stochastic_gd.c File Reference

Demonstrates mini-batch stochastic gradient descent for linear regression using ml-framework. More...

#include <dmlfw.h>
#include <unistd.h>
#include <stdio.h>
#include <stdlib.h>

Include dependency graph for stochastic_gd.c:

dmlfw_mat_double *	buffer_matrix_1 = NULL
dmlfw_mat_double *	buffer_matrix_1_shuffled = NULL
dmlfw_mat_double *	buffer_matrix_2 = NULL
dmlfw_mat_double *	buffer_matrix_2_shuffled = NULL
#define	BUFFER_SIZE   50
void	data_loader (void *x, void *y, uint64_t from_row, uint32_t how_many_rows)
	Supplies training data batches to SGD optimizer via buffer loading.
#define	FREQUENCY_OF_PRINTING_COST   100
dmlfw_gradient_descent_options *	get_gradient_descent_options ()
	Configures options for SGD training.
void	get_one_from_file_buffer (dmlfw_mat_double **target_matrix, uint64_t from_row)
	Retrieves a single row from buffered training data.
FILE *	gnuplot
void	init_buffers ()
	Initializes data buffers used for training data management.
#define	LEARNING_RATE   0.00001
int	main ()
	Main entry point to run buffered stochastic gradient descent linear regression.
#define	MODEL_FILE_NAME   "example-2-model.csv"
dmlfw_column_vec_double *	model_squared_sum_vector = NULL
dmlfw_row_vec_double *	model_transposed = NULL
uint64_t	number_of_columns_in_training_examples = 0
#define	NUMBER_OF_ITERATIONS   100000
uint64_t	number_of_training_examples = 0
int	on_iteration_complete (uint64_t iteration_number, void *y, void *predicted_y, void *model, double regularization_parameter)
	SGD iteration callback to calculate and log cost function.
dmlfw_column_vec_double *	prediction_error = NULL
dmlfw_row_vec_double *	prediction_error_transposed = NULL
void	print_error_and_exit ()
	Prints ml-framework error and terminates execution.
dmlfw_column_vec_double *	product_vector = NULL
#define	REGULARIZATION_PARAMETER   0.5
#define	SHOW_GRAPH   1
#define	TRAINING_DATASET   "IceCreamSales_training_examples.csv"
dmlfw_mat_double *	x_matrix = NULL
dmlfw_mat_double *	xy_matrix = NULL
dmlfw_column_vec_double *	y_vector = NULL

Detailed Description

Demonstrates mini-batch stochastic gradient descent for linear regression using ml-framework.

Author

Mohammed Daniyal

Version

1.0

Date

2025-09-26

This example trains a linear regression model on the IceCreamSales training dataset using mini-batch stochastic gradient descent with specified batch size, learning rate, and regularization. Includes custom data loader, progress callback logging with cost visualization using gnuplot, and buffer management for data loading.

Usage: ./stochastic_gd

Macro Definition Documentation

BUFFER_SIZE

#define BUFFER_SIZE   50

Buffer size used for batch data loading

FREQUENCY_OF_PRINTING_COST

#define FREQUENCY_OF_PRINTING_COST   100

Frequency of printing/logging cost

LEARNING_RATE

#define LEARNING_RATE   0.00001

Learning rate for gradient descent

MODEL_FILE_NAME

#define MODEL_FILE_NAME   "example-2-model.csv"

Output model CSV file path

NUMBER_OF_ITERATIONS

#define NUMBER_OF_ITERATIONS   100000

Maximum number of gradient descent iterations

REGULARIZATION_PARAMETER

#define REGULARIZATION_PARAMETER   0.5

Regularization parameter lambda

SHOW_GRAPH

#define SHOW_GRAPH   1

Enable or disable plotting with gnuplot (1 = enabled)

TRAINING_DATASET

#define TRAINING_DATASET   "IceCreamSales_training_examples.csv"

Training dataset CSV file path

Function Documentation

data_loader()

void data_loader	(	void *	x,
		void *	y,
		uint64_t	from_row,
		uint32_t	how_many_rows
	)

Supplies training data batches to SGD optimizer via buffer loading.

Parameters

x	Pointer to feature matrix pointer to populate.
y	Pointer to target vector pointer to populate.
from_row	Starting row index for batch.
how_many_rows	Number of rows requested.

get_gradient_descent_options()

dmlfw_gradient_descent_options * get_gradient_descent_options

(

)

Configures options for SGD training.

Returns

Pointer to gradient descent options configured for stochastic GD.

get_one_from_file_buffer()

void get_one_from_file_buffer	(	dmlfw_mat_double **	target_matrix,
		uint64_t	from_row
	)

Retrieves a single row from buffered training data.

Parameters

target_matrix	Output pointer to matrix containing the row.
from_row	Row index to retrieve.

init_buffers()

void init_buffers

(

)

Initializes data buffers used for training data management.

main()

int main

(

)

Main entry point to run buffered stochastic gradient descent linear regression.

Returns

0 on success, or error code.

on_iteration_complete()

int on_iteration_complete	(	uint64_t	iteration_number,
		void *	y,
		void *	predicted_y,
		void *	model,
		double	regularization_parameter
	)

SGD iteration callback to calculate and log cost function.

Parameters

iteration_number	Current iteration count.
y	Actual target vector.
predicted_y	Predicted output by model.
model	Current model weights.
regularization_parameter	Regularization factor.

Returns

0 to continue training, -1 to abort.

print_error_and_exit()

void print_error_and_exit

(

)

Prints ml-framework error and terminates execution.

Variable Documentation

buffer_matrix_1

dmlfw_mat_double* buffer_matrix_1 = NULL

Primary buffer matrix holding dataset block for training

buffer_matrix_1_shuffled

dmlfw_mat_double* buffer_matrix_1_shuffled = NULL

Shuffled copy of primary buffer matrix

buffer_matrix_2

dmlfw_mat_double* buffer_matrix_2 = NULL

Secondary buffer matrix holding remaining dataset block

buffer_matrix_2_shuffled

dmlfw_mat_double* buffer_matrix_2_shuffled = NULL

Shuffled copy of secondary buffer matrix

gnuplot

FILE* gnuplot

model_squared_sum_vector

dmlfw_column_vec_double* model_squared_sum_vector = NULL

Vector holding squared sum of model parameters for regularization calculation

model_transposed

dmlfw_row_vec_double* model_transposed = NULL

Transposed vector of current model parameters during callback

number_of_columns_in_training_examples

uint64_t number_of_columns_in_training_examples = 0

Number of columns (features + target) in training dataset

number_of_training_examples

uint64_t number_of_training_examples = 0

Total number of training examples (rows) in dataset

prediction_error

dmlfw_column_vec_double* prediction_error = NULL

Vector holding prediction errors during training progress callback

prediction_error_transposed

dmlfw_row_vec_double* prediction_error_transposed = NULL

Transposed version of prediction_error vector

product_vector

dmlfw_column_vec_double* product_vector = NULL

Temporary vector used for intermediate matrix products

x_matrix

dmlfw_mat_double* x_matrix = NULL

Feature matrix pointer used in data loading for mini-batches

xy_matrix

dmlfw_mat_double* xy_matrix = NULL

Combined feature-target matrix used temporarily for data loading

y_vector

dmlfw_column_vec_double* y_vector = NULL

Target vector pointer used in data loading for mini-batches