-
Notifications
You must be signed in to change notification settings - Fork 27.5k
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
Add ViTDet #25524
Add ViTDet #25524
Conversation
The documentation is not available anymore as the PR was closed or merged. |
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 model!
Very nice and clean PR :) Mostly just nits. The diff on the toctree will need to be resolved before merging.
@amyeroberts thanks for your review, I've addressed all the comments |
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 again for adding and iterating!
There's still an issue with the toctree diff which needs to be addressed
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.
Reminder to resolve this. Once that's done we can merge in the PR
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.
cc @ydshieh who knows the fix for this
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.
If you discard the change in this file of this PR, rebase on main, adding back ViTDet
entry to this file and run the style, it will be fine.
A fix #25661 is merged into main
.
* First draft * Fix READMEs * Update return_dict * Add more tests * Fix docstrings * Address comments * Address more comments * Address more comments * Address more comments, fix test * Fix test
* First draft * Fix READMEs * Update return_dict * Add more tests * Fix docstrings * Address comments * Address more comments * Address more comments * Address more comments, fix test * Fix test
* First draft * Fix READMEs * Update return_dict * Add more tests * Fix docstrings * Address comments * Address more comments * Address more comments * Address more comments, fix test * Fix test
What does this PR do?
This PR adds part 1 of #25051, namely the ViTDet backbone, introduced in Exploring Plain Vision Transformer Backbones for Object Detection.
Note that this PR only adds the backbone, hence there are no compatible checkpoints with the backbone-only. Those can only be added once either VitMatte or Mask R-CNN are added, both of which use VitDet as backbone.