,
      "eval_steps_per_second": 3.559,
      "step": 8540
    },
    {
      "epoch": 3.0,
      "learning_rate": 8.42246252958191e-05,
      "loss": 0.5681,
      "step": 12810
    },
    {
      "epoch": 3.0,
      "eval_cer": 0.10735930735930736,
      "eval_loss": 0.6605939865112305,
      "eval_new_wer": 0.0761904761904762,
      "eval_old_wer": 0.22857142857142856,
      "eval_runtime": 7.6378,
      "eval_samples_per_second": 27.495,
      "eval_steps_per_second": 3.535,
      "step": 12810
    },
    {
      "epoch": 4.0,
      "learning_rate": 7.721042597948987e-05,
      "loss": 0.4336,
      "step": 17080
    },
    {
      "epoch": 4.0,
      "eval_cer": 0.10216450216450217,
      "eval_loss": 0.581123948097229,
      "eval_new_wer": 0.08095238095238096,
      "eval_old_wer": 0.19523809523809524,
      "eval_runtime": 7.6234,
      "eval_samples_per_second": 27.547,
      "eval_steps_per_second": 3.542,
      "step": 17080
    },
    {
      "epoch": 5.0,
      "learning_rate": 7.019458322377071e-05,
      "loss": 0.3415,
      "step": 21350
    },
    {
      "epoch": 5.0,
      "eval_cer": 0.08484848484848485,
      "eval_loss": 0.742690920829773,
      "eval_new_wer": 0.05714285714285714,
      "eval_old_wer": 0.1619047619047619,
      "eval_runtime": 7.6897,
      "eval_samples_per_second": 27.309,
      "eval_steps_per_second": 3.511,
      "step": 21350
    },
    {
      "epoch": 6.0,
      "learning_rate": 6.318202734683146e-05,
      "loss": 0.271,
      "step": 25620
    },
    {
      "epoch": 6.0,
      "eval_cer": 0.09264069264069263,
      "eval_loss": 0.5863712430000305,
      "eval_new_wer": 0.09523809523809523,
      "eval_old_wer": 0.1761904761904762,
      "eval_runtime": 7.7063,
      "eval_samples_per_second": 27.25,
      "eval_steps_per_second": 3.504,
      "step": 25620
    },
    {
      "epoch": 7.0,
      "learning_rate": 5.617111490928215e-05,
      "loss": 0.2626,
      "step": 29890
    },
    {
      "epoch": 7.0,
      "eval_cer": 0.14285714285714285,
      "eval_loss": 1.4601606130599976,
      "eval_new_wer": 0.14285714285714285,
      "eval_old_wer": 0.2714285714285714,
      "eval_runtime": 7.7229,
      "eval_samples_per_second": 27.192,
      "eval_steps_per_second": 3.496,
      "step": 29890
    },
    {
      "epoch": 8.0,
      "learning_rate": 4.9155272153562984e-05,
      "loss": 0.2013,
      "step": 34160
    },
    {
      "epoch": 8.0,
      "eval_cer": 0.10043290043290043,
      "eval_loss": 0.9108497500419617,
      "eval_new_wer": 0.11428571428571428,
      "eval_old_wer": 0.18095238095238095,
      "eval_runtime": 7.7082,
      "eval_samples_per_second": 27.244,
      "eval_steps_per_second": 3.503,
      "step": 34160
    },
    {
      "epoch": 9.0,
      "learning_rate": 4.213778595845386e-05,
      "loss": 0.1484,
      "step": 38430
    },
    {
      "epoch": 9.0,
      "eval_cer": 0.14025974025974025,
      "eval_loss": 1.427252173423767,
      "eval_new_wer": 0.1380952380952381,
      "eval_old_wer": 0.2,
      "eval_runtime": 7.7333,
      "eval_samples_per_second": 27.155,
      "eval_steps_per_second": 3.491,
      "step": 38430
    },
    {
      "epoch": 10.0,
      "learning_rate": 3.5121943202734684e-05,
      "loss": 0.1179,
      "step": 42700
    },
    {
      "epoch": 10.0,
      "eval_cer": 0.1264069264069264,
      "eval_loss": 1.0122177600860596,
      "eval_new_wer": 0.1523809523809524,
      "eval_old_wer": 0.18095238095238095,
      "eval_runtime": 7.7665,
      "eval_samples_per_second": 27.039,
      "eval_steps_per_second": 3.476,
      "step": 42700
    },
    {
      "epoch": 11.0,
      "learning_rate": 2.8106100447015514e-05,
      "loss": 0.0987,
      "step": 46970
    },
    {
      "epoch": 11.0,
      "eval_cer": 0.08831168831168831,
      "eval_loss": 0.7208192944526672,
      "eval_new_wer": 0.0761904761904762,
      "eval_old_wer": 0.14285714285714285,
      "eval_runtime": 7.7985,
      "eval_samples_per_second": 26.928,
      "eval_steps_per_second": 3.462,
      "step": 46970
    },
    {
      "epoch": 12.0,
      "learning_rate": 2.1090257691296347e-05,
      "loss": 0.0771,
      "step": 51240
    },
    {
      "epoch": 12.0,
      "eval_cer": 0.11428571428571428,
      "eval_loss": 1.2593973875045776,
      "eval_new_wer": 0.11428571428571428,
      "eval_old_wer": 0.1761904761904762,
      "eval_runtime": 7.5989,
      "eval_samples_per_second": 27.635,
      "eval_steps_per_second": 3.553,
      "step": 51240
    },
    {
      "epoch": 13.0,
      "learning_rate": 1.4076058374967133e-05,
      "loss": 0.0603,
      "step": 55510
    },
    {
      "epoch": 13.0,
      "eval_cer": 0.10735930735930736,
      "eval_loss": 1.1333523988723755,
      "eval_new_wer": 0.11428571428571428,
      "eval_old_wer": 0.1619047619047619,
      "eval_runtime": 7.5623,
      "eval_samples_per_second": 27.77,
      "eval_steps_per_second": 3.57,
      "step": 55510
    },
    {
      "epoch": 14.0,
      "learning_rate": 7.061859058637918e-06,
      "loss": 0.0489,
      "step": 59780
    },
    {
      "epoch": 14.0,
      "eval_cer": 0.09177489177489177,
      "eval_loss": 0.9029120802879333,
      "eval_new_wer": 0.10476190476190476,
      "eval_old_wer": 0.12857142857142856,
      "eval_runtime": 7.6393,
      "eval_samples_per_second": 27.489,
      "eval_steps_per_second": 3.534,
      "step": 59780
    },
    {
      "epoch": 15.0,
      "learning_rate": 4.7659742308703654e-08,
      "loss": 0.0409,
      "step": 64050
    },
    {
      "epoch": 15.0,
      "eval_cer": 0.09264069264069263,
      "eval_loss": 1.0212522745132446,
      "eval_new_wer": 0.1,
      "eval_old_wer": 0.1380952380952381,
      "eval_runtime": 7.658,
      "eval_samples_per_second": 27.422,
      "eval_steps_per_second": 3.526,
      "step": 64050
    }
  ],
  "max_steps": 64050,
  "num_train_epochs": 15,
  "total_flos": 5.079400634281409e+19,
  "trial_name": null,
  "trial_params": null
}