make OPENCV compile again

src/classifier.c

@@ -32,7 +32,9 @@ void train_classifier(char *datacfg, char *cfgfile, char *weightfile, int *gpus,
     int i;
 
     float avg_loss = -1;
-    //float avg_contrastive_acc = 0;
+#ifdef OPENCV
+    float avg_contrastive_acc = 0;
+#endif
     char *base = basecfg(cfgfile);
     printf("%s\n", base);
     printf("%d\n", ngpus);
@@ -186,8 +188,9 @@ void train_classifier(char *datacfg, char *cfgfile, char *weightfile, int *gpus,
                 else fprintf(stderr, " Tensor Cores are used.\n");
             }
         }
-
-        //int draw_precision = 0;
+#ifdef OPENCV
+        int draw_precision = 0;
+#endif
         if (calc_topk && (i >= calc_topk_for_each || i == net.max_batches)) {
             iter_topk = i;
             if (net.contrastive && l.type != SOFTMAX && l.type != COST) {
@@ -200,7 +203,9 @@ void train_classifier(char *datacfg, char *cfgfile, char *weightfile, int *gpus,
                 topk = validate_classifier_single(datacfg, cfgfile, weightfile, &net, topk_data); // calc TOP-n
                 printf("\n accuracy %s = %f \n", topk_buff, topk);
             }
-            // draw_precision = 1;
+#ifdef OPENCV
+            draw_precision = 1;
+#endif
         }
 
         time_remaining = ((net.max_batches - i) / ngpus) * (what_time_is_it_now() - start) / 60 / 60;
@@ -1280,7 +1285,7 @@ void demo_classifier(char *datacfg, char *cfgfile, char *weightfile, int cam_ind
     int frame_counter = 0;
 
     while(1){
-        struct timeval tval_before, tval_after, tval_result;
+        struct timeval tval_before /*, tval_after, tval_result */;
         gettimeofday(&tval_before, NULL);
 
         //image in = get_image_from_stream(cap);

src/connected_layer.c

@@ -336,12 +336,6 @@ void forward_connected_layer_gpu(connected_layer l, network_state state)
 {
     fill_ongpu(l.outputs*l.batch, 0, l.output_gpu, 1);
 
-//    int m = l.batch;
-//    int k = l.inputs;
-//    int n = l.outputs;
-//    float * a = state.input;
-//    float * b = l.weights_gpu;
-//    float * c = l.output_gpu;
 #ifdef CUDNN
     //float one = 1;    // alpha[0], beta[0]
     float alpha = 1, beta = 0;
@@ -360,6 +354,12 @@ void forward_connected_layer_gpu(connected_layer l, network_state state)
         l.dstTensorDesc,
         l.output_gpu));
 #else // CUDNN
+    int m = l.batch;
+    int k = l.inputs;
+    int n = l.outputs;
+    float * a = state.input;
+    float * b = l.weights_gpu;
+    float * c = l.output_gpu;
     gemm_ongpu(0,1,m,n,k,1,a,k,b,k,1,c,n);
 #endif // CUDNN
 

src/data.c

@@ -1075,7 +1075,7 @@ data load_data_detection(int n, char **paths, int m, int w, int h, int c, int bo
     d.X.vals = (float**)xcalloc(d.X.rows, sizeof(float*));
     d.X.cols = h*w*c;
 
-    float r1 = 0, r2 = 0, r3 = 0, r4 = 0, r_scale = 0;
+    float r1 = 0, r2 = 0, r3 = 0, r4 = 0; //, r_scale = 0;
     float resize_r1 = 0, resize_r2 = 0;
     float dhue = 0, dsat = 0, dexp = 0, flip = 0, blur = 0;
     int augmentation_calculated = 0, gaussian_noise = 0;
@@ -1134,7 +1134,7 @@ data load_data_detection(int n, char **paths, int m, int w, int h, int c, int bo
                     flip = use_flip ? random_gen() % 2 : 0;
                 }
 
-                r_scale = random_float();
+                //r_scale = random_float();
 
                 if (!contrastive || contrastive_color || i % 2 == 0)
                 {
@@ -1288,7 +1288,7 @@ data load_data_detection(int n, char **paths, int m, int w, int h, int c, int bo
                 const int bot_shift = min_val_cmp(h - cut_y[i], max_val_cmp(0, (-pbot*h / oh)));
 
 
-                int k, x, y;
+                int k, y;
                 for (k = 0; k < c; ++k) {
                     for (y = 0; y < h; ++y) {
                         int j = y*w + k*w*h;

src/demo.c

@@ -97,7 +97,7 @@ void *detect_in_thread(void *ptr)
             this_thread_yield();
         }
 
-        layer l = net.layers[net.n - 1];
+        // layer l = net.layers[net.n - 1];
         float *X = det_s.data;
         //float *prediction =
         network_predict(net, X);

src/detector.c

@@ -61,7 +61,9 @@ void train_detector(char *datacfg, char *cfgfile, char *weightfile, int *gpus, i
     char *base = basecfg(cfgfile);
     printf("%s\n", base);
     float avg_loss = -1;
-    // float avg_contrastive_acc = 0;
+#ifdef OPENCV
+    float avg_contrastive_acc = 0;
+#endif
     network* nets = (network*)xcalloc(ngpus, sizeof(network));
 
     srand(time(0));
@@ -324,7 +326,9 @@ void train_detector(char *datacfg, char *cfgfile, char *weightfile, int *gpus, i
         }
         printf("\n %d: %f, %f avg loss, %f rate, %lf seconds, %d images, %f hours left\n", iteration, loss, avg_loss, get_current_rate(net), (what_time_is_it_now() - time), iteration*imgs, avg_time);
         fflush(stdout);
-
+#ifdef OPENCV
+	int draw_precision = 0;
+#endif
         if (calc_map && (iteration >= next_map_calc || iteration == net.max_batches)) {
             if (l.random) {
                 printf("Resizing to initial size: %d x %d ", init_w, init_h);
@@ -370,8 +374,9 @@ void train_detector(char *datacfg, char *cfgfile, char *weightfile, int *gpus, i
                 sprintf(buff, "%s/%s_best.weights", backup_directory, base);
                 save_weights(net, buff);
             }
-
-            // draw_precision = 1;
+#ifdef OPENCV
+            draw_precision = 1;
+#endif
         }
         time_remaining = ((net.max_batches - iteration) / ngpus)*(what_time_is_it_now() - time + load_time) / 60 / 60;
         // set initial value, even if resume training from 10000 iteration
@@ -1799,14 +1804,14 @@ void draw_object(char *datacfg, char *cfgfile, char *weightfile, char *filename,
     }
 
     srand(2222222);
-    char buff[256];
+    char buff[256] = { 0 };
     char *input = buff;
 
     int j;
     float nms = .45;    // 0.4F
     while (1) {
         if (filename) {
-            strncpy(input, filename, 256);
+            strncpy(input, filename, sizeof buff - 1);
             if (strlen(input) > 0)
                 if (input[strlen(input) - 1] == 0x0d) input[strlen(input) - 1] = 0;
         }
@@ -1868,7 +1873,7 @@ void draw_object(char *datacfg, char *cfgfile, char *weightfile, char *filename,
             float avg_loss = get_network_cost(net);
             draw_train_loss(windows_name, img, img_size, avg_loss, max_img_loss, iteration, it_num, 0, 0, "mAP%", 0, dont_show, 0, 0);
 
-            float inv_loss = 1.0 / max_val_cmp(0.01, avg_loss);
+            //float inv_loss = 1.0 / max_val_cmp(0.01, avg_loss);
             //net.learning_rate = *lr_set * inv_loss;
 
             if (*boxonly) {

src/http_stream.cpp

@@ -471,7 +471,7 @@ class MJPG_sender
 #endif
             if (s == sock) // request on master socket, accept and send main header.
             {
-                SOCKADDR_IN address = { 0 };
+                SOCKADDR_IN address = { };
                 SOCKET      client = ::accept(sock, (SOCKADDR*)&address, &addrlen);
                 if (client == SOCKET_ERROR)
                 {

src/image_opencv.cpp

@@ -942,13 +942,15 @@ extern "C" void draw_detections_cv_v3(mat_cv* mat, detection *dets, int num, flo
                 float red = get_color(2, offset, classes);
                 float green = get_color(1, offset, classes);
                 float blue = get_color(0, offset, classes);
+#if 0
                 float rgb[3];
 
                 //width = prob*20+2;
 
                 rgb[0] = red;
                 rgb[1] = green;
                 rgb[2] = blue;
+#endif
                 box b = dets[i].bbox;
                 if (std::isnan(b.w) || std::isinf(b.w)) b.w = 0.5;
                 if (std::isnan(b.h) || std::isinf(b.h)) b.h = 0.5;
@@ -1416,18 +1418,18 @@ extern "C" void cv_draw_object(image sized, float *truth_cpu, int max_boxes, int
 
     int it_trackbar_value = 200;
     std::string const it_trackbar_name = "iterations";
-    int it_tb_res = cv::createTrackbar(it_trackbar_name, window_name, &it_trackbar_value, 1000);
+    (void)cv::createTrackbar(it_trackbar_name, window_name, &it_trackbar_value, 1000);
 
     int lr_trackbar_value = 10;
     std::string const lr_trackbar_name = "learning_rate exp";
-    int lr_tb_res = cv::createTrackbar(lr_trackbar_name, window_name, &lr_trackbar_value, 20);
+    (void)cv::createTrackbar(lr_trackbar_name, window_name, &lr_trackbar_value, 20);
 
     int cl_trackbar_value = 0;
     std::string const cl_trackbar_name = "class_id";
-    int cl_tb_res = cv::createTrackbar(cl_trackbar_name, window_name, &cl_trackbar_value, classes-1);
+    (void)cv::createTrackbar(cl_trackbar_name, window_name, &cl_trackbar_value, classes-1);
 
     std::string const bo_trackbar_name = "box-only";
-    int bo_tb_res = cv::createTrackbar(bo_trackbar_name, window_name, boxonly, 1);
+    (void)cv::createTrackbar(bo_trackbar_name, window_name, boxonly, 1);
 
     int i = 0;
 
@@ -1440,7 +1442,7 @@ extern "C" void cv_draw_object(image sized, float *truth_cpu, int max_boxes, int
         if (pressed_key == 27 || pressed_key == 1048603) break;// break;  // ESC - save & exit
 
         frame_clone = frame.clone();
-        char buff[100];
+        //char buff[100];
         std::string lr_value = "learning_rate = " + std::to_string(1.0 / pow(2, lr_trackbar_value));
         cv::putText(frame_clone, lr_value, cv::Point2i(10, 20), cv::FONT_HERSHEY_SIMPLEX, 0.5, cv::Scalar(10, 50, 10), 3);
         cv::putText(frame_clone, lr_value, cv::Point2i(10, 20), cv::FONT_HERSHEY_SIMPLEX, 0.5, cv::Scalar(20, 120, 60), 2);