all_results.json

{
    "eval_exact_match": 87.5,
    "eval_f1": 90.98214285714286,
    "eval_model_preparation_time": 0.0063,
    "eval_runtime": 9.1144,
    "eval_samples": 64,
    "eval_samples_per_second": 7.022,
    "eval_steps_per_second": 0.439
}