no code implementations • 18 Nov 2021 • Xinnan Du, William Zhang, Jose M. Alvarez
In this paper, we propose a new light-weight self-supervised learning framework that could boost supervised learning performance with minimum additional computation cost.