Deploying an MLflow model with managed online endpoints through the Python SDK V2

In the previous section, we leveraged AMLS Studio to deploy our MLflow model. In this section, we will explore code to deploy an MLflow model to a managed online endpoint through the SDK v2.

In order to leverage the SDK v2 for model deployment, we will leverage the Chapter 6 MLFlow Model Deployment SDK V2.ipynb notebook.

To deploy a managed online endpoint through the SDK V2, follow the next steps:

1. To deploy the model, we will create ManagedOnlineEndpoint with the appropriate configuration. In the case of an MLflow model, we will need to specify name and auth_mode. In addition, we will provide description as well as tags.

Figure 6.18 – Configure ManagedOnlineEndpoint

2. After the endpoint has been configured, we are able to call the create or update method, passing in the endpoint to create the endpoint in the workspace with the create_or_update...