How do I troubleshoot common model registry issues in Sagemaker AI?

Resolution

You can't register a model in the SageMaker AI Model Registry

Note: The following resolution is for a model that's deployed as SageMaker AI, has model cards created, and is associated with a SageMaker AI endpoint.

If you used a pipeline to register your model and use SageMaker AI Python SDK version 2.90.0 or later, then RegisterModel isn't supported. To resolve this issue, use ModelStep to register the model. You can also use Boto3 or SageMaker AI Studio to register your model.

Note: If you need more than 1000 models, then request a quota increase.

An "Access denied" error occurs when you create cross-account models

When you create a model across AWS accounts, you might receive the following error:

"An error occurred (AccessDeniedException) when calling the CreateModel operation: User: arn:aws:sts::123456789**:assumed-role/Role-AA** is not authorized to perform: sagemaker:CreateModel on resource: arn:aws:sagemaker:eu-west-1:123456789**:model-package/sample-model-package-group/20 because no resource-based policy allows the sagemaker:CreateModel action."

The preceding error occurs when you didn't correctly configure your resource policy. To resolve this issue, correctly configure the Amazon Simple Storage Service (Amazon S3) bucket and model package group resource policy. Also, modify the account ID, AWS Region, and model package group as needed. For more information, see Deploy a model version from a different account.

You can't view the model package group in the SageMaker AI Studio environment

Note: If you receive errors when you run AWS Command Line Interface (AWS CLI) commands, then see Troubleshooting errors for the AWS CLI. Also, make sure that you're using the most recent AWS CLI version.

If you use a domain and user profile to create a model package group, then values are shown as tags. The tags are applied to your model package group. To check the tags, run the following list-tags command:

aws sagemaker list-tags --resource-arn example-model-package-group-arn

Note: Replace example-model-package-group-arn with your model package group's ARN.

To view your model package group information, run the following describe-model-package-group command:

aws sagemaker describe-model-package-group --model-package-group-name example-model-package-group-name

Note: Replace example-model-package-group-name with your model package group name.

Note: SageMaker AI uses the sagemaker:domain-arn tag that's attached to SageMaker AI resources for resource filtration. As a result, SageMaker AI only displays resources that are created in the domain and resources that don't have the sagemaker:domain-arn tag. Also, resource filtration is supported only in the Studio UI.