gpt_train_2_768_new / eval_results.json
gokulsrinivasagan's picture
End of training
bc548b6 verified
raw
history blame contribute delete
279 Bytes
{
"epoch": 9.321127579192096,
"eval_accuracy": 0.3345458864628659,
"eval_loss": 3.945629119873047,
"eval_runtime": 233.188,
"eval_samples": 33767,
"eval_samples_per_second": 144.806,
"eval_steps_per_second": 1.51,
"perplexity": 51.70885895206038
}