trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.053367488525989965,
  "global_step": 2000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "learning_rate": 5e-06,
      "loss": 9.1117,
      "step": 500
    },
    {
      "epoch": 0.03,
      "learning_rate": 1e-05,
      "loss": 7.3389,
      "step": 1000
    },
    {
      "epoch": 0.04,
      "learning_rate": 1.5e-05,
      "loss": 6.8415,
      "step": 1500
    },
    {
      "epoch": 0.05,
      "learning_rate": 2e-05,
      "loss": 6.6819,
      "step": 2000
    }
  ],
  "max_steps": 15000,
  "num_train_epochs": 1,
  "total_flos": 1.3476086097431757e+17,
  "trial_name": null,
  "trial_params": null
}