trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 6.8493150684931505,
  "global_step": 1500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 2.28,
      "learning_rate": 3.8584474885844754e-05,
      "loss": 0.3009,
      "step": 500
    },
    {
      "epoch": 4.57,
      "learning_rate": 2.71689497716895e-05,
      "loss": 0.0796,
      "step": 1000
    },
    {
      "epoch": 6.85,
      "learning_rate": 1.5753424657534248e-05,
      "loss": 0.0317,
      "step": 1500
    }
  ],
  "max_steps": 2190,
  "num_train_epochs": 10,
  "total_flos": 660892650498432.0,
  "trial_name": null,
  "trial_params": null
}