Shadowgraph
TrendingBuzzAbout
Sign in

longxiang-ai/awesome-video-diffusions

A curated and auto-updated collection of video diffusion / video generation papers from arXiv, covering text-to-video, image-to-video, controllable generation, world models, video editing, and 16+ research categories.

Python

46

137 commits

updated Oct 5, 2026

See the code

README

Awesome Video Diffusions Awesome

A curated list of latest research papers, projects and resources related to Video Diffusion Models and Video Generation. Content is automatically updated daily.

🗺️ Explore the Paper Atlas: an interactive paper map, monthly trends, topic network, co-author network and author rankings of every tracked paper.

Last Update: 2026-10-05 04:22:28

📰 Latest Updates

🗺️ [2026-09-30] Paper Atlas on GitHub Pages

  • New interactive site built from all daily snapshots, rebuilt after every update

🔧 [2026-08-08] Resilient Scheduled Updates

  • Temporary arXiv rate limits, server errors, and timeouts now preserve the latest valid data and finish with a warning
  • Added stable search exit codes, atomic publication, strict JSON validation, and fallback to any valid historical snapshot

🔎 [2026-08-08] Broader and More Accurate Video Indexing

  • Expanded coverage to 1,000 relevant papers across diffusion, flow matching, autoregressive generation, world models, editing, enhancement, and audio-video generation
  • Added broader arXiv domains, local relevance filtering, and boundary-aware category matching for acronyms such as DiT, T2V, I2V, and V2V

🚀 [2026-02] Project Launched — v1.0

  • Adapted from awesome-gaussians framework for tracking video diffusion research

  • Unified CLI: Single entry point python main.py with subcommands: init, search, suggest, export-bib, readme

  • Interactive Configuration Wizard: Run python main.py init to set up keywords, domains, time range, and API keys step-by-step

  • Custom Time Range Filtering: Support relative periods (6m, 1y, 2y) and absolute date ranges

  • Smart Link Extraction: Automatically extracts and classifies GitHub, project page, dataset, video, demo, and HuggingFace links from paper abstracts

  • BibTeX Export: Fetch BibTeX from arXiv and export to .bib files with category/date filters

  • LLM Keyword Suggestion: Paste a few paper titles or arXiv IDs, and an LLM automatically generates optimized search keywords

  • arXiv Domain Filtering: Restrict searches to specific arXiv categories (e.g., cs.CV, cs.AI, cs.MM)

  • 16 Research Categories: Comprehensive taxonomy covering T2V, I2V, video editing, controllable generation, world models, and more

  • View detailed updates: News.md 📋

Categories

Table of Contents

Categorized Papers

3D-aware Video Generation

Showing the latest 50 out of 52 papers

Applications

Showing the latest 50 out of 204 papers

Architecture & Efficiency

Showing the latest 50 out of 392 papers

Audio & Multi-modal

Showing the latest 50 out of 68 papers

Controllable Generation

Showing the latest 50 out of 328 papers

Human & Character Animation

Showing the latest 50 out of 60 papers

Image-to-Video Generation

Showing the latest 50 out of 93 papers

Long Video Generation

Showing the latest 50 out of 255 papers

Personalization & Customization

Showing the latest 50 out of 164 papers

Physical Understanding

Showing the latest 50 out of 310 papers

Surveys & Benchmarks

Showing the latest 50 out of 313 papers

Text-to-Video Generation

Showing the latest 50 out of 149 papers

Video Editing

Showing the latest 50 out of 93 papers

Video Inpainting & Completion

Video Super-Resolution & Enhancement

Showing the latest 50 out of 171 papers

World Models & Simulation

Showing the latest 50 out of 241 papers

Classic Papers

Open Source Projects

Tutorials & Blogs

📋 Project Features

🛠️ Core Features

  • Unified CLI (main.py): Single entry point with init, search, suggest, export-bib, readme subcommands
  • Interactive Config Wizard: Guided setup for keywords, domains, time range, and API keys via python main.py init
  • Custom Search Keywords: Configure keywords for title, abstract, or both; with arXiv domain filtering (cs.CV, cs.AI, cs.MM, etc.)
  • Time Range Filtering: Relative periods (30d, 6m, 1y, 2y) or absolute date ranges (YYYY-MM-DD to YYYY-MM-DD)
  • Smart Link Extraction: Auto-classifies URLs from abstracts into GitHub, project page, dataset, video, demo, HuggingFace links
  • BibTeX Export: Fetch BibTeX from arXiv official API; export to .bib files with category and date filters
  • LLM Keyword Suggestion: Input paper titles or arXiv IDs to auto-generate optimized search keywords via OpenAI-compatible API
  • Automated Paper Collection: Daily automatic crawling with GitHub Actions
  • Intelligent Classification: Auto-categorize papers into 16 topics (T2V, I2V, Video Editing, Controllable Generation, World Models, etc.)

🛠️ Technical Features

  • Robust Error Handling: Multi-layer retry and fallback strategies ensure stable operation
  • GitHub Actions Integration: Automated CI/CD workflows for daily updates
  • Multi-type Link Badges: README entries display PDF, GitHub (with stars), Project, Dataset, Video, Demo, HuggingFace, and Citation badges
  • Detailed Logging: Comprehensive logging for debugging and monitoring
  • Cross-Platform: Support for Windows/Linux/macOS

📚 Data Output

  • Paper JSON files (data/papers_YYYY-MM-DD.json): Full paper metadata with title, authors, abstract, links, keywords, BibTeX
  • BibTeX files (output/*.bib): Ready-to-use bibliography files for LaTeX
  • Auto-generated README: Categorized and formatted paper listings

🚀 Quick Start

1. Install Dependencies

pip install -r requirements.txt

python main.py init

This wizard walks you through:

  • Setting search keywords (for title, abstract, or both)
  • Selecting arXiv domains (e.g., cs.CV, cs.AI, cs.MM)
  • Configuring time range (relative like 6m/1y, or absolute dates)
  • Setting max results
  • Optionally configuring an OpenAI-compatible API key for keyword suggestion

3. Search Papers

# Search with settings from user_config.json
python main.py search

# Override: fetch 200 papers from the last 6 months, include BibTeX
python main.py search --max-results 200 --recent 6m --bibtex

# Search with absolute date range
python main.py search --date-from 2024-01-01 --date-to 2025-01-01

# Include citation counts from Semantic Scholar
python main.py search --citations

4. Export BibTeX

# Export all papers from the latest data file
python main.py export-bib --output output/references.bib

# Export only "Text-to-Video Generation" papers
python main.py export-bib --category "Text-to-Video Generation" --output output/t2v.bib

# Export papers from a specific date range
python main.py export-bib --date-from 2024-06-01 --date-to 2025-01-01 --output output/recent.bib

5. LLM Keyword Suggestion

# Generate keywords from paper titles
python main.py suggest --titles "Video Diffusion Models" "Stable Video Diffusion"

# Generate from arXiv IDs (auto-fetches titles)
python main.py suggest --arxiv-ids 2204.03458 2311.15127

# Auto-write suggested keywords to config
python main.py suggest --titles "Sora" "CogVideoX" --apply

# Use a custom API endpoint (e.g., DeepSeek)
python main.py suggest --titles "Paper Title" --base-url https://api.deepseek.com/v1 --api-key sk-xxx --model deepseek-chat

6. Generate README

# Basic README
python main.py readme

# Include latest papers section and abstracts
python main.py readme --show-latest --show-abstracts

Configuration File

All settings are stored in data/user_config.json:

{
  "search": {
    "keywords": {
      "both_abstract_and_title": ["video diffusion", "video generation", "text-to-video", "video-to-video"],
      "abstract_only": ["diffusion-based video generation", "flow-based video generation"],
      "title_only": ["world foundation model", "world simulator", "video tokenizer"]
    },
    "domains": ["cs.CV", "cs.AI", "cs.MM", "cs.LG", "cs.RO", "cs.GR", "eess.IV"],
    "time_range": {
      "mode": "relative",
      "relative": "1y"
    },
    "max_results": 1000
  },
  "api_keys": {
    "openai_api_key": "",
    "openai_base_url": "https://api.openai.com/v1",
    "openai_model": "gpt-4o-mini"
  }
}

Contribution Guidelines

Feel free to submit Pull Requests to improve this list! Please follow these formats:

  • Paper entry format: **[Paper Title](link)** - Brief description
  • Project entry format: [Project Name](link) - Project description

License

CC0

arxiv
arxiv-papers
awesome
awesome-list
controllable-generation
diffusion-models
image-to-video
research-paper
text-to-video
video-diffusion
video-editing
world-models

Similar repos

AlonzoLeeeooo/awesome-video-generation

A collection of awesome video generation studies.

786

QuenithAI/Video-Generation-Paper-List

Tracking the latest and greatest research papers on video generation.

171

ziqihuangg/Awesome-From-Video-Generation-to-World-Model

A list of works on video generation towards world model

520

inFaaa/Awesome-Personalized-Video-Creation

📖 This is a repository for organizing papers, codes, and other resources related to personalized…

66

wenhao728/awesome-diffusion-v2v

Awesome diffusion Video-to-Video (V2V). A collection of paper on diffusion model-based video…

291

Eyeline-Labs/Survey-Video-Diffusion

The official paper summary of TMLR'25 paper "Survey of Video Diffusion Models: Foundations,…

43

gracezhao1997/Awesome-Video-World-Models-with-AR-Diffusion

A Curated List of Awesome Video World Models with AR Diffusion: Covering Algorithms, Applications,…

713

minnie-lin/Awesome-Physics-Cognition-based-Video-Generation

A comprehensive list of papers investigating physical cognition in video generation, including…

326

longxiang-ai/awesome-video-diffusions

A curated and auto-updated collection of video diffusion / video generation papers from arXiv, covering text-to-video, image-to-video, controllable generation, world models, video editing, and 16+ research categories.

Python

46

137 commits

updated Oct 5, 2026

See the code

README

Awesome Video Diffusions Awesome

A curated list of latest research papers, projects and resources related to Video Diffusion Models and Video Generation. Content is automatically updated daily.

🗺️ Explore the Paper Atlas: an interactive paper map, monthly trends, topic network, co-author network and author rankings of every tracked paper.

Last Update: 2026-10-05 04:22:28

📰 Latest Updates

🗺️ [2026-09-30] Paper Atlas on GitHub Pages

  • New interactive site built from all daily snapshots, rebuilt after every update

🔧 [2026-08-08] Resilient Scheduled Updates

  • Temporary arXiv rate limits, server errors, and timeouts now preserve the latest valid data and finish with a warning
  • Added stable search exit codes, atomic publication, strict JSON validation, and fallback to any valid historical snapshot

🔎 [2026-08-08] Broader and More Accurate Video Indexing

  • Expanded coverage to 1,000 relevant papers across diffusion, flow matching, autoregressive generation, world models, editing, enhancement, and audio-video generation
  • Added broader arXiv domains, local relevance filtering, and boundary-aware category matching for acronyms such as DiT, T2V, I2V, and V2V

🚀 [2026-02] Project Launched — v1.0

  • Adapted from awesome-gaussians framework for tracking video diffusion research

  • Unified CLI: Single entry point python main.py with subcommands: init, search, suggest, export-bib, readme

  • Interactive Configuration Wizard: Run python main.py init to set up keywords, domains, time range, and API keys step-by-step

  • Custom Time Range Filtering: Support relative periods (6m, 1y, 2y) and absolute date ranges

  • Smart Link Extraction: Automatically extracts and classifies GitHub, project page, dataset, video, demo, and HuggingFace links from paper abstracts

  • BibTeX Export: Fetch BibTeX from arXiv and export to .bib files with category/date filters

  • LLM Keyword Suggestion: Paste a few paper titles or arXiv IDs, and an LLM automatically generates optimized search keywords

  • arXiv Domain Filtering: Restrict searches to specific arXiv categories (e.g., cs.CV, cs.AI, cs.MM)

  • 16 Research Categories: Comprehensive taxonomy covering T2V, I2V, video editing, controllable generation, world models, and more

  • View detailed updates: News.md 📋

Categories

Table of Contents

Categorized Papers

3D-aware Video Generation

Showing the latest 50 out of 52 papers

Applications

Showing the latest 50 out of 204 papers

Architecture & Efficiency

Showing the latest 50 out of 392 papers

Audio & Multi-modal

Showing the latest 50 out of 68 papers

Controllable Generation

Showing the latest 50 out of 328 papers

Human & Character Animation

Showing the latest 50 out of 60 papers

Image-to-Video Generation

Showing the latest 50 out of 93 papers

Long Video Generation

Showing the latest 50 out of 255 papers

Personalization & Customization

Showing the latest 50 out of 164 papers

Physical Understanding

Showing the latest 50 out of 310 papers

Surveys & Benchmarks

Showing the latest 50 out of 313 papers

Text-to-Video Generation

Showing the latest 50 out of 149 papers

Video Editing

Showing the latest 50 out of 93 papers

Video Inpainting & Completion

Video Super-Resolution & Enhancement

Showing the latest 50 out of 171 papers

World Models & Simulation

Showing the latest 50 out of 241 papers

Classic Papers

Open Source Projects

Tutorials & Blogs

📋 Project Features

🛠️ Core Features

  • Unified CLI (main.py): Single entry point with init, search, suggest, export-bib, readme subcommands
  • Interactive Config Wizard: Guided setup for keywords, domains, time range, and API keys via python main.py init
  • Custom Search Keywords: Configure keywords for title, abstract, or both; with arXiv domain filtering (cs.CV, cs.AI, cs.MM, etc.)
  • Time Range Filtering: Relative periods (30d, 6m, 1y, 2y) or absolute date ranges (YYYY-MM-DD to YYYY-MM-DD)
  • Smart Link Extraction: Auto-classifies URLs from abstracts into GitHub, project page, dataset, video, demo, HuggingFace links
  • BibTeX Export: Fetch BibTeX from arXiv official API; export to .bib files with category and date filters
  • LLM Keyword Suggestion: Input paper titles or arXiv IDs to auto-generate optimized search keywords via OpenAI-compatible API
  • Automated Paper Collection: Daily automatic crawling with GitHub Actions
  • Intelligent Classification: Auto-categorize papers into 16 topics (T2V, I2V, Video Editing, Controllable Generation, World Models, etc.)

🛠️ Technical Features

  • Robust Error Handling: Multi-layer retry and fallback strategies ensure stable operation
  • GitHub Actions Integration: Automated CI/CD workflows for daily updates
  • Multi-type Link Badges: README entries display PDF, GitHub (with stars), Project, Dataset, Video, Demo, HuggingFace, and Citation badges
  • Detailed Logging: Comprehensive logging for debugging and monitoring
  • Cross-Platform: Support for Windows/Linux/macOS

📚 Data Output

  • Paper JSON files (data/papers_YYYY-MM-DD.json): Full paper metadata with title, authors, abstract, links, keywords, BibTeX
  • BibTeX files (output/*.bib): Ready-to-use bibliography files for LaTeX
  • Auto-generated README: Categorized and formatted paper listings

🚀 Quick Start

1. Install Dependencies

pip install -r requirements.txt

python main.py init

This wizard walks you through:

  • Setting search keywords (for title, abstract, or both)
  • Selecting arXiv domains (e.g., cs.CV, cs.AI, cs.MM)
  • Configuring time range (relative like 6m/1y, or absolute dates)
  • Setting max results
  • Optionally configuring an OpenAI-compatible API key for keyword suggestion

3. Search Papers

# Search with settings from user_config.json
python main.py search

# Override: fetch 200 papers from the last 6 months, include BibTeX
python main.py search --max-results 200 --recent 6m --bibtex

# Search with absolute date range
python main.py search --date-from 2024-01-01 --date-to 2025-01-01

# Include citation counts from Semantic Scholar
python main.py search --citations

4. Export BibTeX

# Export all papers from the latest data file
python main.py export-bib --output output/references.bib

# Export only "Text-to-Video Generation" papers
python main.py export-bib --category "Text-to-Video Generation" --output output/t2v.bib

# Export papers from a specific date range
python main.py export-bib --date-from 2024-06-01 --date-to 2025-01-01 --output output/recent.bib

5. LLM Keyword Suggestion

# Generate keywords from paper titles
python main.py suggest --titles "Video Diffusion Models" "Stable Video Diffusion"

# Generate from arXiv IDs (auto-fetches titles)
python main.py suggest --arxiv-ids 2204.03458 2311.15127

# Auto-write suggested keywords to config
python main.py suggest --titles "Sora" "CogVideoX" --apply

# Use a custom API endpoint (e.g., DeepSeek)
python main.py suggest --titles "Paper Title" --base-url https://api.deepseek.com/v1 --api-key sk-xxx --model deepseek-chat

6. Generate README

# Basic README
python main.py readme

# Include latest papers section and abstracts
python main.py readme --show-latest --show-abstracts

Configuration File

All settings are stored in data/user_config.json:

{
  "search": {
    "keywords": {
      "both_abstract_and_title": ["video diffusion", "video generation", "text-to-video", "video-to-video"],
      "abstract_only": ["diffusion-based video generation", "flow-based video generation"],
      "title_only": ["world foundation model", "world simulator", "video tokenizer"]
    },
    "domains": ["cs.CV", "cs.AI", "cs.MM", "cs.LG", "cs.RO", "cs.GR", "eess.IV"],
    "time_range": {
      "mode": "relative",
      "relative": "1y"
    },
    "max_results": 1000
  },
  "api_keys": {
    "openai_api_key": "",
    "openai_base_url": "https://api.openai.com/v1",
    "openai_model": "gpt-4o-mini"
  }
}

Contribution Guidelines

Feel free to submit Pull Requests to improve this list! Please follow these formats:

  • Paper entry format: **[Paper Title](link)** - Brief description
  • Project entry format: [Project Name](link) - Project description

License

CC0

arxiv
arxiv-papers
awesome
awesome-list
controllable-generation
diffusion-models
image-to-video
research-paper
text-to-video
video-diffusion
video-editing
world-models

Similar repos

AlonzoLeeeooo/awesome-video-generation

A collection of awesome video generation studies.

786

QuenithAI/Video-Generation-Paper-List

Tracking the latest and greatest research papers on video generation.

171

ziqihuangg/Awesome-From-Video-Generation-to-World-Model

A list of works on video generation towards world model

520

inFaaa/Awesome-Personalized-Video-Creation

📖 This is a repository for organizing papers, codes, and other resources related to personalized…

66

wenhao728/awesome-diffusion-v2v

Awesome diffusion Video-to-Video (V2V). A collection of paper on diffusion model-based video…

291

Eyeline-Labs/Survey-Video-Diffusion

The official paper summary of TMLR'25 paper "Survey of Video Diffusion Models: Foundations,…

43

gracezhao1997/Awesome-Video-World-Models-with-AR-Diffusion

A Curated List of Awesome Video World Models with AR Diffusion: Covering Algorithms, Applications,…

713

minnie-lin/Awesome-Physics-Cognition-based-Video-Generation

A comprehensive list of papers investigating physical cognition in video generation, including…

326