Online Deployment
Explore how to deploy machine learning models online using Azure, including setting up necessary compute resources, managing quotas, and performing deployment updates. Understand the process of testing and scaling your models using online endpoints and MLflow integration to ensure efficient real-time predictions.
We'll cover the following...
Prerequisites for deployment
Online deployment creates a real-time URI for the model. Batch deployment works with low-compute VMs like the Standard-DS11-V2 series. However, for online deployment, we need high-end VCPUs like the Standard-DS2-V2, Standard-DS3-V2, or Standard-Fs2-v2 series. Getting the quota is the first step for online deployments. We can check the quota using the Azure Machine Learning studio by following the steps below:
Select the online endpoint that we created earlier and click “Add Deployment.”
Add the model and environment details.
Click the ...