README.md

---
license: mit
pipeline_tag: image-to-3d
---

<p align="center">
  <img src="https://lhmd.top/zpressor/assets/favicon.svg" alt="ZPressor Logo" style="width: 50px; height: 50px; margin-right: 20px;">
  <h1 align="center">ZPressor: Bottleneck-Aware Compression for Scalable Feed-Forward 3DGS</h1>
  <p align="center">
    <a href="https://lhmd.top">Weijie Wang</a>
    ·
    <a href="https://donydchen.github.io">Donny Y. Chen</a>
    ·
    <a href="https://steve-zeyu-zhang.github.io">Zeyu Zhang</a>
    ·
    <a href="https://openreview.net/profile?id=~Duochao_Shi1">Duochao Shi</a>
    ·
    <a href="https://github.com/AkideLiu">Akide Liu</a>
    ·
    <a href="https://bohanzhuang.github.io">Bohan Zhuang</a>
  </p>
  <h3 align="center">NeurIPS 2025</h3>
  <h3 align="center"><a href="https://arxiv.org/abs/2505.23734">Paper</a> | <a href="https://lhmd.top/zpressor">Project Page</a> | <a href="https://github.com/ziplab/ZPressor">Code</a> | <a href="https://huggingface.co/lhmd/ZPressor">Models</a> </h3>
  <div align="center"></div>
</p>



<p align="center">
  <a href="">
    <img src="https://lhmd.top/zpressor/assets/teaser.jpg" alt="Logo" width="100%">
  </a>
</p>

<p align="center">
<strong>ZPressor is an architecture-agnostic module that compresses multi-view inputs for scalable feed-forward 3DGS.</strong>
</p>

## Citation
If you find our work useful for your research, please consider citing us:

```bibtex
@article{wang2025zpressor,
  title={ZPressor: Bottleneck-Aware Compression for Scalable Feed-Forward 3DGS},
  author={Wang, Weijie and Chen, Donny Y. and Zhang, Zeyu and Shi, Duochao and Liu, Akide and Zhuang, Bohan},
  journal={arXiv preprint arXiv:2505.23734},
  year={2025}
}
```
## Contact
If you have any questions, please contact at wangweijie@zju.edu.cn.

## Acknowledgements

This project is developed with several fantastic repos: [pixelSplat](https://github.com/dcharatan/pixelsplat), [MVSplat](https://github.com/donydchen/mvsplat) and [DepthSplat](https://github.com/cvg/depthsplat). We thank the original authors for their excellent work.