Kai-Wei Chang: dual coordinate descent for L1- and L2-loss dual SVM, multi-class (Crammer & Singer), L1-regularized classifiers.

Cho-Jui Hsieh: dual coordinate descent for L1- and L2-loss dual SVM, multi-class (Crammer & Singer), L1-regularized classifiers.

Rong-En Fan: Big helper on all issues.

Guo-Xun Yuan and Chia-Hu Ho: L1-regularized classifiers.

Hsiang-Fu Yu and Fang-Lan Huang: dual coordinate descent for L2-regularized logistic regression

Hsiang-Fu Yu: python interface

Chia-Hua Ho: solvers for linear support vector regression

Bo-Yu Chu and Chia-Hua Ho: code for parameter selection

Mu-Chu Lee and Wei-Lin Chiang: abstraction of two major operations (dot and axpy) in solvers. Combining two matrix-vector multiplications on Newton to one.

Chih-Yang Hsia and Ya Zhu: code for improved trust-region update rule in the primal-based Newton method (version 2.11)

Hsiang-Fu Yu and Hsin-Yuan Huang: scipy support in python interface (version 2.11)

Chih-Yang Hsia and Wei-Lin Chiang: Preconditioned CG for Newton method (version 2.20)

Jui-Yang Hsia: code for SVR parameter selection (version 2.30)

Hung-Yi Chou and Pin-Yen Lin: one-class SVM solver (version 2.40)

Ching-Pei Lee and Leonardo Galli: line-search Newton method (version 2.40)

Wei-Lin Chiang and Jui-Nan Yen: code for not regularizing bias and preparing 2.40 release (version 2.40)

Hsiang-Fu Yu, Wei-Lin Chiang, Jui-Nan Yen, Yu-Sheng Li, and Leonardo Galli: code and settings for version 2.42

Hsiang-Fu Yu, Jui-Nan Yen, Yu-Sheng Li, and Wei-Lin Chiang helped to create the installation through PyPI for version 2.43.

Guan-Ting Chen helped to improve the implementation of working-set selection in one-class SVM solver and prepared the release of version 2.46.

He-Zhe Lin helped to add a flag for recalculating w in dual solvers and Hung-Chih Chiang helped to prepare the release of version 2.48.

Chih-Jen Lin: Project lead and maintainer.