search

intel / scikit-learn-intelex

Intel(R) Extension for Scikit-learn is a seamless way to speed up your Scikit-learn application
https://intel.github.io/scikit-learn-intelex/
Apache License 2.0
1.21k stars 174 forks source link

Patching not implemented for GradientBoost Algorithms #1736

Open ajay-fuji opened 7 months ago

ajay-fuji commented 7 months ago

While running gradientboost regression and classification algorithms, backend used is sklearn and not onedal. daal4py/sklearn/monkeypatch/dispatcher.py file does not contain mapping for gradientboost. Although while checking for examples, daal4py have gradientboost examples.

Since gradientboost is supported by oneDAL, why patching is not implemented in sklearnex? Is this still in progress or TBD?

napetrov commented 7 months ago

With GBT opts we opted out to focus on XGBoost direct optimization as more influential framework, and for daal4py we focused on model builders - functionality to consume XGBoost, LightGBM, CatBoost models and run faster inference.

scikit gbt was not in focus due to it's limited use, but patching technically can be extended to it as well.