djbp's picture
End of training
a62661a verified
raw
history blame contribute delete
222 Bytes
{
"epoch": 6.885245901639344,
"total_flos": 4.1785806114677883e+18,
"train_loss": 0.4162924766540527,
"train_runtime": 7056.8141,
"train_samples_per_second": 7.675,
"train_steps_per_second": 0.015
}