mathurinm

@fsaforo1 would you be able to send a Pull request to start implementing tweedie datafit ? If this is your first PR, a small contribution guide is at https://github.com/mathurinm/github-assignment/?tab=readme-ov-file#summary-how-to-contribute-to-an-existing-repository

It's better if you start the PR right now, even if not working, this way we can all look at the same code and help you

@fsaforo1 can you show your code, either in your repo either in a pull request, so that we work on it together and integrate it in skglm? Thanks a lot!

I'm late to the party, but instead of graying the unselected solvers, would it be possible to collapse the unselected block, displaying only its title ? If there are many...

@wassimmazouz maybe you can have a look install gsrtopim in the `test` requirement of pyproject.tom by adding `"gsroptim@git+https://github.com/EugeneNdiaye/Gap_Safe_Rules/"` to the list of requirements for test then one needs to install...

Thanks @QB3 that sounds doable How would we pass the current design ? `y` does not have a default value in `Solver.solve`, so it's required. Making it a keyword arg...

There's a working (withoutintercept) implementation with minimal overhead (we inherit most of the stuff from Quadratic) in #250 I need to get the intercept update step right, but it was...

- For the implementation passing `X` and `Xty` see a working draft implementation at #250 - For the one with `XtX` I guess we need to do something else, like...

> However, this comes with a drawback of making the convergence speed very very slow. I think a good enough solution would be to add an `unsafe_stepsize=True` parameter, that uses...

Hi @jeandut, as of #472 instead of `import_from` you can use a regular Python module that is named `benchmark_utils`, accessible under this name from all solvers and datasets. The folder...