-
Notifications
You must be signed in to change notification settings - Fork 27.2k
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
DeviceGuard added to use Deformable Attention more safely on multi-GPU #32910
Conversation
…tr.py Co-authored-by: amyeroberts <[email protected]>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for adding this!
@DonggeunYu We had an issue with our test runners. Could you push an empty commit to trigger a re-run and make everything green? |
It has been completed! |
huggingface#32910) * Update modeling_deformable_detr.py * Update src/transformers/models/deformable_detr/modeling_deformable_detr.py Co-authored-by: amyeroberts <[email protected]> * Update ms_deform_attn_cuda.cu * Update modeling_deformable_detr.py * Update modeling_deformable_detr.py * [empty] this is a empty commit --------- Co-authored-by: amyeroberts <[email protected]>
huggingface#32910) * Update modeling_deformable_detr.py * Update src/transformers/models/deformable_detr/modeling_deformable_detr.py Co-authored-by: amyeroberts <[email protected]> * Update ms_deform_attn_cuda.cu * Update modeling_deformable_detr.py * Update modeling_deformable_detr.py * [empty] this is a empty commit --------- Co-authored-by: amyeroberts <[email protected]>
huggingface#32910) * Update modeling_deformable_detr.py * Update src/transformers/models/deformable_detr/modeling_deformable_detr.py Co-authored-by: amyeroberts <[email protected]> * Update ms_deform_attn_cuda.cu * Update modeling_deformable_detr.py * Update modeling_deformable_detr.py * [empty] this is a empty commit --------- Co-authored-by: amyeroberts <[email protected]>
What does this PR do?
I got zero results from deformable attention when multi-GPU inference.
However, this issue happens only in my environment.
The DeviceGuard has been added to ensure that calculations are performed only on the value device. Thus, my problem has been resolved.
Fixes # (issue)
Before submitting
Pull Request section?
to it if that's the case.
documentation guidelines, and
here are tips on formatting docstrings.
Who can review?
Models: