Dummy proposal_target operator


#1

Hi!

I have been implementing the RCNN inference part in C++ and find myself having to implement the proposal target operator. I realize that it is not loaded in the vgg_test, but just vgg_train - does that mean a dummy proposal target operator registered with mxnet will be okay if I only use vgg_test? I don’t know enough about deep learning to understand if this is a dumb question.

Thanks in advance!


#2

I built a dummy operator and registered it with MXnet and loaded a model that had the proposal target layer. It is able to load the model without any errors.