trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.0472175379426645,
  "global_step": 2400,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.84,
      "learning_rate": 5e-05,
      "loss": 4.2906,
      "step": 500
    },
    {
      "epoch": 1.69,
      "learning_rate": 4.539594843462247e-05,
      "loss": 3.934,
      "step": 1000
    },
    {
      "epoch": 2.53,
      "learning_rate": 4.079189686924494e-05,
      "loss": 3.7996,
      "step": 1500
    },
    {
      "epoch": 3.37,
      "learning_rate": 3.618784530386741e-05,
      "loss": 3.7107,
      "step": 2000
    }
  ],
  "max_steps": 5930,
  "num_train_epochs": 10,
  "total_flos": 5013671510016000.0,
  "trial_name": null,
  "trial_params": null
}