2 ms·

I feel like this is still very useful comparison precisely for the reason that those two models were trained differently.