I am trying to train detectnet within digits. I am using the Kitti format. I am wondering if training detectnet within digits uses all label informations like truncation or occlusion, or just using the bounding box informations. On my case using truncation and occlusion informations will help me more.
Thanks for the help.