{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 20.0,
  "eval_steps": 5,
  "global_step": 20,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 22.02356719970703,
      "learning_rate": 2e-05,
      "loss": 1.2546,
      "step": 1
    },
    {
      "epoch": 2.0,
      "grad_norm": 21.681835174560547,
      "learning_rate": 1.9863613034027224e-05,
      "loss": 1.2546,
      "step": 2
    },
    {
      "epoch": 3.0,
      "grad_norm": 12.661576271057129,
      "learning_rate": 1.9458172417006347e-05,
      "loss": 1.1806,
      "step": 3
    },
    {
      "epoch": 4.0,
      "grad_norm": 9.024881362915039,
      "learning_rate": 1.879473751206489e-05,
      "loss": 1.0682,
      "step": 4
    },
    {
      "epoch": 5.0,
      "grad_norm": 8.294164657592773,
      "learning_rate": 1.789140509396394e-05,
      "loss": 0.9835,
      "step": 5
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9387755102040817,
      "eval_accuracy_<|content|>": 0.0,
      "eval_accuracy_<|from|>": -1,
      "eval_accuracy_<|recipient|>": -1,
      "eval_accuracy_<|stop|>": 0.0,
      "eval_accuracy_first_token": 1.0,
      "eval_accuracy_first_token_get": 1.0,
      "eval_accuracy_first_token_get_total": 2,
      "eval_accuracy_total_num_<|content|>": 2,
      "eval_accuracy_total_num_<|from|>": 0,
      "eval_accuracy_total_num_<|recipient|>": 0,
      "eval_accuracy_total_num_<|stop|>": 2,
      "eval_first_token_param_values": 0.9,
      "eval_first_token_param_values_total": 10,
      "eval_loss": 0.8951861262321472,
      "eval_perplexity": 1.010767880290677,
      "eval_runtime": 0.8904,
      "eval_samples_per_second": 1.123,
      "eval_steps_per_second": 1.123,
      "eval_total_number_first_token": 2,
      "step": 5
    },
    {
      "epoch": 6.0,
      "grad_norm": 8.226706504821777,
      "learning_rate": 1.6772815716257414e-05,
      "loss": 0.8936,
      "step": 6
    },
    {
      "epoch": 7.0,
      "grad_norm": 8.382328987121582,
      "learning_rate": 1.5469481581224274e-05,
      "loss": 0.8304,
      "step": 7
    },
    {
      "epoch": 8.0,
      "grad_norm": 7.861330986022949,
      "learning_rate": 1.4016954246529697e-05,
      "loss": 0.7703,
      "step": 8
    },
    {
      "epoch": 9.0,
      "grad_norm": 7.340066909790039,
      "learning_rate": 1.2454854871407993e-05,
      "loss": 0.7171,
      "step": 9
    },
    {
      "epoch": 10.0,
      "grad_norm": 6.915948390960693,
      "learning_rate": 1.0825793454723325e-05,
      "loss": 0.6641,
      "step": 10
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9489795918367347,
      "eval_accuracy_<|content|>": 0.0,
      "eval_accuracy_<|from|>": -1,
      "eval_accuracy_<|recipient|>": -1,
      "eval_accuracy_<|stop|>": 0.0,
      "eval_accuracy_first_token": 1.0,
      "eval_accuracy_first_token_get": 1.0,
      "eval_accuracy_first_token_get_total": 2,
      "eval_accuracy_total_num_<|content|>": 2,
      "eval_accuracy_total_num_<|from|>": 0,
      "eval_accuracy_total_num_<|recipient|>": 0,
      "eval_accuracy_total_num_<|stop|>": 2,
      "eval_first_token_param_values": 0.9,
      "eval_first_token_param_values_total": 10,
      "eval_loss": 0.6229805946350098,
      "eval_perplexity": 1.0074814049420708,
      "eval_runtime": 0.8828,
      "eval_samples_per_second": 1.133,
      "eval_steps_per_second": 1.133,
      "eval_total_number_first_token": 2,
      "step": 10
    },
    {
      "epoch": 11.0,
      "grad_norm": 6.680131912231445,
      "learning_rate": 9.174206545276678e-06,
      "loss": 0.6224,
      "step": 11
    },
    {
      "epoch": 12.0,
      "grad_norm": 5.9928297996521,
      "learning_rate": 7.545145128592009e-06,
      "loss": 0.5933,
      "step": 12
    },
    {
      "epoch": 13.0,
      "grad_norm": 5.214223861694336,
      "learning_rate": 5.983045753470308e-06,
      "loss": 0.5704,
      "step": 13
    },
    {
      "epoch": 14.0,
      "grad_norm": 4.801523685455322,
      "learning_rate": 4.530518418775734e-06,
      "loss": 0.5552,
      "step": 14
    },
    {
      "epoch": 15.0,
      "grad_norm": 4.650667667388916,
      "learning_rate": 3.2271842837425917e-06,
      "loss": 0.545,
      "step": 15
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.9489795918367347,
      "eval_accuracy_<|content|>": 0.0,
      "eval_accuracy_<|from|>": -1,
      "eval_accuracy_<|recipient|>": -1,
      "eval_accuracy_<|stop|>": 0.0,
      "eval_accuracy_first_token": 1.0,
      "eval_accuracy_first_token_get": 1.0,
      "eval_accuracy_first_token_get_total": 2,
      "eval_accuracy_total_num_<|content|>": 2,
      "eval_accuracy_total_num_<|from|>": 0,
      "eval_accuracy_total_num_<|recipient|>": 0,
      "eval_accuracy_total_num_<|stop|>": 2,
      "eval_first_token_param_values": 0.9,
      "eval_first_token_param_values_total": 10,
      "eval_loss": 0.5376743674278259,
      "eval_perplexity": 1.006453660117816,
      "eval_runtime": 0.8811,
      "eval_samples_per_second": 1.135,
      "eval_steps_per_second": 1.135,
      "eval_total_number_first_token": 2,
      "step": 15
    },
    {
      "epoch": 16.0,
      "grad_norm": 4.648330211639404,
      "learning_rate": 2.1085949060360654e-06,
      "loss": 0.5376,
      "step": 16
    },
    {
      "epoch": 17.0,
      "grad_norm": 4.580282211303711,
      "learning_rate": 1.2052624879351105e-06,
      "loss": 0.5339,
      "step": 17
    },
    {
      "epoch": 18.0,
      "grad_norm": 4.567606449127197,
      "learning_rate": 5.418275829936537e-07,
      "loss": 0.5301,
      "step": 18
    },
    {
      "epoch": 19.0,
      "grad_norm": 4.5482001304626465,
      "learning_rate": 1.3638696597277678e-07,
      "loss": 0.528,
      "step": 19
    },
    {
      "epoch": 20.0,
      "grad_norm": 4.52694034576416,
      "learning_rate": 0.0,
      "loss": 0.5258,
      "step": 20
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.9489795918367347,
      "eval_accuracy_<|content|>": 0.0,
      "eval_accuracy_<|from|>": -1,
      "eval_accuracy_<|recipient|>": -1,
      "eval_accuracy_<|stop|>": 0.0,
      "eval_accuracy_first_token": 1.0,
      "eval_accuracy_first_token_get": 1.0,
      "eval_accuracy_first_token_get_total": 2,
      "eval_accuracy_total_num_<|content|>": 2,
      "eval_accuracy_total_num_<|from|>": 0,
      "eval_accuracy_total_num_<|recipient|>": 0,
      "eval_accuracy_total_num_<|stop|>": 2,
      "eval_first_token_param_values": 0.9,
      "eval_first_token_param_values_total": 10,
      "eval_loss": 0.5257245302200317,
      "eval_perplexity": 1.0063097771319383,
      "eval_runtime": 0.8846,
      "eval_samples_per_second": 1.131,
      "eval_steps_per_second": 1.131,
      "eval_total_number_first_token": 2,
      "step": 20
    },
    {
      "epoch": 20.0,
      "step": 20,
      "total_flos": 7935331915530240.0,
      "train_loss": 0.7579435586929322,
      "train_runtime": 380.9512,
      "train_samples_per_second": 0.053,
      "train_steps_per_second": 0.053
    }
  ],
  "logging_steps": 1.0,
  "max_steps": 20,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 5,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 7935331915530240.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}