mathurinm/celer

ENH: check that we handle warm start properly

Closed this issue · 2 comments

it should increase coverage

@MalcolmWolff how would you go about testing this?

I would fit a first time, then refit with warm_start = True and check that n_iter (in the sense of celer, ie number of subproblems defined) is 1. Currently I'm not sure we return n_iter.

Closed in #49