Rethinking the Reference-based Distinctive Image Captioning | Read Paper on Bytez