trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.6666666666666665,
  "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": 14.034524521131214,
      "learning_rate": 1e-05,
      "loss": 1.741,
      "step": 1
    },
    {
      "epoch": 0.74,
      "grad_norm": 4.618927858178523,
      "learning_rate": 1.8314696123025456e-05,
      "loss": 1.5253,
      "step": 5
    },
    {
      "epoch": 0.89,
      "eval_loss": 1.2364145517349243,
      "eval_runtime": 31.4311,
      "eval_samples_per_second": 11.645,
      "eval_steps_per_second": 0.732,
      "step": 6
    },
    {
      "epoch": 1.48,
      "grad_norm": 1.1785171025823682,
      "learning_rate": 1e-05,
      "loss": 1.2282,
      "step": 10
    },
    {
      "epoch": 1.93,
      "eval_loss": 1.1102970838546753,
      "eval_runtime": 31.4143,
      "eval_samples_per_second": 11.651,
      "eval_steps_per_second": 0.732,
      "step": 13
    },
    {
      "epoch": 2.22,
      "grad_norm": 0.7994256843261388,
      "learning_rate": 1.6853038769745466e-06,
      "loss": 1.1381,
      "step": 15
    },
    {
      "epoch": 2.67,
      "eval_loss": 1.0941128730773926,
      "eval_runtime": 31.5856,
      "eval_samples_per_second": 11.588,
      "eval_steps_per_second": 0.728,
      "step": 18
    },
    {
      "epoch": 2.67,
      "step": 18,
      "total_flos": 3716488888320.0,
      "train_loss": 1.272540317641364,
      "train_runtime": 388.3814,
      "train_samples_per_second": 3.291,
      "train_steps_per_second": 0.046
    }
  ],
  "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
}