checkpoint-24/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.742857142857143,
  "eval_steps": 500,
  "global_step": 24,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.8,
      "learning_rate": 0.00016190476190476192,
      "loss": 2.283,
      "step": 7
    },
    {
      "epoch": 1.6,
      "learning_rate": 9.523809523809524e-05,
      "loss": 2.0752,
      "step": 14
    },
    {
      "epoch": 2.4,
      "learning_rate": 2.857142857142857e-05,
      "loss": 1.9246,
      "step": 21
    }
  ],
  "logging_steps": 7,
  "max_steps": 24,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 3902117643288576.0,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}