GFPGAN README

The project's README from github.com/TencentARC/GFPGAN. As one markdown file: README.md.

English | 简体中文

download PyPI Open issue Closed issue LICENSE python lint Publish-pip

  1. :boom: Updated online demo: Replicate. Here is the backup.
  2. :boom: Updated online demo: Huggingface Gradio
  3. Colab Demo for GFPGAN google colab logo; (Another Colab Demo for the original paper model)

:rocket: Thanks for your interest in our work. You may also want to check our new updates on the tiny models for anime images and videos in Real-ESRGAN :blush:

GFPGAN aims at developing a Practical Algorithm for Real-world Face Restoration.
It leverages rich and diverse priors encapsulated in a pretrained face GAN (e.g., StyleGAN2) for blind face restoration.

:question: Frequently Asked Questions can be found in FAQ.md.

:triangular_flag_on_post: Updates

If GFPGAN is helpful in your photos/projects, please help to :star: this repo or recommend it to your friends. Thanks:blush: Other recommended projects:
:arrow_forward: Real-ESRGAN: A practical algorithm for general image restoration
:arrow_forward: BasicSR: An open-source image and video restoration toolbox
:arrow_forward: facexlib: A collection that provides useful face-relation functions
:arrow_forward: HandyView: A PyQt5-based image viewer that is handy for view and comparison

:book: GFP-GAN: Towards Real-World Blind Face Restoration with Generative Facial Prior

[Paper]   [Project Page]   [Demo]
Xintao Wang, Yu Li, Honglun Zhang, Ying Shan
Applied Research Center (ARC), Tencent PCG

:wrench: Dependencies and Installation

Installation

We now provide a clean version of GFPGAN, which does not require customized CUDA extensions.
If you want to use the original model in our paper, please see PaperModel.md for installation.

  1. Clone repo

    git clone https://github.com/TencentARC/GFPGAN.git
cd GFPGAN

  2. Install dependent packages

    # Install basicsr - https://github.com/xinntao/BasicSR
# We use BasicSR for both training and inference
pip install basicsr

# Install facexlib - https://github.com/xinntao/facexlib
# We use face detection and face restoration helper in the facexlib package
pip install facexlib

pip install -r requirements.txt
python setup.py develop

# If you want to enhance the background (non-face) regions with Real-ESRGAN,
# you also need to install the realesrgan package
pip install realesrgan

:zap: Quick Inference

We take the v1.3 version for an example. More models can be found here.

Download pre-trained models: GFPGANv1.3.pth

wget https://github.com/TencentARC/GFPGAN/releases/download/v1.3.0/GFPGANv1.3.pth -P experiments/pretrained_models

Inference!

python inference_gfpgan.py -i inputs/whole_imgs -o results -v 1.3 -s 2

Usage: python inference_gfpgan.py -i inputs/whole_imgs -o results -v 1.3 -s 2 [options]...

  -h                   show this help
  -i input             Input image or folder. Default: inputs/whole_imgs
  -o output            Output folder. Default: results
  -v version           GFPGAN model version. Option: 1 | 1.2 | 1.3. Default: 1.3
  -s upscale           The final upsampling scale of the image. Default: 2
  -bg_upsampler        background upsampler. Default: realesrgan
  -bg_tile             Tile size for background sampler, 0 for no tile during testing. Default: 400
  -suffix              Suffix of the restored faces
  -only_center_face    Only restore the center face
  -aligned             Input are aligned faces
  -ext                 Image extension. Options: auto | jpg | png, auto means using the same extension as inputs. Default: auto

If you want to use the original model in our paper, please see PaperModel.md for installation and inference.

:european_castle: Model Zoo

Version Model Name Description
V1.3 GFPGANv1.3.pth Based on V1.2; more natural restoration results; better results on very low-quality / high-quality inputs.
V1.2 GFPGANCleanv1-NoCE-C2.pth No colorization; no CUDA extensions are required. Trained with more data with pre-processing.
V1 GFPGANv1.pth The paper model, with colorization.

The comparisons are in Comparisons.md.

Note that V1.3 is not always better than V1.2. You may need to select different models based on your purpose and inputs.

Version Strengths Weaknesses
V1.3 ✓ natural outputs
✓better results on very low-quality inputs
✓ work on relatively high-quality inputs
✓ can have repeated (twice) restorations		 ✗ not very sharp
✗ have a slight change on identity
V1.2 ✓ sharper output
✓ with beauty makeup		 ✗ some outputs are unnatural

You can find more models (such as the discriminators) here: [Google Drive], OR [Tencent Cloud 腾讯微云]

:computer: Training

We provide the training codes for GFPGAN (used in our paper).
You could improve it according to your own needs.

Tips

  1. More high quality faces can improve the restoration quality.
  2. You may need to perform some pre-processing, such as beauty makeup.

Procedures

(You can try a simple version ( options/train_gfpgan_v1_simple.yml) that does not require face component landmarks.)

  1. Dataset preparation: FFHQ

  2. Download pre-trained models and other data. Put them in the experiments/pretrained_models folder.

    1. Pre-trained StyleGAN2 model: StyleGAN2_512_Cmul1_FFHQ_B12G4_scratch_800k.pth
    2. Component locations of FFHQ: FFHQ_eye_mouth_landmarks_512.pth
    3. A simple ArcFace model: arcface_resnet18.pth

  3. Modify the configuration file options/train_gfpgan_v1.yml accordingly.

  4. Training

python -m torch.distributed.launch --nproc_per_node=4 --master_port=22021 gfpgan/train.py -opt options/train_gfpgan_v1.yml --launcher pytorch

:scroll: License and Acknowledgement

GFPGAN is released under Apache License Version 2.0.

BibTeX

@InProceedings{wang2021gfpgan,
    author = {Xintao Wang and Yu Li and Honglun Zhang and Ying Shan},
    title = {Towards Real-World Blind Face Restoration with Generative Facial Prior},
    booktitle={The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
    year = {2021}
}

:e-mail: Contact

If you have any question, please email [email protected] or [email protected].

GFPGAN (TencentARC/GFPGAN)

face-restoration
Searching Database…

37.7K ·GitHub
GFPGAN
What is this?GFPGAN is a practical algorithm for real-world blind face restoration: it rebuilds degraded or low-quality faces in photos using rich priors from a pretrained face GAN like StyleGAN2.
Is GFPGAN an app I open, or a Python tool I run myself?It is a Python package you run yourself — install it with pip and call it from a script or the command line. If you only want to try it, there are hosted Colab and Gradio web demos.
How do I restore my first photo with it?Install Python >= 3.7 with PyTorch >= 1.7, clone the repo and install the dependent packages, then download a pretrained model such as GFPGANv1.3.pth and run the inference script.
Which model version should I pick?Start with the newest V1.4 or the popular V1.3: results are natural and handle very low-quality inputs, but are not very sharp and slightly shift identity. V1.2 is sharper yet some outputs look unnatural.
Is this for developers, or can anyone fix their old photos?Mainly developers and researchers comfortable with Python and PyTorch, but non-coders can restore photos too through the Colab and Hugging Face Gradio demos right in a browser.
Do I need an NVIDIA GPU to run it?An NVIDIA GPU with CUDA is optional. The clean version of GFPGAN needs no customized CUDA extensions, so it runs without a GPU; only the original paper model (V1) requires them — see PaperModel.md.
Does GFPGAN restore only faces or the whole photo?Faces are its core strength; background and other non-face regions can be enhanced too, using Real-ESRGAN, the same authors' general image restoration model.
Can I train GFPGAN on my own data?Yes, the training codes used in the paper are included. The README advises that more high-quality faces improve quality, and offers a simple config that needs no face component landmarks.
Does it work on anime images or videos?Not directly — GFPGAN targets real-world photos. For anime images and videos, the authors point to the tiny models added to their Real-ESRGAN project.
What are similar projects?CodeFormer, lama, latent-diffusion, stylegan, neural-enhance, Real-ESRGAN, InstantID, ControlNet
Which topics is it in?esrgan, face-restoration, face-swap, image-restoration, colab, pip, apache, gan, super-resolution, gradio, pretrained-models, training
Creator
TencentARC
Size
5.3 MB
Stars
37.7K
Updated
CommitApr 02Releasev1.3.8Sep 16
Age
5y
Commits · 90 Days