finetune/qnli/eval_results.json

{
    "epoch": 9.84,
    "eval_accuracy": 0.808398962020874,
    "eval_f1": 0.8294392523364487,
    "eval_loss": 0.5520442128181458,
    "eval_mcc": 0.6169307221841646,
    "eval_runtime": 3.1641,
    "eval_samples": 2286,
    "eval_samples_per_second": 722.485,
    "eval_steps_per_second": 90.39
}