[CVPR2020]对抗性潜在自编码器

ffhq 开源项目推荐(17 个)
ffhq 话题在 GitHub 上没有填写简介,这里按 Star 数汇总最相关的热门仓库。
热门项目 全部 17
GitHub 共 17 个仓库,已全部展示Official implementation of Diffusion Autoencoders
Photometric optimization code for creating the FLAME texture space and other applications
A comprehensive list of recources (papers, repositories etc.) about face restoration methods.
Official Implementation of the paper "A U-Net Based Discriminator for Generative Adversarial Networks" (CVPR 2020)
Image Fine-grained Inpainting (Winner Award of ECCVW AIM 2020 Extreme Inpainting Track1&Track2)
[CVPR-2023] Re-thinking Model Inversion Attacks Against Deep Neural Networks
"Analyzing and Improving the Image Quality of StyleGAN" in TensorFlow 2
An unofficial PyTorch implementation of VQGAN
Wasserstein GAN with gradient penalty tutoria, PyTorch ver.
FFHQ but in jpg
crop/align faces to FFHQ-style using dlib
Lightweight MSRResNet-based blind face super-resolution model trained on FFHQ using the BasicSR framework.
🌙 A modern GAN you can train overnight on one consumer GPU — latent transformer G/D (GAT + CAT), R3GAN loss without double backprop, REPA with DINOv2. Pure PyTorch, no custom CUDA kernels.
얼굴 정렬·파싱 후 주름/모공/홍조 3채널 조건지도(heatmaps + .npy)를 만들어 cGAN 학습에 쓰는 파이프라인. (Face parsing + skin-condition maps (redness/wrinkle/pore) pipeline for conditional GAN training.)
FFHQ-2048: 1,000 sharp 2048x2048 portraits, the first new high-quality public face dataset since FFHQ (2019). Re-mastered by NanoPocket Face Enhance. Full 70k set on request.
Flickr-Portraits-XL (FPXL): 6,834 aligned face portraits cropped from FFHQ at SDXL resolutions (~1 megapixel).
关于这个话题
- 清单取自 GitHub 话题页并按 Star 数排序,每次展示前 17 个项目;原始页面见 https://github.com/topics/ffhq。
- 该话题在 GitHub 上共有 17 个仓库, 本站收录的是其中最热门的 17 个,不是全量列表。
- Star / Fork 每日自动刷新一次,实时数据以 GitHub 为准。
- 当前 17 个项目累计 6306 Star,覆盖 3 种主要语言,数据更新于 2026-11-08。
- 中文介绍来自 GitHub 官方话题说明(github/explore)的机器翻译,可点「查看原文」对照;译文缺失时会先展示原文。