How to select serverless compute as a default in Azure ML Studio Designer

144 views Asked by At

In Microsoft Azure ML Documentation is mentioned: "When using Azure Machine Learning designer, select Serverless as default compute." and "You can also set serverless compute as the default compute in Designer." in documentation: Model training on serverless compute

Where can I set "serverless" as a default compute in Azure ML Studio Designer?

I cann't find. When I setup pipeline job, there is no option "serverless" in Designer for "Select compute type". When I don't select any compute type, error appear "Please select a default compute to run a pipeline.".

1

There are 1 answers

0
JayashankarGS On

In Azure Machine Learning, there are two types of components that the designer supports:

  1. Classic prebuilt components (v1)
  2. Custom components (v2)

enter image description here

The serverless compute is supported only in Custom components (v2),

as it is supported in Azure CLI ml extension v2 (current) and Python SDK azure-ai-ml v2 (current).

enter image description here

Classic prebuilt components (v1) will not be supported for serverless compute. Although the components are supported for designing, no new components will be added.

Therefore, I highly recommend that you use custom components, as they are compatible with AzureML V2 and will continue to receive new updates.

You can refer to the documentation below for more information about designer version 2.

What is the Azure Machine Learning designer(v2)? - Azure Machine Learning | Microsoft Learn