Cartinoe
Oscar: Object-Semantics Aligned Pre-training for Vision-Language Tasks ๋…ผ๋ฌธ ๋ฆฌ๋ทฐ