[ENH] exploring integration with `lipo`

[fkiraly]

Random discovery today, one of the sktime core developers - @jdb78, the author of pytorch-forecasting - has also written a gradient-free optimization package 5 years ago:

https://github.com/jdb78/lipo

Given the similarity in scope, this might be an interesting test case for integration.

[Abhishek9639]

Hi @fkraly,
This looks like an interesting enhancement. I’d love to explore the integration with lipo and evaluate how it can fit into the project.
If it’s okay, could you please assign this issue to me?
Thanks!

[direkkakkar319-ops]

Hi @fkiraly, great find! The fact that lipo is parameter-free makes it a really compelling addition ,no tuning the tuner is a genuinely nice property to offer users.

I'd like to take a stab at this if it's still open.
My plan is to write a wrapper that adapts lipo's continuous bounds API to Hyperactive's discrete search space format, handle categoricals, and wire it up following the existing extension templates.
I'll open a draft PR once I have something working and would love early feedback if there are any gotchas to watch out for.

Thanks!

[direkkakkar319-ops]

Hi, @SimonBlanke i have raised a PR for the following issue would appreciate your feedback .