Add DETR object detection head. #176
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Henry suggested trying it with other backbone (not Helios). |
|
I should also try longer training run with Helios+DETR with dropout enabled, most recently I tried disabling it but maybe it works better enabled (previously I was seeing very different train/eval results so that's why I disabled it). |
|
I will mark this as draft until we have a chance to debug why DETR code isn't working for training on COCO dataset with current version of the various libraries like pytorch and such. |
|
Unmarking as draft, I think we should just merge it in for now since it's already used by some configs in rslearn_projects I think, and then can continue debugging later. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I haven't been able to train the head successfully but it seems to be correct? I guess I should try it on COCO.
Most of the code is from here:
https://github.com/facebookresearch/detr
I wasn't able to use it as a library because their repository is not structured as a module. Also it was too tightly connected to the backbone.