Sigmoid: Sigmoid activation
Description Super class Public fields Methods See Also
Description
Each activation module has a forward method that takes in a batch of pre-activations Z and returns a batch of activations A.
Each activation module has a backward method that takes in dLdA and returns dLdZ, with the exception of SoftMax, where we assume dLdZ is passed in.
Super class
neuralnetr::ClassModule -> Sigmoid
Public fields
Athe activation vector
Methods
Public methods
Inherited methods
Method forward()
Usage
Sigmoid$forward(Z)
Arguments
Za vector of pre-activations
Returns
a vector of activations.
Method backward()
Usage
Sigmoid$backward(dLdA)
Arguments
dLdAvector of gradients.
Returns
a vector gradeints.
Method clone()
The objects of this class are cloneable with this method.
Usage
Sigmoid$clone(deep = FALSE)
Arguments
deepWhether to make a deep clone.
See Also
Other activation:
ReLU,
SoftMax,
Tanh
