AI & Machine Learning Will Be Part of Your Device
Growth of AI & Machine Learning
Artificial Intelligence (AI) & Machine Learning had really became so hot topic lately. With the goal to empower every organization and people on the planet, Microsoft definitely is on the race for AI & Machine Learning.
Next Major Windows 10 Update to Support ONNX with Hardware VPU
During Windows Developer Day yesterday, Microsoft has announced the major upcoming update for Windows 10 – “Spring Creator Update” will include Windows AI Platform – the ability to run Open Neural Network Exchange (ONNX) models natively with hardware acceleration.
Wait… Hardware acceleration? Yes! Windows Machine Learning will use the hardware for any given AI/ML tasks and distributes it across multiple hardware types including CPU, GPU and Intel’s Vision Processing Units (VPU) intelligently. Today, Intel also announced their dedicated Movidius Myriad X Vision Processing Unit (VPU) for Windows Machine Learning Platform.
ONNX is an open framework for interchangeable standard for ML models that is developed by Microsoft, Facebook, and Amazon Web Services. With the ONNX format, developers can choose the right framework for their task, framework authors can focus on innovative enhancements and hardware vendors can streamline optimizations.
What’s the Excitements for Developers?
There are new AI capabilities in Windows, and Azure AI services, infrastructure and tools such as Azure Machine Learning and Visual Studio Tools for AI, developers and data scientists will be able to deliver new and exciting AI innovations faster.
WIth ONNX, Windows can then support both classic machine learning and deep learning, enabling a spectrum of AI models and scenarios. Developers can get their ONNX models to be included in their apps in several ways:
- Create and train ONNX models using any of the frameworks that support ONNX including Caffe2, Chainer, PyTorch, and Microsoft Cognitive Toolkit (CNTK).
- Convert models from various sources including SciKit-Learn and CoreML into ONNX models by using ONNXML tools.
- Grab pre-trained ONNX models from the growing ONNX model zoo.
Windows ML API
Microsoft also had announced the Windows Machine Learning API for Win32 & WinRT Layers. It has the capabilities to convert images to Tensor resources. Therefore, developers are able to develop machine-learning models, trained in Azure, and import it to their apps using Visual Studio and run them on their PCs. Developers will be able to add an ONNX file to a UWP app and generate model interfaces in the project with Visual Studio Preview 15.7.
I am quite excited to see how it will continue to develop, especially AI & ML has been around for Microsoft to support the community through Cortana, Cognitive Services, Bot Service, Azure Machine Learning and CNTK. Let’s see these will boost the development of AI & ML in tech industry!