sl82's picture
End of training
bf4523d
raw
history blame
210 Bytes
{
"epoch": 3.0,
"total_flos": 1.8124066505760768e+18,
"train_loss": 0.33203531901041666,
"train_runtime": 1146.1315,
"train_samples_per_second": 63.605,
"train_steps_per_second": 0.497
}