-
Notifications
You must be signed in to change notification settings - Fork 419
Open
Description
We are using this repo for a big project in university. We have succesfully implemented the model however it is handicapped by the fact that it can only have one label per image. In our setting we have multiple objects. We recon that making it able train on multiple object would greatly increase its performance.
So the questions is:
If you have multiple objects in an image, how can this be implemented in the traning using our current setup?
Best regards,
Christain
Metadata
Metadata
Assignees
Labels
No labels