I have to work with tensorflow 1.15 and need a custom layer. A very simplistic layer can look like this:
class Dummy(keras.layers.Layer): def __init__(self, u