---
license: fair-noncommercial-research-license
library_name: transformers
pipeline_tag: image-text-to-text
tags:
- qwen3-vl
- video-language-model
- region-understanding
- motion-captioning
---

# MotionAtlas-4B

This repository contains the MotionAtlas-4B model for **MotionAtlas: Detailed Region Captioning for Motion-Centric Videos**.

**TL; DR:** MotionAtlas shifts motion captioning from global video descriptions to region-aware motion captions, enabling precise evaluation with MotionAtlas-Bench and scalable training with MotionAtlas-Data. The model is designed for detailed motion-centric video understanding over referred regions.

## Usage

For detailed usage of this model, please refer to our [GitHub repo](https://github.com/Kagura-0001/MotionAtlas) and [project page](https://kagura-0001.github.io/projects/MotionAtlas/?v=2).