{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 76.19047619047619,
  "eval_steps": 200,
  "global_step": 1200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 12.698412698412698,
      "grad_norm": 0.8574791550636292,
      "learning_rate": 4e-05,
      "loss": 3.1963,
      "step": 200
    },
    {
      "epoch": 12.698412698412698,
      "eval_loss": 0.4380848705768585,
      "eval_runtime": 9.7951,
      "eval_samples_per_second": 5.105,
      "eval_steps_per_second": 0.715,
      "eval_wer": 0.5369978858350951,
      "step": 200
    },
    {
      "epoch": 25.396825396825395,
      "grad_norm": 1.0231786966323853,
      "learning_rate": 8e-05,
      "loss": 0.5929,
      "step": 400
    },
    {
      "epoch": 25.396825396825395,
      "eval_loss": 0.12973153591156006,
      "eval_runtime": 9.3583,
      "eval_samples_per_second": 5.343,
      "eval_steps_per_second": 0.748,
      "eval_wer": 0.23255813953488372,
      "step": 400
    },
    {
      "epoch": 38.095238095238095,
      "grad_norm": 1.3454779386520386,
      "learning_rate": 8.571428571428571e-05,
      "loss": 0.3401,
      "step": 600
    },
    {
      "epoch": 38.095238095238095,
      "eval_loss": 0.07642678916454315,
      "eval_runtime": 9.2054,
      "eval_samples_per_second": 5.432,
      "eval_steps_per_second": 0.76,
      "eval_wer": 0.15010570824524314,
      "step": 600
    },
    {
      "epoch": 50.79365079365079,
      "grad_norm": 0.8762056231498718,
      "learning_rate": 5.714285714285714e-05,
      "loss": 0.2357,
      "step": 800
    },
    {
      "epoch": 50.79365079365079,
      "eval_loss": 0.05995969474315643,
      "eval_runtime": 8.2294,
      "eval_samples_per_second": 6.076,
      "eval_steps_per_second": 0.851,
      "eval_wer": 0.11205073995771671,
      "step": 800
    },
    {
      "epoch": 63.492063492063494,
      "grad_norm": 0.8204830288887024,
      "learning_rate": 2.857142857142857e-05,
      "loss": 0.1865,
      "step": 1000
    },
    {
      "epoch": 63.492063492063494,
      "eval_loss": 0.05562509596347809,
      "eval_runtime": 9.2039,
      "eval_samples_per_second": 5.432,
      "eval_steps_per_second": 0.761,
      "eval_wer": 0.09936575052854123,
      "step": 1000
    },
    {
      "epoch": 76.19047619047619,
      "grad_norm": 0.6718953847885132,
      "learning_rate": 0.0,
      "loss": 0.1665,
      "step": 1200
    },
    {
      "epoch": 76.19047619047619,
      "eval_loss": 0.05189620330929756,
      "eval_runtime": 8.8728,
      "eval_samples_per_second": 5.635,
      "eval_steps_per_second": 0.789,
      "eval_wer": 0.08879492600422834,
      "step": 1200
    },
    {
      "epoch": 76.19047619047619,
      "step": 1200,
      "total_flos": 5.022634798903542e+18,
      "train_loss": 0.7863523006439209,
      "train_runtime": 4539.8167,
      "train_samples_per_second": 4.405,
      "train_steps_per_second": 0.264
    }
  ],
  "logging_steps": 200,
  "max_steps": 1200,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 80,
  "save_steps": 200,
  "total_flos": 5.022634798903542e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}