---
license: apache-2.0
---
# FaceScore

<p align="center">
   📃 <a href="https://arxiv.org/abs/2406.17100" target="_blank">Paper</a> • 🌐 <a href="https://github.com/OPPO-Mente-Lab/FaceScore" target="_blank">Repo</a> 
</p>

**FaceScore: Benchmarking and Enhancing Face Quality in Human Generation**

Traditional facial quality assessment focuses on whether a face is suitable for recognition, while image aesthetic scorers emphasize overall aesthetics rather than details. FaceScore is the first reward model that focuses on faces in text-to-image models, designed to score the faces generated in images. It is fine-tuned on positive and negative sample pairs generated using an inpainting pipeline based on real face images and surpasses previous models in predicting human preferences for generated faces.

- [LoRA base on SDXL](#lora-based-on-sdxl)
- [Citation](#citation)


## LoRA based on SDXL
We leverage FaceScore to filter data and perform direct preference optimization on SDXL.
The LoRA weight is [here](https://huggingface.co/AIGCer-OPPO/FaceScore-dpo-SDXL-LoRA/tree/main).
Here we provide a quick example:
```
from diffusers import StableDiffusionXLPipeline, UNet2DConditionModel
import torch

# load pipeline
inference_dtype = torch.float16
pipe = StableDiffusionXLPipeline.from_pretrained(
    "stabilityai/stable-diffusion-xl-base-1.0",
    torch_dtype=inference_dtype,
)
vae = AutoencoderKL.from_pretrained(
    'madebyollin/sdxl-vae-fp16-fix',
    torch_dtype=inference_dtype,
)
pipe.vae = vae
# You can load it locally
pipe.load_lora_weights("AIGCer-OPPO/FaceScore-dpo-SDXL-LoRA")
pipe.to('cuda')

generator=torch.Generator(device='cuda').manual_seed(42)
image = pipe(
    prompt='A woman in a costume standing in the desert',
    guidance_scale=5.0,
    generator=generator,
    output_type='pil',
).images[0]
image.save('A woman in a costume standing in the desert.png')
```
We provide some examples generated by ours (right) and compare with the original SDXL (left) below.
<div style="display: flex; justify-content: space-around;">
    <div style="text-align: center;">
        <img src="https://huggingface.co/AIGCer-OPPO/FaceScore-dpo-SDXL-LoRA/resolve/main/assets/desert.jpg" alt="图片1" style="width: 600px;" />
        <p>A woman in a costume standing in the desert. </p>
    </div>
    <div style="text-align: center;">
        <img src="https://huggingface.co/AIGCer-OPPO/FaceScore-dpo-SDXL-LoRA/resolve/main/assets/scarf.jpg" alt="图片2" style="width: 600px;" />
        <p>A woman wearing a blue jacket and scarf.</p>
    </div>
</div>
<div style="display: flex; justify-content: space-around;">
    <div style="text-align: center;">
        <img src="https://huggingface.co/AIGCer-OPPO/FaceScore-dpo-SDXL-LoRA/resolve/main/assets/stage.jpg" alt="图片1" style="width: 600px;" />
        <p>A young woman in a blue dress performing on stage. </p>
    </div>
    <div style="text-align: center;">
        <img src="https://huggingface.co/AIGCer-OPPO/FaceScore-dpo-SDXL-LoRA/resolve/main/assets/striped.jpg" alt="图片2" style="width: 600px;" />
        <p>A woman with black hair and a striped shirt.</p>
    </div>
</div>
<div style="display: flex; justify-content: space-around;">
    <div style="text-align: center;">
        <img src="https://huggingface.co/AIGCer-OPPO/FaceScore-dpo-SDXL-LoRA/resolve/main/assets/sword.jpg" alt="图片1" style="width: 600px;" />
        <p>A woman with white hair and white armor is holding a sword. </p>
    </div>
    <div style="text-align: center;">
        <img src="https://huggingface.co/AIGCer-OPPO/FaceScore-dpo-SDXL-LoRA/resolve/main/assets/white.jpg" alt="图片2" style="width: 600px;" />
        <p>A woman with long black hair and a white shirt.</p>
    </div>
</div>

## Citation

```
@misc{liao2024facescorebenchmarkingenhancingface,
      title={FaceScore: Benchmarking and Enhancing Face Quality in Human Generation}, 
      author={Zhenyi Liao and Qingsong Xie and Chen Chen and Hannan Lu and Zhijie Deng},
      year={2024},
      eprint={2406.17100},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2406.17100}, 
}
```