---
title: UCM
canonical_url: "https://www.modelscope.cn/models/DAMOXR/UCM"
md_url: "https://www.modelscope.cn/models/DAMOXR/UCM.md"
repository: DAMOXR/UCM
chinese_name: UCM
last_updated: 2026-08-07
downloads: 5
stars: 0
---

# UCM

> UCM - DAMOXR 在 ModelScope 开源的模型。Tian-Xing Xu 1 , Zi-Xuan Wang 1 , Guangyuan Wang †2 , Li Hu †2 , Zhongyi Zhang 3 , Peng Zhang 2 , Bang Zhang ‡2 , Song-Hai Zhang ‡1

- **Repository**: DAMOXR/UCM
- **Downloads**: 5
- **Stars**: 0
- **Last updated**: 2026-08-07

Source: https://www.modelscope.cn/models/DAMOXR/UCM

---

<!-- ## ___***UCM: Unified Modeling of Camera Control and Memory with Time-aware Positional Encoding Warping for World Models***___
<div align="center">

_**[Tian-Xing Xu<sup>*1</sup>](https://scholar.google.com/citations?user=zHp0rMIAAAAJ), 
[Zi-Xuan Wang<sup>*1</sup>](https://scholar.google.com/citations?user=4JKwZRsAAAAJ), 
[Guangyuan Wang<sup>*&dagger;2</sup>](https://scholar.google.com/citations?user=OYwm0GsAAAAJ&hl=zh-CN), 
[Li Hu<sup>&dagger;2</sup>](https://scholar.google.com/citations?user=Arz3iGUAAAAJ), 
[Zhongyi Zhang<sup>3</sup>](https://scholar.google.com/citations?user=dOWmKOUAAAAJ), 
[Peng Zhang<sup>2</sup>](https://scholar.google.com/citations?user=QTgxKmkAAAAJ), 
Bang Zhang<sup>&ddagger;2</sup>, 
[Song-Hai Zhang<sup>&ddagger;1</sup>](https://scholar.google.com/citations?user=AWtV-EQAAAAJ)**_

<sup>\*</sup>Co-first authors
<sup>&dagger;</sup>Project leaders
<sup>&ddagger;</sup>Corresponding authors

<sup>1</sup>Tsinghua University
<sup>2</sup>Alibaba
<sup>3</sup>UCST


![Version](https://img.shields.io/badge/version-1.0.0-blue) &nbsp;<a href='https://arxiv.org/pdf/2602.22960'><img src='https://img.shields.io/badge/arXiv-2504.01016-b31b1b.svg'></a> &nbsp; <a href='https://humanaigc.github.io/ucm-webpage/'><img src='https://img.shields.io/badge/Project-Page-Green'></a> &nbsp;
</div>

<div align="center">
    <img width="320" alt="pipeline" src="assets/teaser.gif">
    <img width="320" alt="pipeline" src="assets/teaser2.gif">
</div> -->

<div align="center">

<h2><i><b>UCM: Unified Modeling of Camera Control and Memory with Time-aware Positional Encoding Warping for World Models</b></i></h2>

[***Tian-Xing Xu***](https://scholar.google.com/citations?user=zHp0rMIAAAAJ)^∗¹^, [***Zi-Xuan Wang***](https://scholar.google.com/citations?user=4JKwZRsAAAAJ)^∗¹^, [***Guangyuan Wang***](https://scholar.google.com/citations?user=OYwm0GsAAAAJ&hl=zh-CN)^∗†²^, [***Li Hu***](https://scholar.google.com/citations?user=Arz3iGUAAAAJ)^†²^, [***Zhongyi Zhang***](https://scholar.google.com/citations?user=dOWmKOUAAAAJ)^³^, [***Peng Zhang***](https://scholar.google.com/citations?user=QTgxKmkAAAAJ)^²^, ***Bang Zhang***^‡²^, [***Song-Hai Zhang***](https://scholar.google.com/citations?user=AWtV-EQAAAAJ)^‡¹^

¹Tsinghua University &nbsp; ²Alibaba &nbsp; ³UCST

<sup>∗</sup>Co-first authors &nbsp; <sup>†</sup>Project leaders &nbsp; <sup>‡</sup>Corresponding authors

[![Version](https://img.shields.io/badge/version-1.0.0-blue)](https://humanaigc.github.io/ucm-webpage/) &nbsp; [![arXiv](https://img.shields.io/badge/arXiv-2504.01016-b31b1b.svg)](https://arxiv.org/pdf/2602.22960) &nbsp; [![Project Page](https://img.shields.io/badge/Project-Page-Green)](https://humanaigc.github.io/ucm-webpage/) &nbsp; [![License](https://img.shields.io/badge/license-Apache%202.0-blue)](LICENSE)

<br/>

<img width="320" alt="teaser1" src="assets/teaser.gif"/> <img width="320" alt="teaser2" src="assets/teaser2.gif"/>

</div>
<!-- ## 🔆 Notice

<!-- We recommend that everyone use English to communicate on issues, as this helps developers from around the world discuss, share experiences, and answer questions together. For further implementation details, please contact `xutx21@mails.tsinghua.edu.cn` or `wangzixu21@mails.tsinghua.edu.cn`. For business licensing and other related inquiries, don't hesitate to contact `yixuan.wgy@alibaba-inc.com`. -->

<!-- If you find UCM useful, **please help ⭐ this repo**, which is important to Open-Source projects. Thanks! -->

## 📝 Introduction

We present UCM, a novel framework to explore 4D world of a reference image following the user-specified camera trajectory, which unifies long-term memory and precise camera control via a time-aware positional encoding warping mechanism.

Release Notes:
- `[2026/08/07]` 🔥🔥🔥**UCM** is released now, have fun!

## 🚀 Quick Start

### Installation
1. Clone this repo:
```bash
git clone --recursive https://github.com/HumanAIGC/UCM.git
```
<!-- TODO -->
2. Install dependencies (please refer to [requirements.txt](requirements.txt)):
```bash
pip install -r requirements.txt
```
### Model Download

| Model | Download Links |
| :-----  | :----- |
| **UCM** | 🤗 [HuggingFace](https://huggingface.co/HumanAIGC-Team/UCM) 🤖 [ModelScope](https://modelscope.cn/models/DAMOXR/UCM) |

Download models using huggingface-cli:
```sh
pip install "huggingface_hub[cli]"
huggingface-cli download HumanAIGC-Team/UCM --local-dir ./workspace/pretrained/
```
Download models using modelscope-cli:
 ```sh
pip install modelscope
modelscope download --model DAMOXR/UCM --local_dir ./workspace/pretrained/
```

### Inference
Run inference code on our provided demo videos, which requires a GPU with ~42GB memory and ~5min to generate a 12s video (241 frames):

```bash
python main.py \
    --img_path examples/images/frame_0000.png \
    --traj_path examples/cameras/cameras_0000.json \
    --prompt "The video captures a serene and picturesque scene of a traditional Dutch village on a bright, sunny day. The sky is a vibrant blue with scattered white clouds, creating a perfect backdrop for the charming architecture and lush greenery. The camera pans slowly across the village, revealing a row of quaint houses with red-tiled roofs and brick facades, typical of Dutch design. Some houses have green-painted wooden shutters and doors, adding a touch of color to the scene.  A narrow cobblestone street runs through the village, lined with parked cars on both sides, indicating a peaceful residential area."
```

To obtain all demo videos, you can use the following instruction:

```bash
python main.py --metafile examples/examples.csv
```

**Parameters**
- `--img_path`: Path to your reference image.
- `--traj_path`: Path to your specific camera trajectory file (.json).
- `--prompt`: Text prompt.
- `--save_folder`: Path to your folder for saving generated videos.
- `--camera_scale_factor`: Scales the camera center within the trajectory to match the scale of the 3D scene representation. 
- `--num_denoising_steps`: The number of denoising iterations. `20` reaches a balance for the Gradio demo. `50` is used in our paper.
- `--guidance_scale`: Classifier-Free Guidance scale. The default value `5.0` is recommended.
- `--seed`: Seed for initializing the random number generator, controlling the randomness of Gaussian noise sampling.
- `--duration`: Only the first `Duration` frames of the camera trajectory will be processed. `-1` represents the whole trajectory.

## 🤖 Gradio Demo

<!-- - Online demo: [**Huggingface Demo**](https://huggingface.co/spaces/) TODO -->
- Local demo:
  ```bash
  gradio app.py
  ```

## 🤝 Contributing

- Welcome to open issues and pull requests.
- Welcome to optimize the inference speed and memory usage, e.g., through model quantization, distillation, or other acceleration techniques.

## ❤️ Acknowledgement

We have used codes from other great research work, including [STream3R](https://github.com/NIRVANALAN/STream3R) and [Wan2.1](https://github.com/Wan-Video/Wan2.1). We sincerely thank the authors for their awesome works!

## 📄 License

This project is licensed under the [Apache License 2.0](LICENSE).

## 📜 Citation

If you find this work helpful, please consider citing:

```BibTeX
@article{xu2026ucm,
  title={UCM: Unified Modeling of Camera Control and Memory with Time-aware Positional Encoding Warping for World Models},
  author={Xu, Tianxing and Wang, Zixuan and Wang, Guangyuan and Hu, Li and Zhang, Zhongyi and Zhang, Peng and Zhang, Bang and Zhang, Songhai},
  journal={arXiv preprint arXiv:2602.22960},
  year={2026}
}
```
<!-- TODO -->