unable to access GPU resources on LLM_Sandbox_number1

Question

While trying to setup an Azure Machine Learning Workspace, I was unable to assign GPUs as a part of the compute. I received access to only CPUs. However, I have been using GPUs in other parts of my application pipeline. Is there anyway to add these resources to Azure ML workspace?

Answer 1

Hi Ken Rybarczyk,

You are unable to access GPU resources in your Azure Machine Learning workspace because your subscription likely does not have GPU quotas enabled, or the sandbox environment you are using does not support GPU-based compute. By default, many Azure subscriptions, especially sandbox, educational, or trial accounts, have zero quota for GPU virtual machines (such as NC, ND, or NV series) and may also have policy restrictions preventing GPU usage. To add GPU resources, you need to check the quota for your subscription in the Azure Portal under “Usage + quotas” and, if the quota is zero, submit a quota increase request specifying the GPU VM family and region.