run-11/checkpoint-1000/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.680965147453083,
  "eval_steps": 500,
  "global_step": 1000,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7582970261573792,
      "eval_loss": 0.47316744923591614,
      "eval_runtime": 9.4669,
      "eval_samples_per_second": 315.098,
      "eval_steps_per_second": 19.753,
      "step": 373
    },
    {
      "epoch": 1.3404825737265416,
      "grad_norm": 10.698078155517578,
      "learning_rate": 1.3190684345219897e-05,
      "loss": 0.5103,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7700302004814148,
      "eval_loss": 0.48503395915031433,
      "eval_runtime": 9.4683,
      "eval_samples_per_second": 315.05,
      "eval_steps_per_second": 19.75,
      "step": 746
    },
    {
      "epoch": 2.680965147453083,
      "grad_norm": 21.843738555908203,
      "learning_rate": 1.114878274379267e-05,
      "loss": 0.2684,
      "step": 1000
    }
  ],
  "logging_steps": 500,
  "max_steps": 3730,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 5530928427021096.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 1.5232585946647127e-05,
    "per_device_train_batch_size": 32
  }
}