nttrain.h File Reference

Training Utilities.
https://tituxdev.github.io/NeuroTIC/src/CPU/nttrain.c. More...

#include "ntcore.h"

Include dependency graph for nttrain.h:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Data Structures
struct	traindata_t
	Structure to hold training dataset and parameters. More...

Typedefs
typedef data_t	precision_t
typedef uint64_t	sample_t
typedef uint64_t	attempts_t
typedef struct traindata_t	traindata_t
	Structure to hold training dataset and parameters.

Functions
void	newtraindata (traindata_t *train_data, net_s *net)
	Allocates memory for training data arrays.
attempts_t	backpropagation (net_s *net, traindata_t *data)
	Trains a network using backpropagation.

Detailed Description

Training Utilities.
https://tituxdev.github.io/NeuroTIC/src/CPU/nttrain.c.

Provides functions to allocate training datasets and train feedforward networks using standard backpropagation.

Author

Oscar Sotomayor

Date

2026

Definition in file nttrain.h.

Typedef Documentation

attempts_t

typedef uint64_t attempts_t

precision_t

typedef data_t precision_t

sample_t

typedef uint64_t sample_t

traindata_t

typedef struct traindata_t traindata_t

Structure to hold training dataset and parameters.

Function Documentation

backpropagation()

attempts_t backpropagation	(	net_s *	net,
		traindata_t *	train_data
	)

Trains a network using backpropagation.

Parameters

net	Pointer to the network to train.
train_data	Pointer to the training data.

Returns

attempts_t Number of iterations performed.

Implements the backpropagation algorithm to train the network. For each training sample:

• Computes outputs via feedforward.
• Calculates errors for each neuron.
• Propagates deltas backward and updates weights and biases.
• Repeats until error is below tolerance or max_attempts is reached.

References feedforward(), net_s::layers, traindata_t::max_attempts, net_s::neurons, traindata_t::samples, and traindata_t::tolerance.

newtraindata()

void newtraindata	(	traindata_t *	train_data,
		net_s *	net
	)

Allocates memory for training data arrays.

Parameters

traindata_t	Pointer to training data structure to initialize.
net_s	Pointer to the network being trained.

Allocates memory for training data arrays.
Initializes in and results arrays based on the network structure and number of samples.

References net_s::layers, and traindata_t::samples.