KO | EN

GitLite

Minimalist Open Source Repository Search Engine

gitlite — search

#developer-tools #typescript #codex #ai-agents #claude-code #dsh #dsh-plugin #cli #macos #ai-agent #deepseek-harness #open-source

Video-R4

★ 28 Open GitHub ↗

Reinforcing Text-Rich Video Reasoning with Visual Rumination

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Video-R4: Reinforcing Text-Rich Video Reasoning with Visual Rumination

> *[Yolo Yunlong Tang](https://yunlong10.github.io/)1, [Daiki Shimada](https://scholar.google.com/citations?user=1uAwouQAAAAJ&hl=en)2, [Hang Hua](https://hanghuacs.notion.site/)3, [Chao Huang](https://wikichao.github.io/)1, [Jing Bi](https://jing.vision/)1, [Rogerio Feris](https://www.rogerioferis.org/)3, [Chenliang Xu](https://www.cs.rochester.edu/u/cxu22/)1*

> *1University of Rochester, 2Sony Group Corporation, 3MIT-IBM Watson AI Lab*

## 🌟 News
- [2025-11-23] Introducing Video-R4, a reinforced video agent with visual rumination for text-rich video reasoning. The [arXiv paper](https://arxiv.org/abs/2511.17490) has been released. Code, model, and dataset are coming soon.

## 🚀 Video-R4 Training Framework
<img src="assets/training_framework.png" width="100%">

## 📊 Data Curation Pipeline
<img src="assets/data_pipeline.png" width="100%">

## 📈 Performance
<img src="assets/bar.png" width="100%">

## 📦 Installation
```bash
conda create -n video-r4 python=3.10
conda activate video-r4
git clone https://github.com/yunlong10/Video-R4.git
cd Video-R4
pip install -r requirements.txt
```

## 📖 Citation
If you find this work useful, please consider citing:
```bibtex
@article{tang2025video-r4,
  title={Video-R4: Reinforcing Text-Rich Video Reasoning with Visual Rumination},
  author={Tang, Yunlong and Shimada, Daiku and Hua, Hang and Huang, Chao and Bi, Jing and Feris, Rogerio and Xu, Chenliang},
  journal={arXiv preprint arXiv:2511.17490},
  year={2025}
}
```

## 🤝 Acknowledgments

This work was supported by Sony Group Corporation. We would like to thank Sayaka Nakamura and Jerry Jun Yokono for their insightful discussion.

We also thank the authors of the following projects for their contributions:
- [M4-ViteVQA](https://github.com/bytedance/VTVQA)
- [SlideVQA](https://github.com/nttmdlab-nlp/SlideVQA)
- [MP-DocVQA](https://github.com/rubenpt91/MP-DocVQA-Framework)
- [Open-R1](https://github.com/huggingface/open-r1)
- [OpenRLHF](https://github.com/OpenRLHF/OpenRLHF)
- [Ray](https://github.com/ray-project/ray)
- [Qwen2.5-VL](https://huggingface.co/collections/Qwen/qwen25-vl)
- [Video-R1](https://github.com/tulerfeng/Video-R1)
- [Pixel-Reasoner](https://github.com/TIGER-AI-Lab/Pixel-Reasoner)
- [DeepSeek-R1](https://github.com/deepseek-ai/DeepSeek-R1)
- [MVBench](https://huggingface.co/datasets/OpenGVLab/MVBench)
- [Video-MME](https://github.com/MME-Benchmarks/Video-MME)
- [Video-MMMU](https://videommmu.github.io/)

Video-R4

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README