Foundation Models: A New Approach to Image Classification,ICCV21

mPLUG-2: A Modularized Multi-modal Foundation Model Across Text, Image and Video