Skip to content
New issue

Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? # to your account

How to use the swin-transformer as the backbone for DAB-DETR #23

Open
yuan738 opened this issue Apr 25, 2022 · 5 comments
Open

How to use the swin-transformer as the backbone for DAB-DETR #23

yuan738 opened this issue Apr 25, 2022 · 5 comments

Comments

@yuan738
Copy link

yuan738 commented Apr 25, 2022

Thank you for help!

@yuan738 yuan738 changed the title How to use the swim-transformer as the backbone for DAB-DETR How to use the swin-transformer as the backbone for DAB-DETR Apr 26, 2022
@yuan738
Copy link
Author

yuan738 commented Apr 28, 2022

When I use swin-transformer as the backbone of the model, the error is ValueError: return_layers are not present in model.

So how can I modify the code of return_layers in backbone.py?

Thank you.

@SaifeiYan
Copy link

When I use swin-transformer as the backbone of the model, the error is ValueError: return_layers are not present in model.

So how can I modify the code of return_layers in backbone.py?

Thank you.
do you solve this problem,can you help me

@yuan738
Copy link
Author

yuan738 commented Jun 13, 2022

sry, I didn't solve this problem.

@fernandorovai
Copy link

I'm trying to fix this one as well

@rentainhe
Copy link
Collaborator

rentainhe commented Sep 28, 2022

Thank you for help!

Hi~ We've released our new codebase for DETR-like models called detrex, we've already supported Swin-Backbone for DAB-DETR in detrex now. And we've released new baselines with better mAP for DAB-DETR in detrex.

By using detrex, it's easier for you to use other backbones instead of Swin, ResNet, etc. @SaifeiYan @yuan738 @fernandorovai

# for free to join this conversation on GitHub. Already have an account? # to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants