Software Open Access

Lightning: large-scale linear classification, regression and ranking in Python

Blondel, Mathieu; Pedregosa, Fabian

Lightning is a Python library for large-scale machine learning. More specifically, the library focuses on linear models for classification, regression and ranking. Lightning is the first project to integrate scikit-learn-contrib, a repository of high-quality projects that follow the same API conventions as scikit-learn. Compared to scikit-learn, the main advantages of lightning are its scalability and its flexibility. Indeed, lightning implements cutting-edge optimization algorithms that allow to train models with millions of samples within seconds on commodity hardware. Furthermore, lightning can leverage prior knowledge thanks to so-called structured penalties, an area of research that has recently found applications in domains as diverse as biology, neuroimaging, finance or text processing. Lightning is available under the 3-clause BSD license at http://contrib.scikit-learn.org/lightning/.

Files (764.6 kB)
Name Size
sklearn-contrib-lightning-0.4.0.tar.gz
md5:7324683380bb8a5d9257817bd549728d
764.6 kB Download
255
12
views
downloads
All versions This version
Views 255255
Downloads 1212
Data volume 9.2 MB9.2 MB
Unique views 247247
Unique downloads 1111

Share

Cite as