Alexander Andreev

Fixed in https://github.com/intel/scikit-learn-intelex/pull/980 and 2021.6 release

Hi @guankaer, could you provide some code + OS, python, sklearn and sklearnex versions to reproduce this issue, please?

Case when number of rows is less than number of features is specific even for streaming mode. We will take time to find out is this case meaningful from point...

I guess CI fails might be ignored for docs only PRs. @outoftardis, can you verify that change is meaningful and suitable for README.md (not another file like CONTRIBUTING.md, for example)?

`py310` wheel will be available in next release. We will think about possibility to add `sdist`, but currently I don't see much difficulties for its addition except requirement to download...

2021.6 PyPI and intel conda releases with correct py310 support are available

I will add run of some faiss benchmarks later. Can you provide details of case where MKL didn't improve performance?

Hi @eddiebergman, Can you share the reason why short sklearnex mention with link to manual can't be included in README.md or start section of docs? Currently, sklearnex doesn't support distributed...

`patch_sklearn` is recommended and not connected to distributed mode setup. sklearnex doesn't directly support dask but works in subprocess. For example, this code will be accelerated: ```python from sklearnex import...

@eddiebergman, any updates from Monday meeting?