scikit-learn tip #37: issue with set_config() function

[Xavier-F]

Dear all,

Thanks to the fantastic work from Kevin to spread his data science knowledges accross the web, I discovered this new pipeline interactive diagramm in sklearn v0.23. I tried to use it on one of my jupyter notebook after ruing a conda update --all, to ensure no compatibility issues, and I am really struggeling with it.
I can import the package. But when I just try to set display parameter to 'diagram', it is just telling me "set_config() got an unexpected keyword argument 'display'"

I am searching to fix this issue on my notebook since nearly 1 h, but impossible to find what went wrong. Does anyone have an idea on how to run this new function?

Thanks a lot for your support, and take good care!

[ericabrauer]

The first thing I would do is to run help(set_config). It may be that the function only accepts positional arguments. Hope this helps!

[akanz1]

Can you check the scikit-learn version being used by the currently active ipython kernel?

sklearn.__version__

Also make sure the package is up to date in your currently active virtual environment or switch to base if that's where you updated to 0.0.23.

[justmarkham]

@Xavier-F Thanks for your kind words!

I agree with @akanz1, definitely check your sklearn version, since you will need 0.23:

import sklearn
sklearn.__version__

And with conda, you can check your versions with conda list

Hope that helps!

[Xavier-F]

Dear Erica, Andreas and Kevin

Thanks a lot for your super prompt support. Indeed, seems that the conda update --all only update scikit learn package toward 0.22 version.
I just tried again with pip install --upgrade scikit-learn, and looks like all is runing smoothly now on my notebook.

Once again thanks a lot, and take good care