turtschinba100320/info2Praktikum-NeuronalesNetzBastiBjoern
2
0
Fork 0
forked from freudenreichan/info2Praktikum-NeuronalesNetz
Code Pull Requests Activity

matrix fehler behoben

Browse Source
This commit is contained in:
Björn 2025-11-18 01:52:43 +01:00
parent a031bb0b7a
commit 6c26652744
3 changed files with 74 additions and 53 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
main.c
View File

@ -31,7 +31,7 @@ int main(int argc, char *argv[])
unsigned char *predictions = NULL;
printf("Processing %u images ...\n", series->count);
predictions = predict(model, series->images, series->count);
if(predictions != NULL)

121
matrix.c
View File

@ -1,8 +1,8 @@
#include <stdlib.h>
#include <string.h>
#include "matrix.h"
#include <stdio.h>
// TODO Matrix-Funktionen implementieren
// Matrix erzeugen
Matrix createMatrix(unsigned int rows, unsigned int cols)
@ -11,29 +11,33 @@ Matrix createMatrix(unsigned int rows, unsigned int cols)
matrix.buffer = NULL;
matrix.rows = 0;
matrix.cols = 0;
// Wenn die Dimensionen gültig sind, Speicher reservieren
if (rows > 0 && cols > 0)
{
matrix.buffer = (MatrixType *)malloc(rows * cols * sizeof(MatrixType));
if (matrix.buffer != NULL)
{
matrix.rows = rows;
matrix.cols = cols;
}
}
if (rows == 0 || cols == 0)
return matrix; // leere Matrix
matrix.buffer = (MatrixType *)malloc(rows * cols * sizeof(MatrixType));
if (!matrix.buffer)
return matrix; // Speicher konnte nicht reserviert werden
matrix.rows = rows;
matrix.cols = cols;
// Initialisiere alle Werte auf UNDEFINED_MATRIX_VALUE
for (unsigned int i = 0; i < rows * cols; i++)
matrix.buffer[i] = UNDEFINED_MATRIX_VALUE;
return matrix;
}
// Matrix Speicher freigeben
void clearMatrix(Matrix *matrix)
{
if (matrix->buffer != NULL)
{
if (!matrix) return;
if (matrix->buffer)
free(matrix->buffer);
matrix->buffer = NULL;
}
matrix->buffer = NULL;
matrix->rows = 0;
matrix->cols = 0;
}
@ -41,67 +45,84 @@ void clearMatrix(Matrix *matrix)
// Wert setzen
void setMatrixAt(MatrixType value, Matrix matrix, unsigned int rowIdx, unsigned int colIdx)
{
if (rowIdx < matrix.rows && colIdx < matrix.cols)
{
matrix.buffer[rowIdx * matrix.cols + colIdx] = value;
}
if (!matrix.buffer) return;
if (rowIdx >= matrix.rows || colIdx >= matrix.cols) return;
matrix.buffer[rowIdx * matrix.cols + colIdx] = value;
}
// Wert auslesen
MatrixType getMatrixAt(const Matrix matrix, unsigned int rowIdx, unsigned int colIdx)
{
if (rowIdx < matrix.rows && colIdx < matrix.cols)
{
return matrix.buffer[rowIdx * matrix.cols + colIdx];
}
return 0; // Fallback
}
if (!matrix.buffer) return UNDEFINED_MATRIX_VALUE;
if (rowIdx >= matrix.rows || colIdx >= matrix.cols) return UNDEFINED_MATRIX_VALUE;
return matrix.buffer[rowIdx * matrix.cols + colIdx];
}
// Matrizen addieren
Matrix add(const Matrix m1, const Matrix m2)
{
if (m1.rows != m2.rows || m1.cols != m2.cols)
if (!m1.buffer || !m2.buffer) return createMatrix(0,0);
// gleiche Dimension
if (m1.rows == m2.rows && m1.cols == m2.cols)
{
return createMatrix(0, 0); // Falls Matrix-Dimensionen nicht passen
Matrix result = createMatrix(m1.rows, m1.cols);
if (!result.buffer) return result;
for (unsigned int r = 0; r < m1.rows; r++)
for (unsigned int c = 0; c < m1.cols; c++)
result.buffer[r * result.cols + c] = m1.buffer[r * m1.cols + c] + m2.buffer[r * m2.cols + c];
return result;
}
Matrix result = createMatrix(m1.rows, m1.cols);
if (result.buffer == NULL) return result;
for (unsigned int r = 0; r < m1.rows; r++)
// Matrix2 ist ein Spaltenvektor
if (m1.rows == m2.rows && m2.cols == 1)
{
for (unsigned int c = 0; c < m1.cols; c++)
{
result.buffer[r * m1.cols + c] =
getMatrixAt(m1, r, c) + getMatrixAt(m2, r, c);
}
Matrix result = createMatrix(m1.rows, m1.cols);
if (!result.buffer) return result;
for (unsigned int r = 0; r < m1.rows; r++)
for (unsigned int c = 0; c < m1.cols; c++)
result.buffer[r * result.cols + c] = m1.buffer[r * m1.cols + c] + m2.buffer[r];
return result;
}
return result;
// Matrix1 ist ein Spaltenvektor
if (m1.rows == m2.rows && m1.cols == 1)
{
Matrix result = createMatrix(m2.rows, m2.cols);
if (!result.buffer) return result;
for (unsigned int r = 0; r < m2.rows; r++)
for (unsigned int c = 0; c < m2.cols; c++)
result.buffer[r * result.cols + c] = m1.buffer[r] + m2.buffer[r * m2.cols + c];
return result;
}
// passt nicht
return createMatrix(0,0);
}
// Matrizen multiplizieren
Matrix multiply(const Matrix m1, const Matrix m2)
{
if (m1.cols != m2.rows)
{
return createMatrix(0, 0); // Falls Matrix-Dimensionen nicht passen
}
if (!m1.buffer || !m2.buffer) return createMatrix(0,0);
if (m1.cols != m2.rows) return createMatrix(0,0);
Matrix result = createMatrix(m1.rows, m2.cols);
if (result.buffer == NULL) return result;
if (!result.buffer) return result;
for (unsigned int r = 0; r < m1.rows; r++)
{
for (unsigned int c = 0; c < m2.cols; c++)
{
MatrixType sum = 0;
for (unsigned int k = 0; k < m1.cols; k++)
{
sum += getMatrixAt(m1, r, k) * getMatrixAt(m2, k, c);
}
result.buffer[r * m2.cols + c] = sum;
sum += m1.buffer[r * m1.cols + k] * m2.buffer[k * m2.cols + c];
result.buffer[r * result.cols + c] = sum;
}
}
return result;
}

4
neuralNetwork.c
View File

@ -197,7 +197,7 @@ static Matrix forward(const NeuralNetwork model, Matrix inputBatch)
if(result.buffer != NULL)
{
for(int i = 0; i < model.numberOfLayers; i++)
for(int i = 0; i < model.numberOfLayers; i++)
{
Matrix biasResult;
Matrix weightResult;
@ -246,7 +246,7 @@ unsigned char *predict(const NeuralNetwork model, const GrayScaleImage images[],
{
Matrix inputBatch = imageBatchToMatrixOfImageVectors(images, numberOfImages);
Matrix outputBatch = forward(model, inputBatch);
unsigned char *result = argmax(outputBatch);
clearMatrix(&outputBatch);