def __init__(self):
ActivationFunction.__init__(
self,
lambda x: x*math.sin(x),
# This could be optimized as the derivative
# will always be called after the fn.
# And will be called with the same x value.
lambda x: x*math.cos(x) + math.sin(x))
def __init__(self):
ActivationFunction.__init__(
self,
lambda x: math.tanh(x),
# This could be optimized as the derivative
# will always be called after the fn.
# And will be called with the same x value.
lambda x: 1 - math.pow(math.tanh(x), 2))
def __init__(self):
ActivationFunction.__init__(
self,
lambda x: sigmoid(x),
# This could be optimized as the derivative
# will always be called after the fn.
# And will be called with the same x value.
lambda x: sigmoid(x) * (1 - sigmoid(x)))
def __init__(self):
ActivationFunction.__init__(
self, lambda x: x, lambda x: 1.0)
