Azure ML Deploy Inference Endpoint
In this video, we'll walk through how to take a machine learning model and turn it into a working web service. We'll start by setting up the necessary resources in Azure, then move into the ML Studio to create a compute cluster and train our model. We'll cover the full process, from uploading data and training the model all the way through registering it and deploying it to a REST endpoint, finishing up by showing you how to test the endpoint and fetch the necessary authentication token.
Description
When a machine learning model is validated and effective at predicting outcomes, the next step is putting them to use in production applications. An excellent way to integrate AI and ML models with other applications is providing predictions as a standard web service endpoint--accepting JSON inputs and emitting predictions as JSON objects.
This video is a end-to-end tutorial showing how to deploy a Machine Learning model created in Microsoft Azure Machine Learning to a REST endpoint callable from another application via HTTP.
This video is based on the blog post https://robkerr.ai/deploy-azure-machine-learning-model-to-rest
00:00 - Introduction
00:39 - Create Resources
01:58 - Launch ML Studio
02:40 - Create Compute
04:15 - Upload Data
05:51 - Create Notebook
07:51 - Train Model
09:00 - Review Model
10:31 - Register Model
11:21 - Deploy Endpoint
14:32 - Fetch Bearer Token
16:12 - Test REST Endpoint