Learning Disentangled Avatars with Hybrid 3D Representations. (Face, Body, Hair and Clothing) - View it on GitHub