TextAttack
Framework for adversarial attacks, data augmentation, and model training in NLP.
About
TextAttack by the QData lab is a Python framework for adversarial attacks, data augmentation, and model training in natural language processing. It includes more than a dozen attack recipes for probing how well a model holds up under adversarial input, augmentation methods to improve generalization, and a single-command training module with downloads included. A GPU is optional but speeds it up. Released under the MIT license.
Reviews (0)
Leave a Review
No reviews yet. Be the first to review!
Details
- Category
- Natural Language Processing
- Price
- Free
- Platform
- Local/Desktop
- Difficulty
- Intermediate (3/5)
- License
- MIT
- Added
- Apr 3, 2026
Related Tools
Generalist model for named entity recognition using bidirectional transformer.
Language detection library ported from Google language-detection.
Comprehensive NLP toolkit for education and research in Python.
Data protection and PII anonymization framework by Microsoft.
Topic modeling and document similarity library for unsupervised NLP.
Simple framework for state-of-the-art NLP by Zalando Research.