trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.999463464194797,
  "global_step": 2052,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.73,
      "learning_rate": 0.0002269005847953216,
      "loss": 0.2878,
      "step": 500
    },
    {
      "epoch": 1.46,
      "learning_rate": 0.00015380116959064327,
      "loss": 0.1983,
      "step": 1000
    },
    {
      "epoch": 2.19,
      "learning_rate": 8.07017543859649e-05,
      "loss": 0.1427,
      "step": 1500
    },
    {
      "epoch": 2.92,
      "learning_rate": 7.6023391812865485e-06,
      "loss": 0.0947,
      "step": 2000
    }
  ],
  "max_steps": 2052,
  "num_train_epochs": 3,
  "total_flos": 1.137732354048e+18,
  "trial_name": null,
  "trial_params": null
}