Hacker News new | ask | show | jobs
by olup 574 days ago
They say the models are under 3b parameters. If only for voice generation it sounds pretty good, no ?