Class LayerDimlp2

Inheritance Relationships

Base Type

Class Documentation

class LayerDimlp2 : public Layer

The LayerDimlp2 class represents a Dimlp layer in a neural network using a staircase activation function.

This class inherits from the Layer class and overrides several functions to implement specific behaviors for the forward pass and delta computation. This layer does not adapt weights or biases during training.

Public Functions

LayerDimlp2(float eta, float mu, float flat, int nbDown, int nbUp, int nbWeights, int nbWeightsForInit, int nbKnots)

Constructor for the LayerDimlp2 class.

Initializes the layer with the specified parameters and sets up the staircase activation function.

Parameters:
  • eta – Learning rate.

  • mu – Momentum factor.

  • flat – Flatness factor.

  • nbDown – Number of neurons in the previous layer.

  • nbUp – Number of neurons in this layer.

  • nbWeights – Number of weights in this layer.

  • nbWeightsForInit – Number of weights for initialization.

  • nbKnots – Number of knots for the StairObj.