Is there a way to convert detectron2 to a more compact model for faster inferencing ?
I moved your post Detectron2: faster inferencing.