This code is no longer maintained.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.
It's a hanging PR so far in celer (https://github.com/mathurinm/celer/pull/169, not merged due to speed regression and it does not handle 0 weights). I think it would be very useful for non convex penalties too.
I was wondering if we could do a generic l1 reweighting for an penalty with weights arguments, this way could think of reweighted Huber or reweighted lad lasso
It's a hanging PR so far in celer (https://github.com/mathurinm/celer/pull/169, not merged due to speed regression and it does not handle 0 weights). I think it would be very useful for non convex penalties too.