Viewing a single comment thread. View all comments

DukkyDrake t1_jduwklh wrote

You would be better off with a copy of the source training text dataset.

1