{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 681,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.2,
      "learning_rate": 1.8695652173913045e-05,
      "loss": 1.7027,
      "step": 45
    },
    {
      "epoch": 0.4,
      "learning_rate": 2.911764705882353e-05,
      "loss": 1.4439,
      "step": 90
    },
    {
      "epoch": 0.59,
      "learning_rate": 2.6911764705882354e-05,
      "loss": 1.3038,
      "step": 135
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.4705882352941174e-05,
      "loss": 1.3647,
      "step": 180
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.25e-05,
      "loss": 1.3154,
      "step": 225
    },
    {
      "epoch": 1.19,
      "learning_rate": 2.0294117647058825e-05,
      "loss": 1.3009,
      "step": 270
    },
    {
      "epoch": 1.39,
      "learning_rate": 1.8088235294117645e-05,
      "loss": 1.2918,
      "step": 315
    },
    {
      "epoch": 1.59,
      "learning_rate": 1.5882352941176473e-05,
      "loss": 1.2529,
      "step": 360
    },
    {
      "epoch": 1.78,
      "learning_rate": 1.3676470588235295e-05,
      "loss": 1.289,
      "step": 405
    },
    {
      "epoch": 1.98,
      "learning_rate": 1.1470588235294117e-05,
      "loss": 1.2225,
      "step": 450
    },
    {
      "epoch": 2.18,
      "learning_rate": 9.264705882352942e-06,
      "loss": 1.2486,
      "step": 495
    },
    {
      "epoch": 2.38,
      "learning_rate": 7.058823529411765e-06,
      "loss": 1.1968,
      "step": 540
    },
    {
      "epoch": 2.58,
      "learning_rate": 4.852941176470589e-06,
      "loss": 1.2706,
      "step": 585
    },
    {
      "epoch": 2.78,
      "learning_rate": 2.647058823529412e-06,
      "loss": 1.2685,
      "step": 630
    },
    {
      "epoch": 2.97,
      "learning_rate": 4.411764705882353e-07,
      "loss": 1.2211,
      "step": 675
    }
  ],
  "logging_steps": 45,
  "max_steps": 681,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 5.955994886602752e+16,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}