Image Classification on ImageNet 1K (1%)
80.7Top-1 AccSemi-SST
Evaluation results
| Method | Links | |
|---|---|---|
| Semi-SSTArchitecture Type=Transformer, Architecture=ViT-Huge, Param=632M2025.05 | 80.7 | |
| Semi-SSTArchitecture=ViT-Huge, Param=632M2025.05 | 80.7 | |
| Super-SSTArchitecture Type=Transformer, Architecture=ViT-Huge, Param=632M2025.05 | 80.3 | |
| Super-SSTArchitecture=ViT-Huge, Param=632M2025.05 | 80.3 | |
| Semi-ViTArchitecture Type=Transformer, Architecture=ViT-Huge, Param=632M2025.05 | 80 | |
| Semi-ViTArchitecture Type=Transformer, Architecture=ViT-Large, Param=307M2025.05 | 77.3 | |
| Super-SSTArchitecture Type=Transformer, Architecture=ViT-Small, Param=22M, distilled=true2025.05 | 76.9 | |
| Super-SST (distilled)Architecture=ViT-Small, Param=22M2025.05 | 76.9 | |
| SimCLR v2Arch.=RN152w3+SK, Param.=795M2022.02 | 74.9 | |
| SimMatchV2Architecture Type=CNN, Architecture=ResNet-50, Param=26M2025.05 | 71.9 | |
| Semi-SSTArchitecture Type=Transformer, Architecture=ViT-Small, Param=22M2025.05 | 71.4 | |
| Semi-SSTArchitecture=ViT-Small, Param=22M2025.05 | 71.4 | |
| SimMatchV2Architecture Type=CNN, Architecture=ConvNeXt-T, Param=28M2025.05 | 71.3 | |
| BYOLArch.=RN200, Param.=250M2022.02 | 71.2 | |
| Semi-ViTArchitecture Type=Transformer, Architecture=ViT-Base, Param=86M2025.05 | 71 | |
| Super-SSTArchitecture Type=Transformer, Architecture=ViT-Small, Param=22M2025.05 | 70.4 | |
| Super-SSTArchitecture=ViT-Small, Param=22M2025.05 | 70.4 | |
| Semi-ViTArchitecture Type=Transformer, Architecture=ViT-Small, Param=22M2025.05 | 68 | |
| SimMatchArchitecture Type=CNN, Architecture=ResNet-50, Param=26M2025.05 | 67.2 | |
| PAWSArchitecture Type=CNN, Architecture=ResNet-50, Param=26M2025.05 | 66.5 | |
| CoMatchArchitecture Type=CNN, Architecture=ResNet-50, Param=26M2025.05 | 66 | |
| SimMatchV2Architecture Type=Transformer, Architecture=ViT-Small, Param=22M2025.05 | 63.7 | |
| EMANArchitecture Type=CNN, Architecture=ResNet-50, Param=26M2025.05 | 63 | |
| SEERArch.=RG10B, Param.=10B2022.02 | 62.4 | |
| USL-CLIPSSL Protocol=FixMatch, Features=CLIP2021.10 | 62.2 | |
| iBOTArch.=ViT-S/162021.11 | 61.9 | |
| USL-MoCoSSL Protocol=FixMatch, Features=MoCov22021.10 | 61.6 | |
| Stratified SelectionSSL Protocol=FixMatch2021.10 | 60.9 | |
| SEERArch.=RG256, Param.=1.5B2022.02 | 60.5 | |
| DINOArch.=ViT-S/162021.11 | 60.3 | |
| SimCLRv2Arch.=RN50, self-distillation=true2021.11 | 60 | |
| SimCLRv2Architecture Type=CNN, Architecture=ResNet-50, Param=26M2025.05 | 60 | |
| Random SelectionSSL Protocol=FixMatch2021.10 | 58.8 | |
| SimCLRv2Arch.=RN50, self-distillation=false2021.11 | 57.9 | |
| SEERArch.=RG128, Param.=693M2022.02 | 57.5 | |
| SwAVArch.=RN502021.11 | 53.9 | |
| SwAVArch.=RN50, Param.=24M2022.02 | 53.9 | |
| BYOLArch.=RN502021.11 | 53.2 | |
| USL-CLIPSSL Protocol=SimCLRv2, Features=CLIP2021.10 | 52.6 | |
| Stratified SelectionSSL Protocol=SimCLRv22021.10 | 52 | |
| USL-MoCoSSL Protocol=SimCLRv2, Features=MoCov22021.10 | 51.5 | |
| DMAE-BPre-training epochs=100, Supervised training / fine-tuning epochs=1002022.08 | 50.3 | |
| Random SelectionSSL Protocol=SimCLRv22021.10 | 49.7 | |
| MAE-BPre-training epochs=1600, Supervised training / fine-tuning epochs=1002022.08 | 49.6 | |
| SimCLRArch.=RN50, Param.=24M2022.02 | 48.3 | |
| DeiT-B-Hard DistillationPre-training epochs=1600, Supervised training / fine-tuning epochs=1002022.08 | 37.3 | |
| DeiT-B-Soft DistillationPre-training epochs=1600, Supervised training / fine-tuning epochs=1002022.08 | 36 | |
| Supervised Feature AlignmentPre-training epochs=1600, Supervised training / fine-tuning epochs=1002022.08 | 34.2 | |
| MAE-BPre-training epochs=100, Supervised training / fine-tuning epochs=1002022.08 | 33.9 |