updates

detect.py

@@ -34,7 +34,7 @@ def detect(cfg, weights, images, output='output', img_size=416, conf_thres=0.3,
     classes = load_classes(parse_data_cfg('cfg/coco.data')['names'])  # Extracts class labels from file
     colors = [[random.randint(0, 255), random.randint(0, 255), random.randint(0, 255)] for _ in range(len(classes))]
 
-    for i, (path, img, img0) in enumerate(dataloader):
+    for i, (path, img, im0) in enumerate(dataloader):
         print("%g/%g '%s': " % (i + 1, len(dataloader), path), end='')
         t = time.time()
 
@@ -54,11 +54,10 @@ def detect(cfg, weights, images, output='output', img_size=416, conf_thres=0.3,
             if detections is not None:
                 save_img_path = os.path.join(output, path.split('/')[-1])
                 save_txt_path = save_img_path + '.txt'
-                img = img0
 
                 # The amount of padding that was added
-                pad_x = max(img.shape[0] - img.shape[1], 0) * (img_size / max(img.shape))
-                pad_y = max(img.shape[1] - img.shape[0], 0) * (img_size / max(img.shape))
+                pad_x = max(im0.shape[0] - im0.shape[1], 0) * (img_size / max(im0.shape))
+                pad_y = max(im0.shape[1] - im0.shape[0], 0) * (img_size / max(im0.shape))
                 # Image height and width after padding is removed
                 unpad_h = img_size - pad_y
                 unpad_w = img_size - pad_x
@@ -70,34 +69,29 @@ def detect(cfg, weights, images, output='output', img_size=416, conf_thres=0.3,
 
                 for x1, y1, x2, y2, conf, cls_conf, cls_pred in detections:
                     # Rescale coordinates to original dimensions
-                    box_h = ((y2 - y1) / unpad_h) * img.shape[0]
-                    box_w = ((x2 - x1) / unpad_w) * img.shape[1]
-                    y1 = (((y1 - pad_y // 2) / unpad_h) * img.shape[0]).round()
-                    x1 = (((x1 - pad_x // 2) / unpad_w) * img.shape[1]).round()
+                    box_h = ((y2 - y1) / unpad_h) * im0.shape[0]
+                    box_w = ((x2 - x1) / unpad_w) * im0.shape[1]
+                    y1 = (((y1 - pad_y // 2) / unpad_h) * im0.shape[0]).round()
+                    x1 = (((x1 - pad_x // 2) / unpad_w) * im0.shape[1]).round()
                     x2 = (x1 + box_w).round()
                     y2 = (y1 + box_h).round()
                     x1, y1, x2, y2 = max(x1, 0), max(y1, 0), max(x2, 0), max(y2, 0)
 
-                    # write to file
-                    if save_txt:
+                    if save_txt:  # Write to file
                         with open(save_txt_path, 'a') as file:
-                            file.write(('%g %g %g %g %g %g\n') % (x1, y1, x2, y2, cls_pred, cls_conf * conf))
+                            file.write('%g %g %g %g %g %g\n' % (x1, y1, x2, y2, cls_pred, cls_conf * conf))
 
-                    if save_images:
-                        # Add bbox to the image
+                    if save_images:  # Add bbox to the image
                         label = '%s %.2f' % (classes[int(cls_pred)], conf)
-                        plot_one_box([x1, y1, x2, y2], img, label=label, color=colors[int(cls_pred)])
+                        plot_one_box([x1, y1, x2, y2], im0, label=label, color=colors[int(cls_pred)])
 
-                if save_images:
-                    # Save generated image with detections
-                    cv2.imwrite(save_img_path, img)
+                if save_images:  # Save generated image with detections
+                    cv2.imwrite(save_img_path, im0)
 
         print(' Done. (%.3fs)' % (time.time() - t))
 
     if platform == 'darwin':  # MacOS
-        os.system('open ' + output)
-        os.system('open ' + save_img_path)
-
+        os.system('open ' + output + '&& open ' + save_img_path)
 
 
 if __name__ == '__main__':

test.py

@@ -27,8 +27,7 @@ def test(cfg, data_cfg, weights, batch_size=16, img_size=416, iou_thres=0.5, con
     model.to(device).eval()
 
     # Get dataloader
-    # dataset = load_images_with_labels(test_path)
-    # dataloader = torch.utils.data.DataLoader(dataset, batch_size=batch_size, shuffle=False, num_workers=n_cpus)
+    # dataloader = torch.utils.data.DataLoader(load_images_with_labels(test_path), batch_size=batch_size)  # pytorch
     dataloader = load_images_and_labels(test_path, batch_size=batch_size, img_size=img_size)
 
     mean_mAP, mean_R, mean_P = 0.0, 0.0, 0.0

train.py

@@ -45,8 +45,7 @@ def train(
     model = Darknet(cfg, img_size)
 
     # Get dataloader
-    dataloader = load_images_and_labels(train_path, batch_size=batch_size, img_size=img_size,
-                                        multi_scale=multi_scale, augment=True)
+    dataloader = load_images_and_labels(train_path, batch_size, img_size, multi_scale=multi_scale, augment=True)
 
     lr0 = 0.001
     if resume:

utils/datasets.py

@@ -41,7 +41,8 @@ class load_images():  # for inference
         assert img0 is not None, 'Failed to load ' + img_path
 
         # Padded resize
-        img, _, _, _ = letterbox(img0, height=self.height, color=(127.5, 127.5, 127.5))
+        img, ratio, padw, padh = letterbox(img0, height=self.height, color=(127.5, 127.5, 127.5))
+        print(ratio, padw, padh)
 
         # Normalize RGB
         img = img[:, :, ::-1].transpose(2, 0, 1)