squarerun / train_results.json
corranm's picture
End of training
9207f1d verified
raw
history blame contribute delete
208 Bytes
{
"epoch": 45.0,
"total_flos": 1.611130661124526e+18,
"train_loss": 0.5009927850430724,
"train_runtime": 1277.117,
"train_samples_per_second": 16.279,
"train_steps_per_second": 1.022
}