Artificial intelligence (AI) is taking a new direction with the integration of WebGPU in ONNX Runtime Web, a JavaScript library that allows web developers to implement machine learning models directly in browsers. This collaboration promises an era of more efficient and powerful web applications, capable of performing AI tasks in real time without relying on external servers.
What is WebGPU?
WebGPU is a web API that unlocks the power of your device’s graphics processing unit (GPU) for high performance tasks. With advanced features like compute shader and support for half-precision (FP16), WebGPU not only reduces GPU memory and bandwidth usage, but also speeds up arithmetic calculations. This results in more efficient and scalable machine learning applications running directly in your browser.
What ONNX Runtime Web means to us
With ONNX Runtime Web, you can now enjoy faster and more powerful AI in your browser, with real-time applications such as image generation. This ensures that your data remains private and secure, as it does not need to be sent to servers for processing.
It seems that browsers are going to evolve to be able to work with AI. An additional workload that they have to deal with very well to avoid worse performance and users getting angry with this.
WebGPU is already included by default in Chrome 113 and Edge 113 for Mac, Windows, ChromeOS and Chrome 121 for Android. Be sure to check your browser compatibility and find resources for WebGPU acceleration with ONNX Runtime Web. We remind you that Edge and Chrome are already on a higher version so you should have it available.
This advance without a doubt paves the way for a future where powerful AI capabilities are available and accessible within your web browser. Of course, as we have indicated, it should not affect the performance of the team.