Face Restoration with Generative Facial Prior

The original model was completed as part of the work by Xintao Wang, Yu Li, Honglun Zhang and Ying Shan. You can see their full repo at GFPGAN.This repo is my container/Docker friendly version which conforms to my docker-compose set up you'll see across all my github projects. The original work requires PyTorch so I've pulled an optimized base image from Nvidia's container repo. nvcr.io/nvidia/pytorch:21.11-py3

This build provides a containerized environment with all the requirements needed to run the GFPGAN model on NVIDIA GPUs.

Original Image	Restored Image

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}
}