DropMAE: Masked Autoencoders With Spatial-Attention Dropout for Tracking Tasks

Devs

DropMAE: Masked Autoencoders With Spatial-Attention Dropout for Tracking Tasks | Read Paper on Bytez