In Keras, to predict class of a datatest, the predict_classes()
is used.
For example:
classes = model.predict_classes(X_test, batch_size=32)
My question is, I know the usage of batch_size
in training, but why does it need a batch_size
for prediction? how does it work?