mathurinm

This one supports groups while the merged PR doesn't, so it does not hurt to leave it open IMO

note: this is a PR from the `skglm` repo, instead it should be done from a branch of your fork @PABannier

No longer able to reproduce, closing.

Do you see an easy way of computing alpha_max for non convex penalites ? Is it worth the cost and code complication ?

We'd need a flexible way, because we handle penalties such as IndicatorBox. What's the pathological case we want to warn again in this case? For the record, sklearn fits an...

We could instead raise a warning when we exit the solver without having updated any coefficient

@Klopfe maybe you know what's going on ?

@Badr-MOUFAD thanks! the example name must start with `plot_` otherwise it is not run (that's why there is a running time of 0s here: https://output.circle-artifacts.com/output/job/ce6fb6a7-a421-48ad-b1e3-554d3621231d/artifacts/0/dev/auto_examples/grid_search_cv.html#sphx-glr-auto-examples-grid-search-cv-py) In addition can you add...

@jasonge27 any chance to merge this ?

Thanks for the quick reply Cedric, this is happening with the latest dev version : ``` In [1]: import ot ot.__version__ In [2]: ot.__version__ Out[2]: '0.9.3dev' ``` Contrary to #516,...