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

Reproduced CenterNet-S4_ DLA_ 8x with RTX3090, but the FPS is only 26.4, far from 71 #83

Open
FH-JAVA opened this issue May 4, 2022 · 3 comments
Labels
documentation Improvements or additions to documentation

Comments

@FH-JAVA
Copy link

FH-JAVA commented May 4, 2022

image

@FH-JAVA FH-JAVA added the documentation Improvements or additions to documentation label May 4, 2022
@xingyizhou
Copy link
Owner

Can you show your pytorch/ cuda/ cudnn/ detectron2 version? You can find them on the very beginning of the output log. The output log should also show the runtime breakdown if you use a recent version of detectron2, can you report them as well?

Best,
Xingyi

@FH-JAVA
Copy link
Author

FH-JAVA commented May 5, 2022

Can you show your pytorch/ cuda/ cudnn/ detectron2 version? You can find them on the very beginning of the output log. The output log should also show the runtime breakdown if you use a recent version of detectron2, can you report them as well?

Best, Xingyi
Thank you for your reply.
The previously rented GPU was released, so I rented another rtx3090 and reproduced CenterNet-S4_ DLA_ 8x, but FPS is now 14.4, lower than before.
image
the log is available here.

@MoaazAbdulrahman
Copy link

@xingyizhou
@FH-JAVA
I am facing exactly the same issue on RTX 3090 the inference time is way far from the reported numbers.
Were you able to solve this issue?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
documentation Improvements or additions to documentation
Projects
None yet
Development

No branches or pull requests

3 participants