The visualization showed nothing on the kitti dataset image

[steveice]

Hi authors,

I used both All-in-one fashion and visualization on the trained model on # ResNet-18 backbone and DLA-3D backbone. But none of them showed any bounding box in the kitti dataset image. However, if using the pre-trained dataset, there are showing the correct result with bounding box on the image. Can you provide any hints to fix this issue?