Use model interprebility in Azure ML to explain model predictions and provide feature importances at inference time.
- create_explanations generate and pickle the explainer object for the best model of an AutoML run.
- deploy_explanations deploy a web service with the model explanations and model predictions. Also generate the scoring script that goes with the deployment.