Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Supported operations query mechanism #504

Closed
ibelem opened this issue Jan 12, 2024 · 2 comments
Closed

Supported operations query mechanism #504

ibelem opened this issue Jan 12, 2024 · 2 comments

Comments

@ibelem
Copy link
Contributor

ibelem commented Jan 12, 2024

We have seen ongoing WebNN implementation on Windows and ChromeOS, with WebNN prototype development underway for Android and macOS. In addition to computation on CPU and GPU devices, we have also seen leading hardware manufacturers increase their support for NPUs.

Ideally, all APIs and operations defined by the WebNN specification would be fully supported on WebNN-enabled operating systems, devices. However, due to differences in software stacks and hardware, it is inevitable that some operations cannot be fully supported and require fallback via JavaScript machine learning frameworks or web applications.

Given this, web developers want to know the current device and operating system, such as WebNN GPU on Windows or WebNN CPU on Android, and which operations are supported. It should be nice to have the query mechanism in the WebNN specification to allow JavaScript code to query which operations are supported by which operating system, device and and power preference after getting the MLContext object.

@ibelem
Copy link
Contributor Author

ibelem commented Feb 5, 2024

Based on https://issues.chromium.org/issues/41496343, detecting data types of WebNN is the other feature requirement besides the ops.

@inexorabletash
Copy link
Member

This seems to be a duplicate of #463 (including both ops and data types) so I'm going to close it, but if I missed any distinction please comment and we can re-open. Or add to the other issue.

@inexorabletash inexorabletash closed this as not planned Won't fix, can't repro, duplicate, stale Apr 9, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

2 participants