This is the code for building custom keras layer
class MyLayer(Layer):
def __init__(self, output_dim, **kwargs):
self.output_dim = output_dim
super(MyLayer, self).__init__(**kwargs)
what is super doing here and is it absolutely necessary ?