Bisher's picture
End of training
6cd5852 verified
raw
history blame contribute delete
226 Bytes
{
"epoch": 0.19583843329253367,
"total_flos": 6.602768364719419e+17,
"train_loss": 0.021594215556979178,
"train_runtime": 3210.596,
"train_samples_per_second": 193.241,
"train_steps_per_second": 0.318
}