Underdog shows off a 32MB speech model in new demo
Despite its small size, Underdog's new model can generate both speech and music.
This is a summary curated by AIFuture. Read the complete article at the original source:
Read the full story on TechCrunch