S2V, an open-source speech-to-video model that transforms portrait photos into film-quality avatars capable of speaking, singing, and performing, now available on Hugging Face, GitHub, and ModelScope.