We are happy to announce that SHAP has joined the German Center for Open Source AI (GC.OS)!
SHAP is an open source Python library that helps explain the predictions of machine learning models. Using Shapley values, it shows how individual features contribute to a prediction, helping users understand and investigate their models.
A first milestone together
GC.OS has already helped us tackle a major maintenance challenge by funding a stipend for work to remove SHAP’s Numba and llvmlite dependencies. Zhihao Dai drove the implementation, moving performance-critical code from Numba to C++.
This gives SHAP more control over support for new Python versions and removes dependency constraints for users. It also opens a clearer path towards running SHAP in the browser. You can read more about the work in Removing numba and llvmlite dependencies from SHAP.
Looking ahead
Thank you to GC.OS for supporting the work and to everyone who contributed to making it happen. We are excited to build more amazing things together and keep making open source explainability easier to use.
