advertorch v0.1: An Adversarial Robustness Toolbox based on PyTorch

TL;DR

advertorch v0.1 is an open-source PyTorch-based toolbox designed to facilitate adversarial robustness research by providing implementations of attacks, defenses, and robust training methods, leveraging dynamic computation for efficiency.

Contribution

It introduces a comprehensive, easy-to-use toolbox for adversarial robustness research built on PyTorch, enabling rapid development and testing of adversarial methods.

Findings

Provides a variety of attack and defense implementations

Leverages PyTorch's dynamic graph for concise code

Open sourced for community use and development

Abstract

advertorch is a toolbox for adversarial robustness research. It contains various implementations for attacks, defenses and robust training methods. advertorch is built on PyTorch (Paszke et al., 2017), and leverages the advantages of the dynamic computational graph to provide concise and efficient reference implementations. The code is licensed under the LGPL license and is open sourced at https://github.com/BorealisAI/advertorch .