diff --git a/src/cnn/include/train.h b/src/cnn/include/train.h
index 2cfe564..9057f8b 100644
--- a/src/cnn/include/train.h
+++ b/src/cnn/include/train.h
@@ -35,6 +35,6 @@ void* train_thread(void* parameters);
 /*
  * Fonction principale d'entraînement du réseau neuronal convolutif
 */
-void train(int dataset_type, char* images_file, char* labels_file, char* data_dir, int epochs, char* out);
+void train(int dataset_type, char* images_file, char* labels_file, char* data_dir, int epochs, char* out, char* recover);
 
 #endif
\ No newline at end of file
diff --git a/src/cnn/main.c b/src/cnn/main.c
index 07915df..ba90b6f 100644
--- a/src/cnn/main.c
+++ b/src/cnn/main.c
@@ -24,6 +24,7 @@ void help(char* call) {
     printf("\t(mnist)\t--images  | -i [FILENAME]\tFichier contenant les images.\n");
     printf("\t(mnist)\t--labels  | -l [FILENAME]\tFichier contenant les labels.\n");
     printf("\t (jpg) \t--datadir | -dd [FOLDER]\tDossier contenant les images.\n");
+    printf("\t\t--recover | -r [FILENAME]\tRécupérer depuis un modèle existant.\n");
     printf("\t\t--epochs  | -e [int]\t\tNombre d'époques.\n");
     printf("\t\t--out     | -o [FILENAME]\tFichier où écrire le réseau de neurones.\n");
     printf("\trecognize:\n");
@@ -55,6 +56,7 @@ int main(int argc, char* argv[]) {
         int epochs = EPOCHS;
         int dataset_type = 0;
         char* out = NULL;
+        char* recover = NULL;
         int i = 2;
         while (i < argc) {
             if ((! strcmp(argv[i], "--dataset"))||(! strcmp(argv[i], "-d"))) {
@@ -80,6 +82,9 @@ int main(int argc, char* argv[]) {
             else if ((! strcmp(argv[i], "--out"))||(! strcmp(argv[i], "-o"))) {
                 out = argv[i+1];
                 i += 2;
+            } else if ((! strcmp(argv[i], "--recover"))||(! strcmp(argv[i], "-r"))) {
+                recover = argv[i+1];
+                i += 2;
             } else {
                 printf("Option choisie inconnue: %s\n", argv[i]);
                 i++;
@@ -111,7 +116,7 @@ int main(int argc, char* argv[]) {
             printf("Pas de fichier de sortie spécifié, défaut: out.bin\n");
             out = "out.bin";
         }
-        train(dataset_type, images_file, labels_file, data_dir, epochs, out);
+        train(dataset_type, images_file, labels_file, data_dir, epochs, out, recover);
         return 0;
     }
     if (! strcmp(argv[1], "test")) {
diff --git a/src/cnn/train.c b/src/cnn/train.c
index e86434b..bb807de 100644
--- a/src/cnn/train.c
+++ b/src/cnn/train.c
@@ -75,8 +75,9 @@ void* train_thread(void* parameters) {
 }
 
 
-void train(int dataset_type, char* images_file, char* labels_file, char* data_dir, int epochs, char* out) {
+void train(int dataset_type, char* images_file, char* labels_file, char* data_dir, int epochs, char* out, char* recover) {
     srand(time(NULL));
+    Network* network;
     int input_dim = -1;
     int input_depth = -1;
     float accuracy;
@@ -111,7 +112,12 @@ void train(int dataset_type, char* images_file, char* labels_file, char* data_di
     }
 
     // Initialisation du réseau
-    Network* network = create_network_lenet5(1, 0, TANH, GLOROT, input_dim, input_depth);
+    if (!recover) {
+        network = create_network_lenet5(1, 0, TANH, GLOROT, input_dim, input_depth);
+    } else {
+        network = read_network(recover);
+    }
+    
 
     shuffle_index = (int*)malloc(sizeof(int)*nb_images_total);
     for (int i=0; i < nb_images_total; i++) {
@@ -184,6 +190,9 @@ void train(int dataset_type, char* images_file, char* labels_file, char* data_di
         knuth_shuffle(shuffle_index, nb_images_total);
         batches_epoques = div_up(nb_images_total, BATCHES);
         nb_images_total_remaining = nb_images_total;
+        #ifndef USE_MULTITHREADING
+        train_params->nb_images = BATCHES;
+        #endif
         for (int j=0; j < batches_epoques; j++) {
             #ifdef USE_MULTITHREADING
             if (j == batches_epoques-1) {
@@ -222,6 +231,11 @@ void train(int dataset_type, char* images_file, char* labels_file, char* data_di
             (void)nb_images_total_remaining; // Juste pour enlever un warning
 
             train_params->start = j*BATCHES;
+
+            // Ne pas dépasser le nombre d'images à cause de la partie entière
+            if (j == batches_epoques-1) {
+                train_params->nb_images = nb_images_total - j*BATCHES;
+            }
             
             train_thread((void*)train_params);