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.

Sign up for GitHub

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

Jump to bottom

WebGPU support for Vision Tasks #5826

Open
pgn-dev opened this issue Jan 15, 2025 · 2 comments
Open

WebGPU support for Vision Tasks #5826

pgn-dev opened this issue Jan 15, 2025 · 2 comments
Assignees
@schmidt-sebastian
Labels
platform:javascript MediaPipe Javascript issues stat:awaiting googler Waiting for Google Engineer's Response task:face landmarker Issues related to Face Landmarker: Identify facial features for visual effects and avatars. task:hand landmarker Issues related to hand landmarker: Identify and track hands and fingers task:pose landmarker Issues related to Pose Landmarker: Find people and body positions type:feature Enhancement in the New Functionality or Request for a New Solution

Comments

@pgn-dev
Copy link

pgn-dev commented Jan 15, 2025

MediaPipe Solution (you are using)

FaceLandmarker, HandLandmarker, PoseLandmarker

Programming language

typescript

Are you willing to contribute it

None

Describe the feature and the current behaviour/state

Use WebGPU for running vision tasks. Currently the TaskRunners appears to be running WebGL with OffScreenCanvas

Will this change the current API? How?

No response

Who will benefit with this feature?

Everyone

Please specify the use cases for this feature

Should make prediction on browser-based solutions faster.

Any Other info

No response
@pgn-dev pgn-dev added the type:feature Enhancement in the New Functionality or Request for a New Solution label Jan 15, 2025
@kuaashish kuaashish assigned kuaashish and unassigned kalyan2789g Jan 15, 2025
@kuaashish kuaashish added task:face landmarker Issues related to Face Landmarker: Identify facial features for visual effects and avatars. task:hand landmarker Issues related to hand landmarker: Identify and track hands and fingers task:pose landmarker Issues related to Pose Landmarker: Find people and body positions platform:javascript MediaPipe Javascript issues labels Jan 15, 2025
@kuaashish kuaashish added the stat:awaiting googler Waiting for Google Engineer's Response label Jan 16, 2025
@schmidt-sebastian
Copy link
Collaborator

Our Vision tasks do use WebGL at the moment, as our WebGPU support is currently focused on our LLM Inference API. Please do watch our announcement over the coming weeks and months in this space.

@pgn-dev
Copy link
Author

pgn-dev commented Jan 16, 2025

Thanks for confirming @schmidt-sebastian

Do you expect a significant improvement in latency over WebGL with WebGPU?

TFLite already seems to support WebGPU. Would running the TFLite models from the task files work out of the box?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@schmidt-sebastian schmidt-sebastian

Labels
platform:javascript MediaPipe Javascript issues stat:awaiting googler Waiting for Google Engineer's Response task:face landmarker Issues related to Face Landmarker: Identify facial features for visual effects and avatars. task:hand landmarker Issues related to hand landmarker: Identify and track hands and fingers task:pose landmarker Issues related to Pose Landmarker: Find people and body positions type:feature Enhancement in the New Functionality or Request for a New Solution
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
4 participants
@pgn-dev @schmidt-sebastian @kuaashish @kalyan2789g