mathurinm

Thanks to @josephsalmon  Celer first point is around 0.8 s and then convergence is fast. it seems like some first computation takes a long time and Blitz does not...

There is a buffer type mismatch so far with these two tests

1) avoid repetition of the block: ``` @cython.boundscheck(False) @cython.wraparound(False) @cython.cdivision(True) ``` 2) Not having these options when running on CI may catch more bugs (as #136)

int[:] X_indptr everywhere when calling check_estimator(clf) instead of check_estimator(Lasso) the test fails for pseudo large csr matrices easy to implement

ENH: test datasets module

2

It's doable as in benchopt

follow up to #99

Github now support math rendering in markdown with MathJax (https://docs.github.com/en/get-started/writing-on-github/working-with-advanced-formatting/writing-mathematical-expressions), so in the readme the math can be typed directly instead of embedded as images. Currently the rendering is not...

ENH improve speed of fitting Cox model by relying on the fast skglm solver

6

I am one of the developers of ``skglm``, a python package that improves ``scikit-learn`` for Generalized Linear Models by providing more functionalities and faster solvers. We have recently worked on...