Fetching the paper…

PerceptionCLIP: Visual Classification by Inferring and Conditioning on Contexts · Around