← all repositories

NorskRegnesentral/skweak

A Python toolkit that applies weak supervision techniques to automatically label documents for NLP tasks.

926 stars Python ML FrameworksData Tooling
skweak
Velocity · 7d
+0.5
★ / day
Trend
steady
star history

skweak enables users to define labelling functions (heuristics, gazetteers, ML models, or crowd annotations) that automatically label documents, then aggregates their predictions using a statistical model that estimates each function’s accuracy and confusion matrix. It supports both sequence labelling and text classification tasks, providing a complete API built around spaCy.

heatdrop uses Google Analytics to see which pages get read — nothing else. Your call. How we handle data.