sedrickkeh's picture
End of training
a4908cb verified
raw
history blame contribute delete
209 Bytes
{
"epoch": 2.896,
"total_flos": 5.682962857616998e+16,
"train_loss": 0.6458813110987346,
"train_runtime": 3188.9509,
"train_samples_per_second": 0.937,
"train_steps_per_second": 0.014
}