Speech-driven 3D facial animation synthesis has been a challenging task both in industry
and research. Recent methods mostly focus on deterministic deep learning methods …