Easily train a good VC model with voice data <= 10 mins!
1 min voice data can also be used to train a good TTS model! (few shot voice cloning)
A generative speech model for daily dialogue.