8 Deploying the model

 

This chapter covers

  • Overview of model deployment
  • Deployment versus one-off scoring
  • Why deployment is a difficult topic
  • Steps for deploying your model
  • Introduction to pipelines
  • Maintaining a model after it has been deployed

In chapter 6, we went through the process of iteratively training the deep learning model for predicting streetcar delays, and in chapter 7, we went through a further set of experiments to explore the behavior of the model. Now that we have a trained model, we are going to look at two ways to deploy the model or, in other words, make it possible for streetcar users to get predictions about whether their streetcar trips are going to be delayed. First, we’ll get an overview of the deployment process. Next, we’ll contrast the one-off scoring we introduced in chapter 6 with deployment. Then we will go through the specific steps to deploy the model by using two approaches: a web page and Facebook Messenger. Next, we will describe how to encapsulate the data preparation process by using pipelines and go over the details of implementing pipelines for the streetcar delay prediction model. We will wrap up the chapter with a review of how to maintain a model when it has been deployed.

8.1 Overview of model deployment

8.2 If deployment is so important, why is it so hard?

8.3 Review of one-off scoring

8.4 The user experience with web deployment

8.5 Steps to deploy your model with web deployment

8.6 Behind the scenes with web deployment

8.7 The user experience with Facebook Messenger deployment

8.8 Behind the scenes with Facebook Messenger deployment

8.9 More background on Rasa

8.10 Steps to deploy your model in Facebook Messenger with Rasa

8.11 Introduction to pipelines