trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 14.915254237288135,
  "eval_steps": 200,
  "global_step": 1320,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 2.26,
      "eval_loss": 3.592435359954834,
      "eval_runtime": 87.4726,
      "eval_samples_per_second": 15.159,
      "eval_steps_per_second": 1.898,
      "eval_wer": 1.0,
      "step": 200
    },
    {
      "epoch": 4.52,
      "eval_loss": 3.494601249694824,
      "eval_runtime": 91.8794,
      "eval_samples_per_second": 14.432,
      "eval_steps_per_second": 1.807,
      "eval_wer": 1.0,
      "step": 400
    },
    {
      "epoch": 5.65,
      "learning_rate": 0.001988,
      "loss": 5.7152,
      "step": 500
    },
    {
      "epoch": 6.78,
      "eval_loss": 3.463015079498291,
      "eval_runtime": 89.7613,
      "eval_samples_per_second": 14.773,
      "eval_steps_per_second": 1.849,
      "eval_wer": 1.0,
      "step": 600
    },
    {
      "epoch": 9.04,
      "eval_loss": 3.452516555786133,
      "eval_runtime": 31.1573,
      "eval_samples_per_second": 42.558,
      "eval_steps_per_second": 5.328,
      "eval_wer": 1.0,
      "step": 800
    },
    {
      "epoch": 11.3,
      "learning_rate": 0.0007902439024390245,
      "loss": 3.5048,
      "step": 1000
    },
    {
      "epoch": 11.3,
      "eval_loss": 3.4328794479370117,
      "eval_runtime": 33.6159,
      "eval_samples_per_second": 39.446,
      "eval_steps_per_second": 4.938,
      "eval_wer": 1.0,
      "step": 1000
    },
    {
      "epoch": 13.56,
      "eval_loss": 3.407379388809204,
      "eval_runtime": 31.2641,
      "eval_samples_per_second": 42.413,
      "eval_steps_per_second": 5.31,
      "eval_wer": 1.0,
      "step": 1200
    },
    {
      "epoch": 14.92,
      "step": 1320,
      "total_flos": 5.202162859259014e+18,
      "train_loss": 4.313969837535511,
      "train_runtime": 3673.9724,
      "train_samples_per_second": 11.542,
      "train_steps_per_second": 0.359
    }
  ],
  "logging_steps": 500,
  "max_steps": 1320,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 15,
  "save_steps": 400,
  "total_flos": 5.202162859259014e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}