What is awesome-video-diffusion?

This is a curated reference list tracking research papers and open-source tools in the area of AI-generated video. The underlying technology, called diffusion modeling, is a method for training AI systems to create realistic video from text descriptions, edit existing footage, restore low-quality video, and generate animations of people or objects moving. The list is organized into categories covering different applications of this technology. Some sections focus on creating video from scratch using text prompts. Others cover editing existing video clips, controlling how subjects or camera motion behave, generating long-form video or film-length content, and specialized uses like talking-head generation (making a still face appear to speak), virtual try-on (showing clothing on a person), and 3D or 4D content creation. Additional sections cover video understanding (having AI analyze what is happening in a video rather than create it), health and biology applications, game world generation, and AI safety research related to video models. Each entry in the list links to a research paper on arXiv, a GitHub repository, or a project website, so readers can explore the actual code or technical details behind any entry. The repo is maintained by researchers at the Show Lab and is updated frequently as new work appears. It does not contain runnable code itself. It is a navigation tool for anyone trying to understand what exists in this field, from open-source toolkits to commercial products to academic benchmarks. The full README is longer than what was shown.