This repository is a PyTorch Implementation of generating adversarial images using various types of attacks schemes.
Clone this repository into any place you want.
git clone https://github.com/aamir-mustafa/generate-adv-examples
cd generate-adv-examples
Download the NIPS 2017 Adversarial Learning Development Dataset comprising of 1000 ImageNet alike images of size 299 x 299.
Kaggle Link.
Download the file images.zip
to the directory.
Use the following python file to generate and save the adversarial examples against various ImageNet pre-trained models. We provide implementations for four types of attacks : FGSM, BIM, MIM, PGD
generate.py