diff --git a/example/ssd_coco2017/dataset.py b/example/ssd_coco2017/dataset.py
index c725c14e006a806fd817fecc85722aac38c4f4e8..b88b22c8626d520d142a9570b345f359a4ef3468 100644
--- a/example/ssd_coco2017/dataset.py
+++ b/example/ssd_coco2017/dataset.py
@@ -137,7 +137,7 @@ def ssd_bboxes_encode(boxes):
     num_match_num = np.array([len(np.nonzero(t_label)[0])], dtype=np.int32)
     return bboxes, t_label.astype(np.int32), num_match_num
 
-def ssd_bboxes_decode(boxes, index, image_shape):
+def ssd_bboxes_decode(boxes, index):
     """Decode predict boxes to [x, y, w, h]"""
     boxes_t = boxes[index]
     default_boxes_t = default_boxes[index]
diff --git a/example/ssd_coco2017/util.py b/example/ssd_coco2017/util.py
index 0b276c69f2304001fe0bc4fb1f99b979eebc68ed..6e1028537576df231b6d820096eeb148858c3bea 100644
--- a/example/ssd_coco2017/util.py
+++ b/example/ssd_coco2017/util.py
@@ -110,14 +110,12 @@ def metrics(pred_data):
         pred_boxes = sample['boxes']
         boxes_scores = sample['box_scores']
         annotation = sample['annotation']
-        image_shape = sample['image_shape']
 
         annotation = np.squeeze(annotation, axis=0)
-        image_shape = np.squeeze(image_shape, axis=0)
 
         pred_labels = np.argmax(boxes_scores, axis=-1)
         index = np.nonzero(pred_labels)
-        pred_boxes = ssd_bboxes_decode(pred_boxes, index, image_shape)
+        pred_boxes = ssd_bboxes_decode(pred_boxes, index)
 
         pred_boxes = pred_boxes.clip(0, 1)
         boxes_scores = np.max(boxes_scores, axis=-1)
diff --git a/example/yolov3_coco2017/train.py b/example/yolov3_coco2017/train.py
index 6a9729f2000d00546d1a88dba96c25f11afdda15..999cbdc62d4d72d448937b1059c58c1cbc3127b5 100644
--- a/example/yolov3_coco2017/train.py
+++ b/example/yolov3_coco2017/train.py
@@ -60,7 +60,7 @@ def init_net_param(net, init='ones'):
             p.set_parameter_data(initializer(init, p.data.shape(), p.data.dtype()))
 
 
-if __name__ == '__main__':
+def main():
     parser = argparse.ArgumentParser(description="YOLOv3 train")
     parser.add_argument("--only_create_dataset", type=bool, default=False, help="If set it true, only create "
                                                                                 "Mindrecord, default is false.")
@@ -153,3 +153,6 @@ if __name__ == '__main__':
             dataset_sink_mode = True
         print("Start train YOLOv3, the first epoch will be slower because of the graph compilation.")
         model.train(args_opt.epoch_size, dataset, callbacks=callback, dataset_sink_mode=dataset_sink_mode)
+
+if __name__ == '__main__':
+    main()