Fetching the paper…

CO2: Consistent Contrast for Unsupervised Visual Representation Learning · Around