trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 1113,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.45,
      "grad_norm": 5.323787212371826,
      "learning_rate": 2.75381850853549e-05,
      "loss": 3.2259,
      "step": 500
    },
    {
      "epoch": 0.9,
      "grad_norm": 5.575705528259277,
      "learning_rate": 5.0763701707097935e-06,
      "loss": 2.9056,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "step": 1113,
      "total_flos": 159583189651080.0,
      "train_loss": 3.04393067064311,
      "train_runtime": 385.498,
      "train_samples_per_second": 92.382,
      "train_steps_per_second": 2.887
    }
  ],
  "logging_steps": 500,
  "max_steps": 1113,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 159583189651080.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}