cnn/train: add newline & change initialisation

src/cnn/include/config.h

@@ -8,6 +8,8 @@
 #define LEARNING_RATE 3e-4 // Taux d'apprentissage
 #define USE_MULTITHREADING // Commenter pour utiliser un seul coeur durant l'apprentissage (meilleur pour des tailles de batchs traités rapidement)
 
+//#define DETAILED_TRAIN_TIMINGS // Afficher le temps de forward/ backward
+
 //* Paramètres d'ADAM optimizer
 #define ALPHA 3e-4
 #define BETA_1 0.9

src/cnn/train.c

@@ -62,6 +62,10 @@ void* train_thread(void* parameters) {
     float accuracy = 0.;
     float loss = 0.;
 
+    #ifdef DETAILED_TRAIN_TIMINGS
+        double start_time;
+    #endif
+
     pthread_t tid;
     LoadImageParameters* load_image_param = (LoadImageParameters*)malloc(sizeof(LoadImageParameters));
     if (dataset_type != 0) {
@@ -74,7 +78,20 @@ void* train_thread(void* parameters) {
     for (int i=start;  i < start+nb_images; i++) {
         if (dataset_type == 0) {
             write_image_in_network_32(images[index[i]], height, width, network->input[0][0], true);
+
+            #ifdef DETAILED_TRAIN_TIMINGS
+                start_time = omp_get_wtime();
+            #endif
+
             forward_propagation(network);
+
+            #ifdef DETAILED_TRAIN_TIMINGS
+                printf("Temps de forward: ");
+                printf_time(omp_get_wtime() - start_time);
+                printf("\n");
+                start_time = omp_get_wtime();
+            #endif
+
             maxi = indice_max(network->input[network->size-1][0][0], 10);
             if (maxi == -1) {
                 printf("\n");
@@ -88,6 +105,13 @@ void* train_thread(void* parameters) {
 
             backward_propagation(network, labels[index[i]]);
 
+            #ifdef DETAILED_TRAIN_TIMINGS
+                printf("Temps de backward: ");
+                printf_time(omp_get_wtime() - start_time);
+                printf("\n");
+                start_time = omp_get_wtime();
+            #endif
+
             if (maxi == labels[index[i]]) {
                 accuracy += 1.;
             }
@@ -104,9 +128,30 @@ void* train_thread(void* parameters) {
                 pthread_create(&tid, NULL, load_image, (void*) load_image_param);
             }
             write_256_image_in_network(param->dataset->images[index[i]], width, param->dataset->numComponents, network->width[0], network->input[0]);
+
+            #ifdef DETAILED_TRAIN_TIMINGS
+                start_time = omp_get_wtime();
+            #endif
+
             forward_propagation(network);
+
+            #ifdef DETAILED_TRAIN_TIMINGS
+                printf("Temps de forward: ");
+                printf_time(omp_get_wtime() - start_time);
+                printf("\n");
+                start_time = omp_get_wtime();
+            #endif
+
             maxi = indice_max(network->input[network->size-1][0][0], param->dataset->numCategories);
             backward_propagation(network, param->dataset->labels[index[i]]);
+            
+            #ifdef DETAILED_TRAIN_TIMINGS
+                printf("Temps de backward: ");
+                printf_time(omp_get_wtime() - start_time);
+                printf("\n");
+                start_time = omp_get_wtime();
+            #endif
+
 
             if (maxi == (int)param->dataset->labels[index[i]]) {
                 accuracy += 1.;
@@ -160,9 +205,9 @@ void train(int dataset_type, char* images_file, char* labels_file, char* data_di
     int nb_images_total_remaining; // Images restantes dans un batch
     int batches_epoques; // Batches par époque
 
-    int*** images; // Images sous forme de tableau de tableaux de tableaux de pixels (degré de gris, MNIST)
-    unsigned int* labels; // Labels associés aux images du dataset MNIST
-    jpegDataset* dataset; // Structure de données décrivant un dataset d'images jpeg
+    int*** images = NULL; // Images sous forme de tableau de tableaux de tableaux de pixels (degré de gris, MNIST)
+    unsigned int* labels = NULL; // Labels associés aux images du dataset MNIST
+    jpegDataset* dataset = NULL; // Structure de données décrivant un dataset d'images jpeg
     if (dataset_type == 0) { // Type MNIST
         // Chargement des images du set de données MNIST
         int* parameters = read_mnist_images_parameters(images_file);
@@ -186,10 +231,14 @@ void train(int dataset_type, char* images_file, char* labels_file, char* data_di
     Network* network;
     if (!recover) {
         if (dataset_type == 0) {
-            network = create_network_lenet5(LEARNING_RATE, 0, RELU, NORMALIZED_XAVIER, input_width, input_depth);
+            network = create_network_lenet5(LEARNING_RATE, 0, LEAKY_RELU, HE, input_width, input_depth);
             //network = create_simple_one(LEARNING_RATE, 0, RELU, GLOROT, input_width, input_depth);    
         } else {
-            network = create_network_VGG16(LEARNING_RATE, 0, RELU, NORMALIZED_XAVIER, dataset->numCategories);
+            network = create_network_VGG16(LEARNING_RATE, 0, RELU, HE, dataset->numCategories);
+
+            #ifdef USE_MULTITHREADING
+                printf_warning("Utilisation de VGG16 avec multithreading. La quantité de RAM utilisée peut devenir excessive\n");
+            #endif
         }
     } else {
         network = read_network(recover);