trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.999910370171193,
  "global_step": 5578,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 1e-05,
      "loss": 0.4337,
      "step": 5578
    },
    {
      "epoch": 1.0,
      "step": 5578,
      "total_flos": 1.7627111644725248e+16,
      "train_loss": 0.43365896504683576,
      "train_runtime": 43080.098,
      "train_samples_per_second": 33.148,
      "train_steps_per_second": 0.129
    }
  ],
  "max_steps": 5578,
  "num_train_epochs": 1,
  "total_flos": 1.7627111644725248e+16,
  "trial_name": null,
  "trial_params": null
}