Deployments#

When deploying to production, it is usually desirable to include as few dependencies as possible. For this scenario we offer the interpret-core package, which includes only the dependencies required for EBM fitting, predicting, editing, serialization, and generating explanations.

The interpret package includes all the dependencies of interpret-core, plus additional dependencies required for visualizing EBMs, generating other glassbox models, and blackbox explanations.

For further customization, we support extra tags on pip installs to fine tune the dependencies.

Install with every dependency (default)

The package interpret installs every dependency needed to run any part of the package.

pip install interpret

conda install -c conda-forge interpret

git clone interpretml/interpret.git && cd interpret/scripts && make install

Install with minimal dependencies

When you only want the required dependencies, or you wish to customize the dependencies, install the package interpret-core instead.

pip install interpret-core

conda install -c conda-forge interpret-core

git clone interpretml/interpret.git && cd interpret/scripts && make install-core

Install with some official dependencies (pip)

This scenario is not covered in all package managers we support. If you are installing with pip, you can take advantage of extra tags that are exposed for interpret-core.

Installing with all extra tags is equivalent to installing the interpret package, and can be done with:

pip install interpret-core[debug,notebook,plotly,lime,sensitivity,shap,linear,skoperules,treeinterpreter,aplr,dash,testing]
  • interpret utilizing interpret-core tags found here.

  • interpret-core extra tags found here.