Opencv4.6 Problems

[GinkgoX]

when I test the pyhton script: File "...\yolact-opencv-dnn-cpp-python-main\main_yolact.py", line 152, in decode priors[:, 2:] * np.exp(loc[:, 2:] * self.variances[1]), ValueError: operands could not be broadcast together with shapes (10597,2) (10597,79)

the dimension is not match

When I test the cpp script: the result as follows:

[74284853]

I have the same problem as you. Is there any solution，please？

[74284853]

？

[by011]

Same problem

[74284853]

Same problem

The problem is the same. I solved it by modifying the opencv version to 4.5.5.

[by011]

tq very much

[by011]

Same problem

The problem is the same. I solved it by modifying the opencv version to 4.5.5.

May i know how to downgrade opencv 4.6.0 to 4.5.5 ? I tried conda uninstall opencv and conda install opencv 4.5.5 but didnt work

[74284853]

Same problem

The problem is the same. I solved it by modifying the opencv version to 4.5.5.

May i know how to downgrade opencv 4.6.0 to 4.5.5 ? I tried conda uninstall opencv and conda install opencv 4.5.5 but didnt work

pip install opencv-python==4.5.1.18 (As long as the version number is 4.5.x）

[74284853]

(first):pip uninstall opencv-python

[by011]

(first):pip uninstall opencv-python

will try it out, thank you very much

edited: yes, it worked, tq very much, but the outcome is bad compare to original pytorch version of YOLACT.