microsoft/Windows-Machine-Learning
🔧 Toolmicrosoft
Samples and tools for deploying AI models on Windows devices using Windows ML.
Windows-Machine-Learning (WinML) is a platform for deploying trained machine learning models on Windows devices. This repository offers a collection of C++ samples demonstrating how to load, bind, and evaluate models using the Windows ML API. It also includes command-line tools and scripts for model conversion and optimization, supporting formats like ONNX, TensorFlow, PyTorch, Caffe2, CoreML, Keras, and scikit-learn. Key features include hardware acceleration via DirectML, support for custom operators, and integration with Windows app development. The samples cover image classification, object detection, and natural language processing scenarios. The tools, such as WinMLTools, help convert models from various frameworks to ONNX format. This repository serves as a starting point for developers to leverage AI on Windows without needing deep expertise in low-level GPU programming.
💡Highlights
- ├─C++ samples for Windows ML API
- ├─Supports ONNX, TensorFlow, PyTorch
- └─Hardware acceleration via DirectML
🎯For
- ├─Windows developers
- ├─Machine learning engineers
- └─AI application developers