Errors While Creating SageMaker Domain and Issues with Unified Studio Project Profile

0

Hello,

I'm encountering multiple issues while trying to create a SageMaker domain and use the Unified Studio environment.

Domain Creation Errors When creating the domain, I received the following validation error related to model modalities:

ValidationException: 2 validation errors detected:
Value '[SPEECH]' at 'generativeAIGovernance.models.10.member.inputModalities' failed to satisfy constraint: [Member must satisfy enum value set: [TEXT, EMBEDDING, VIDEO, IMAGE]];
Value '[SPEECH, TEXT]' at 'generativeAIGovernance.models.10.member.outputModalities' failed to satisfy constraint: [Member must satisfy enum value set: [TEXT, EMBEDDING, VIDEO, IMAGE]] I didn't manually configure any models with SPEECH modality, so I'm unsure where this is being set.

Additionally, I encountered duplicate policy errors:

EntityAlreadyExistsException: A policy called AmazonSageMaker Bedrock Model Consumption Policy-fe260cc8791040 already exists. Duplicate names are not allowed.
EntityAlreadyExistsException: A policy called AmazonSageMaker Bedrock Model Consumption Policy-6a0def8acdcffa already exists. Duplicate names are not allowed.

How should I resolve these issues? Is there a way to reset or clean up previous failed attempts?

Unified Studio – "No Data to Show" in Project Profile After the domain was successfully created and I opened Unified Studio, I attempted to create a new project, but the "Project Profile" section shows No data to show.

Why is this happening, and how can I populate the available project profiles? Any help would be appreciated. Thank you!

below things are raw text of error:

There was an error while provisioning your domain ValidationException: 2 validation errors detected: Value '[SPEECH]' at 'generativeAIGovernance.models.10.member.inputModalities' failed to satisfy constraint: Member must satisfy constraint: [Member must satisfy enum value set: [TEXT, EMBEDDING, VIDEO, IMAGE]]; Value '[SPEECH, TEXT]' at 'generativeAIGovernance.models.10.member.outputModalities' failed to satisfy constraint: Member must satisfy constraint: [Member must satisfy enum value set: [TEXT, EMBEDDING, VIDEO, IMAGE]]

Error while creating role EntityAlreadyExistsException: A policy called AmazonSageMaker Bedrock Model Consumption Policy-fe260cc8791040 already exists. Duplicate names are not allowed.

Error while creating role EntityAlreadyExistsException: A policy called AmazonSageMaker Bedrock Model Consumption Policy-6a0def8acdcffa already exists. Duplicate names are not allowed.

asked 24 days ago73 views
1 Answer
1

It seems you're facing multiple issues with your SageMaker domain creation and Unified Studio setup. Let's address these problems one by one:

  1. ValidationException for generative AI models: The error suggests there's a mismatch in the input and output modalities configured for one of the generative AI models. The system is detecting 'SPEECH' as an input/output modality, which is not supported. This could be due to an incorrect configuration in your domain settings. To resolve this:
  • Review your domain configuration, particularly the generative AI settings.
  • Ensure that only supported modalities (TEXT, EMBEDDING, VIDEO, IMAGE) are specified for input and output.
  • If you didn't manually configure these settings, there might be a default configuration causing this issue. Try creating a new domain with minimal configurations to see if the problem persists.
  1. EntityAlreadyExistsException for Bedrock Model Consumption Policy: This error indicates that policies with specific names already exist in your account. This often happens when previous domain creation attempts have partially succeeded or failed. To resolve:
  • Check your IAM console for these policies and delete them if they're not associated with any active resources.
  • If you can't delete them, try using a different name for your domain or contact AWS support to help clean up these resources.
  1. "No data to show" in Project Profile: This issue could be related to incomplete domain setup or permissions problems. To troubleshoot:
  • Verify that your domain is in the "InService" status.
  • Check that the necessary IAM roles and permissions are correctly set up for your domain and user profile.
  • Ensure your AWS account is properly configured to use SageMaker projects in your specific region.
  • Review the domain details for any failed updates or configurations.

If these steps don't resolve your issues, you might need to:

  • Delete the partially created domain and start fresh.
  • Double-check all configurations and permissions.
  • Consider using the Quick setup option in SageMaker Unified Studio to create a new domain with default settings.
  • If problems persist, reach out to AWS Support for further assistance, as there might be account-specific issues or regional limitations affecting your setup.

Remember to review and adjust your IAM roles and policies to ensure they have the appropriate permissions for creating and managing SageMaker domains and projects.

Sources
SageMaker Domain Setup Fails: "No data to show" in Project Profile & ValidationException Error | AWS re:Post
Troubleshooting in Amazon SageMaker Unified Studio - Amazon SageMaker Unified Studio
Create a Amazon SageMaker Unified Studio domain - quick setup - Amazon SageMaker Unified Studio
An integrated experience for all your data and AI with Amazon SageMaker Unified Studio | AWS Big Data Blog

profile picture
answered 24 days ago
profile pictureAWS
EXPERT
reviewed 22 days ago

    Bedrock model access - 42 approved models - Grant model access

    how remove it?

    1. Is there a way to fully reset my account’s SageMaker and Bedrock-related configurations and IAM resources to a clean state? Alternatively, is there a recommended method to remove all auto-created roles, policies, and domain metadata so I can start fresh?

    Any guidance or support would be appreciated. Thank you.

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Guidelines for Answering Questions