boschresearch/unetgan
A PyTorch implementation of a U-Net-based discriminator architecture for GANs in image generation, from CVPR 2020.
Not currently ranked — collecting fresh signals.
star history
This repository provides the official implementation of a CVPR 2020 research paper introducing a U-Net based discriminator for Generative Adversarial Networks. The discriminator uses skip connections to preserve spatial information, improving gradient flow during training. The project includes training scripts with options for CutMix augmentation, warmup scheduling, and supports both conditional and unconditional image generation on datasets like CelebA and FFHQ.
Frequently asked
- What is boschresearch/unetgan?
- A PyTorch implementation of a U-Net-based discriminator architecture for GANs in image generation, from CVPR 2020.
- Is unetgan open source?
- Yes — boschresearch/unetgan is open source, released under the AGPL-3.0 license.
- What language is unetgan written in?
- boschresearch/unetgan is primarily written in Python.
- How popular is unetgan?
- boschresearch/unetgan has 403 stars on GitHub.
- Where can I find unetgan?
- boschresearch/unetgan is on GitHub at https://github.com/boschresearch/unetgan.