4 ms·

Loading ML models into GPU memory.