"Bottleneck Transformers for Visual Recognition."

Aravind Srinivas et al. (2021)
a service of  Schloss Dagstuhl - Leibniz Center for Informatics