This code is no longer maintained. The codebase has been moved to https://github.com/scikit-learn-contrib/skglm. This repository only serves to reproduce the results of the AISTATS 2021 paper "Anderson acceleration of coordinate descent" by Quentin Bertrand and Mathurin Massias.
[x] find new name for class as it will handle enet
[x] add support for generic triplet penalty(w, alpha), prox(w[j], alpha, stepsize), kkt_violation(XtR, alpha) to handle any penalty (maybe we should pass a Penalty jit class instead ?)
[ ] be aware that WeightedLasso needs a array to work correctly
[ ] pass scikitlearn's check_estimator
[x] in path, keep XtR (to be called gradient later on, for generic datafits) up to date (it's computed at first iteration for WS creation, and before exit to check KKT)
@QB3 I have edited the todolist for monday's sprint, feel free to add stuff. I think having a functional enet at the end of the day would be nice and useful for the community !
penalty(w, alpha), prox(w[j], alpha, stepsize), kkt_violation(XtR, alpha)
to handle any penalty (maybe we should pass aPenalty
jit class instead ?)check_estimator