As part of Azure Machine Learning (AzureML) service capabilities, Azure Arc enabled ML extends Azure ML service capabilities seamlessly to Kubernetes clusters, and enables customer to train and deploy models on Kubernetes anywhere at scale. With a simple AzureML extension deployment, customer can instantly onboard their data science team with productivity tools for full ML lifecycle, and have access to both Azure managed compute and customer managed Kubernetes anywhere. Customer is flexible to train and deploy models wherever and whenever business requires so. With built-in AzureML pipeline and MLOps support for Kubernetes, customer can scale machine learning adoption in their organization easily. Built on top of Azure Arc hybrid cloud platform, Arc enabled ML natively supports on-premise machine learning with Kubernetes.

This repository is intended to serve as an information hub for customers and partners who are interested in Azure Arc enabled ML public preview. Use this repository for onboarding and testing instructions as well as an avenue to provide feedback, issues, enhancement requests and stay up to date as the preview progresses.

1. An Azure subscription. If you don't have an Azure subscription, create a free account before you begin.
2. A Kubernetes cluster up and running - We recommend minimum of 4 vCPU cores and 8GB memory, around 2 vCPU cores and 3GB memory will be used by Azure Arc agent and AzureML extension components.
3. A kubeconfig file and context pointing to the Kubernetes cluster.
4. Install the latest release of Helm 3.
5. The Kubernetes cluster is connected to Azure Arc. For Azure Kubernetes Services (AKS) cluster, connection to Azure Arc is optional.
6. (Optional) If the Kubernetes cluster is behind of an outbound proxy server or firewall, please ensure to meet network requirements.
7. Meet the pre-requisites listed under the generic cluster extensions documentation.
   • Azure CLI version >=2.24.0
   • Azure CLI extension k8s-extension version >=1.0.0.
8. Create an AzureML workspace if you don't have one already.
   • Install and setup the latest AzureML CLI v2.

Getting started with ArcML is easy with following steps:

1. Install AzureML extension
2. Attach Kubernetes cluster to AzureML workspace and create a compute target
3. Train image classification model with AzureML CLI v2
4. Train image classification model with AzureML Python SDK 1.30 or above
5. Deploy an image classification model - create an endpoint with blue/green deployment

Using Azure Stack HCI:

• Use AKS on Azure Stack HCI for on-premises machine learning

Trouble shooting:

• TroubleShootingGuide

Azure Arc enabled ML essentially brings a new compute target to Azure Machine Learning service. With this new Kubernetes compute target, customer can use existing Azure ML tools and service capabilities to train or deploy model on Kubernetes anywhere. You can git clone available Azure ML examples on the internet, with simple change of compute target name, and those examples will run seamlessly on Kubernetes.

In addition to above built-in AzureML features, ArcML also supports following unique machine learning features:

• Target different instance typs (node types) for training or inference workload deployment

• Train model with PV/PVC storage mount
• Assign managed identity to the compute
• Use managed identity for your endpoint
• Custom container registry support
• Multiple AzureML workspaces share the same Kubernetes cluster
• Deploy model using customer container with built-in model or entry script. In this case, the model and the entry script will not be saved at the cloud, but in local.
• Interactive job to access your training compute using VS Code, Jupyter Notebook, Jupyter Lab, and summarize metrics with Tensorboard. Sign up here to get access to its Github repo.

• Azure Kubernetes Services (AKS)
• AKS Engine
• AKS on Azure Stack HCI
• Azure RedHat OpenShift Service (ARO)
• OpenShift Container Platform (OCP)
• Google GKE
• Canonical Kubernetes Distribution
• Amazon EKS
• Kind
• K3s-Lightweight Kubernetes
• Kubernetes 1.18.x, 1.19.x, 1.20.x, 1.21.x, 1.22.x

We are always looking for feedback on our current experiences and what we should work on next. If there is anything you would like us to prioritize, please feel free to suggest so via our GitHub Issue Tracker. You can submit a bug report, a feature suggestion or participate in discussions.

Or reach out to us: [email protected] if you have any questions or feedback.

• Check for known issues
• View general feedback
• Browse roadmap items
• Open a bug, provide feedback, or suggest an improvement

This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit https://cla.opensource.microsoft.com.

When you submit a pull request, a CLA bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., status check, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repos using our CLA.

This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact [email protected] with any additional questions or comments.

The lifecycle management (health, kubernetes version upgrades, security updates to nodes, scaling, etc.) of the AKS or Arc enabled Kubernetes cluster is the responsibility of the customer.

For AKS, read what is managed and what is shared responsibility here

All preview features are available on a self-service, opt-in basis and are subject to breaking design and API changes. Previews are provided "as is" and "as available," and they're excluded from the service-level agreements and limited warranty.