COPE: End-to-end trainable Constant Runtime Object Pose Estimation | Read Paper on Bytez