Spaces:

sklearn-docs

/

MNIST_classification_using_multinomial_logistic_L1

like 8

Sleeping

Hi @eswardivi thanks for the nice contribution! I really like the possibility to sketch a number :)
Just a comment, I think the custom solver could be fixed to saga, because standard scaling the data can be detrimental depending on the solver (see https://github.com/scikit-learn/scikit-learn/pull/25334#issuecomment-1470697577) and having too many options can get the user a bit lost IMHO.
I think elasticnet is also not working properly :(