whsprite / AdvGAN_WGAN-GP Public

Notifications You must be signed in to change notification settings
Fork 0
Star 1

Unoffical implmentation of Using Wasserstein GAN to Generate Adversarial Examples. Modified from https://github.com/mathcbc/advGAN_pytorch.

1 star 0 forks Branches Tags Activity

Notifications

Name		Name	Last commit message	Last commit date
Latest commit History 14 Commits
README.md		README.md
WGAN.py		WGAN.py
main.py		main.py
models.py		models.py
test_WGAN_examples.py		test_WGAN_examples.py

Repository files navigation

AdvGAN_WGAN-GP

Unoffical implmentation of "Using Wasserstein GAN to Generate Adversarial Examples".

Usage

1. Edit `main.py` to load your target model(trained) for attack.

2. Train AdvGAN_WAGAN-GP

   python main.py

3.Edit `test_WGAN_examples.py` to load your target model and AdvGAN.

4. Test model

   python test_WGAN_examples.py

Generated Adversarial Examples On MNIST(Acc: 4.15%)

About

Unoffical implmentation of Using Wasserstein GAN to Generate Adversarial Examples. Modified from https://github.com/mathcbc/advGAN_pytorch.

Report repository

Releases

No releases published

Packages

No packages published

Languages

Python 100.0%