Despite its small size, Underdog's new model can generate both speech and music....
-
Underdog shows off a 32MB speech model in new demo
techcrunch.com
techcrunch.com