Image Classification on Stanford Dogs
94.1AccuracyOSCS-SupCon
Evaluation results
| Method | Links | ||
|---|---|---|---|
| OSCS-SupConBackbone=ConvNeXt-tiny2026.05 | 94.1 | — | |
| SCS-SupConBackbone=ConvNeXt-tiny2025.12 | 93.8 | — | |
| CS-SupCon w. ov.Backbone=ConvNeXt-tiny2025.12 | 93.1 | — | |
| CS-SupCon w. ov.Backbone=ConvNeXt-tiny2026.05 | 93.1 | — | |
| CSTCNBackbone=ConvNeXt-tiny2025.12 | 93 | — | |
| PCLBackbone=ConvNeXt-tiny2025.12 | 92.9 | — | |
| TimeSCLBackbone=ConvNeXt-tiny2025.12 | 92.9 | — | |
| SupConBackbone=ConvNeXt-tiny2025.12 | 92.8 | — | |
| CS-SupConBackbone=ConvNeXt-tiny2025.12 | 92.8 | — | |
| SupConBackbone=ConvNeXt-tiny2026.05 | 92.8 | — | |
| CS-SupConBackbone=ConvNeXt-tiny2026.05 | 92.8 | — | |
| PaCoBackbone=ConvNeXt-tiny2025.12 | 92.7 | — | |
| CSA-RSICBackbone=ConvNeXt-tiny2025.12 | 92.7 | — | |
| DACLBackbone=ConvNeXt-tiny2025.12 | 92.4 | — | |
| BaselineBackbone=ConvNeXt-tiny2025.12 | 92.1 | — | |
| Circle LossBackbone=ConvNeXt-tiny2025.12 | 92.1 | — | |
| BaselineBackbone=ConvNeXt-tiny2026.05 | 92.1 | — | |
| ARCBackbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.252023.10 | 91.9 | — | |
| FNCLBackbone=ConvNeXt-tiny2025.12 | 91.9 | — | |
| BiasBackbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.282023.10 | 91.2 | — | |
| LoRABackbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.442023.10 | 91 | — | |
| VPT-ShallowBackbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.252023.10 | 90.7 | — | |
| VPT-DeepBackbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.852023.10 | 90.2 | — | |
| ARCattBackbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.222023.10 | 90.1 | — | |
| AdapterBackbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.412023.10 | 89.8 | — | |
| SSF*Backbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.39, Data Augmentation=Advanced (cutmix, mixup, label smoothing)2023.10 | 89.6 | — | |
| Full fine-tuningBackbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=85.982023.10 | 89.4 | — | |
| ARC*Backbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.25, Data Augmentation=Advanced (cutmix, mixup, label smoothing)2023.10 | 89.1 | — | |
| BYOLBackbone=ConvNeXt-tiny2025.12 | 88.2 | — | |
| BYOLBackbone=ConvNeXt-tiny2026.05 | 88.2 | — | |
| SSFBackbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.39, Data Augmentation=Basic2023.10 | 87.7 | — | |
| OSCS-SupConBackbone=TinyViT (5M)2026.05 | 87 | — | |
| SCS-SupConBackbone=TinyViT (5M)2025.12 | 86.6 | — | |
| Linear probingBackbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.182023.10 | 86.2 | — | |
| CS-SupCon w. ov.Backbone=TinyViT (5M)2025.12 | 86.2 | — | |
| CS-SupCon w. ov.Backbone=TinyViT (5M)2026.05 | 86.2 | — | |
| SimCLRBackbone=ConvNeXt-tiny2025.12 | 86.1 | — | |
| SimCLRBackbone=ConvNeXt-tiny2026.05 | 86.1 | — | |
| CSTCNBackbone=TinyViT (5M)2025.12 | 86 | — | |
| CSA-RSICBackbone=TinyViT (5M)2025.12 | 86 | — | |
| CS-SupConBackbone=TinyViT (5M)2025.12 | 86 | — | |
| CS-SupConBackbone=TinyViT (5M)2026.05 | 86 | — | |
| PCLBackbone=TinyViT (5M)2025.12 | 85.8 | — | |
| TimeSCLBackbone=TinyViT (5M)2025.12 | 85.8 | — | |
| PaCoBackbone=TinyViT (5M)2025.12 | 85.7 | — | |
| FNCLBackbone=TinyViT (5M)2025.12 | 85.6 | — | |
| SupConBackbone=TinyViT (5M)2025.12 | 85.5 | — | |
| SupConBackbone=TinyViT (5M)2026.05 | 85.5 | — | |
| Circle LossBackbone=TinyViT (5M)2025.12 | 85.4 | — | |
| DACLBackbone=TinyViT (5M)2025.12 | 84.8 | — | |
| VT-FSLLearning Paradigm=Training-based2026.03 | 84.63 | — | |
| SARELearning Paradigm=Training-based2026.03 | 84.29 | — | |
| BaselineBackbone=TinyViT (5M)2025.12 | 83.4 | — | |
| BaselineBackbone=TinyViT (5M)2026.05 | 83.4 | — | |
| EvoAug (Learned Clustering)Model=ViT-Small, Evaluation Protocol=5-way, 1-shot2026.02 | 79.7 | — | |
| EvoAug (Learned Clustering)Model=ResNet50, Evaluation Protocol=5-way, 1-shot2026.02 | 78.86 | — | |
| RandAugmentModel=ViT-Small, Evaluation Protocol=5-way, 1-shot2026.02 | 78.47 | — | |
| BYOLBackbone=TinyViT (5M)2025.12 | 78.3 | — | |
| BYOLBackbone=TinyViT (5M)2026.05 | 78.3 | — | |
| L2TTeacher architecture=ResNet34, Student architecture=ResNet18, Source domain=ImageNet, Number of experiment repeats=32021.02 | 78.08 | — | |
| AutoAugmentModel=ViT-Small, Evaluation Protocol=5-way, 1-shot2026.02 | 77.83 | — | |
| SimCLRBackbone=TinyViT (5M)2025.12 | 77.6 | — | |
| SimCLRBackbone=TinyViT (5M)2026.05 | 77.6 | — | |
| IOTAShot=16-shot, Backbone=ViT-B/162026.01 | 77.51 | — | |
| AutoAugmentModel=ResNet50, Evaluation Protocol=5-way, 1-shot2026.02 | 77.22 | — | |
| Random TreeModel=ResNet50, Evaluation Protocol=5-way, 1-shot2026.02 | 76.58 | — | |
| AFDTeacher architecture=ResNet34, Student architecture=ResNet18, Source domain=ImageNet, Number of experiment repeats=3, Beta=1,0002021.02 | 76.06 | — | |
| RandAugmentModel=ResNet50, Evaluation Protocol=5-way, 1-shot2026.02 | 75.86 | — | |
| NoOp / Classical TreeModel=ResNet50, Evaluation Protocol=5-way, 1-shot2026.02 | 75.79 | — | |
| Naive BaselineModel=ViT-Small, Evaluation Protocol=5-way, 1-shot2026.02 | 75.55 | — | |
| MapleShot=16-shot, Backbone=ViT-B/162026.01 | 75.08 | — | |
| KD+ATS+EnsTeacher=RNX101-32-8d, Student=MV22022.10 | 74.67 | 63.54 | |
| KgCoOpShot=16-shot, Backbone=ViT-B/162026.01 | 74.59 | — | |
| TCPShot=16-shot, Backbone=ViT-B/162026.01 | 74.17 | — | |
| ProTextShot=16-shot, Backbone=ViT-B/162026.01 | 73.46 | — | |
| FineDeficsLearning Paradigm=Training-based2026.03 | 73.36 | — | |
| KD+ATS+EnsTeacher=RNX101-32-8d, Student=SFV22022.10 | 73.22 | 63.54 | |
| KD+ATSTeacher=RNX101-32-8d, Student=MV22022.10 | 73.16 | 60.88 | |
| LP++Shot=16-shot, Backbone=ViT-B/162026.01 | 73.08 | — | |
| ResKDTeacher=RNX101-32-8d, Student=MV22022.10 | 72.85 | 62.26 | |
| CoOpShot=16-shot, Backbone=ViT-B/162026.01 | 72.65 | — | |
| EnsTeacher=RNX101-32-8d, Student=MV22022.10 | 72.53 | 61.82 | |
| MapleShot=8-shot, Backbone=ViT-B/162026.01 | 72.45 | — | |
| ST-KDTeacher=RNX101-32-8d, Student=MV22022.10 | 72.06 | 59.77 | |
| ProTextShot=8-shot, Backbone=ViT-B/162026.01 | 71.77 | — | |
| EnsTeacher=RNX101-32-8d, Student=SFV22022.10 | 71.65 | 61.82 | |
| TCPShot=8-shot, Backbone=ViT-B/162026.01 | 71.61 | — | |
| ESKDTeacher=RNX101-32-8d, Student=MV22022.10 | 71.56 | 59.01 | |
| Tmax-pooling dilation=12026.01 | 71.5 | — | |
| KgCoOpShot=8-shot, Backbone=ViT-B/162026.01 | 71.43 | — | |
| KDTeacher=RNX101-32-8d, Student=MV22022.10 | 71.25 | 58.89 | |
| Co-CoOpShot=16-shot, Backbone=ViT-B/162026.01 | 71.15 | — | |
| LP++Shot=8-shot, Backbone=ViT-B/162026.01 | 71.05 | — | |
| KD+ATSTeacher=RNX101-32-8d, Student=SFV22022.10 | 70.92 | 60.88 | |
| ResKDTeacher=RNX101-32-8d, Student=SFV22022.10 | 70.73 | 62.26 | |
| TAKDTeacher=RNX101-32-8d, Student=MV22022.10 | 70.61 | 58.87 | |
| ProGradShot=16-shot, Backbone=ViT-B/162026.01 | 70.33 | — | |
| Naive BaselineModel=ResNet50, Evaluation Protocol=5-way, 1-shot2026.02 | 70.3 | — | |
| SimSiam + AugSelfFine-tuning strategy=L2SP, Pre-trained on=ImageNet100, Backbone=ResNet-502021.11 | 70.26 | — | |
| SCKDTeacher=RNX101-32-8d, Student=MV22022.10 | 70.13 | 59.04 |