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

Why the code is taking much time on GPU than CPU? #3

Open
zobeirraisi opened this issue Mar 5, 2021 · 2 comments
Open

Why the code is taking much time on GPU than CPU? #3

zobeirraisi opened this issue Mar 5, 2021 · 2 comments

Comments

@zobeirraisi
Copy link

zobeirraisi commented Mar 5, 2021

I run the code on my local GPU and CPU:
for CPU it takes 18 sec
but for GPU it takes 59 sec?
How it can be solved?

@zobeirraisi zobeirraisi changed the title Why the code it taking much time on GPU than CPU? Why the code is taking much time on GPU than CPU? Mar 5, 2021
@acheman
Copy link

acheman commented Apr 6, 2023

I run the code on my local GPU and CPU: for CPU it takes 18 sec but for GPU it takes 59 sec? How it can be solved?

Hello, I wonder if this code base is used to convert DETR into a detection box that can detect rotation? I modified it in the DETR code according to this code, what else do I need to modify so that I can detect the rotation detection frame (xc, yc, w, h, angle), thank you very much.

@craymichael
Copy link

This is in part due to all of the .to(device) calls in box_ops.py. They can be changed to e.g. torch.zeros(..., device=device) calls for each reference respectively

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

No branches or pull requests

3 participants