trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 4,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_logits/chosen": -17.38763427734375,
      "eval_logits/rejected": -17.937782287597656,
      "eval_logps/chosen": -1147.5987548828125,
      "eval_logps/rejected": -1209.950439453125,
      "eval_loss": 0.0057400548830628395,
      "eval_rewards/accuracies": 1.0,
      "eval_rewards/chosen": 8.250604629516602,
      "eval_rewards/margins": 20.80431365966797,
      "eval_rewards/rejected": -12.553709030151367,
      "eval_runtime": 20.8787,
      "eval_samples_per_second": 0.239,
      "eval_steps_per_second": 0.096,
      "step": 2
    },
    {
      "epoch": 2.0,
      "eval_logits/chosen": -17.409107208251953,
      "eval_logits/rejected": -17.894569396972656,
      "eval_logps/chosen": -1115.763427734375,
      "eval_logps/rejected": -1251.18212890625,
      "eval_loss": 0.0001020339404931292,
      "eval_rewards/accuracies": 1.0,
      "eval_rewards/chosen": 11.434137344360352,
      "eval_rewards/margins": 28.111000061035156,
      "eval_rewards/rejected": -16.676864624023438,
      "eval_runtime": 22.1862,
      "eval_samples_per_second": 0.225,
      "eval_steps_per_second": 0.09,
      "step": 4
    }
  ],
  "logging_steps": 10,
  "max_steps": 6,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}