How to attach a Microsoft.HybridCompute/machines to ML studio as compute instance

Question

I have onboarded a local machine to the Azure resources group using the Azure Arc solution. However, I am facing issues when attaching it as a compute to ML Studio.

It is not accepting Microsoft.HybridCompute key on Resource ID. In Notes, it is said as VMs only. But wanna know if there is any option to perform the action,

Answer 1

If your local machine registered via Azure Arc meets the requirements for Azure ML supported compute instances, try to ensure that it's registered as a compatible VM type rather than directly as a Microsoft.HybridCompute/machines resource.

Another option is to use an Azure Arc-enabled Kubernetes cluster, which can serve as a compute target for Azure ML. This approach is more suited if you are working with containerized workloads and want to use local resources for training.

If there is a need to use the local machine as a compute target, a custom script may be required to create the connection between Azure ML Studio and the machine through a more manual process. This isn't natively supported but can involve using Azure CLI or REST APIs to manage compute connections.