sklvq.activations.ActivationBaseClass

class sklvq.activations.ActivationBaseClass[source]

Activation base class

Abstract class for implementing activation functions. Provides abstract methods with expected call signatures.

Custom activation function ‘__init__’ should accept any parameters as key-value pairs.

See also

Identity, Sigmoid, SoftPlus, Swish
abstract __call__(x: numpy.ndarray) → numpy.ndarray[source]

Should implement an activation function

Parameters
xndarray of any shape
Returns
ndarray of shape (x.shape)

Should perform an elementwise evaluation of some activation function.

abstract gradient(x: numpy.ndarray) → numpy.ndarray[source]

Should implement the activation function’s gradient

Parameters
xndarray of any shape
Returns
ndarray of shape (x.shape)

Should return the elementwise evaluation of the activation function’s gradient.

Examples using sklvq.activations.ActivationBaseClass