trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "global_step": 1000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 6.666666666666668e-08,
      "loss": 1.1133,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "eval_cer": 416.4573186362471,
      "eval_loss": 1.3322362899780273,
      "eval_runtime": 1200.065,
      "eval_samples_per_second": 0.886,
      "eval_steps_per_second": 0.443,
      "eval_wer": 343.562937357727,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "step": 1000,
      "total_flos": 4.2463420416e+18,
      "train_loss": 1.1132755126953124,
      "train_runtime": 2073.0676,
      "train_samples_per_second": 0.965,
      "train_steps_per_second": 0.482
    }
  ],
  "max_steps": 1000,
  "num_train_epochs": 9223372036854775807,
  "total_flos": 4.2463420416e+18,
  "trial_name": null,
  "trial_params": null
}