4 ms·Correct, we may have a few MB of text that is training data. But the trained models are several GB'sby olavgg 10y agoCorrect, we may have a few MB of text that is training data. But the trained models are several GB's