---
title: windowseat-reflection-removal-v1-0
canonical_url: "https://www.modelscope.cn/models/huawei-bayerlab/windowseat-reflection-removal-v1-0"
md_url: "https://www.modelscope.cn/models/huawei-bayerlab/windowseat-reflection-removal-v1-0.md"
repository: huawei-bayerlab/windowseat-reflection-removal-v1-0
last_updated: 2026-09-14
license: apache-2.0
pipeline_tag: image-to-image
tasks:
  - image-to-image
base_model:
  - Qwen/Qwen-Image-Edit-2509
base_model_relation: adapter
parameters: 853.0M
tensor_type:
  - BF16
  - I64
library_name:
  - safetensors
  - pytorch
frameworks:
  - pytorch
language:
  - en
downloads: 18
stars: 1
tags:
  - "reflection removal"
  - "computational photography"
  - image-to-image
  - diffusion
  - zero-shot
---

# windowseat-reflection-removal-v1-0

> windowseat-reflection-removal-v1-0 - huawei-bayerlab 在 ModelScope 开源的模型。2026-09-13: Mirrored on ModelScope: 2026-06-04: Presented at the NTIRE workshop at CVPR 2026. 2025-12-05: Initial release: inference code, the released checkpoint, and the demo.

huawei-bayerlab/windowseat-reflection-removal-v1-0 是 ModelScope 魔搭社区上的 853.0M 参数image-to-image模型，采用 apache-2.0 许可，基于 Qwen/Qwen-Image-Edit-2509 构建。

- **Repository**: huawei-bayerlab/windowseat-reflection-removal-v1-0
- **License**: apache-2.0
- **Tasks**: image-to-image
- **Parameters**: 853.0M
- **Base model**: Qwen/Qwen-Image-Edit-2509
- **Tags**: reflection removal, computational photography, image-to-image, diffusion, zero-shot
- **Downloads**: 18
- **Stars**: 1
- **Last updated**: 2026-09-14

Source: https://www.modelscope.cn/models/huawei-bayerlab/windowseat-reflection-removal-v1-0

---

<p align="center">
  <img src="doc/images/windowseat_teaser.png" alt="Logo" style="max-width: 100%; height: auto;">
</p>

<p align="center">
  <a title="Website" href="https://hf.co/spaces/huawei-bayerlab/windowseat-reflection-removal-web" target="_blank" rel="noopener noreferrer" style="display: inline-block; margin: 2px;">
      <img src="assets/shields/project-website.svg" alt="Website" style="display: inline-block; vertical-align: middle; margin: 0;">
  </a>
  <a title="arXiv" href="https://arxiv.org/abs/2512.05000" target="_blank" rel="noopener noreferrer" style="display: inline-block; margin: 2px;">
      <img src="assets/shields/arxiv-pdf.svg" alt="arXiv" style="display: inline-block; vertical-align: middle; margin: 0;">
  </a>
  <a title="Github" href="https://github.com/huawei-bayerlab/windowseat-reflection-removal/" target="_blank" rel="noopener noreferrer" style="display: inline-block;">
    <img src="https://img.shields.io/github/stars/huawei-bayerlab/windowseat-reflection-removal?label=GitHub%20%E2%98%85&logo=github&color=C8C" alt="Github" style="display: inline-block; vertical-align: middle; margin: 0;">
  </a>
  <a title="License" href="https://huggingface.co/huawei-bayerlab/windowseat-reflection-removal-v1-0/resolve/main/LICENSE.txt" target="_blank" rel="noopener noreferrer" style="display: inline-block; margin: 2px;">
      <img src="assets/shields/license-apache-2.0.svg" alt="License" style="display: inline-block; vertical-align: middle; margin: 0;">
  </a>
</p>

## News

2026-09-13: Mirrored on ModelScope: <a href="https://www.modelscope.cn/studios/huawei-bayerlab/windowseat-reflection-removal-web" target="_blank" rel="noopener noreferrer" style="display: inline-block; margin: 2px;"><img src="assets/shields/modelscope-website.svg" style="display: inline-block; vertical-align: middle; margin: 0; height: 17px;" alt="ModelScope website"></a> <a href="https://www.modelscope.cn/models/huawei-bayerlab/windowseat-reflection-removal-v1-0" target="_blank" rel="noopener noreferrer" style="display: inline-block; margin: 2px;"><img src="assets/shields/modelscope-model.svg" style="display: inline-block; vertical-align: middle; margin: 0; height: 17px;" alt="ModelScope model"></a> <a href="https://www.modelscope.cn/studios/huawei-bayerlab/windowseat-reflection-removal" target="_blank" rel="noopener noreferrer" style="display: inline-block; margin: 2px;"><img src="assets/shields/modelscope-demo.svg" style="display: inline-block; vertical-align: middle; margin: 0; height: 17px;" alt="ModelScope demo"></a><br>
2026-06-04: Presented at the NTIRE workshop at CVPR 2026.<br>
2025-12-05: Initial release: inference code, the released checkpoint, and the demo.<br>

# WindowSeat Model Card

This is a model card for the `windowseat-reflection-removal-v1-0` model for single-image reflection removal. The model is derived from `Qwen/Qwen-Image-Edit-2509` using LoRA adaptation as described in our paper titled 
"Reflection Removal through Efficient Adaptation of Diffusion Transformers" by Daniyar Zakarin, Thiemo Wandel, Anton Obukhov, Dengxin Dai.

**See the [Quick Start section](https://github.com/huawei-bayerlab/windowseat-reflection-removal?tab=readme-ov-file#quick-start) of the paper's code repository for instructions on how to set up the environment and process photos with this model.**

- **Model Name:** `windowseat-reflection-removal-v1-0`
- **Task:** Single Image Reflection Removal
- **Base Model:** [`Qwen/Qwen-Image-Edit-2509`](https://huggingface.co/Qwen/Qwen-Image-Edit-2509)
- **Model Type:** End-to-end single-step latent diffusion reflection removal from a single image.
- **Resources for more information:** [Project Website](https://hf.co/spaces/huawei-bayerlab/windowseat-reflection-removal-web), [Paper](https://arxiv.org/abs/2512.05000), [Code](https://github.com/huawei-bayerlab/windowseat-reflection-removal).
- **Framework:** PyTorch, Transformers, PEFT
- **Language:** English.
- **License:** [Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0)
- **Developed by:** HUAWEI Bayer Lab
- **Cite as:**

```bibtex
@InProceedings{Zakarin_2026_CVPR,
  author    = {Zakarin, Daniyar and Wandel, Thiemo and Obukhov, Anton and Dai, Dengxin},
  title     = {Reflection Removal through Efficient Adaptation of Diffusion Transformers},
  booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},
  year      = {2026},
  pages     = {2776--2785}
}
```