kg59's picture
End of training
0dceb78
{
"epoch": 4.0,
"total_flos": 2.1362991696950723e+18,
"train_loss": 0.4688039951854282,
"train_runtime": 1140.3926,
"train_samples_per_second": 24.174,
"train_steps_per_second": 0.189
}