Pushing the Limits of Self-Supervised ResNets: DeepMind’s ReLICv2 Beats Strong Supervised Baselines on ImageNet
A DeepMind research team proposes ReLICv2, which demonstrates for the first time that representations learned without labels can consistently outperform a strong, supervised baseline on ImageNet and even achieve comparable results to state-of-the-art self-supervised vision transformers (ViTs).