ifm

Representations trained with contrastive learning are adept at solving various vision tasks including classification, object detection, instance segmentation, and more [ 5, 15, 44 ].