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.

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

How to train unregular box? #294

Open
Joyako opened this issue Aug 7, 2023 · 5 comments
Open

How to train unregular box? #294

Joyako opened this issue Aug 7, 2023 · 5 comments

Comments

@Joyako
Copy link

Joyako commented Aug 7, 2023

Hi, my data set labels are unregular, where to modify the data preprocessing code?
The format of the label is like this:

截屏2023-08-07 上午10 40 43
@rentainhe
Copy link
Collaborator

rentainhe commented Aug 7, 2023

You can try to register your own data following detectron2's official documentation here: https://detectron2.readthedocs.io/en/latest/tutorials/datasets.html

Then you can try to update the data config to load your own dataset.

@Joyako
Copy link
Author

Joyako commented Aug 7, 2023

@rentainhe thank you for your reply, does data augmentation need to be modified?

@rentainhe
Copy link
Collaborator

@rentainhe thank you for your reply, does data augmentation need to be modified?

Yes, we use detectron2's default data augmentation, but it only works on the normal box, you should double-check this in detectron2's source code.

@Joyako
Copy link
Author

Joyako commented Aug 7, 2023

@rentainhe thank you for your reply, does data augmentation need to be modified?

Yes, we use detectron2's default data augmentation, but it only works on the normal box, you should double-check this in detectron2's source code.

thanks, I got it

@rentainhe
Copy link
Collaborator

@rentainhe thank you for your reply, does data augmentation need to be modified?

Yes, we use detectron2's default data augmentation, but it only works on the normal box, you should double-check this in detectron2's source code.

thanks, I got it

You're welcome, detectron2's documentation has detailed tutorial on how to register your own dataset and augmentations, it may be helpful

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

No branches or pull requests

2 participants