NNBuilder Class Reference

Implements a neural network builder for constructing and registering neural network layers. More...

#include <NNBuilder.h>

Public Member Functions
	NNBuilder ()
	Default constructor for NNBuilder.
	NNBuilder (std::string weightsFile, std::string biasFile)
	Constructor for NNBuilder with weights and bias files.
	~NNBuilder ()
	Destructor for NNBuilder.
std::string	build ()
	Builds and returns the computation string representing the neural network.
NNBuilder &	denseLayer (int units, int input_size, float *weights, float *bias, Activation ac)
	Adds a dense layer to the neural network.
NNBuilder &	denseLayer (int units, int input_size, Activation ac)
	Adds a dense layer to the neural network using weights and biases from files.
NNBuilder &	convLayer (int num_filters, int *dims, float *weights, float *bias, Activation ac)
	Adds a convolutional layer to the neural network.
NNBuilder &	maxPoolLayer (int side, int height, int width)
	Adds a max pooling layer to the neural network.

Detailed Description

Implements a neural network builder for constructing and registering neural network layers.

Constructor & Destructor Documentation

NNBuilder()

NNBuilder::NNBuilder ( std::string weightsFile, std::string biasFile )

inline

Constructor for NNBuilder with weights and bias files.

Parameters

weightsFile	Path to the file containing weights.
biasFile	Path to the file containing biases.

Member Function Documentation

build()

std::string NNBuilder::build ( )

inline

Builds and returns the computation string representing the neural network.

Returns

The computation string.

convLayer()

NNBuilder & NNBuilder::convLayer ( int num_filters, int * dims, float * weights, float * bias, Activation ac )

inline

Adds a convolutional layer to the neural network.

Parameters

num_filters	Number of filters in the convolutional layer.
dims	Dimensions of the filters.
weights	Pointer to the weights array.
bias	Pointer to the bias array.
ac	Activation function to apply after the convolutional layer.

Returns

Reference to the NNBuilder object for method chaining.

denseLayer() [1/2]

NNBuilder & NNBuilder::denseLayer ( int units, int input_size, Activation ac )

inline

Adds a dense layer to the neural network using weights and biases from files.

Parameters

units	Number of units in the dense layer.
input_size	Size of the input to the dense layer.
ac	Activation function to apply after the dense layer.

Returns

Reference to the NNBuilder object for method chaining.

denseLayer() [2/2]

NNBuilder & NNBuilder::denseLayer ( int units, int input_size, float * weights, float * bias, Activation ac )

inline

Adds a dense layer to the neural network.

Parameters

units	Number of units in the dense layer.
input_size	Size of the input to the dense layer.
weights	Pointer to the weights array.
bias	Pointer to the bias array.
ac	Activation function to apply after the dense layer.

Returns

Reference to the NNBuilder object for method chaining.

maxPoolLayer()

NNBuilder & NNBuilder::maxPoolLayer ( int side, int height, int width )

inline

Adds a max pooling layer to the neural network.

Parameters

side	Size of the pooling window.
height	Height of the input feature map.
width	Width of the input feature map.

Returns

Reference to the NNBuilder object for method chaining.

---

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

• NNBuilder.h