Comparisons anchored on ART, the LF AI & Data library for testing classic machine learning models against evasion, poisoning, extraction and inference attacks.
A maintained library against Microsoft's automation layer built on top of it: what Counterfit's CLI adds for assessments, its last release in 2022, and why most teams testing fraud, vision or tabular models now script ART directly.
A multi-framework library for images, tables, audio and video against an NLP-specific attack and augmentation framework for Hugging Face models: which fits a text classifier, which fits everything else, and how they combine.
The classic adversarial example shortlist judged on maintenance first: ART's 2025 release and LF AI & Data home, Foolbox's last release in 2024 and CleverHans' in 2021, then framework support and threat coverage beyond evasion.
Tell us what you're working on, whether it's a deployment, an audit, a security test or a cyber range. You'll speak with an engineer who can help you scope it.