trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.9850746268656714,
  "eval_steps": 500,
  "global_step": 200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.75,
      "learning_rate": 0.00017289686274214118,
      "loss": 1.1122,
      "step": 50
    },
    {
      "epoch": 1.49,
      "learning_rate": 0.00010314107590781284,
      "loss": 0.4744,
      "step": 100
    },
    {
      "epoch": 2.24,
      "learning_rate": 3.154528940713113e-05,
      "loss": 0.3524,
      "step": 150
    },
    {
      "epoch": 2.99,
      "learning_rate": 4.934396342684e-08,
      "loss": 0.2938,
      "step": 200
    }
  ],
  "logging_steps": 50,
  "max_steps": 200,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 149284221566976.0,
  "trial_name": null,
  "trial_params": null
}