Harry24k/adversarial-attacks-pytorch

Health v2 · maintenance only

Adoption not part of the score

2177 stars · 368 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Torchattacks is a PyTorch library providing implementations of adversarial attacks to generate adversarial examples against deep learning models. It offers a PyTorch-like interface so users can easily apply attacks like PGD to their models.

Use cases

• generate adversarial examples for a pytorch model
• evaluate model robustness against adversarial attacks
• run PGD attacks on image classifiers
• benchmark adversarial attack methods in research
• create adversarial training data
• test deep learning model security

When to choose

• you use PyTorch and need ready-made adversarial attack implementations
• you are researching adversarial robustness and want a citable, widely used library
• you want a simple PyTorch-like API to generate adversarial examples

When to avoid

• you need adversarial defenses or training frameworks rather than attacks
• you work in TensorFlow or JAX instead of PyTorch
• you need attacks for non-image modalities not covered by the library

Facets

library · maturity active

machine-learning deep-learning security machine-learning deep-learning security python adversarial-attacks adversarial-examples pytorch robustness torchattacks

Member repositories

For agents

markdown · JSON · MCP: product_card(name="Harry24k/adversarial-attacks-pytorch")