ml.h File Reference

Simple Machine Learning library using 4x4 Matrices and Vectors. More...

#include "matrix.h"
#include "vec.h"

Include dependency graph for ml.h:

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

Go to the source code of this file.

Classes
struct	ML_Layer
	Represents a single fully connected layer 4 inputs -> 4 outputs. More...
struct	ML_Network
	A neural network composed of multiple layers. More...

Functions
void	ml_init (ML_Network *net, int num_layers, float learning_rate)
	Initialize a new network.
Vec4	ml_forward (ML_Network *net, Vec4 input)
	Perform forward propagation.
float	ml_train_step (ML_Network *net, Vec4 input, Vec4 target)
	Train the network on a single sample.

Detailed Description

Simple Machine Learning library using 4x4 Matrices and Vectors.

Demonstrates a real-world use case for the vector and matrix libraries. Implements a simple Feed-Forward Neural Network (MLP) with Sigmoid activation.

Limitations:

• Fixed layer width of 4 neurons (due to Mat4/Vec4 usage).
• Fixed network depth (configurable up to MAX_LAYERS).

Definition in file ml.h.

Function Documentation

ml_forward()

Vec4 ml_forward	(	ML_Network *	net,
		Vec4	input
	)

Perform forward propagation.

Parameters

net	Pointer to the network.
input	Input vector (4 floats).

Returns

Output vector (4 floats).

Definition at line 47 of file ml.c.

References ml_forward().

Referenced by ml_forward(), and ml_train_step().

Here is the call graph for this function:

Here is the caller graph for this function:

ml_init()

void ml_init	(	ML_Network *	net,
		int	num_layers,
		float	learning_rate
	)

Initialize a new network.

Parameters

net	Pointer to the network struct.
num_layers	Number of layers to create.
learning_rate	Learning rate for gradient descent.

Definition at line 36 of file ml.c.

References ml_init().

Referenced by ml_init().

Here is the call graph for this function:

Here is the caller graph for this function:

ml_train_step()

float ml_train_step	(	ML_Network *	net,
		Vec4	input,
		Vec4	target
	)

Train the network on a single sample.

Parameters

net	Pointer to the network.
input	Input vector.
target	Target output vector.

Returns

MSE Loss for this sample.

Definition at line 74 of file ml.c.

References ml_forward(), ml_train_step(), Vec4::w, Vec4::x, Vec4::y, and Vec4::z.

Referenced by ml_train_step().

Here is the call graph for this function:

Here is the caller graph for this function: