distilgpt2-multiprompt / all_results.json
pszemraj's picture
End of training
ddd029f
raw
history blame contribute delete
403 Bytes
{
"epoch": 2.0,
"eval_loss": 2.0213370323181152,
"eval_runtime": 38.4218,
"eval_samples": 12974,
"eval_samples_per_second": 337.672,
"eval_steps_per_second": 42.216,
"perplexity": 7.548410658690356,
"train_loss": 2.144662865950036,
"train_runtime": 4258.8623,
"train_samples": 247096,
"train_samples_per_second": 116.039,
"train_steps_per_second": 0.453
}