In this example we will design a "measure" layer, that outputs the accuracy and a confusion matrix for a binary problem during training and the accuracy, false positive rate and false negative rate during test/validation. Although Caffe already has a Accuracy layer, sometimes you want some...