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

Does anyone replace FL with GHM-C in FCOS ？It seems to be very low #74

Open
convolutionROC opened this issue Jun 27, 2019 · 5 comments
Open

Does anyone replace FL with GHM-C in FCOS ？It seems to be very low #74

convolutionROC opened this issue Jun 27, 2019 · 5 comments

Comments

@convolutionROC
Copy link

The mAP seems to be almost the same by replacing Focal loss with CE.
@BlueBlueFF
Copy link

same problem

@tianzhi0549
Copy link
Owner

@convolutionROC @BlueBlueFF Did CE mean the standard cross-entropy?

@BlueBlueFF
Copy link

@convolutionROC @BlueBlueFF Did CE mean the standard cross-entropy?

Yes~~

@tianzhi0549
Copy link
Owner

@BlueBlueFF Are the experimental results on COCO benchmark? On COCO, focal loss should be much better than CE.

@milliema
Copy link

May I ask how to replace the focal loss in FCOS with GHMC? I want to try that as well, but I am not familiar with the mmdetection framework used by GHM-loss. Appreciate if you could provide some tips on that.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
4 participants
@tianzhi0549 @convolutionROC @BlueBlueFF @milliema