ONNX Runtime WebGPU Demos

Welcome to the ONNX Runtime WebGPU demo repository! This repository contains various examples showcasing the capabilities of ONNX Runtime with WebGPU.

Contents

• Stable Diffusion Turbo
• Segment Anything
• Phi3-mini
• TinyLlama
• YoloV9

Overview of ONNX Runtime

ONNX Runtime (ORT) is a cross-platform, high-performance machine learning model accelerator. It provides a flexible interface to integrate hardware-specific libraries, making it compatible with models from various frameworks like PyTorch, TensorFlow/Keras, TFLite, and scikit-learn. ORT is designed to improve inference performance for a wide variety of machine learning models and can run on different hardware and operating systems. It's widely used in Microsoft products and services, such as Office, Azure, and Bing.

Overview of WebGPU

WebGPU is a web graphics API that allows web applications to efficiently use the GPU (Graphics Processing Unit) of a device for graphics rendering and computation. It is the successor to WebGL and provides better compatibility with modern GPUs, support for general-purpose GPU computations, and access to more advanced GPU features. WebGPU is designed with the web platform in mind, featuring an idiomatic JavaScript API, integration with promises, and support for importing videos. It aims to provide more powerful, faster, and safer graphics performance than WebGL.

How to Test

To test these demos, please visit the ONNX Runtime WebGPU demo page.

Enjoy exploring the power of ONNX Runtime with WebGPU!