checkpoint-897/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.1531762295081966,
  "eval_steps": 500,
  "global_step": 897,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.05,
      "learning_rate": 1.9965870307167238e-05,
      "loss": 2.5473,
      "step": 390
    },
    {
      "epoch": 2.09,
      "learning_rate": 2.8895635673624287e-05,
      "loss": 2.3861,
      "step": 780
    }
  ],
  "logging_steps": 390,
  "max_steps": 5856,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 7.832002086567936e+16,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}