distilled-mt5-small-0.02-0.5 / train_results.json
Lvxue's picture
End of training
c2cda95
{
"epoch": 5.0,
"train_loss": 4.472965654296875,
"train_runtime": 2699.6495,
"train_samples": 10000,
"train_samples_per_second": 18.521,
"train_steps_per_second": 4.63
}