List your company

Semi-supervised Image Classification on ImageNet 1k (1%)

77.3Top-1 AccI-JEPA
23.32437.33751.3565.363Nov 23, 2020Apr 24, 2021Sep 24, 2021Feb 23, 2022Jul 26, 2022Dec 25, 2022May 27, 2023

Evaluation results

MethodLinks
I-JEPAArch.=ViT-H/16, Resolution=448x448, Epochs=300, View Data Augmentation=without2023.0177.3—
MSN + IbM2Backbone=ViT-B/4, IbM2=true2023.0576.1—
MSNArch.=ViT-B/4, Epochs=300, View Data Augmentation=extra2023.0175.7—
MSN + IbM2Backbone=ViT-L/7, IbM2=true2023.0575.6—
MSNBackbone=ViT-B/4, IbM2=false2023.0575.4—
MSNBackbone=ViT-L/7, IbM2=false2023.0574.8—
data2vecArch.=ViT-L/16, Epochs=1600, View Data Augmentation=without2023.0173.3—
I-JEPAArch.=ViT-H/14, Epochs=300, View Data Augmentation=without2023.0173.3—
MAEArch.=ViT-H/14, Epochs=1600, View Data Augmentation=without2023.0171.5—
BYOLArch.=R200 (x2), Param.=250M, Repr.=40962021.0571.289.5
BYOLArch.=RN200 (2x), Epochs=800, View Data Augmentation=extra2023.0171.2—
DINO + IbM2Backbone=ViT-S/8, IbM2=true2023.0570.6—
SimCLR v2Arch.=RN151 (2x), Epochs=800, View Data Augmentation=extra2023.0170.2—
DINOBackbone=ViT-S/8, IbM2=false2023.0570.1—
DINOArch.=ViT-B/8, Epochs=300, View Data Augmentation=extra2023.0170—
iBOTArch.=ViT-B/16, Epochs=400, View Data Augmentation=extra2023.0169.7—
I-JEPAArch.=ViT-L/16, Epochs=600, View Data Augmentation=without2023.0169.4—
BYOLArch.=R50 (x4), Param.=375M, Repr.=81922021.0569.187.9
VICRegArch.=R200 (x2), Param.=250M, Repr.=40962021.0568.888.2
MSN + IbM2Backbone=ViT-S/16, IbM2=true2023.0568—
MSNBackbone=ViT-S/16, IbM2=false2023.0567.3—
MAEArch.=ViT-L/16, Epochs=1600, View Data Augmentation=without2023.0167.1—
DINO + IbM2Backbone=ViT-S/16, IbM2=true2023.0565.2—
DINOBackbone=ViT-S/16, IbM2=false2023.0564.7—
FNCBackbone=ResNet-50, Learning Type=Representation Learning / Contrastive2020.1163.785.3
SimCLRArch.=R50 (x4), Param.=375M, Repr.=81922021.056385.8
VICRegArch.=R50 (x2), Param.=93M, Repr.=40962021.0562.684.5
BYOLArch.=R50 (x2), Param.=93M, Repr.=40962021.0562.284.1
C-BYOLBackbone=ResNet-502022.0860.683.4
MoCov3 + IbM2Backbone=ViT-S/16, IbM2=true2023.0559.1—
SimCLRArch.=R50 (x2), Param.=93M, Repr.=40962021.0558.583
MoCov3Backbone=ViT-S/16, IbM2=false2023.0558.1—
SimCLR v2Backbone=ResNet-50, Learning Type=Representation Learning / Contrastive2020.1157.982.5
BYOL + IbM2Backbone=ResNet50, IbM2=true2023.0556.9—
NNCLRBackbone=ResNet-502022.0856.480.7
RényiCLBackbone=ResNet-502022.0856.480.6
BYOLBackbone=ResNet50, IbM2=false2023.0555.7—
Barlow TwinsBackbone=ResNet-502022.085579.2
SwAVBackbone=ResNet-502020.1153.978.5
SwAVBackbone=ResNet-502022.0853.978.5
BYOLBackbone=ResNet-502020.1153.278.4
BYOLBackbone=ResNet-502022.0853.278.4
SimCLR + IbM2Backbone=ResNet50, IbM2=true2023.0551.3—
SimCLRBackbone=ResNet50, IbM2=false2023.0550.5—
SimCLR v1Backbone=ResNet-50, Learning Type=Representation Learning / Contrastive2020.1148.375.5
SimCLRBackbone=ResNet-502022.0848.375.5
PIRLBackbone=ResNet-50, Learning Type=Representation Learning / Contrastive2020.1130.760.4
SupervisedBackbone=ResNet-502020.1125.456.4
SupervisedBackbone=ResNet-502022.0825.448.4
FixMatchBackbone=ResNet-50, Learning Type=Semi-supervised2020.11—71.5
PCLBackbone=ResNet-50, Learning Type=Representation Learning / Contrastive2020.11—75.6
UDABackbone=ResNet-50, Learning Type=Semi-supervised2020.11—68.8