trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.769230769230769,
  "eval_steps": 500,
  "global_step": 18,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.15,
      "grad_norm": 14.895696980777833,
      "learning_rate": 1e-05,
      "loss": 1.73,
      "step": 1
    },
    {
      "epoch": 0.77,
      "grad_norm": 4.5055799500485305,
      "learning_rate": 1.8314696123025456e-05,
      "loss": 1.5074,
      "step": 5
    },
    {
      "epoch": 0.92,
      "eval_loss": 1.5827996730804443,
      "eval_runtime": 3.9915,
      "eval_samples_per_second": 11.524,
      "eval_steps_per_second": 0.752,
      "step": 6
    },
    {
      "epoch": 1.54,
      "grad_norm": 1.2354478167407779,
      "learning_rate": 1e-05,
      "loss": 1.2223,
      "step": 10
    },
    {
      "epoch": 2.0,
      "eval_loss": 1.545841097831726,
      "eval_runtime": 4.1419,
      "eval_samples_per_second": 11.106,
      "eval_steps_per_second": 0.724,
      "step": 13
    },
    {
      "epoch": 2.31,
      "grad_norm": 0.7577227867864917,
      "learning_rate": 1.6853038769745466e-06,
      "loss": 1.1253,
      "step": 15
    },
    {
      "epoch": 2.77,
      "eval_loss": 1.5395549535751343,
      "eval_runtime": 4.1393,
      "eval_samples_per_second": 11.113,
      "eval_steps_per_second": 0.725,
      "step": 18
    },
    {
      "epoch": 2.77,
      "step": 18,
      "total_flos": 3716488888320.0,
      "train_loss": 1.2632259196705289,
      "train_runtime": 307.3031,
      "train_samples_per_second": 3.954,
      "train_steps_per_second": 0.059
    }
  ],
  "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
}