Mapping the video diffusion paper explosion

What it does This repository is the companion awesome-list to an ACM Computing Surveys paper on video diffusion models. It organizes papers and projects into dense markdown tables, covering datasets, text-to-video generation, conditional synthesis, editing, and understanding. Think of it as a literature map for a field that publishes faster than most people can read.

The interesting bit The list goes beyond the usual text-to-video suspects, cataloging brain-guided generation, sound-guided editing, and even non-diffusion baselines alongside the core methods. It also mixes closed APIs like Sora and Pika with open-source repos such as CogVideoX and Open-Sora in the same tables, giving a rare side-by-side view of what is public versus proprietary.

Key highlights

• Tied to a peer-reviewed ACM CSUR survey rather than an informal blog post.
• Tracks the full research pipeline: caption datasets, training-based and training-free generation, video completion, and modality-guided editing.
• Maintains an up-to-date roster of open-source toolboxes and foundation models, including Stable Video Diffusion, AnimateDiff, and VideoCraft.
• Covers niche control modalities like pose, sound, and brain signals for video generation.
• Includes recent work through May 2025, suggesting active maintenance.

Caveats

• The README is a collection of markdown tables, not a runnable framework or library.
• Several typos (“Genetation,” “gudied”) appear in the table of contents and headers.

Verdict Researchers and engineers who need to navigate the video diffusion literature without reading every arXiv abstract will save time here. If you are looking for a unified training framework, keep scrolling—this is strictly a curated reading list.