Class representing the N2CMU device. More...

#include <n2cmu.h>

Public Member Functions
	N2Coprocessor (uint8_t rx=N2CMU_RX_PIN, uint8_t tx=N2CMU_TX_PIN)
	Constructor for N2Coprocessor class. More...

bool	begin ()
	Initialize the N2CMU device. More...

bool	handshake ()
	Perform handshake with the N2CMU device. More...

bool	cpuReset ()
	Reset the CPU of the N2CMU device. More...

void	createNetwork (uint8_t inputCount, uint8_t hiddenCount, uint8_t outputCount)
	Create a neural network with specified input, hidden, and output neuron counts. More...

bool	train (float data, float output, uint16_t len, float learningRate)
	Train the neural network with provided data and output. More...

bool	infer (float input, float output)
	Make inference with the neural network using provided input data. More...

void	resetNetwork ()
	Reset the neural network parameters. More...

void	setInputCount (uint8_t inputCount)
	Set the number of input neurons. More...

void	setHiddenCount (uint8_t hiddenCount)
	Set the number of hidden neurons. More...

void	setOutputCount (uint8_t outputCount)
	Set the number of output neurons. More...

void	setEpochCount (uint16_t epoch)
	Set the epoch count for training. More...

uint8_t	getInputCount ()
	Get the number of input neurons. More...

uint8_t	getHiddenCount ()
	Get the number of hidden neurons. More...

uint8_t	getOutputCount ()
	Get the number of output neurons. More...

uint16_t	getEpochCount ()
	Get the epoch count for training. More...

bool	setHiddenNeuron (float *hiddenNeuron)
	Set hidden neuron values. More...

bool	setOutputNeuron (float *outputNeuron)
	Set output neuron values. More...

bool	setHiddenWeights (float *hiddenWeights)
	Set hidden neuron weights. More...

bool	setOutputWeights (float *outputWeights)
	Set output neuron weights. More...

bool	setHiddenBias (float *hiddenBias)
	Set hidden neuron biases. More...

bool	setOutputBias (float *outputBias)
	Set output neuron biases. More...

bool	setHiddenGradient (float *hiddenGrad)
	Set hidden neuron gradients. More...

bool	setOutputGradient (float *outputGrad)
	Set output neuron gradients. More...

void	getHiddenNeuron (float *hiddenNeuron)
	Get hidden neuron values. More...

void	getOutputNeuron (float *outputNeuron)
	Get output neuron values. More...

void	getHiddenWeights (float *hiddenWeights)
	Get hidden neuron weights. More...

void	getOutputWeights (float *outputWeights)
	Get output neuron weights. More...

void	getHiddenBias (float *hiddenBias)
	Get hidden neuron biases. More...

void	getOutputBias (float *outputBias)
	Get output neuron biases. More...

void	getHiddenGradient (float *hiddenGrad)
	Get hidden neuron gradients. More...

void	getOutputGradient (float *outputGrad)
	Get output neuron gradients. More...

Detailed Description

Class representing the N2CMU device.

The N2Coprocessor class offers an interface for controlling and interacting with the N2CMU shield device. It encapsulates functions for initialization, configuration, training, inference, and parameter manipulation of neural networks.

Constructor & Destructor Documentation

◆ N2Coprocessor()

N2Coprocessor::N2Coprocessor	(	uint8_t	rx = `N2CMU_RX_PIN`,
		uint8_t	tx = `N2CMU_TX_PIN`
	)

inline

Constructor for N2Coprocessor class.

Constructs a new N2Coprocessor object with the specified RX and TX pins for serial communication.

Parameters

rx	Pin number for receiving data from N2CMU.
tx	Pin number for transmitting data to N2CMU.

Member Function Documentation

◆ begin()

bool N2Coprocessor::begin ( )

Initialize the N2CMU device.

This function initializes the N2CMU device by establishing communication and performing necessary setup operations. It must be called before any other operations are performed on the N2CMU device.

Returns: True if initialization was successful, false otherwise.

◆ cpuReset()

bool N2Coprocessor::cpuReset ( )

Reset the CPU of the N2CMU device.

This function resets the CPU of the N2CMU device, restoring it to a known state. It can be useful for recovering from unexpected errors or initializing the device before starting a new operation.

Returns: True if CPU reset was successful, false otherwise.

◆ createNetwork()

void N2Coprocessor::createNetwork	(	uint8_t	inputCount,
		uint8_t	hiddenCount,
		uint8_t	outputCount
	)

Create a neural network with specified input, hidden, and output neuron counts.

This function creates a neural network model with the specified number of input, hidden, and output neurons. It initializes the network architecture for subsequent training and inference tasks.

Parameters

inputCount	Number of input neurons.
hiddenCount	Number of hidden neurons.
outputCount	Number of output neurons.

◆ getEpochCount()

uint16_t N2Coprocessor::getEpochCount ( )

Get the epoch count for training.

This function retrieves the number of training epochs currently set for the neural network.

Returns: Epoch count for training.

◆ getHiddenBias()

void N2Coprocessor::getHiddenBias ( float * hiddenBias )

Get hidden neuron biases.

This function retrieves the biases of the hidden neurons in the neural network. It allows obtaining the current biases of the hidden neurons during or after training or inference.

Parameters

hiddenBias Array to store hidden neuron biases.

◆ getHiddenCount()

uint8_t N2Coprocessor::getHiddenCount ( )

Get the number of hidden neurons.

This function retrieves the number of hidden neurons currently set for the neural network.

Returns: Number of hidden neurons.

◆ getHiddenGradient()

void N2Coprocessor::getHiddenGradient ( float * hiddenGrad )

Get hidden neuron gradients.

This function retrieves the gradients of the hidden neurons in the neural network. It allows obtaining the current gradients of the hidden neurons during or after training.

Parameters

hiddenGrad Array to store hidden neuron gradients.

◆ getHiddenNeuron()

void N2Coprocessor::getHiddenNeuron ( float * hiddenNeuron )

Get hidden neuron values.

This function retrieves the values of the hidden neurons in the neural network. It allows obtaining the current state of the hidden neurons during or after training or inference.

Parameters

hiddenNeuron Array to store hidden neuron values.

◆ getHiddenWeights()

void N2Coprocessor::getHiddenWeights ( float * hiddenWeights )

Get hidden neuron weights.

This function retrieves the weights of the connections between input and hidden neurons in the neural network. It allows obtaining the current weights of these connections during or after training or inference.

Parameters

hiddenWeights Array to store hidden neuron weights.

◆ getInputCount()

uint8_t N2Coprocessor::getInputCount ( )

Get the number of input neurons.

This function retrieves the number of input neurons currently set for the neural network.

Returns: Number of input neurons.

◆ getOutputBias()

void N2Coprocessor::getOutputBias ( float * outputBias )

Get output neuron biases.

This function retrieves the biases of the output neurons in the neural network. It allows obtaining the current biases of the output neurons during or after training or inference.

Parameters

outputBias Array to store output neuron biases.

◆ getOutputCount()

uint8_t N2Coprocessor::getOutputCount ( )

Get the number of output neurons.

This function retrieves the number of output neurons currently set for the neural network.

Returns: Number of output neurons.

◆ getOutputGradient()

void N2Coprocessor::getOutputGradient ( float * outputGrad )

Get output neuron gradients.

This function retrieves the gradients of the output neurons in the neural network. It allows obtaining the current gradients of the output neurons during or after training.

Parameters

outputGrad Array to store output neuron gradients.

◆ getOutputNeuron()

void N2Coprocessor::getOutputNeuron ( float * outputNeuron )

Get output neuron values.

This function retrieves the values of the output neurons in the neural network. It allows obtaining the current state of the output neurons during or after training or inference.

Parameters

outputNeuron Array to store output neuron values.

◆ getOutputWeights()

void N2Coprocessor::getOutputWeights ( float * outputWeights )

Get output neuron weights.

This function retrieves the weights of the connections between hidden and output neurons in the neural network. It allows obtaining the current weights of these connections during or after training or inference.

Parameters

outputWeights Array to store output neuron weights.

◆ handshake()

bool N2Coprocessor::handshake ( )

Perform handshake with the N2CMU device.

This function performs a handshake with the N2CMU device to ensure proper communication and synchronization. It verifies the connection and readiness of the N2CMU device for subsequent operations.

Returns: True if handshake was successful, false otherwise.

◆ infer()

bool N2Coprocessor::infer	(	float *	input,
		float *	output
	)

Make inference with the neural network using provided input data.

This function performs inference with the trained neural network using the provided input data. It computes the output of the network based on the input and returns the result.

Parameters

input	Pointer to the input data array.
output	Pointer to store the output data array.

Returns: True if inference was successful, false otherwise.

◆ resetNetwork()

void N2Coprocessor::resetNetwork ( )

Reset the neural network parameters.

This function resets the parameters of the neural network to their initial values. It clears any previously learned weights, biases, or gradients, effectively resetting the network to its initial state.

◆ setEpochCount()

void N2Coprocessor::setEpochCount ( uint16_t epoch )

Set the epoch count for training.

This function sets the number of training epochs for the neural network. An epoch represents one complete pass through the entire training dataset.

Parameters

epoch Epoch count for training.

◆ setHiddenBias()

bool N2Coprocessor::setHiddenBias ( float * hiddenBias )

Set hidden neuron biases.

This function sets the biases of the hidden neurons in the neural network. It allows specifying the initial biases of the hidden neurons before training or inference.

Parameters

hiddenBias Array of hidden neuron biases.

Returns: True if setting was successful, false otherwise.

◆ setHiddenCount()

void N2Coprocessor::setHiddenCount ( uint8_t hiddenCount )

Set the number of hidden neurons.

This function sets the number of hidden neurons for the neural network. It allows dynamically configuring the hidden layer of the network to adjust its capacity and complexity.

Parameters

hiddenCount Number of hidden neurons.

Note: After calling this function, all the network data such as weights, biases, and gradients will return to default 0 values.

◆ setHiddenGradient()

bool N2Coprocessor::setHiddenGradient ( float * hiddenGrad )

Set hidden neuron gradients.

This function sets the gradients of the hidden neurons in the neural network. It allows specifying the initial gradients of the hidden neurons before training or inference.

Parameters

hiddenGrad Array of hidden neuron gradients.

Returns: True if setting was successful, false otherwise.

◆ setHiddenNeuron()

bool N2Coprocessor::setHiddenNeuron ( float * hiddenNeuron )

Set hidden neuron values.

This function sets the values of the hidden neurons in the neural network. It allows specifying the initial values of the hidden neurons before training or inference.

Parameters

hiddenNeuron Array of hidden neuron values.

Returns: True if setting was successful, false otherwise.

◆ setHiddenWeights()

bool N2Coprocessor::setHiddenWeights ( float * hiddenWeights )

Set hidden neuron weights.

This function sets the weights of the connections between input and hidden neurons in the neural network. It allows specifying the initial weights of these connections before training or inference.

Parameters

hiddenWeights Array of hidden neuron weights.

Returns: True if setting was successful, false otherwise.

◆ setInputCount()

void N2Coprocessor::setInputCount ( uint8_t inputCount )

Set the number of input neurons.

This function sets the number of input neurons for the neural network. It allows dynamically configuring the input layer of the network to match the requirements of the input data.

Parameters

inputCount Number of input neurons.

Note: After calling this function, all the network data such as weights, biases, and gradients will return to default 0 values.

◆ setOutputBias()

bool N2Coprocessor::setOutputBias ( float * outputBias )

Set output neuron biases.

This function sets the biases of the output neurons in the neural network. It allows specifying the initial biases of the output neurons before training or inference.

Parameters

outputBias Array of output neuron biases.

Returns: True if setting was successful, false otherwise.

◆ setOutputCount()

void N2Coprocessor::setOutputCount ( uint8_t outputCount )

Set the number of output neurons.

This function sets the number of output neurons for the neural network. It allows dynamically configuring the output layer of the network to match the requirements of the output data.

Parameters

outputCount Number of output neurons.

Note: After calling this function, all the network data such as weights, biases, and gradients will return to default 0 values.

◆ setOutputGradient()

bool N2Coprocessor::setOutputGradient ( float * outputGrad )

Set output neuron gradients.

This function sets the gradients of the output neurons in the neural network. It allows specifying the initial gradients of the output neurons before training or inference.

Parameters

outputGrad Array of output neuron gradients.

Returns: True if setting was successful, false otherwise.

◆ setOutputNeuron()

bool N2Coprocessor::setOutputNeuron ( float * outputNeuron )

Set output neuron values.

This function sets the values of the output neurons in the neural network. It allows specifying the initial values of the output neurons before training or inference.

Parameters

outputNeuron Array of output neuron values.

Returns: True if setting was successful, false otherwise.

◆ setOutputWeights()

bool N2Coprocessor::setOutputWeights ( float * outputWeights )

Set output neuron weights.

This function sets the weights of the connections between hidden and output neurons in the neural network. It allows specifying the initial weights of these connections before training or inference.

Parameters

outputWeights Array of output neuron weights.

Returns: True if setting was successful, false otherwise.

◆ train()

bool N2Coprocessor::train	(	float *	data,
		float *	output,
		uint16_t	len,
		float	learningRate
	)

Train the neural network with provided data and output.

This function trains the neural network using the provided input data and corresponding output. It adjusts the network parameters based on the training data to improve its performance and accuracy.

Parameters

data	Pointer to the input data array.
output	Pointer to the output data array.
len	Length of the data arrays.
learningRate	Learning rate for training.

Returns: True if training was successful, false otherwise.

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

src/n2cmu.h
src/n2cmu.cpp

Public Member Functions

Detailed Description

Constructor & Destructor Documentation

◆ N2Coprocessor()

Member Function Documentation

◆ begin()

◆ cpuReset()

◆ createNetwork()

◆ getEpochCount()

◆ getHiddenBias()

◆ getHiddenCount()

◆ getHiddenGradient()

◆ getHiddenNeuron()

◆ getHiddenWeights()

◆ getInputCount()

◆ getOutputBias()

◆ getOutputCount()

◆ getOutputGradient()

◆ getOutputNeuron()

◆ getOutputWeights()

◆ handshake()

◆ infer()

◆ resetNetwork()

◆ setEpochCount()

◆ setHiddenBias()

◆ setHiddenCount()

◆ setHiddenGradient()

◆ setHiddenNeuron()

◆ setHiddenWeights()

◆ setInputCount()

◆ setOutputBias()

◆ setOutputCount()

◆ setOutputGradient()

◆ setOutputNeuron()

◆ setOutputWeights()

◆ train()