{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.0015333895576171127,
  "eval_steps": 3,
  "global_step": 10,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.00015333895576171127,
      "grad_norm": NaN,
      "learning_rate": 2e-05,
      "loss": 0.5493,
      "step": 1
    },
    {
      "epoch": 0.00015333895576171127,
      "eval_loss": NaN,
      "eval_runtime": 66.8613,
      "eval_samples_per_second": 41.07,
      "eval_steps_per_second": 20.535,
      "step": 1
    },
    {
      "epoch": 0.00030667791152342254,
      "grad_norm": NaN,
      "learning_rate": 4e-05,
      "loss": 0.0,
      "step": 2
    },
    {
      "epoch": 0.0004600168672851338,
      "grad_norm": NaN,
      "learning_rate": 6e-05,
      "loss": 0.0,
      "step": 3
    },
    {
      "epoch": 0.0004600168672851338,
      "eval_loss": NaN,
      "eval_runtime": 65.8107,
      "eval_samples_per_second": 41.726,
      "eval_steps_per_second": 20.863,
      "step": 3
    },
    {
      "epoch": 0.0006133558230468451,
      "grad_norm": NaN,
      "learning_rate": 8e-05,
      "loss": 0.0,
      "step": 4
    },
    {
      "epoch": 0.0007666947788085563,
      "grad_norm": NaN,
      "learning_rate": 0.0001,
      "loss": 1.1171,
      "step": 5
    },
    {
      "epoch": 0.0009200337345702676,
      "grad_norm": NaN,
      "learning_rate": 0.00012,
      "loss": 0.1073,
      "step": 6
    },
    {
      "epoch": 0.0009200337345702676,
      "eval_loss": NaN,
      "eval_runtime": 65.7313,
      "eval_samples_per_second": 41.776,
      "eval_steps_per_second": 20.888,
      "step": 6
    },
    {
      "epoch": 0.0010733726903319788,
      "grad_norm": NaN,
      "learning_rate": 0.00014,
      "loss": 0.0,
      "step": 7
    },
    {
      "epoch": 0.0012267116460936902,
      "grad_norm": NaN,
      "learning_rate": 0.00016,
      "loss": 0.0,
      "step": 8
    },
    {
      "epoch": 0.0013800506018554013,
      "grad_norm": NaN,
      "learning_rate": 0.00018,
      "loss": 0.0,
      "step": 9
    },
    {
      "epoch": 0.0013800506018554013,
      "eval_loss": NaN,
      "eval_runtime": 65.6685,
      "eval_samples_per_second": 41.816,
      "eval_steps_per_second": 20.908,
      "step": 9
    },
    {
      "epoch": 0.0015333895576171127,
      "grad_norm": NaN,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "step": 10
    }
  ],
  "logging_steps": 1,
  "max_steps": 10,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 3,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 706576901996544.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}