Create a machine learning model for offline use on a Windows device


Experience the benefits of local client-side evaluation with the intelligent edge. Download a pretrained image classification model and add it to a UWP app. By using Windows Machine Learning (ML), models can run evaluations offline on any Windows 10 device.

Learning objectives

In this module, you take the following steps:

  • Compare using a machine learning model locally with Windows ML to a cloud-based system.
  • Set up a simple UWP app with C#.
  • Add a pretrained machine learning model to your app.
  • Load the model, bind the inputs and outputs, and use it to evaluate input.
  • Test the model. Draw digits into a Windows InkCanvas to see whether the model recognizes what number we drew.


  • A basic understanding of UWP app development with C#.