mathurinm

There seems to be an error in the `.xy` file that prevents this from working (see also https://github.com/mathurinm/libsvmdata/issues/33#issuecomment-1159092937) ``` In [3]: libsvmdata.fetch_libsvm("SUSY", verbose=True) Dataset: SUSY Downloading data from https://www.csie.ntu.edu.tw/~cjlin/libsvmtools/datasets/binary/SUSY.xz (585.8...

You may have a cached version already; try forcing the download or deleting the file to reproduce Le jeu. 16 févr. 2023 à 20:26, Badr MOUFAD ***@***.***> a écrit :...

@wangqiuoe can you send a PR to fix this ?

Hello, This seems non immediate to fix: ``` import lzma from sklearn.datasets import load_svmlight_file f = lzma.open("SUSY.xz") res = load_svmlight_file(f) ``` yields ``` ValueError: could not convert string to float:...

Welcome back @tomaszkacprzak ! $\epsilon$ is a vector, so I'm not sure how $1[\epsilon > 0]$ is to be interpreted, when multiplying $\Vert \epsilon \Vert^2$? Is it not something like:...

Yes that sounds like a plan, and no I don't foresee any difficulty!

@tomaszkacprzak could you also please let us know in which context you use skglm, and what it brings you compared to alternatives?

Hi @tomaszkacprzak, any news on this?

I have an issue with the datafit update step, because the gradient of `0.5 || Xw - y + a 1_n||` with respect to `a` is `n a - (Xw...

@bgirault-inria is the matrix to store the inverse of $C + \sum x_i a_i a_i^\top$ ? That you can update with Shermann-Morrison after a change of a single coordinate of...