2 ms·Significantly faster generation and training for AI-based audio systems7 points by el_duderino 8y ago