HandFixer/README.MD

# HandFixer一键手部修复
## Introduction
This project addresses the issue of numerous hand defects in AI-generated portrait images by using a workflow based on FLUX/FLUX-Fill to correct hand deformities. It is effective for both real-life and anime images, including those generated by DiT/SDXL/Midjourney. It supports diffusers, gradio, ComfyUI, and more

本项目针对AI生成人像图片存在大量手部缺陷的问题，使用了一种基于FLUX/FLUX-Fill修复手部畸形的工作流，对真人图像和动漫图像均有效，包括由DiT/SDXL/Midjourney生成的图片；支持diffusers、gradio、ComfyUI等

## Results
- realistic images 真实感图片

|source| ORI | FIX | ORI | FIX |ORI | FIX  |
|------|-------|-------|-------|-------|-------|-------|
|sdxl|![](assets/readme/sdxl/1.jpg)|![](assets/readme/sdxl/1_fix.jpg) |![](assets/readme/sdxl/2.jpg)|![](assets/readme/sdxl/2_fix.jpg)|![](assets/readme/sdxl/3.jpg)|![](assets/readme/sdxl/3_fix.jpg)|
|DiT|![](assets/readme/sdxl/4.jpg)|![](assets/readme/sdxl/4_fix.jpg) |![](assets/readme/sdxl/5.jpg)|![](assets/readme/sdxl/5_fix.jpg)|![](assets/readme/sdxl/6.jpg)|![](assets/readme/sdxl/6_fix.jpg)|
- anime images 动漫图片

|source| ORI | FIX  | ORI | FIX  |ORI | FIX  |
|----|-------|-------|-------|-------|-------|-------|
|midjourney-v6.1|![mj1](assets/readme/midjourney_anime/mj1.jpg)|![mj1_fix](assets/readme/midjourney_anime/mj1_fix.jpg) |![mj2](assets/readme/midjourney_anime/mj2.jpg)|![mj2_fix](assets/readme/midjourney_anime/mj2_fix.jpg)|![](assets/readme/midjourney_anime/6.jpg)|![](assets/readme/midjourney_anime/6_fix.jpg)
|SFT-SDXL|![](assets/readme/midjourney_anime/3.jpg)|![](assets/readme/midjourney_anime/3_fix.jpg)|![](assets/readme/midjourney_anime/4.jpg)|![](assets/readme/midjourney_anime/4_fix.jpg)|![](assets/readme/midjourney_anime/5.jpg)|![](assets/readme/midjourney_anime/5_fix.jpg)|


## TODO
- [ ] Hand Pose/Depth Controlnet: use structure/anatomy conditions to reinforce genreation
- [x] ComfyUI workflow

## Quick Start 
```

pip3 install requirements.txt



python3 diffusers_pipeline.py \

    --pretrained_model_name_or_path='black-forest-labs/FLUX.1-dev' \

    --input_dir 'assets/examples' \

    --output_dir 'outputs'

```

## ComfyUI Workflow
1. "Download HandFixer to `ComfyUI/custom_nodes/` and install the requirements." 
下载HandFixer到`ComfyUI/custom_nodes/`并安装requirements
```

put HandFixer to ComfyUI/custom_nodes/

pip install -r requirements.txt

```
2. Ensure that the model components have been downloaded.
 确保模型组件已经下载

| Model Name  | File Name              | Installation Location            | Download Link      |
|-------------|------------------------|----------------------------------|--------------------|
| VAE | ae.safetensors	| ComfyUI/models/vae| [Download](https://huggingface.co/black-forest-labs/FLUX.1-schnell/blob/main/ae.safetensors)|
| CLIP Model  | clip_l.safetensors     | ComfyUI/models/clip              | [Download](https://huggingface.co/comfyanonymous/flux_text_encoders/blob/main/clip_l.safetensors)      |

|             | t5xxl_fp16.safetensors | ComfyUI/models/clip              | [Download](https://huggingface.co/comfyanonymous/flux_text_encoders/blob/main/t5xxl_fp16.safetensors)      |
| Flux Fill Model | fluxl-fill-dev.safetensors | ComfyUI/models/unet | [Flux Fill Release](https://huggingface.co/black-forest-labs/FLUX.1-Fill-dev) |

3. drag [workflow](workflow/HandFixer-fill.json) to ComfyUI. 将[workflow](workflow/HandFixer-fill.json)拖动到comfyUI


## Streamlit UI
models will be downloaded from huggingface automatically
```

streamlit run app_streamlit.py

```
![Streamlit](assets/readme/streamlit_UI.jpg)

## Gradio
```

python3 app_gradio.py

```

## Limitations
- The FLUX-dev model is not a custom inpainting model and may occasionally have boundary artifacts; FLUX-Fill is a dedicated inpainting model with better boundary handling, but the performance on generation decreases