Text-to-Image Diffusion Models are Zero-Shot Video Generators - View it on GitHub