OmniVoice Singing

A multilingual zero-shot TTS model that extends OmniVoice with controllable singing and expressive tags, enabling voice-consistent sung phrases and stylized speech synthesis from text.