AIlocal AIunderdog

Underdog shows off a 32MB speech model in new demo

TechCrunch AIIvan Mehta2026年11月6日0 次浏览来源

Despite its small size, Underdog's new model can generate both speech and music.