We are unable to access Resnet models under AI Foundry

Question

We are unable to access Resnet models under AI Foundry

Preetham Johnson Vedala 0

We created an AI Foundry resource under West US2. When searching for Resent Models under Model Catalog, we are unable to find the service. Our org is on pay-as-you-go subscription.

SRILAKSHMI C 19,715 Reputation points Microsoft External Staff Moderator

2026-07-16T08:44:42.15+00:00
Hi @Preetham Johnson Vedala

Thank you for reaching out to Microsoft Q&A.

We understand that you've created an Azure AI Foundry resource in West US 2 using a Pay-As-You-Go subscription, but you're unable to locate the ResNet models in the Model Catalog.

Even with a paid subscription, models may not appear in the Model Catalog due to regional availability, project context, access permissions, gated model access, or temporary portal synchronization issues. We recommend reviewing the following checks:

1. Verify regional availability

Model availability is region-specific and may also vary by deployment type. If the ResNet models are not currently supported in West US 2 (or only supported for specific deployment types), they may not appear in the catalog.

Please ensure that:

Your Azure AI Foundry project is created in West US 2.

You're viewing the Model Catalog for the correct project and region.

No region or deployment-type filters are limiting the displayed models.

Verify your project context

Ensure that you're signed in with the Azure account associated with the correct subscription and that you've selected the appropriate Azure AI Foundry project before browsing the Model Catalog.

Verify your permissions

Please confirm that your account has sufficient permissions on the Azure AI Foundry project.

Reader permissions are generally sufficient to browse the Model Catalog.

If you also intend to deploy models, additional roles such as Contributor or other deployment-related roles may be required.

Check for gated model access

Some models require subscription-level approval before they become visible in the Model Catalog.

Even with a Pay-As-You-Go subscription, gated models may remain hidden until access has been approved and the permissions have fully propagated.

Rule out portal caching or synchronization issues

In some cases, the Model Catalog may not refresh correctly due to browser caching or catalog synchronization delays.

We recommend:

Performing a hard refresh of the browser.

Signing out of Azure AI Foundry and signing back in.

Clearing any active filters in the Model Catalog.

Searching again for ResNet or browsing the available Vision models.

Please verify the following:

Confirm that the ResNet models are supported in West US 2 for your intended deployment type.

Verify that you're viewing the correct Azure AI Foundry project and region.

Ensure you have at least Reader access to the project.

Confirm whether the models require gated access and that your subscription has the necessary approvals.

Refresh the portal and try searching again after signing out and back in.

I Hope this helps. Do let me know if you have any further queries.

Thank you!
SRILAKSHMI C 19,715 Reputation points Microsoft External Staff Moderator

2026-07-19T00:45:14.2333333+00:00

Hi @Preetham Johnson Vedala

Did you get any chance to review the above response. Do let me know if you have any further queries.

Thank you!
Preetham Johnson Vedala 0 Reputation points

2026-07-20T04:25:48.0766667+00:00

We signed into ai.azure.com and created a project by selecting AI hub resource type. Under model catalog, we are able to see Resnet 50 but when trying to 'use the model', it is not allowing the vm selection, all the options seemed to be greyed out. We are on pay-as-you-go subscription.

1 answer

Your answer

SRILAKSHMI C 19,715 Reputation points Microsoft External Staff Moderator

2026-07-19T00:45:14.2333333+00:00

Hi @Preetham Johnson Vedala

Did you get any chance to review the above response. Do let me know if you have any further queries.

Thank you!
Preetham Johnson Vedala 0 Reputation points

2026-07-20T04:25:48.0766667+00:00

We signed into ai.azure.com and created a project by selecting AI hub resource type. Under model catalog, we are able to see Resnet 50 but when trying to 'use the model', it is not allowing the vm selection, all the options seemed to be greyed out. We are on pay-as-you-go subscription.

Answer 1

The Pay-As-You-Go subscription is probably not the blocker. A project created directly on a Microsoft Foundry resource shows models that are available for standard deployment to that resource. ResNet is a classical computer-vision architecture; it is not exposed there as an Azure-sold standard model/service in the same way as GPT, Phi, or other generative models.

Use one of these paths instead:

Train or fine-tune ResNet: create an Azure Machine Learning workspace (or a hub-based Foundry project), open Designer, add the ResNet component, choose ResNet18/34/50/etc., and connect a labeled image directory to Train PyTorch Model.
Use a pretrained ResNet: obtain a trusted PyTorch/ONNX model, register it in the Machine Learning workspace, and deploy it to a managed online endpoint with the required inference environment and scoring script.

Microsoft's Foundry documentation explains that project type controls which models appear: Foundry-resource projects show standard-deployment models, while hub-based projects expose managed-compute and broader catalog options. So West US 2 and PAYG alone do not make ResNet appear in the new catalog.

Also search using the exact spelling ResNet and clear catalog filters, but if it remains absent in a Foundry-resource project, that is expected catalog scoping.

References: Foundry model/project scope and Azure ML ResNet component.

Preetham Johnson Vedala 0 Reputation points

2026-07-17T09:47:49.4133333+00:00

We signed into ai.azure.com and created a project by selecting AI hub resource type. Under model catalog, we are able to see Resnet 50 but when trying to 'use the model', it is not allowing the vm selection, all the options seemed to be greyed out. We are on pay-as-you-go subscription.