https://github.com/lucidrains/unet-stylegan2

A Pytorch implementation of Stylegan2 with UNet Discriminator
https://github.com/lucidrains/unet-stylegan2

artificial-intelligence deep-learning generative-adversarial-network u-net

Last synced: 6 days ago
JSON representation

A Pytorch implementation of Stylegan2 with UNet Discriminator

• Host: GitHub
• URL: https://github.com/lucidrains/unet-stylegan2
• Owner: lucidrains
• License: mit
• Created: 2020-07-25T01:14:35.000Z (over 4 years ago)
• Default Branch: master
• Last Pushed: 2020-11-10T05:39:21.000Z (about 4 years ago)
• Last Synced: 2025-01-21T18:09:29.918Z (14 days ago)
• Topics: artificial-intelligence, deep-learning, generative-adversarial-network, u-net
• Language: Python
• Homepage:
• Size: 595 KB
• Stars: 206
• Watchers: 10
• Forks: 24
• Open Issues: 7
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        


## UNet Stylegan2

An implementation of Stylegan2 with UNet Discriminator. This repository works largely the same way as Stylegan2 Pytorch. Simply replace all the `stylegan2_pytorch` command with `unet_stylegan2` instead.





Update: Results have been very good. Will need to investigate combining this with a few other techniques, and then I will write up full instructions for use.

## Install

```bash

$ pip install unet-stylegan2

```

## Usage

```bash

$ unet_stylegan2 --data ./path/to/data

```

## Citations

```bibtex

@misc{karras2019analyzing,

    title={Analyzing and Improving the Image Quality of StyleGAN},

    author={Tero Karras and Samuli Laine and Miika Aittala and Janne Hellsten and Jaakko Lehtinen and Timo Aila},

    year={2019},

    eprint={1912.04958},

    archivePrefix={arXiv},

    primaryClass={cs.CV}

}

```

```bibtex

@misc{schnfeld2020unet,

    title={A U-Net Based Discriminator for Generative Adversarial Networks},

    author={Edgar Schönfeld and Bernt Schiele and Anna Khoreva},

    year={2020},

    eprint={2002.12655},

    archivePrefix={arXiv},

    primaryClass={cs.CV}

}

```