trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.88,
  "eval_steps": 500,
  "global_step": 18,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.16,
      "grad_norm": 10.58998792449542,
      "learning_rate": 1e-05,
      "loss": 1.7036,
      "step": 1
    },
    {
      "epoch": 0.8,
      "grad_norm": 6.985998929295557,
      "learning_rate": 1.8314696123025456e-05,
      "loss": 1.7004,
      "step": 5
    },
    {
      "epoch": 0.96,
      "eval_loss": 1.889893651008606,
      "eval_runtime": 8.0882,
      "eval_samples_per_second": 7.418,
      "eval_steps_per_second": 0.495,
      "step": 6
    },
    {
      "epoch": 1.6,
      "grad_norm": 5.121326011067327,
      "learning_rate": 1e-05,
      "loss": 1.2864,
      "step": 10
    },
    {
      "epoch": 1.92,
      "eval_loss": 1.9471222162246704,
      "eval_runtime": 8.0881,
      "eval_samples_per_second": 7.418,
      "eval_steps_per_second": 0.495,
      "step": 12
    },
    {
      "epoch": 2.4,
      "grad_norm": 4.922775683291317,
      "learning_rate": 1.6853038769745466e-06,
      "loss": 0.8787,
      "step": 15
    },
    {
      "epoch": 2.88,
      "eval_loss": 2.021740198135376,
      "eval_runtime": 8.0512,
      "eval_samples_per_second": 7.452,
      "eval_steps_per_second": 0.497,
      "step": 18
    },
    {
      "epoch": 2.88,
      "step": 18,
      "total_flos": 3716488888320.0,
      "train_loss": 1.186334901385837,
      "train_runtime": 420.5441,
      "train_samples_per_second": 2.782,
      "train_steps_per_second": 0.043
    }
  ],
  "logging_steps": 5,
  "max_steps": 18,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 3716488888320.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}