Closed Dwrety closed 2 years ago
Yes, there is some improvement, not sure significant. You can see the reason in the ViLD paper. Regards.
@dyabel The conclusion of my experiment is that the result of detpro_base.py is very poor. Cuold you share the results of detpro_base.py?
Is there any significant improvement by adding a separate head network for the roi feature alignment by calling
forward_embedding_for_image()
function in the ensemble setup?