phunc20's picture
End of training
2338bfd
raw
history blame
209 Bytes
{
"epoch": 50.0,
"total_flos": 3.989386232229888e+17,
"train_loss": 0.12627770403089622,
"train_runtime": 1656.8438,
"train_samples_per_second": 9.687,
"train_steps_per_second": 0.091
}