Jérémie du Boisberranger

Thanks @jjerphan !

I see a lot of -1 in this discussion. I'm not in favor of adding a warning mechanism either. I'd rather explain in the eps field of the docstring that...

Working on `NMF` and `MiniBatchNMF`

> @jeremiedbb > Wondering why sklearn/linear_model/_glm/glm.py doesn't appear on this list? @reshamas The estimators of this file are listed here: `PoissonRegressor`, `TweedieRegressor` and `GammaRegressor`. `GeneralizedLinearRegressor` is now a private class...

@devkaranjoshi the checklist only references the ones for which there is a PR and this PR is merged. Here's a list of all current PRs not merged yet: https://github.com/scikit-learn/scikit-learn/pulls?page=1&q=is%3Apr+label%3AValidation+is%3Aopen

@kasmith11 indeed a PR adding param validation has been submitted a few weeks ago and was merged a few days ago, so this estimator is done. But there are still...

@stefmolin the validation here can't depend on another parameter or on the input. I'd just put a generic ``[str, callable]``.

There's a PR to add this kind of constraint but it's not merged yet https://github.com/scikit-learn/scikit-learn/pull/24048

Looks very interesting. Do you think it would provide a solution to this issue https://github.com/xianyi/OpenBLAS/issues/3187 ?

> could you please try with the OpenBLAS "develop" branch ? sorry I said master but it was against the develop branch. I didn't notice the main branch name was...