Cross validation in DL4J

@agibsonccc Hello, I have trained a model using 5 fold cross validation for image segmentation, so I saved 5 trained networks. Now I want to test the trained networks in a new test dataset, please, how to average the outputs of the networks in the test data set, in order to have one output segmented image in DL4J.
Another thing is that the output of the networks is an INDArray object. So is it possible to use the add method of INDArray to add all the 5 outputs and calculate the average using the div method?