ResearchBenchmarksImage Classification on ImageNet Occlusion (val)For AI assistantsMarkdownContent indexImage Classification on ImageNet Occlusion (val)Follow65.9S_{.5}iBOT64.65264.97665.365.624Nov 15, 2021Evaluation resultsMethodMethodLinksS_{.5}NS_{.5}iBOTBackbone=ViT-S/16, Pre...Backbone=ViT-S/16, Pre-training epochs=800, Evaluation protocol=Linear Probing, Information loss ratio=0.52021.1165.943.4DINOBackbone=ViT-S/16, Pre...Backbone=ViT-S/16, Pre-training epochs=800, Evaluation protocol=Linear Probing, Information loss ratio=0.52021.1164.742