MLP01SCD Tiny Demo code

Reference

1. Gradient free sign activation zero one loss neural networks for adversarially robust classification

2. Paper code for reproducing results in the paper

3. Full code for usage

Prepare

1. Install pytorch

conda install pytorch torchvision torchaudio cudatoolkit=11.3 -c pytorch

2. Install advertorch

pip install git+https://github.com/BorealisAI/advertorch.git

3. You may also need sklearn and pandas as well.

Guide

1. Training neural network in main.
2. Run for PGD/FGSM attack in attack_demo.