MicrosoftDocs / azure-docs

Open source documentation of Microsoft Azure

https://docs.microsoft.com/azure

Creative Commons Attribution 4.0 International

10.1k stars 21.14k forks source link

Difference between local and online deployments #111243

Closed natke closed 7 months ago

natke commented 1 year ago

It's great to have the tools to develop and debug endpoints locally. The document is a little light on differences between the paths that you use for environments and models locally vs in the cloud. Is there a single configuration that can be used for both?

Document Details

⚠ Do not edit this section. It is required for learn.microsoft.com ➟ GitHub issue linking.

ID: 88bf509a-2f62-0585-e007-e1eaba60f0cf
Version Independent ID: 7c1ea59e-d234-d931-226e-cf701df8c259
Content: Deploy machine learning models to online endpoints - Azure Machine Learning
Content Source: articles/machine-learning/how-to-deploy-online-endpoints.md
Service: machine-learning
Sub-service: mlops
GitHub Login: @dem108
Microsoft Alias: sehan

Naveenommi-MSFT commented 1 year ago

@natke Thanks for your feedback! We will investigate and update as appropriate.

Naveenommi-MSFT commented 1 year ago

@natke Thank you for bringing this to our attention. I've delegated this to content author @dem108, who will review it and offer their insightful opinions.

Naveenommi-MSFT commented 1 year ago

@dem108 Could you please review add comments on this, update as appropriate.

dem108 commented 7 months ago

Thanks for your feedback @natke . Local deployment does have limitation that it cannot use model or environment registered in the workspace, but only can use model/environment from local. This is basically because local deployment would run locally isolated from cloud. This is described here: https://learn.microsoft.com/en-us/azure/machine-learning/how-to-deploy-online-endpoints?view=azureml-api-2&tabs=azure-cli#:~:text=Local%20endpoints%20have%20the%20following%20limitations

We recognize it only mentions local model. Let us follow up and update it to reflect on environment as well.

Thanks!

dem108 commented 7 months ago

The article is updated to mention environment as well. Let us close this, but feel free to comment further.

dem108 commented 7 months ago

please-close

natke commented 7 months ago

Thank you for making the update. So do I conclude from this that I need a different deployment yml file for local and cloud deployments?

dem108 commented 3 months ago

Thank you for making the update. So do I conclude from this that I need a different deployment yml file for local and cloud deployments?

Not necessarily. You can use the same file but override as needed by using --set.