Fetching the paper…

FaceTalk: Audio-Driven Motion Diffusion for Neural Parametric Head Models · Around