Efficient Video Diffusion Models via Content-Frame Motion-Latent Decomposition (ICLR 2024) - View it on GitHub