I have a task to detect humans and cars separately. My dataset is annotated separately for humans and cars, meaning some images are only annotated with humans, while others are only annotated with cars. I want to use different detection heads and different loss functions for training them separately.
how can i do that?