Alexander Andreev

/intelci: run

Reading [MklLapack code](https://github.com/oneapi-src/oneDAL/blob/master/cpp/daal/src/externals/service_lapack_mkl.h#L103) you can find `xx` functions set number of threads to 1 locally: ```cpp template struct MklLapack { typedef DAAL_INT SizeType; static void xgetrf(DAAL_INT * m, DAAL_INT *...

I can't reproduce this issue with 2024.0.1 and 2024.1.0 versions of sklearnex. The only visible difference is usage of Windows 11 with conda pkgs manager on my side.

Support of `pypy` interpreter is not implemented currently.

> So far I am also running into an issue downloading datasets in 3.9 with urllib, specifically with the SSL certs. To get it to download I had to export...

> Also, some of the errors seen in the PR associated with KNN/sklearn in the CI checks have been solved with this PR: [scikit-learn/scikit-learn@f473d7e](https://github.com/scikit-learn/scikit-learn/commit/f473d7e5bb4aaaf2bb292aa7b0c6195b470daf39) > > But likely a handling...

Hi, this error highly likely is caused by different number of input column between fitting and prediction. Please, check shape of data provided for prediction.

@Kange2014, what is the type (numpy.ndarray, pd.DataFrame, etc.) and shape of input data? It's essential to reproduce this issue, and some additional information might be required later.

I run mostly similar code (except loading from 'model.pkl' because it's unsafe to use pickle on externally provided models) and it finished successfully: ```python import pandas as pd import numpy...