topic Re: Performance benchmarking data for Intel distributed deep learning/machine learning libraries in Intel® Distribution for Python*

Performance benchmarking data for Intel distributed deep learning/machine learning libraries

Parag_Jain — Fri, 03 Sep 2021 04:30:06 GMT

I need information about how well does ML/DL libraries distributed by Intel performs as compared to non-Intel distributed version of same ML/DL libraries. Few of the libraries that I am referring to are numpy, sklearn, tensorflow, pytorch and xgboost.

I also need to find out if Spacy is planned in future to be optimized and distributed by Intel. If not, is there an existing alternative to Spacy which is already in distribution by Intel?

Re: Performance benchmarking data for Intel distributed deep learning/machine learning libraries

AthiraM_Intel — Thu, 09 Sep 2021 11:12:30 GMT

Hi,

Thanks for posting in Intel Forums.

You could refer the following links to know more about intel optimizations.

https://github.com/IntelPython/scikit-learn_bench

https://github.com/intel/scikit-learn-intelex

https://medium.com/intel-analytics-software/new-optimizations-for-cpu-in-xgboost-1-1-81144ea21115

https://www.intel.in/content/www/in/en/analytics/artificial-intelligence/mlperf-validates-cpu-for-deep-learning-article.html

You can use Intel® oneAPI AI Analytics Toolkit (AI Kit) which includes intel optimized frameworks and libraries. Please find the below links for more details:

https://github.com/oneapi-src/oneAPI-samples/tree/master/AI-and-Analytics/Features-and-Functionality

https://software.intel.com/content/www/us/en/develop/tools/oneapi/ai-analytics-toolkit.html#gs.a5p2k3

Regarding Spacy, currently it is not intel optimized. We are checking internally whether there is any plan to optimize Spacy by Intel and let you know.

Hope this helps. If you have any further issue, please let us know.

Thanks.

Re: Performance benchmarking data for Intel distributed deep learning/machine learning libraries

AthiraM_Intel — Thu, 14 Oct 2021 11:20:00 GMT

Hi,

As discussed internally, we are closing this case. If you need any additional information, please post a new question as this thread will no longer be monitored by Intel.

Thanks.