Structural unit representing a computation node. More...

#include <ntcore.h>

Data Fields
input_t	inputs

index_t	bff_idx

data_t **	in

weight_t *	w

bias_t	b

index_t	fn

data_t	out

Detailed Description

Structural unit representing a computation node.

A neuron binds input references, weight coefficients, a bias term, and an activation selector. It does not define memory ownership beyond structural integrity requirements.

Field Documentation

◆ b

bias_t neuron_s::b

Bias term.

Referenced by loadnet(), randnet(), and weighing().

◆ bff_idx

index_t neuron_s::bff_idx

Selected buffer set index.

Referenced by buildnet(), loadnet(), newdense(), newfeedforward(), and savenet().

◆ fn

index_t neuron_s::fn

Activation function selector.

Referenced by activate(), loadnet(), randnet(), and savenet().

◆ in

data_t** neuron_s::in

Resolved input references.

Referenced by buildnet(), and weighing().

◆ inputs

input_t neuron_s::inputs

Logical number of inputs.

Referenced by buildnet(), loadnet(), randnet(), savenet(), and weighing().

◆ out

data_t neuron_s::out

Last computed output value.

Referenced by activate(), and buildnet().

◆ w

weight_t* neuron_s::w

Weight coefficients.