List your company
ResearchFormerly WizWand

Image Classification on ImageNet 10% labels 1k (train val)

76Top-1 AccuracyDINO
58.00862.67967.3572.021Aug 4, 2024
Updated 8mo ago

Evaluation results

MethodLinks
DINOarch.=ViT-B/16, params=85M, Training Protocol=Semi-supervised2024.0876
BAMarch.=ViT-B/16, params=85M, Training Protocol=Semi-supervised2024.0875.9
MoCoV3arch.=ViT-B/16, params=85M, Training Protocol=Semi-supervised2024.0874.5
DINOarch.=ViT-S/16, params=21M, Training Protocol=Semi-supervised2024.0873.9
MAEarch.=ViT-B/16, params=85M, Training Protocol=Semi-supervised2024.0873.7
BAMarch.=ViT-S/16, params=21M, Training Protocol=Semi-supervised2024.0873.2
MoCoV3arch.=ViT-S/16, params=21M, Training Protocol=Semi-supervised2024.0858.7