how can I use data assets in azure ml registries in azure ml designer?

Question

I have created some uri_folder data asset in my custom Azure ml registry.
Also, I have created some components in my registry.

When I want to create a new custom pipeline in Azure ML Designer, I can use my custom component, but I can't see my registered data assets.

How can I use shared/Registered data assets in designer in multiple workspaces?
Thanks.

Answer 1

Hello @Hossein Ansari

Thanks for reaching out to us, could you please try below workflow to see if that works for you?

To use data assets that are stored in your Azure ML Registry in Azure ML Designer, you can follow these steps:

1. In the Azure ML Designer interface, click on the "Datastores" tab on the left-hand side of the screen.
2. Select "Azure ML" as the datastore type, and choose the appropriate Azure ML workspace from the dropdown menu.
3. Enter the authentication information for the workspace, including the subscription ID, resource group, and workspace name.
4. Once you have connected to the workspace, you should be able to see your registered data assets under the "Data assets" section. If you don't see your data assets, ensure that they are registered in the same workspace that you are connecting to.
5. Drag and drop the data asset you want to use onto the canvas to create a data reference node. You can then connect this node to other nodes in your pipeline to use the data asset as an input.

To use shared or registered data assets in multiple workspaces, you can follow these steps:

1. Register the data asset in the Azure ML Registry of the workspace where it was created.
2. In the other workspace where you want to use the data asset, create a new datastore as described above and connect to the appropriate workspace.
3. In the pipeline, drag and drop a "Data reference" node onto the canvas and select "Registered dataset" as the reference type.
4. Choose the appropriate workspace and dataset from the dropdown menus.
5. Connect the data reference node to other nodes in your pipeline as needed.

By following these steps, you should be able to use your registered data assets in your Azure ML Designer pipelines and share them across multiple workspaces.

Let us know if any issues arise. Thanks.