trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.5128205128205128,
  "eval_steps": 10,
  "global_step": 20,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.26,
      "learning_rate": 1.8e-06,
      "loss": 2.4415,
      "step": 10
    },
    {
      "epoch": 0.26,
      "eval_loss": 3.103879928588867,
      "eval_runtime": 16.1503,
      "eval_samples_per_second": 12.384,
      "eval_steps_per_second": 3.096,
      "eval_wer": 0.4556672760511883,
      "step": 10
    },
    {
      "epoch": 0.51,
      "learning_rate": 3.8e-06,
      "loss": 2.0569,
      "step": 20
    },
    {
      "epoch": 0.51,
      "eval_loss": 3.0585784912109375,
      "eval_runtime": 15.679,
      "eval_samples_per_second": 12.756,
      "eval_steps_per_second": 3.189,
      "eval_wer": 0.4547531992687386,
      "step": 20
    },
    {
      "epoch": 0.51,
      "step": 20,
      "total_flos": 1.746851843427936e+16,
      "train_loss": 2.2491682052612303,
      "train_runtime": 96.9606,
      "train_samples_per_second": 0.794,
      "train_steps_per_second": 0.206
    }
  ],
  "logging_steps": 10,
  "max_steps": 20,
  "num_train_epochs": 1,
  "save_steps": 10,
  "total_flos": 1.746851843427936e+16,
  "trial_name": null,
  "trial_params": null
}