Prot10's picture
End of training
285604a
{
"epoch": 30.0,
"total_flos": 4.840276186658304e+18,
"train_loss": 1.0325976332028708,
"train_runtime": 3493.542,
"train_samples_per_second": 17.492,
"train_steps_per_second": 0.137
}