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

Add fp16 unit tests to the CUDA ops when pytorch >= 1.6 #1028

Closed

AronLin opened this issue May 17, 2021 · 1 comment

Contributor

AronLin commented May 17, 2021 •

edited

Loading

Describe the Issue

I notice that there was a runtime error #1004 and it was fixed in mmcv 1.3.4.
However, the same problem still exists when using DCNv2.

So I suggest to check the support of CUDA ops for fp16 or add unit tests to CUDA ops to support fp16，when pytorch >= 1.6.

Environment
pytorch >= 1.6.0
mmcv >= 1.3.2

I will send a PR to fix the bug when using DCNv2 with the same way in PR #1014

AronLin mentioned this issue

[Fix] Support amp (pytorch >= 1.6.0) on DCN and DCNv2/ Add unit tests on DCN/DCNv2 amp #1029

Merged

Collaborator

zhouzaida commented May 17, 2021

thanks

zhouzaida closed this as completed

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