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": 1.422909488057392,
      "learning_rate": 1e-05,
      "loss": 1.172,
      "step": 1
    },
    {
      "epoch": 0.77,
      "grad_norm": 2.560781280087234,
      "learning_rate": 1.8314696123025456e-05,
      "loss": 1.1713,
      "step": 5
    },
    {
      "epoch": 0.92,
      "eval_loss": 1.3682807683944702,
      "eval_runtime": 4.0789,
      "eval_samples_per_second": 11.277,
      "eval_steps_per_second": 0.735,
      "step": 6
    },
    {
      "epoch": 1.54,
      "grad_norm": 0.8237183280783895,
      "learning_rate": 1e-05,
      "loss": 1.0185,
      "step": 10
    },
    {
      "epoch": 2.0,
      "eval_loss": 1.3434566259384155,
      "eval_runtime": 4.2029,
      "eval_samples_per_second": 10.945,
      "eval_steps_per_second": 0.714,
      "step": 13
    },
    {
      "epoch": 2.31,
      "grad_norm": 0.7082335303612385,
      "learning_rate": 1.6853038769745466e-06,
      "loss": 0.9011,
      "step": 15
    },
    {
      "epoch": 2.77,
      "eval_loss": 1.3273903131484985,
      "eval_runtime": 4.2006,
      "eval_samples_per_second": 10.951,
      "eval_steps_per_second": 0.714,
      "step": 18
    },
    {
      "epoch": 2.77,
      "step": 18,
      "total_flos": 3716488888320.0,
      "train_loss": 0.996643046538035,
      "train_runtime": 314.035,
      "train_samples_per_second": 3.869,
      "train_steps_per_second": 0.057
    }
  ],
  "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
}