CAPE: Camera View Position Embedding for Multi-View 3D Object Detection | Read Paper on Bytez