checkpoint-200/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.9408866995073892,
  "eval_steps": 500,
  "global_step": 200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.79,
      "learning_rate": 0.00017777777777777779,
      "loss": 2.1357,
      "step": 40
    },
    {
      "epoch": 1.58,
      "learning_rate": 0.00013333333333333334,
      "loss": 1.7736,
      "step": 80
    },
    {
      "epoch": 2.36,
      "learning_rate": 8.888888888888889e-05,
      "loss": 1.6946,
      "step": 120
    },
    {
      "epoch": 3.15,
      "learning_rate": 4.4444444444444447e-05,
      "loss": 1.6754,
      "step": 160
    },
    {
      "epoch": 3.94,
      "learning_rate": 0.0,
      "loss": 1.6676,
      "step": 200
    }
  ],
  "logging_steps": 40,
  "max_steps": 200,
  "num_train_epochs": 4,
  "save_steps": 500,
  "total_flos": 2.6476384309424947e+17,
  "trial_name": null,
  "trial_params": null
}