Apr-15-2019, 06:04 AM
The main idea is to pass predicted and original class label arrays to the
confusion_matrix
function. Arrays should have the same length. Currently, I haven't installed keras
framework and can't reproduce the problem...