optimize yolov3 darknet53

official/cv/yolov3_darknet53/src/yolo.py

@@ -211,8 +211,6 @@ class DetectionBlock(nn.Cell):
 
         box_xy = prediction[:, :, :, :, :2]
         box_wh = prediction[:, :, :, :, 2:4]
-        box_confidence = prediction[:, :, :, :, 4:5]
-        box_probs = prediction[:, :, :, :, 5:]
 
         # gridsize1 is x
         # gridsize0 is y
@@ -220,11 +218,13 @@ class DetectionBlock(nn.Cell):
                                                                                 grid_size[0])), ms.float32)
         # box_wh is w->h
         box_wh = ops.Exp()(box_wh) * self.anchors / input_shape
-        box_confidence = self.sigmoid(box_confidence)
-        box_probs = self.sigmoid(box_probs)
 
         if self.conf_training:
             return grid, prediction, box_xy, box_wh
+        box_confidence = prediction[:, :, :, :, 4:5]
+        box_probs = prediction[:, :, :, :, 5:]
+        box_confidence = self.sigmoid(box_confidence)
+        box_probs = self.sigmoid(box_probs)
         return self.concat((box_xy, box_wh, box_confidence, box_probs))