In the image branch, the downsample factor of image neck is asserted to 2 after dtransform and depthnet. But why?
If I changed the lidar voxel size and grid size the bev feature map size will change so the image bev feature map size should be changed accordingly.
I understand I can do that through changing xbound, ybound, but I can also do that througn changing downsample factor, so why is it set to be a fixed number?
Thank you for your interest in our project. This repository is no longer actively maintained, so we will be closing this issue. Please refer to the amazing implementation at MMDetection3D. Thank you again!
In the image branch, the downsample factor of image neck is asserted to 2 after dtransform and depthnet. But why? If I changed the lidar voxel size and grid size the bev feature map size will change so the image bev feature map size should be changed accordingly. I understand I can do that through changing xbound, ybound, but I can also do that througn changing downsample factor, so why is it set to be a fixed number?