BYOL -The Alternative to Contrastive Self-Supervised Learning

GLIP: Introducing Language-Image Pre-Training to Object Detection