Unable to Test Model in Azure ML Studio - No Dataset Listed + Quota Errors Despite No Usage

Question

Hi team,

I'm currently following Microsoft's Azure Machine Learning Studio learning path using my personal Microsoft account and encountered multiple blocking issues during the "Test model (preview)" step after training and registering a model with AutoML.

Issue Summary:

After training a model and registering it successfully, when trying to use the "Test model (preview)" feature:

• The "Select a dataset" dropdown shows no available datasets, despite having two properly registered MLTable datasets in the workspace (see attached screenshots).

Datasets are visible and previewable in the "Data" pane, but do not show in the testing interface.

Environment:

Account: Personal Microsoft account (evandrocruzhotmail.onmicrosoft.com)

Subscription: "Azure subscription 1"

Workspace: GPGAz

Region: Australia Central

Additional Blockers:

Quota Limitation: My subscription has 0 usage but still shows no available quota for standard compute SKUs (e.g., Standard DSv2 Family). Quota request is not eligible, and test clusters fail due to:

Ineligible for quota adjustment.

  Compute provisioning fails with: **`ResourceOperationFailure: Resource provider [N/A] isn't registered with Subscription [N/A]`** (Even after confirming that `Microsoft.MachineLearningServices`, `Microsoft.Storage`, and `Microsoft.Compute` are registered.)

  
  **Testing Blocked:** Even after model registration and cluster setup, I can't complete the tutorial because the platform doesn't allow me to assign a dataset for testing.
```Screenshots:

Dataset registered and visible

Quota dashboard showing 0 usage and ineligibility

Compute failure error

"Select a dataset" dropdown with nothing listed

Request:

I’m trying to complete Microsoft’s own certification prep steps, but the experience has been very frustrating due to these limitations that block core functionality like dataset access and compute provisioning.

Is there a workaround or configuration that can help me unlock dataset selection in model testing?

Is there any way to escalate or enable minimal quota needed to complete this learning path?

Any help or official guidance would be greatly appreciated!

Thank you,Hi team,

I'm currently following Microsoft's Azure Machine Learning Studio learning path using my personal Microsoft account and encountered multiple blocking issues during the **"Test model (preview)"** step after training and registering a model with AutoML.

Issue Summary:

After training a model and registering it successfully, when trying to use the **"Test model (preview)"** feature:

The "Select a dataset" dropdown shows **no available datasets**, despite having two properly registered MLTable datasets in the workspace (see attached screenshots).

Datasets are visible and previewable in the "Data" pane, but do not show in the testing interface.

Environment:

Account: Personal Microsoft account (evandrocruzhotmail.onmicrosoft.com)

Subscription: "Azure subscription 1"

Workspace: GPGAz

Region: Australia Central

Additional Blockers:

**Quota Limitation:** My subscription has 0 usage but still shows no available quota for standard compute SKUs (e.g., Standard DSv2 Family). Quota request is **not eligible**, and test clusters fail due to:

   `Ineligible for quota adjustment.`
   
```yaml
  Compute provisioning fails with:  
   **`ResourceOperationFailure: Resource provider [N/A] isn't registered with Subscription [N/A]`**  
   (Even after confirming that `Microsoft.MachineLearningServices`, `Microsoft.Storage`, and `Microsoft.Compute` are registered.)

  
  **Testing Blocked:** Even after model registration and cluster setup, I can't complete the tutorial because the platform doesn't allow me to assign a dataset for testing.
```Screenshots:

Dataset registered and visible

Quota dashboard showing 0 usage and ineligibility

Compute failure error

"Select a dataset" dropdown with nothing listed

Request:

I’m trying to complete Microsoft’s own certification prep steps, but the experience has been very frustrating due to these limitations that block core functionality like dataset access and compute provisioning.

Is there a workaround or configuration that can help me unlock dataset selection in model testing?

Is there any way to escalate or enable minimal quota needed to complete this learning path?

Any help or official guidance would be greatly appreciated!

Thank you,

Answer 1

Hi Evandro Cruz,

If you frequently encounter these limits and need more capacity for your projects, you might consider upgrading your subscription from Free Trial to a Pay-As-You-Go plan. This transition can help remove many of the restrictions you are currently facing.

Here are a few additional suggestions that may help you resolve the issues:

1. Understanding Your Quotas: The Free Trial subscription comes with specific resource limits, including quotas for compute instances and other services. You can check your current quotas by navigating to the Azure Machine Learning Studio and reviewing the management settings.
2. Requesting a Quota Increase: If you find that you require additional resources beyond your existing quota, you can submit a request for an increase. To do this, please follow these steps:
   • Visit the Azure Management portal.
   • Navigate to Management > Quota.
     • Use the "Request Quota" option to submit your request. Please note that requests are processed in the order they are received.
3. Default Limits: It’s also important to be aware that default limits may vary based on the specific resources you are trying to utilize. I recommend consulting the Azure Machine Learning quotas documentation for detailed information.

Thank you for your understanding