dl::activation::Sigmoid Class Reference

Sigmoid activation function. More...

#include <functions.hpp>

Inheritance diagram for dl::activation::Sigmoid:

Collaboration diagram for dl::activation::Sigmoid:

Public Member Functions
double	forward (double x)
	Compute sigmoid forward pass.
double	backward (double x)
	Compute sigmoid derivative.
Public Member Functions inherited from dl::activation::ActivationFunction
virtual	~ActivationFunction ()=default
	Virtual destructor for proper cleanup.

Detailed Description

Sigmoid activation function.

The sigmoid function maps any real number to a value between 0 and 1, making it useful for binary classification problems.

Mathematical definition:

• Forward: f(x) = 1 / (1 + exp(-x))
• Derivative: f'(x) = f(x) * (1 - f(x))

Warning

Can suffer from vanishing gradient problem for large |x|

Definition at line 81 of file functions.hpp.

Member Function Documentation

backward()

double dl::activation::Sigmoid::backward

(

double

x

)

Compute sigmoid derivative.

Parameters

x	Input value

Returns

sigmoid(x) * (1 - sigmoid(x))

Definition at line 26 of file functions.cpp.

forward()

double dl::activation::Sigmoid::forward

(

double

x

)

Compute sigmoid forward pass.

Parameters

x	Input value

Returns

1 / (1 + exp(-x))

Definition at line 20 of file functions.cpp.

---

The documentation for this class was generated from the following files:

• include/activation/functions.hpp
• src/activation/functions.cpp