New
qwenaudio/cosyvoice — Fun-CosyVoice 3.0 multilingual speech synthesis model
qwenaudio/cosyvoice is a repository of multilingual speech generation models, and the Fun-CosyVoice 3.0 introduced in its README supports 9 languages, more than 18 Chinese dialects and accents, and zero-shot voice cloning.