trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.5998171289241085,
  "eval_steps": 500,
  "global_step": 1476,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 1.1004,
      "step": 246
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.9532,
      "step": 492
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.9327,
      "step": 738
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.9272,
      "step": 984
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.9225,
      "step": 1230
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.9088,
      "step": 1476
    }
  ],
  "logging_steps": 246,
  "max_steps": 2460,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 246,
  "total_flos": 7.5472623308587e+16,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}