HOLD: Category-agnostic 3D Reconstruction of Interacting Hands and Objects from Video | Read Paper on Bytez