Skip to content

Conversation

@NingW101
Copy link
Contributor

@NingW101 NingW101 commented Mar 10, 2025

@ibelem
Copy link
Collaborator

ibelem commented Mar 11, 2025

Thanks @NingW101 , will not merge this PR until we support these NPU demos.

Copy link
Member

@anssiko anssiko left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'll approve this to be merged when the NPU demos are functional, but please be prepared to do another update to align with the new design once ready:

MLDeviceType was removed in PR webmachinelearning/webnn#809 and a new query mechanism is being defined in webmachinelearning/webnn#815

This suggests further changes to the urlArgs passing and other code paths that use device types is needed when that new design is finalized and reflected in the implementation.

@ibelem
Copy link
Collaborator

ibelem commented Mar 11, 2025

Thanks @anssiko for the reminder. Yes, we are still using MLDeviceType in demos, samples, WebNN EP of ONNX Runtime Web, will update the test framework code when we use new design.

@anssiko
Copy link
Member

anssiko commented Mar 11, 2025

@ibelem feel free to share your use cases in webmachinelearning/webnn#815 from the test framework perspective. I understand more explicit control over device types would be very helpful for this framework. Perhaps a specific runtime flag could allow more control over device types for testing purposes.

@zhibo1-wang
Copy link
Contributor

@ibelem This PR can be merged now.

@ibelem
Copy link
Collaborator

ibelem commented Aug 6, 2025

Thanks @zhibo1-wang, merge it.

@ibelem ibelem merged commit 23ea418 into webmachinelearning:main Aug 6, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants