---
pipeline_tag: image-to-image
base_model: OpenMOSS-Team/MOSS-Image-SFT
language:
  - en
  - zh
tags:
  - lora
  - image-editing
  - moss-image
  - comfyui
---

<p align="center">
  <img src="assets/moss-image-edit-banner-cases.png" alt="MOSS Image Edit" width="1280">
</p>

<p align="center">
  <strong>Give your images a new expression.</strong><br>
  MOSS Image Edit is an <strong>image-editing LoRA for MOSS-Image</strong>. Edit colors, outfits, lighting, seasons, and visual style with text instructions.<br>
  This repository provides LoRA weights and configuration. Use them with the MOSS-Image Base or SFT model.
</p>

<p align="center">
  <a href="https://github.com/Chenkin-x/Moss-Image-Edit">GitHub · 中文介绍</a>
  &nbsp; · &nbsp;
  <a href="https://github.com/Chenkin-x/Moss-Image-Edit/blob/main/README.en.md">English</a>
  &nbsp; · &nbsp;
  <a href="https://github.com/Chenkin-x/ComfyUI-Moss-Image-Edit/blob/main/README.en.md">ComfyUI guide</a>
</p>

## Editing showcase

Explore changes in color, outfits, lighting, seasons, and visual style. Each group starts with the original image, followed by its edits.

![MOSS Image Edit showcase](assets/moss-image-edit-showcase.jpg)

## Get started

Download this editing LoRA together with either [MOSS-Image-Base](https://huggingface.co/OpenMOSS-Team/MOSS-Image-Base) or [MOSS-Image-SFT](https://huggingface.co/OpenMOSS-Team/MOSS-Image-SFT), then open a ComfyUI workflow. **Both base models are supported; choose either one.**

| Resource | Link |
|:---|:---|
| Installation and model placement | [English guide](https://github.com/Chenkin-x/ComfyUI-Moss-Image-Edit/blob/main/README.en.md) · [中文教程](https://github.com/Chenkin-x/ComfyUI-Moss-Image-Edit) |
| Edit with one reference | [Single-reference workflow](https://github.com/Chenkin-x/ComfyUI-Moss-Image-Edit/blob/main/workflows/moss_edit_single_reference.json) |
| Edit with multiple references | [Multi-reference workflow](https://github.com/Chenkin-x/ComfyUI-Moss-Image-Edit/blob/main/workflows/moss_edit_multi_reference.json) |

Upload your reference image, describe the change, and generate the result. With multiple references, explain the role of each image in your instruction; up to four images can be connected.

To download the LoRA and base-model files from a terminal, using SFT as the example:

```bash
hf download chinoll/Moss-Image-Edit adapter_config.json adapter_model.safetensors --local-dir models/Moss-Image-Edit
hf download OpenMOSS-Team/MOSS-Image-SFT --local-dir models/MOSS-Image-SFT
```

For Base, use this base-model download command instead:

```bash
hf download OpenMOSS-Team/MOSS-Image-Base --local-dir models/MOSS-Image-Base
```

Set the ComfyUI loader's `snapshot_dir` to the directory containing your chosen base model.

## Credits

Thanks to [@onglon114514](https://github.com/onglon114514) for generating the showcase images.

Built on [MOSS-Image](https://huggingface.co/OpenMOSS-Team/MOSS-Image-SFT). Project code is available under [GPLv3](https://github.com/Chenkin-x/Moss-Image-Edit/blob/main/LICENSE).