trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.984819734345351,
  "global_step": 2100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.95,
      "learning_rate": 2.715370018975332e-05,
      "loss": 1.2474,
      "step": 500
    },
    {
      "epoch": 1.9,
      "learning_rate": 2.4307400379506644e-05,
      "loss": 0.9239,
      "step": 1000
    },
    {
      "epoch": 2.85,
      "learning_rate": 2.146110056925996e-05,
      "loss": 0.8101,
      "step": 1500
    },
    {
      "epoch": 3.8,
      "learning_rate": 1.861480075901328e-05,
      "loss": 0.731,
      "step": 2000
    }
  ],
  "max_steps": 5270,
  "num_train_epochs": 10,
  "total_flos": 126741142877995008,
  "trial_name": null,
  "trial_params": null
}