Training an Open-Vocabulary Monocular 3D Detection Model without 3D Data | Read Paper on Bytez