bbox_vote funtion problem

[Spiralpill]

你好
我在運行demo.py時，出現了錯誤：

Traceback (most recent call last):
  File "demo.py", line 252, in <module>
    test_oneimage()
  File "demo.py", line 245, in test_oneimage
    det = bbox_vote(det)
  File "demo.py", line 73, in bbox_vote
    det_accu_sum[:, 0:4] = np.sum(det_accu[:, 0:4], axis=0) / np.sum(det_accu[:, -1:])
  File "/usr/local/lib/python3.6/dist-packages/torch/tensor.py", line 367, in __rdiv__
    return self.reciprocal() * other
TypeError: mul(): argument 'other' (position 1) must be Tensor, not numpy.ndarray

我於前一行查看了det_accu：

print(det_accu)
[[884.7191162109375 687.6126098632812 996.62158203125 823.8545532226562
  0.9999959468841553]
 [884.49133 687.3231 996.8358 823.6817 tensor(1.0000)]
 [884.73834 686.1709 996.74756 822.86896 tensor(1.0000)]]

錯誤是否與det_accu[0]有關？
如果有關，這錯誤怎麼產生？
我該如何處理這問題？

[Spiralpill]

查看det_accu結果修正為：

print(det_accu)
[[884.7155303418695 687.6098228897172 996.6175426085247 823.8512140447128
  0.9999959468841553]
 [tensor(884.4860) tensor(687.3190) tensor(996.8298) tensor(823.6767)
  tensor(1.0000)]
 [tensor(884.7256) tensor(686.1610) tensor(996.7332) tensor(822.8571)
  tensor(1.0000)]]

[oyerst]

add 'score = score.cpu().numpy' after score definition.

[Spiralpill]

Thank oyerst, You solved my problem.
Answer:

while detections[0, i, j, 0] >= thresh:
                score = detections[0, i, j, 0]
                score = score.cpu().numpy()  #Plus here
                pt = (detections[0, i, j, 1:]*scale).cpu().numpy()
                coords = (pt[0], pt[1], pt[2], pt[3]) 
                det.append([pt[0], pt[1], pt[2], pt[3], score])
                j += 1

[Spiralpill]

score = score.cpu().numpy Add on or near line 110 aslam6892 <notifications@github.com> 於 2019年4月20日 週六 上午5:54寫道：

…

Can you please specify line number in 'demo.py' where we need to place this line score = score.cpu().numpy I also stuck in this line Thanks for help — You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub <https://github.com/TencentYoutuResearch/FaceDetection-DSFD/issues/17#issuecomment-485023495>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AI5DKXSGWMUGFF2B6ZOI3I3PRI5RVANCNFSM4HEFRRGA> .