NNFS::NeuralNetwork Class Reference

A neural network model. More...

#include <NeuralNetwork.hpp>

Inheritance diagram for NNFS::NeuralNetwork:

Public Member Functions
	NeuralNetwork (std::shared_ptr< Loss > loss=nullptr, std::shared_ptr< Optimizer > optimizer=nullptr)
	Constructor for NeuralNetwork.
void	fit (const Eigen::MatrixXd &examples, const Eigen::MatrixXd &labels, const Eigen::MatrixXd &test_examples, const Eigen::MatrixXd &test_labels, int epochs, int batch_size, bool verbose=true) override
	Fit the neural network model to the given data.
void	add_layer (std::shared_ptr< Layer > layer)
	Adds a layer to the neural network model.
void	compile ()
	Compiles the neural network model.
void	save (std::string path)
	Saves the model to a file in a custom binary format. The model can be loaded using the NNFS::load method.
void	load (std::string path)
	Loads a model from a file in a custom binary format. The model must have been saved using the NNFS::save method.
void	accuracy (double &accuracy, const Eigen::MatrixXd &examples, const Eigen::MatrixXd &labels)
	Calculates the accuracy of the neural network on the provided examples and labels.
Eigen::MatrixXd	predict (const Eigen::MatrixXd &sample)
	Predicts the class of the provided sample(s).
Public Member Functions inherited from NNFS::Model
virtual	~Model ()=default
	Basic destructor.
virtual void	fit (const Eigen::MatrixXd &examples, const Eigen::MatrixXd &labels, const Eigen::MatrixXd &test_examples, const Eigen::MatrixXd &test_labels, int epochs, int batch_size, bool verbose=false)=0
	Evaluate the model on the given examples.

Detailed Description

A neural network model.

This class represents a neural network model capable of training on data and making predictions.

Constructor & Destructor Documentation

NeuralNetwork()

NNFS::NeuralNetwork::NeuralNetwork ( std::shared_ptr< Loss >  loss = nullptr, std::shared_ptr< Optimizer >  optimizer = nullptr  )

inline

Constructor for NeuralNetwork.

Parameters

[in]	loss	The loss function, must be a subclass of Loss, defaults to nullptr. If nullptr, training will not be possible.
[in]	optimizer	The optimizer, must be a subclass of Optimizer, defaults to nullptr. If nullptr, training will not be possible.

Member Function Documentation

accuracy()

void NNFS::NeuralNetwork::accuracy ( double &  accuracy, const Eigen::MatrixXd &  examples, const Eigen::MatrixXd &  labels  )

inline

Calculates the accuracy of the neural network on the provided examples and labels.

Parameters

[out]	accuracy	Accuracy of the neural network on the provided examples and labels.
[in]	examples	Examples to calculate the accuracy on.
[in]	labels	Labels to calculate the accuracy on.

add_layer()

void NNFS::NeuralNetwork::add_layer ( std::shared_ptr< Layer >  layer )

inline

Adds a layer to the neural network model.

This method adds a layer to the neural network model. The layer is added to the end of the neural network model.

Parameters

[in]

layer

The layer to be added to the neural network model

compile()

void NNFS::NeuralNetwork::compile ( )

inline

Compiles the neural network model.

This method compiles the neural network model by initializing the weights and biases of the layers and setting the input and output dimensions of each layer.

fit()

void NNFS::NeuralNetwork::fit ( const Eigen::MatrixXd &  examples, const Eigen::MatrixXd &  labels, const Eigen::MatrixXd &  test_examples, const Eigen::MatrixXd &  test_labels, int  epochs, int  batch_size, bool  verbose = true  )

inlineoverridevirtual

Fit the neural network model to the given data.

This function trains the neural network model on the given data for the given number of epochs.

Parameters

[in]	examples	Examples to evaluate the model on
[in]	labels	Labels of the examples
[in]	test_examples	Examples to validate the model on
[in]	test_labels	Labels of the validation examples
[in]	epochs	The number of epochs to train for
[in]	batch_size	The batch size, i.e. the number of examples to train on in each batch
[in]	verbose	Whether to print out information about the training process

Implements NNFS::Model.

load()

void NNFS::NeuralNetwork::load ( std::string  path )

inline

Loads a model from a file in a custom binary format. The model must have been saved using the NNFS::save method.

Parameters

[in]

path

The path to load the model from

predict()

Eigen::MatrixXd NNFS::NeuralNetwork::predict ( const Eigen::MatrixXd &  sample )

inline

Predicts the class of the provided sample(s).

Parameters

[in]

sample

Sample(s) to predict the class of.

Returns

Predictions of the neural network for the provided sample(s).

save()

void NNFS::NeuralNetwork::save ( std::string  path )

inline

Saves the model to a file in a custom binary format. The model can be loaded using the NNFS::load method.

Parameters

[in]

path

The path to save the model to

---

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

• include/NNFS/Model/NeuralNetwork.hpp