What is Windows* ML
Windows* ML is Microsoft's unified local AI inferencing framework that allows developers to run machine learning models directly on Windows devices. Powered by ONNX Runtime, it automatically accelerates AI workloads across CPUs, GPUs, and NPUs while keeping user data private and reducing operational costs.
For more information, please refer to: What is Windows ML?
System Requirements
| Device Acceleration | System Generation | System Memory |
|---|---|---|
| CPU | 11th Generation Intel® Core™ processors (formerly codenamed Tiger Lake) or newer | Minimum 8GB |
| GPU | 12th Generation Intel Core processors (formerly codenamed Alder Lake) or newer with latest driver | 16 GB recommended |
| NPU | Intel Core Ultra Series 1 processors (formerly codenamed Meteor Lake) or newer with latest driver | 16 GB recommended |
| Discrete GPU | Intel® Arc™ A310 (formerly codenamed Alchemist) or newer | 8 GB recommended |
| Discrete GPU | Intel Arc B580 (formerly codenamed Alchemist) or newer | 10 GB recommended |
Automated Driver Installation
Please install the latest GPU and NPU drivers using Intel Driver Support Assistant (IDSA) that can be downloaded at the following link: Intel® Driver & Support Assistant (Intel® DSA)
Support
Please request support by creating GIT issues at: OpenVINO All Issues
Microsoft Foundry* Samples
Explore comprehensive samples demonstrating each of the three Windows AI Foundry pillars that use Windows ML as the underlying machine learning runtime:
- Built-in Models: Leverage pre-integrated Windows models
- Ready-to-use Open Source Models: Deploy popular open-source AI models via Foundry Local
- Bring Your Own Model: Import and optimize custom models and use Windows ML inferencing Api’s
Samples Repo Link
Samples for Microsoft AI Foundry
Featured Reference Application: Microsoft Foundry Hybrid AI App
This application demonstrates the complete Microsoft Foundry* on Windows stack through a real-world technician assistant use case. It showcases how developers can integrate all three pillars of Microsoft Foundry on Windows: Windows AI APIs, Foundry Local, and Windows ML Inferencing APIs in a single, cohesive application with escalation to cloud if needed.