Skip to content

Instantly share code, notes, and snippets.

@WittmannF

WittmannF/deploy-automl-azure.md

Created October 18, 2020 20:31
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save WittmannF/579425c3f9b58a82fbf33472c34146e5 to your computer and use it in GitHub Desktop.
Save WittmannF/579425c3f9b58a82fbf33472c34146e5 to your computer and use it in GitHub Desktop.
Download ZIP
Raw
deploy-automl-azure.md

Solution: Deploy an Azure Machine Learning Model

Part 1: Configure deployment settings

  1. Create a new Automated ML run

  1. Next, make sure you have the dataset uploaded. If you don't, upload and select it. This solution uses the bike-no.csv dataset.

  1. Create and configure your new compute cluster.

  1. Once the new compute cluster is successfully created, use this cluster to run the autoML experiment. Make sure you fill the name and target column.

  1. You will see the experiment in the experiment section and a new model is created.

Part 2: Deploy an Azure ML model

  1. Go to the Automated ML section and find the recent experiment with a completed status. Click on it.

  1. Go to the "Model" tab and select a model from the list and click it. Above it, a triangle button (or Play button) will show with the "Deploy" word. Click on it.

Then

  1. Fill out the form with a meaningful name and description. For Compute Type use Azure Container Instance (ACI)

  2. Enable Authentication

  3. Do not change anything in the Advanced section.

  1. Deployment takes a few seconds. After a successful deployment, a green checkmark will appear on the "Run" tab and the "Deploy status" will show as succeed.

Next

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment