finale Änderungen

info2praktikum-neuronalesnetz/neuralNetwork.c

@@ -118,56 +118,60 @@ static void assignActivations(NeuralNetwork model)
     if(model.numberOfLayers > 0)
         model.layers[model.numberOfLayers-1].activation = softmax;
 }
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include "neuralNetwork.h"
 
 NeuralNetwork loadModel(const char *path)
 {
-    NeuralNetwork model = {NULL, 0};
+    NeuralNetwork nn = {.layers = NULL, .numberOfLayers = 0};
     FILE *file = fopen(path, "rb");
+    if (!file) return nn;
 
-    if(file != NULL)
+    const char *header = "__info2_neural_network_file_format__";
-    {
+    size_t headerLen = strlen(header);
-        if(checkFileHeader(file))
+    char buffer[64] = {0};
-        {
-            unsigned int inputDimension = readDimension(file);
-            unsigned int outputDimension = readDimension(file);
 
-            while(inputDimension > 0 && outputDimension > 0)
+    if (fread(buffer, sizeof(char), headerLen, file) != headerLen ||
-            {
+        strncmp(buffer, header, headerLen) != 0) {
-                Layer layer = readLayer(file, inputDimension, outputDimension);
-                Layer *layerBuffer = NULL;
-
-                if(isEmptyLayer(layer))
-                {
-                    clearLayer(&layer);
-                    clearModel(&model);
-                    break;
-                }
-
-                layerBuffer = (Layer *)realloc(model.layers, (model.numberOfLayers + 1) * sizeof(Layer));
-
-                if(layerBuffer != NULL)
-                    model.layers = layerBuffer;
-                else
-                {
-                    clearModel(&model);
-                    break;
-                }
-
-                model.layers[model.numberOfLayers] = layer;
-                model.numberOfLayers++;
-
-                inputDimension = outputDimension;
-                outputDimension = readDimension(file);
-            }
-        }
         fclose(file);
-
+        return nn;
-        assignActivations(model);
     }
 
-    return model;
+    Layer *layers = NULL;
+    unsigned int inputDim = 0, outputDim = 0;
+
+    while (fread(&inputDim, sizeof(unsigned int), 1, file) == 1 && inputDim != 0) {
+        fread(&outputDim, sizeof(unsigned int), 1, file);
+
+        layers = realloc(layers, (nn.numberOfLayers + 1) * sizeof(Layer));
+
+        Matrix weights;
+        weights.rows = outputDim;
+        weights.cols = inputDim;
+        weights.buffer = malloc(weights.rows * weights.cols * sizeof(MatrixType));
+        fread(weights.buffer, sizeof(MatrixType), weights.rows * weights.cols, file);
+
+        Matrix biases;
+        biases.rows = outputDim;
+        biases.cols = 1;
+        biases.buffer = malloc(biases.rows * biases.cols * sizeof(MatrixType));
+        fread(biases.buffer, sizeof(MatrixType), biases.rows * biases.cols, file);
+
+        layers[nn.numberOfLayers].weights = weights;
+        layers[nn.numberOfLayers].biases = biases;
+        layers[nn.numberOfLayers].activation = NULL;
+
+        nn.numberOfLayers++;  // **WICHTIG**
+    }
+
+    nn.layers = layers;
+    fclose(file);
+    return nn;
 }
 
+
 static Matrix imageBatchToMatrixOfImageVectors(const GrayScaleImage images[], unsigned int count)
 {
     Matrix matrix = {NULL, 0, 0};
@@ -254,15 +258,16 @@ unsigned char *predict(const NeuralNetwork model, const GrayScaleImage images[],
     return result;
 }
 
-void clearModel(NeuralNetwork *model)
+
+void clearModel(NeuralNetwork *nn)
 {
-    if(model != NULL)
+    if (!nn || !nn->layers) return;
-    {
+
-        for(int i = 0; i < model->numberOfLayers; i++)
+    for (unsigned int i = 0; i < nn->numberOfLayers; i++) {
-        {
+        free(nn->layers[i].weights.buffer);
-            clearLayer(&model->layers[i]);
+        free(nn->layers[i].biases.buffer);
-        }
-        model->layers = NULL;
-        model->numberOfLayers = 0;
     }
-}
+    free(nn->layers);
+    nn->layers = NULL;
+    nn->numberOfLayers = 0;
+}

info2praktikum-neuronalesnetz/neuralNetworkTests.c

@@ -38,6 +38,7 @@ static void prepareNeuralNetworkFile(const char *path, const NeuralNetwork nn)
         fwrite(layer.biases.buffer, sizeof(MatrixType), biasCount, file);
     }
 
+
     // End-Marker (inputDim = 0)
     unsigned int zero = 0;
     fwrite(&zero, sizeof(unsigned int), 1, file);