diff --git a/README_CN.md b/README_CN.md index 880f2063..846ee64f 100644 --- a/README_CN.md +++ b/README_CN.md @@ -4,4 +4,200 @@ ##
English | 简体中文
-还未完工，欢迎贡献！ +
+ + +[![download](https://img.shields.io/github/downloads/TencentARC/GFPGAN/total.svg)](https://github.com/TencentARC/GFPGAN/releases) +[![PyPI](https://img.shields.io/pypi/v/gfpgan)](https://pypi.org/project/gfpgan/) +[![Open issue](https://img.shields.io/github/issues/TencentARC/GFPGAN)](https://github.com/TencentARC/GFPGAN/issues) +[![Closed issue](https://img.shields.io/github/issues-closed/TencentARC/GFPGAN)](https://github.com/TencentARC/GFPGAN/issues) +[![LICENSE](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](https://github.com/TencentARC/GFPGAN/blob/master/LICENSE) +[![python lint](https://github.com/TencentARC/GFPGAN/actions/workflows/pylint.yml/badge.svg)](https://github.com/TencentARC/GFPGAN/blob/master/.github/workflows/pylint.yml) +[![Publish-pip](https://github.com/TencentARC/GFPGAN/actions/workflows/publish-pip.yml/badge.svg)](https://github.com/TencentARC/GFPGAN/blob/master/.github/workflows/publish-pip.yml) + +
+ +1. :boom: **已更新** 在线演示: [![Replicate](https://img.shields.io/static/v1?label=Demo&message=Replicate&color=blue)](https://replicate.com/tencentarc/gfpgan). [备份链接](https://replicate.com/xinntao/gfpgan). +2. :boom: **已更新** 在线演示: [![Huggingface Gradio](https://img.shields.io/static/v1?label=Demo&message=Huggingface%20Gradio&color=orange)](https://huggingface.co/spaces/Xintao/GFPGAN) +3. [Colab演示](https://colab.research.google.com/drive/1sVsoBd9AjckIXThgtZhGrHRfFI6UUYOo)google colab logo; (另一个根据原版论文模型实现的[Colab演示](https://colab.research.google.com/drive/1Oa1WwKB4M4l1GmR7CtswDVgOCOeSLChA?usp=sharing) ) + + + +> :rocket: **感谢您对我们工作的兴趣. 也许您想了解一下我们最新的基于*小模型*的*动画图片与视频*[Real-ESRGAN](https://github.com/xinntao/Real-ESRGAN/blob/master/docs/anime_video_model.md)模型** :blush: + +GFPGAN旨在开发一个**针对现实世界中面部修复的实用算法**.
+本算法利用丰富多样的先验数据，封装在一个预训练的人脸GAN（*例如* StyleGAN2）网络中，进行盲面部修复. + +:question: 常见问题请参考[FAQ.md](FAQ.md). + +:triangular_flag_on_post: **更新日志** + +- :white_check_mark: 增加[RestoreFormer](https://github.com/wzhouxiff/RestoreFormer)推理代码. +- :white_check_mark: 增加[V1.4模型](https://github.com/TencentARC/GFPGAN/releases/download/v1.3.0/GFPGANv1.4.pth)，相比V1.3模型会生成更多细节并提高辨识度. +- :white_check_mark: 增加[V1.3模型](https://github.com/TencentARC/GFPGAN/releases/download/v1.3.0/GFPGANv1.3.pth), 此版本生成更**自然**的修复结果，并在*极低质量*/*高质量*的输入时生成更好的效果.详情移步至[模型库](#european_castle-model-zoo), [对比](Comparisons.md) +- :white_check_mark: 集成至[Huggingface Spaces](https://huggingface.co/spaces)和[Gradio](https://github.com/gradio-app/gradio). 查看[Gradio Web Demo](https://huggingface.co/spaces/akhaliq/GFPGAN). +- :white_check_mark: 支持增强非面部区域（背景）通过[Real-ESRGAN](https://github.com/xinntao/Real-ESRGAN). +- :white_check_mark: 我们提供一个*纯净*版GFPGAN，运行无需CUDA. +- :white_check_mark: 我们提供了一个更新的模型，不对人脸进行着色. + +--- +如果GFPGAN对您的照片/项目有帮助, 请帮忙:star:本项目或推荐给您的朋友们, 感谢:blush: +其他推荐项目:
+:arrow_forward: [Real-ESRGAN](https://github.com/xinntao/Real-ESRGAN): 一种实用的通用图像修复算法
+:arrow_forward: [BasicSR](https://github.com/xinntao/BasicSR): 一个开源的图像和视频修复工具箱
+:arrow_forward: [facexlib](https://github.com/xinntao/facexlib): 一个提供面部相关函数的集合库
+:arrow_forward: [HandyView](https://github.com/xinntao/HandyView): 一个基于PyQt5的图片查看器, 方便查看和比较
+ +--- + +### :book: GFP-GAN: Towards Real-World Blind Face Restoration with Generative Facial Prior + +> [[论文](https://arxiv.org/abs/2101.04061)]   [[项目主页](https://xinntao.github.io/projects/gfpgan)]   [演示]
+> [Xintao Wang](https://xinntao.github.io/), [Yu Li](https://yu-li.github.io/), [Honglun Zhang](https://scholar.google.com/citations?hl=en&user=KjQLROoAAAAJ), [Ying Shan](https://scholar.google.com/citations?user=4oXBp9UAAAAJ&hl=en)
+> Applied Research Center (ARC), Tencent PCG + +

+ +

+ +--- + +## :wrench: 依赖与安装 + +- Python >= 3.7 (推荐使用[Anaconda](https://www.anaconda.com/download/#linux)或[Miniconda](https://docs.conda.io/en/latest/miniconda.html)) +- [PyTorch >= 1.7](https://pytorch.org/) +- 可选: NVIDIA GPU + [CUDA](https://developer.nvidia.com/cuda-downloads) +- 可选: Linux + +### 安装 + +我们现在提供一个*纯净*版GFPGAN, 运行无需CUDA.
+如果您想使用我们论文的原版模型, 请参考[论文模型](PaperModel.md)安装. + +1. 克隆代码 + + ```bash + git clone https://github.com/TencentARC/GFPGAN.git + cd GFPGAN + ``` + +2. 安装依赖包 + + ```bash + # 安装 basicsr - https://github.com/xinntao/BasicSR + # 我们使用BaiscSR进行训练和推理 + pip install basicsr + + # 安装 facexlib - https://github.com/xinntao/facexlib + # 我们使用facexlib软件包中的面部检测和面部修复辅助工具 + pip install facexlib + + pip install -r requirements.txt + python setup.py develop + + # 如果您想使用Real-ESRGAN增强背景（非面部）区域 + # 您需要安装 realesrgan 包 + pip install realesrgan + ``` + +## :zap: 快速推理 + +我们以V1.3模型为例. 更多模型可以查看[此处](#european_castle-model-zoo). + +下载预训练模型: [GFPGANv1.3.pth](https://github.com/TencentARC/GFPGAN/releases/download/v1.3.0/GFPGANv1.3.pth) + +```bash +wget https://github.com/TencentARC/GFPGAN/releases/download/v1.3.0/GFPGANv1.3.pth -P experiments/pretrained_models +``` + +**推理!** + +```bash +python inference_gfpgan.py -i inputs/whole_imgs -o results -v 1.3 -s 2 +``` + +```console +Usage: python inference_gfpgan.py -i inputs/whole_imgs -o results -v 1.3 -s 2 [options]... + + -h 显示此帮助 + -i input 输入图片或目录. 默认: inputs/whole_imgs + -o output 输出目录. 默认: results + -v version GFPGAN模型版本. 可选: 1 | 1.2 | 1.3. 默认: 1.3 + -s upscale 图像的最终超采样比例. 默认: 2 + -bg_upsampler 背景超采样方式. 默认: realesrgan + -bg_tile 背景采样器的片尺寸, 0表示无. 默认: 400 + -suffix 输出文件的后缀 + -only_center_face 仅修复中心面部 + -aligned 输入是对齐的面孔 + -ext 输入图像类型. 可选: auto | jpg | png, auto 意味着使用相同类型作为输入, 默认: auto +``` + +如果您想使用我们论文中的原版模型, 请参考[论文模型](PaperModel.md)进行安装和推理. + +## :european_castle: 模型库 + +| 版本 | 模型名称 | 描述 | +|:-------:|:--------------------------------------------------------------------------------------------------------------------:|:--------------------------------------------------------------------------------------------------------------:| +| V1.3 | [GFPGANv1.3.pth](https://github.com/TencentARC/GFPGAN/releases/download/v1.3.0/GFPGANv1.3.pth) | 基于V1.2; **更自然**的修复结果; 在*极低质量*/*高质量*的输入时生成更好的效果. | +| V1.2 | [GFPGANCleanv1-NoCE-C2.pth](https://github.com/TencentARC/GFPGAN/releases/download/v0.2.0/GFPGANCleanv1-NoCE-C2.pth) | 无上色;无必须CUDA支持. 通过更多预处理后的数据进行训练. | +| V1 | [GFPGANv1.pth](https://github.com/TencentARC/GFPGAN/releases/download/v0.1.0/GFPGANv1.pth) | 论文原版模型, 有上色. | + +对比请参考[此处](Comparisons.md). + +请注意V1.3并不是总比V1.2效果好. 您也许需要根据您的需求和输入而选择不同的模型. + +| 版本 | 优势 | 劣势 | +|:-------:|:-----------------------------------------------------------------------------------------------------------------------------------------------------------:|:--------------------------------------------------------:| +| V1.3 | ✓ 自然的输出
✓在低质量输入时有更好的效果
✓ 可在相对高质量输入时工作
✓ 可以重复（两次）进行修复 | ✗ 不是非常清晰
✗ 对面容有明显改变 | +| V1.2 | ✓ 更清晰的输出
✓ 自带美颜效果 | ✗ 一些输出非常不自然 | + +您可以在下面链接找到**更多模型(例如discriminators)**: [[Google Drive](https://drive.google.com/drive/folders/17rLiFzcUMoQuhLnptDsKolegHWwJOnHu?usp=sharing)], 或[[腾讯微云](https://share.weiyun.com/ShYoCCoc)] + +## :computer: 训练 + +我们提供GFPGAN的训练代码(在论文中所使用的).
+您可以根据您的需求自行调节. + +**提示** + +1. 高质量的面部输入可以提高修复质量. +2. 您也许需要一些预处理, 例如美颜. + +**流程** + +(您可以尝试一个简单的版本( `options/train_gfpgan_v1_simple.yml`)并不需要面部标记.) + +1. 准备数据集: [FFHQ](https://github.com/NVlabs/ffhq-dataset) + +2. 下载预训练的模型和其他数据. 把它们放在`experiments/pretrained_models`目录. + + 1. [Pre-trained StyleGAN2 model: StyleGAN2_512_Cmul1_FFHQ_B12G4_scratch_800k.pth](https://github.com/TencentARC/GFPGAN/releases/download/v0.1.0/StyleGAN2_512_Cmul1_FFHQ_B12G4_scratch_800k.pth) + 2. [Component locations of FFHQ: FFHQ_eye_mouth_landmarks_512.pth](https://github.com/TencentARC/GFPGAN/releases/download/v0.1.0/FFHQ_eye_mouth_landmarks_512.pth) + 3. [A simple ArcFace model: arcface_resnet18.pth](https://github.com/TencentARC/GFPGAN/releases/download/v0.1.0/arcface_resnet18.pth) + +3. 对配置文件`options/train_gfpgan_v1.yml`进行相应修改. + +4. 训练 + +> python -m torch.distributed.launch --nproc_per_node=4 --master_port=22021 gfpgan/train.py -opt options/train_gfpgan_v1.yml --launcher pytorch + +## :scroll: 版权与致谢 + +GFPGAN以Apache License Version 2.0许可证发布. + +## 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: 联系方式 + +如果您有任何疑问, 请邮件联系 `xintao.wang@outlook.com` 或 `xintaowang@tencent.com`.