roberta-base-qnli / eval_results.json
Jeremiah Zhou
End of training
af9690f
raw
history blame contribute delete
227 Bytes
{
"epoch": 6.0,
"eval_accuracy": 0.9245835621453414,
"eval_loss": 0.2992143929004669,
"eval_runtime": 8.9543,
"eval_samples": 5463,
"eval_samples_per_second": 610.1,
"eval_steps_per_second": 76.277
}