SageMaker Studio UI fails to deploy BYO model: image URI & model artifact fields accepted but deployment stalls and/or errors without actionable feedback

### Product Version

- [ ] Amazon SageMaker Studio Classic
- [x] Amazon SageMaker Studio
- [ ] Issue is not related to SageMaker Studio

### Issue Description

When using SageMaker Studio (UI) to deploy a Bring Your Own (BYO) model, the UI accepts a valid inference image URI and model artifacts S3 path but:

The deploy endpoint step stalls and eventually fails with a generic error in a manually created CloudWatch log group, without surfacing the underlying error from SageMaker. This prevented us from diagnosing configuration issues directly from the Studio UI and contradicts expected behavior where validation errors and service exceptions should be clearly surfaced.

The only way our model was able to be deployed was through a notebook instance, where we received more clarity by way of error messages and logs and were able to correct the issues and deploy the model. However, our model is only able to be deployed through a notebook and we are still unable to deploy through the Studio UI, even after working with both AWS SageMaker SME's and engineers.

### Expected Behavior

We were instructed by AWS SageMaker SME's to use the SageMaker Studio UI to deploy our pre-trained model to a real-time endpoint. The expected behavior is such that the configuration options available in the Studio UI are sufficient to package, deploy, and register a pre-trained model, and deploy the model to a real-time endpoint.

### Observed Behavior

When we got to the endpoint deployment step, the endpoint consistently failed with generic errors that were insufficient for troubleshooting.

### Product Category

Models

### Feedback Category

User Interface

### Other Details

_No response_

No repository files, tests, or implementation entry points are identified in the report. Start by reproducing the BYO model deployment in SageMaker Studio using the valid image URI and S3 artifact path, then compare the Studio UI failure with the notebook deployment; done means the underlying validation or service error is surfaced instead of a generic failure.

Written by the indexing model from the issue text.