Rémi Lafage

Maybe... But there is no problem in the first place if you just clone `CountVectorizerParams` so no need of a special serde code complication.

Thanks for the heads up!

I've just took a look at scikit-learn to see what they have done. SOMs failed to make it [ten years ago](https://github.com/scikit-learn/scikit-learn/pull/2996). More recently, [this discussion](https://github.com/scikit-learn/scikit-learn/pull/19842) seems to suggest that recent...

> What, precisely, would you like me to research/investigate next before potentially proceeding with a PR? > It would be interesting to benchmark the method against kmeans for instance and/or...

I would say let's start with `linfa-kohonen`. At the moment, the use of `linfa-` prefix seems enough to show a project is linfa-friendly. It will be always possible to move...

Related to https://github.com/codecov/feedback/issues/358. The action should be updated to v4 but a token seems to be required.

Fixed by #368

Actually, `KRG` is not meant to handle multiple outputs. It has been recently added (too discreetly!) at the end of the [Kriging section](https://smt.readthedocs.io/en/latest/_src_docs/surrogate_models/gpr/krg.html#kriging-krg) but we should definitely raise an error...

After discussion, we start by adding a warning when using several training outputs instead of a hard error (PR #686). Though it is not the best one, a surrogate trained...

Below, just an example to see how multi-outputs kriging can go wrong. I've just used sphere and cos(10*x) functions. If we train against both at once with `KRG()`, we get...