{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 522,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.29,
      "learning_rate": 1.8084291187739463e-05,
      "loss": 1.8363,
      "step": 50
    },
    {
      "epoch": 0.57,
      "learning_rate": 1.6168582375478928e-05,
      "loss": 1.7907,
      "step": 100
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4252873563218392e-05,
      "loss": 1.7897,
      "step": 150
    },
    {
      "epoch": 1.15,
      "learning_rate": 1.2337164750957855e-05,
      "loss": 1.7892,
      "step": 200
    },
    {
      "epoch": 1.44,
      "learning_rate": 1.0421455938697318e-05,
      "loss": 1.7858,
      "step": 250
    },
    {
      "epoch": 1.72,
      "learning_rate": 8.505747126436782e-06,
      "loss": 1.7702,
      "step": 300
    },
    {
      "epoch": 2.01,
      "learning_rate": 6.590038314176246e-06,
      "loss": 1.7661,
      "step": 350
    },
    {
      "epoch": 2.3,
      "learning_rate": 4.674329501915709e-06,
      "loss": 1.7512,
      "step": 400
    },
    {
      "epoch": 2.59,
      "learning_rate": 2.7586206896551725e-06,
      "loss": 1.7275,
      "step": 450
    },
    {
      "epoch": 2.87,
      "learning_rate": 8.429118773946361e-07,
      "loss": 1.661,
      "step": 500
    },
    {
      "epoch": 2.87,
      "eval_accuracy": 0.3688046647230321,
      "eval_loss": 1.633470058441162,
      "eval_runtime": 1.6169,
      "eval_samples_per_second": 424.268,
      "eval_steps_per_second": 26.594,
      "step": 500
    },
    {
      "epoch": 3.0,
      "step": 522,
      "total_flos": 1735431130580256.0,
      "train_loss": 1.7622158079768506,
      "train_runtime": 215.5621,
      "train_samples_per_second": 77.212,
      "train_steps_per_second": 2.422
    }
  ],
  "logging_steps": 50,
  "max_steps": 522,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 1735431130580256.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}