Please explain this error: ValueError operands could not be broadcast together with shapes (6400,) (1600,)

[Abdul Samad]

I'm running main.py and getting this error.

yolov8-face-landmarks-opencv-dnn on main via 🐍 v3.8.10
❯ py main.py
Traceback (most recent call last):
  File "main.py", line 171, in <module>
    boxes, scores, classids, kpts = YOLOv8_face_detector.detect(srcimg)
  File "main.py", line 75, in detect
    det_bboxes, det_conf, det_classid, landmarks = self.post_process(outputs, scale_h, scale_w, padh, padw)
  File "main.py", line 92, in post_process
    bbox = self.distance2bbox(anchors, bbox_pred, max_shape=(self.input_height, self.input_width)) * stride
  File "main.py", line 134, in distance2bbox
    x1 = points[:, 0] - distance[:, 0]
ValueError: operands could not be broadcast together with shapes (6400,) (1600,)

[hpc203]

yolov8-face-landmarks-opencv-dnn on main via 🐍 v3.8.10
❯ py main.py
Traceback (most recent call last):
  File "main.py", line 171, in <module>
    boxes, scores, classids, kpts = YOLOv8_face_detector.detect(srcimg)
  File "main.py", line 75, in detect
    det_bboxes, det_conf, det_classid, landmarks = self.post_process(outputs, scale_h, scale_w, padh, padw)
  File "main.py", line 92, in post_process
    bbox = self.distance2bbox(anchors, bbox_pred, max_shape=(self.input_height, self.input_width)) * stride
  File "main.py", line 134, in distance2bbox
    x1 = points[:, 0] - distance[:, 0]
ValueError: operands could not be broadcast together with shapes (6400,) (1600,)

6400=80x80，1600=40x40，说明特征图的尺寸跟grid_anchor的尺寸不一致，原因是在python版本的opencv4.7.0里，forward函数输出特征图的尺寸，分别是40x40，20x20，80x80，我在main.py的注释里有说明这一点的。不过我已经更新了main.py，在初始化函数里生成的anchors是一个字典，字典的key是下采样尺度stride，这样在后处理函数里，根据特征图的尺寸跟输入分辨率计算出当前特征图的下采样尺度stride，根据当前的stride从anchors里提取到当前anchor，然后decode

[Abdul Samad]

非常感谢，它解决了错误。