Image Classification on ImageNet-1k 1% labels
75.7Top-1 AccMSN
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MSNArchitecture=ViT-B/4, Params.=86M2022.04 | 75.7 | — | |
| MSNArchitecture=ViT-L/7, Params.=304M2022.04 | 75.1 | — | |
| SimCLRv2Architecture=RN151+SK (3×), Params.=795M2022.04 | 74.9 | — | |
| BYOLArchitecture=RN200 (2×), Params.=250M2022.04 | 71.2 | — | |
| BYOLBackbone=RN200(2×)2026.04 | 71.2 | — | |
| SimCLRv2Backbone=RN152(2×)2026.04 | 70.2 | — | |
| DINOArchitecture=ViT-B/8, Params.=86M, Re-evaluated=true2022.04 | 70 | — | |
| DINOBackbone=ViT-B/82026.04 | 70 | — | |
| iBOTArchitecture=ViT-B/16, Params.=86M, Re-evaluated=true2022.04 | 69.7 | — | |
| iBOTBackbone=ViT-B/162026.04 | 69.7 | — | |
| ST-CBackbone=ResNet502026.04 | 69.1 | — | |
| ST-S+CBackbone=ResNet50 + Encodeur-Pyramidale-Style2026.04 | 68.8 | — | |
| MSNArchitecture=ViT-S/16, Params.=22M2022.04 | 67.2 | — | |
| PAWSArchitecture=RN50, Params.=24M2022.04 | 66.5 | — | |
| iBOTArchitecture=ViT-S/16, Params.=22M2022.04 | 65.9 | — | |
| I-JEPABackbone=ViT-S/162026.04 | 65 | — | |
| DINOArchitecture=ViT-S/16, Params.=22M2022.04 | 64.5 | — | |
| C-BYOLBackbone=ResNet-50, Evaluation Protocol=Semi-supervised2026.02 | 60.6 | 83.4 | |
| BYOL multi-taskBackbone=ResNet-50, Evaluation Protocol=Semi-supervised2026.02 | 60.4 | 84 | |
| ReLIC v2Backbone=ResNet-50, Evaluation Protocol=Semi-supervised2026.02 | 58.1 | 81.3 | |
| SimCLRv2Architecture=RN50, Params.=24M2022.04 | 57.9 | — | |
| NNCLRBackbone=ResNet-50, Evaluation Protocol=Semi-supervised2026.02 | 56.4 | 80.7 | |
| Barlow TwinsBackbone=ResNet-50, Evaluation Protocol=Semi-supervised2026.02 | 55 | 79.2 | |
| Barlow-Tw.Architecture=RN50, Params.=24M2022.04 | 55 | — | |
| SwAVBackbone=ResNet-50, Evaluation Protocol=Semi-supervised2026.02 | 53.9 | 78.5 | |
| BYOLBackbone=ResNet-50, Evaluation Protocol=Semi-supervised2026.02 | 53.2 | 78.4 | |
| SimCLRBackbone=ResNet-50, Evaluation Protocol=Semi-supervised2026.02 | 48.3 | 75.5 | |
| supervisedBackbone=ResNet-50, Evaluation Protocol=Semi-supervised2026.02 | 25.4 | 48.4 |