How to use an existing machine learning model with Azure Machine Learning?

Question

I have a Keras ML model .h5 file that I would like to publish as a web-service. This model was created in databricks. I want to use Azure ML for this purpose. I am following the steps given in this Azure documentation - https://learn.microsoft.com/en-us/azure/machine-learning/how-to-deploy-existing-model

One of the prerequisites is to have "Azure Machine Learning SDK".

My question is how to install "Azure Machine Learning SDK" in my Azure ml workspace? Do I need to type the commands in the Cloud Shell?

Any pointer would be helpful. Thanks.

Ram · Accepted Answer · 2020-02-03 10:00:29Z

2

If you are running in your own environment, follow SDK installation instructions. If you are running in Azure Notebooks or another Microsoft managed environment, the SDK is already installed.

An Azure Machine Learning workspace. To create the workspace, see Create an Azure Machine Learning workspace. A workspace is all you need to get started with your own cloud-based notebook server, a DSVM, or Azure Databricks.

edited Feb 3, 2020 at 10:00

answered Feb 3, 2020 at 9:49

Ram

2,7641 gold badge9 silver badges15 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Marius Constantinescu · Accepted Answer · 2023-04-10 01:08:01Z

There are a few prerequisites to get your model ready for inferencing. Assuming you have :

A compute instance needed for training, registration of the model or both (e.g. use Azure CLI as explained in the article at the end)
You must somehow bring that model into the Azure Machine Learning Workspace and into your Compute instance, e.g. a GIT repo could work to clone it into the instance.

And finally, the most straightforward way to publish for inference is to use Online Managed Endpoint using something like explained at https://learn.microsoft.com/en-us/azure/machine-learning/how-to-deploy-online-endpoints?view=azureml-api-2&tabs=azure-cli

In short, you'll need to define a deployment referencing the path to a model (e.g. in Data Assets), an environment, a scoring script (unless you use MLFlow and train locally, but that's a longer process) and the type of currate image to use, e.g. Standard_DS3_v2 and how many instances to scale up to when receiving requests.

Collectives™ on Stack Overflow

How to use an existing machine learning model with Azure Machine Learning?

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related