Fetching the paper…

FancyVideo: Towards Dynamic and Consistent Video Generation via Cross-frame Textual Guidance · Around