The official PyTorch implementation of paper: Switching Gradient Directions for Query-Efficient Black-Box Adversarial Attacks
This repository is not active