Abstract: Text-to-Video generation has achieved remarkable progress with the rise of diffusion models. In this work, we introduce Cached Memory-Guided Video Generation (Corgi), aiming to generate ...