Tagged: 自监督

Align before Fuse: Vision and Language Representation Learning with Momentum Distillation 背景 VLP（Vis...