julienmercier's picture
End of training
d352e4e
raw
history blame
209 Bytes
{
"epoch": 2.97,
"total_flos": 1.6261064842200515e+18,
"train_loss": 0.18832773187933446,
"train_runtime": 875.8058,
"train_samples_per_second": 24.183,
"train_steps_per_second": 0.25
}