Is there a way for inferencing more than one task with a unified checkpoint only. #14
Description
The paper suggests that the model can infer more than one task on a single image without loading weights from two different checkpoints. Is it possible with the current Codebase, if yes please guide.