The Web Neural Network (WebNN) API is a web-friendly, hardware-agnostic abstraction layer that allows web developers to harness the power of neural networks within the browser environment regardless of underlying hardware.

This session focuses on how to use WebNN to achieve near-native AI inference performance and power characteristics from CPU, GPU, and NPU clients through the API’s implementation in Chromium-based browsers such as Chrome* and Microsoft Edge*.

The following topics are covered in this session:

• An overview of the WebNN API unified programming model
• Using WebNN to run AI model inference on CPUs, GPUs, and NPUs and achieve near-native performance
• Loading and running ONNX* (Open Neural Network Exchange) models in web browsers with an execution provider
• How different browsers implement the API across platforms

This session includes coding for an online WebNN Hello World example.

Skill level: Intermediate
 

• WebNN Install Guide for AI PC
   

• WebNN Hello World
• WebNN GitHub*