trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 20,
  "global_step": 15,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.06701570680628273,
      "grad_norm": 2.5355777593627256,
      "learning_rate": 0.0,
      "logits/chosen": -3.060791015625,
      "logits/rejected": NaN,
      "logps/chosen": -281.296875,
      "logps/rejected": -261.578125,
      "loss": 0.6914,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 1.0,
      "step": 15,
      "total_flos": 0.0,
      "train_loss": 0.6713338216145833,
      "train_runtime": 13324.2114,
      "train_samples_per_second": 4.768,
      "train_steps_per_second": 0.001
    }
  ],
  "logging_steps": 20,
  "max_steps": 15,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 30,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}