trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.07357810315650062,
  "global_step": 2000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "learning_rate": 2.9448164226326247e-05,
      "loss": 4.808,
      "step": 500
    },
    {
      "epoch": 0.04,
      "learning_rate": 2.8896328452652492e-05,
      "loss": 4.7761,
      "step": 1000
    },
    {
      "epoch": 0.06,
      "learning_rate": 2.8344492678978738e-05,
      "loss": 4.6768,
      "step": 1500
    },
    {
      "epoch": 0.07,
      "learning_rate": 2.7792656905304984e-05,
      "loss": 4.6186,
      "step": 2000
    }
  ],
  "max_steps": 27182,
  "num_train_epochs": 1,
  "total_flos": 1010362220544000.0,
  "trial_name": null,
  "trial_params": null
}