Machine learning service instances

Last updated: Oct 09, 2024

You can use a machine learning service instance to govern what resources to use to train machine learning models and deploy assets. In this tutorial, you learn how to create, modify, and delete a machine learning service instance and associate it with a deployment space.

Creating a machine learning service instance

You can use the service instance to associate with projects, experiments, and deployment spaces. For more information, see Associating a service instance with a deployment space

Prerequisites:

You must have the service-level administrator role to be able to create a Watson Machine Learning service instance to use with deployment spaces. Refer to Checking your IBM Cloud IAM account access roles.

Follow these steps to create a machine learning service instance:

From the navigation menu, click Services > Service instances to access your service instances list.
Click Add service to open the IBM Cloud service catalog.
Select a machine learning service instance. For example, you can select Watson Machine Learning in the AI/Machine Learning category.
Select a region to deploy in. For example, if you are based in North America, you can choose Dallas.
Select a pricing plan. For example, you can choose the Lite free version and click Create.

Service instances list

Creating a deployment space

You can use a deployment space for your MLOps process for organizing all supporting assets for a deployment, such as creating and managing deployments, evaluating online deployments, and managing access for collaborators.

Follow these steps to create a deployment space:

From the navigation menu, click Deployments.
From the Deployments page, click New deployment space.
Enter a name for your deployment space and select the storage service.
Optional: Enter a description, deployment space tags, select a machine learning service instance, and upload space assets.
Click Create.

Associating a machine learning service instance with a deployment space

Associate your machine learning service instance with your deployment space to deploy assets from your space.

Prerequisites:

To associate a machine learning service instance with your space, you must:

Be a space administrator
Have admin access to the machine learning service instance

Follow these steps to associate your machine learning service instance with your deployment space:

From the Deployments page, select your deployment space and go to the Manage tab. If you don't have an existing deployment space, you must create a deployment space. For more information, see Creating a deployment space.
In the Machine learning service section, click Associate instance.
From the dropdown list, select an existing machine learning service instance or create a new machine learning service instance. For more information, see Creating a Watson Machine Learning service instance.
Click Save.

You can also create a space and associate it with a Watson Machine Learning instance by using the Python API. For more information, see the Space management sample notebook.

Modify a machine learning service instance

Follow these steps to modify the machine learning service instance associated with your deployment space:

From the Deployments page, select your deployment space and go to the Manage tab. If you don't have an existing deployment space, you must create a deployment space. For more information, refer to Creating a deployment space.
In the Machine learning service section, click the Edit () icon.
From the dropdown list, select an existing machine learning service instance or create a new machine learning service instance. For more information, see Creating a machine learning service instance.
Click Save.

Deleting a machine learning service instance

Important: Before you remove a service instance from a deployment space, delete any deployments that are associated with the instance.

Follow these steps to delete a service instance:

From the navigation menu, click Services > Service instances to access your service instances list.
Click the Open and close list of options () icon and click Delete.

Delete service

Checking your IBM Cloud IAM account access roles

Service instances are managed as part of your IBM Cloud account. When a platform administrator sets up Cloud accounts, you are assigned Cloud IAM account access roles and service-level roles.

To check your IAM account access roles and service-level access roles in IBM Cloud:

From the navigation menu, click Administration > Access (IAM).
Click Users, then click your name.
Click the Access tab.

If you want the IBM Cloud account administrator role or another role, contact the IBM Cloud account owner or administrator. You can find the list of your account administrators from the Access (IAM) > Users page.

Learn more

For details on the services available from IBM watsonx, see IBM Cloud services.

Parent topic: Deploying models with Watson Machine Learning