MLOps on Azure End-to-End (E2E) Playbook (Ep. 2)

Original article was published on Artificial Intelligence on Medium

In Ep.1, I have demonstrated how you can set up the MLOps quickstart code from a GitHub Repository in Azure DevOps. I am going to switch gears to provisioning Azure Machine Learning on Azure, a tool to “empowering developers and data scientists with a wide range of productive experiences for building, training, and deploying machine learning models. Create an Azure Machine Learning workspace to train, manage, and deploy machine-learning experiments and web services.”

Let’s supply the following details:

  • Workspace name: Enter a unique name that identifies your workspace. Names must be unique across the resource group. Use a name that’s easy to recall and to differentiate from workspaces created by others.
  • Subscription: Select the Azure subscription that you want to use.
  • Resource group: Use an existing resource group in your subscription or enter a name to create a new resource group. A resource group holds related resources for an Azure solution.
  • Location: Select the location closest to your users and the data resources to create your workspace.
  • Workspace edition: Basic vs. Enterprise comparison

Give Azure some time for deploying resources via Azure Resource Manager (ARM). Afterwards, please launch and “try the new Azure Machine Learning Studio”.

Here’s a landing view that you see:

Primarily, Microsoft Azure Machine Learning consists of functionalities to author data science notebooks, control machine learning assets, and manage machine learning underlying compute infrastructure, e.g. CPU, GPU, or FPGA etc.

Author

Assets

Manage

For this episode, let’s deep dive on the notebook portion.

As you can see, you are allowed to:

  1. Create new file (Python Notebook or Text File)
  2. Create new folder
  3. Upload files (local only)
  4. Upload folders (local only)
  5. Refresh the view
  6. Collapse file explorer

In addition, you can:

  • Collaborate with your colleagues
  • Use RStudio, JupyterLab, Jupyter Notebook, or SSH.

Treat this notebook authoring experience as your development workspace. When you are ready for deployment, you should bring this code to Azure DevOps. This place is not meant for triggering build & release pipelines. Please also be reminded that Microsoft Azure Notebooks is an obsolete view, therefore you should be utilizing the notebook from Microsoft Azure Machine Learning platform.

Cheers! It works magically!