andrecastro's picture
End of training
62f6163
raw
history blame
208 Bytes
{
"epoch": 3.0,
"total_flos": 7.802294942660198e+17,
"train_loss": 0.1717462847037782,
"train_runtime": 1149.2673,
"train_samples_per_second": 27.333,
"train_steps_per_second": 0.854
}