Michael Mayer

Results 161 comments of Michael Mayer

> * Could you try to make CI green? > * Then, en entry in the user guide, https://scikit-learn.org/stable/inspection.html, would be good. > * Could you add a test for...

No idea how to fix the failing checks... might it be due to some different pandas on Debian?

> > No idea how to fix the failing checks... might it be due to some different pandas on Debian? > > You can see in the [CI](https://dev.azure.com/scikit-learn/scikit-learn/_build/results?buildId=66118&view=logs&j=aabdcdc3-bb64-5414-b357-ed024fe8659e&t=b7b3ba55-d585-563b-a032-f235636c22b0&l=4911) that Linux_Docker...

@ogrisel Thanks for your assisstance. I had a look at _safe_assign(): It actually supports only pandas and numpy. So no wonder it fails for polars :-). ```py import polars as...

> Is this now blocked by #28931? I don't think so. But there seems to be other old-release Pandas related failures that are tricky to investigate.

The predict function should return only one numeric value: https://github.com/bgreenwell/fastshap/issues/14

Rerunning above code with version 0.7.1 clearly shows some improvement, but there is still a factor of 20: Maybe your dataset does not contain compressable columns? The difference comes from...

Thanks for the explanations @Mytherin . A random double column and a random string column shows almost equal parquet size (using duckdb 0.7.1): ``` import numpy as np import pandas...

Since it seems to be a standard workflow to store the result as parquet, I'd love to keep the issue open.

Same comment as last time