image-to-3d-bbox
image-to-3d-bbox copied to clipboard
It predicts all the same orientation angle
Thanks for the wonderful work.
I've tried training the model and it seemed successfully worked. But when I ran the 3D orientation regression part, it appeared that always same angle is predicted. I found a weight that you provided from one of the issues, and it seems it correctly predicts the angle when I used the weight. Thus, it is considered the training has a problem.
Have you ever had this kind of issue? Thank you in advance.