Image Classification on Oxford Flowers 102
99.56AccuracyCRATE-α
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| CRATE-αModel variant=L/142024.05 | 99.56 | — | — | — | — | |
| LaCLIPArchitecture=ViT-B/16, Pre-training Data=LAION-400M, Protocol=5-way 5-shot2023.05 | 99.5 | — | — | — | — | |
| ImageNet-1kModel=ViT-T, Type=SL2023.03 | 99.4 | — | — | — | — | |
| CLIPArchitecture=ViT-B/16, Pre-training Data=LAION-400M, Protocol=5-way 5-shot2023.05 | 99.4 | — | — | — | — | |
| CRATE-αModel variant=B/162024.05 | 99.27 | — | — | — | — | |
| LaCLIPArchitecture=ViT-B/32, Pre-training Data=LAION-400M, Protocol=5-way 5-shot2023.05 | 99.2 | — | — | — | — | |
| LaCLIPArchitecture=ViT-B/16, Pre-training Data=RedCaps, Protocol=5-way 5-shot2023.05 | 99.2 | — | — | — | — | |
| CaiT-M-36 ↑ 224Resolution=224, Crop-ratio=0.875, Training=Longer schedules2021.03 | 99.1 | — | — | — | — | |
| CLIP-LoRA + FMAShots=162025.10 | 99.1 | — | — | — | — | |
| CRATE-αModel variant=L/322024.05 | 99.01 | — | — | — | — | |
| CaiT-M-36 224Resolution=224, Crop-ratio=0.8752021.03 | 99 | — | — | — | — | |
| CaiT-S-36 ↑ 224Resolution=224, Crop-ratio=0.875, Training=Longer schedules2021.03 | 99 | — | — | — | — | |
| Grafit RegNetY-8GFFLOPs=8.0G, Pre-training=ImageNet2021.03 | 99 | — | — | — | — | |
| RCDB-21kModel=ViT-B, Type=FDSL2023.03 | 99 | — | — | — | — | |
| VisualAtom-21kModel=ViT-B, Type=FDSL2023.03 | 99 | — | — | — | — | |
| CLIPArchitecture=ViT-B/32, Pre-training Data=LAION-400M, Protocol=5-way 5-shot2023.05 | 99 | — | — | — | — | |
| CLIPArchitecture=ViT-B/16, Pre-training Data=RedCaps, Protocol=5-way 5-shot2023.05 | 99 | — | — | — | — | |
| iBOTBackbone=ViT-B/162022.03 | 98.9 | — | — | — | — | |
| MugsBackbone=ViT-B/162022.03 | 98.9 | — | — | — | — | |
| ExFractalDB-1kModel=ViT-T, Type=FDSL2023.03 | 98.9 | — | — | — | — | |
| RCDB-1kModel=ViT-T, Type=FDSL2023.03 | 98.9 | — | — | — | — | |
| VisualAtom-1kModel=ViT-T, Type=FDSL2023.03 | 98.9 | — | — | — | — | |
| EfficientNet-B7Resolution=224, Crop-ratio=0.8752021.03 | 98.8 | — | — | — | — | |
| CaiT-S-36 224Resolution=224, Crop-ratio=0.8752021.03 | 98.8 | — | — | — | — | |
| EfficientNet-B7FLOPs=37.3G, Pre-training=ImageNet2021.03 | 98.8 | — | — | — | — | |
| DINOBackbone=ViT-B/16, Pre-training=Self-supervised, Evaluation Protocol=finetuning2021.04 | 98.8 | — | — | — | — | |
| MugsBackbone=ViT-S/162022.03 | 98.8 | — | — | — | — | |
| DINOBackbone=ViT-B/162022.03 | 98.8 | — | — | — | — | |
| CRISPPEFT=CRISP (ours), Params(%)=21M2026.03 | 98.8 | — | — | — | — | |
| CeiT-SFLOPs=12.9G, Resolution=384x384, Pre-training=ImageNet2021.03 | 98.6 | — | — | — | — | |
| iBOTBackbone=ViT-S/162022.03 | 98.6 | — | — | — | — | |
| PASSModel=ViT-T, Type=SSL (DINO)2023.03 | 98.6 | — | — | — | — | |
| EfficientNet-B5FLOPs=10.3G, Pre-training=ImageNet2021.03 | 98.5 | — | — | — | — | |
| Deit-BFLOPs=52.8G, Resolution=384x384, Pre-training=ImageNet2021.03 | 98.5 | — | — | — | — | |
| DINOBackbone=ViT-S/16, Pre-training=Self-supervised, Evaluation Protocol=finetuning2021.04 | 98.5 | — | — | — | — | |
| DINOBackbone=ViT-S/162022.03 | 98.5 | — | — | — | — | |
| ImageNet-1kModel=ViT-T, Type=SSL (DINO)2023.03 | 98.5 | — | — | — | — | |
| DeiT-B 224Resolution=224, Crop-ratio=0.8752021.03 | 98.4 | — | — | — | — | |
| Deit-BFLOPs=17.3G, Resolution=224x224, Pre-training=ImageNet2021.03 | 98.4 | — | — | — | — | |
| SupervisedBackbone=ViT-B/16, Pre-training=Supervised, Evaluation Protocol=finetuning2021.04 | 98.4 | — | — | — | — | |
| Sup.Backbone=ViT-B/162022.03 | 98.4 | — | — | — | — | |
| LaCLIPArchitecture=ViT-B/16, Pre-training Data=CC12M, Protocol=5-way 5-shot2023.05 | 98.4 | — | — | — | — | |
| FractalDB-1kModel=ViT-T, Type=FDSL2023.03 | 98.3 | — | — | — | — | |
| Grafit ResNet-50FLOPs=4.1G, Pre-training=ImageNet2021.03 | 98.2 | — | — | — | — | |
| CeiT-SFLOPs=4.5G, Resolution=224x224, Pre-training=ImageNet2021.03 | 98.2 | — | — | — | — | |
| SupervisedBackbone=ViT-S/16, Pre-training=Supervised, Evaluation Protocol=finetuning2021.04 | 98.2 | — | — | — | — | |
| Sup.Backbone=ViT-S/162022.03 | 98.2 | — | — | — | — | |
| Skip TuningShot=16, Time (s)=962, Mem. (M)=5282024.12 | 98.17 | — | — | — | — | |
| SupervisedProtocol=Fine-tuned, Backbone=ResNet-50 (4x), Pre-training=ImageNet2020.02 | 98 | — | — | — | — | |
| BEITBackbone=ViT-B/162022.03 | 98 | — | — | — | — | |
| SFDATop-k=k=32022.07 | 97.99 | — | — | — | — | |
| CLIP-LoRAShots=162025.10 | 97.9 | — | — | — | — | |
| CeiT-TFLOPs=3.6G, Resolution=384x384, Pre-training=ImageNet2021.03 | 97.8 | — | — | — | — | |
| LaSLIPArchitecture=ViT-B/16, Pre-training Data=CC12M, Protocol=5-way 5-shot2023.05 | 97.8 | — | — | — | — | |
| CRATE-αModel variant=B/322024.05 | 97.77 | — | — | — | — | |
| MoCo-v3Backbone=ViT-B/162022.03 | 97.7 | — | — | — | — | |
| NLEEPTop-k=k=32022.07 | 97.65 | — | — | — | — | |
| SupervisedEvaluation protocol=Fine-tuned, Pre-training=ImageNet, Backbone=ResNet-502020.02 | 97.6 | — | — | — | — | |
| SimCLRProtocol=Fine-tuned, Backbone=ResNet-50 (4x), Pre-training=ImageNet2020.02 | 97.6 | — | — | — | — | |
| CLIPArchitecture=ViT-B/16, Pre-training Data=CC12M, Protocol=5-way 5-shot2023.05 | 97.6 | — | — | — | — | |
| ProSRCShot=16, Time (s)=14403, Mem. (M)=33732024.12 | 97.6 | — | — | — | — | |
| PLOT++Shots=162025.10 | 97.6 | — | — | — | — | |
| SFDATop-k=k=22022.07 | 97.59 | — | — | — | — | |
| LogMETop-k=k=32022.07 | 97.56 | — | — | — | — | |
| EsViTBackbone=Swin-T, Evaluation Protocol=Linear probe2021.06 | 97.4 | — | — | — | — | |
| SFDAcomTop-k=k=32022.07 | 97.35 | — | — | — | — | |
| SFDAcomTop-k=k=22022.07 | 97.24 | — | — | — | — | |
| CoOpShots=162025.10 | 97.2 | — | — | — | — | |
| EfficientViT-M5Throughput (imgs/s)=106212023.05 | 97.1 | — | — | — | — | |
| SLIPArchitecture=ViT-B/16, Pre-training Data=CC12M, Protocol=5-way 5-shot2023.05 | 97.1 | — | — | — | — | |
| CoOpShot=16, Time (s)=3879, Mem. (M)=45512024.12 | 97.07 | — | — | — | — | |
| NLEEPTop-k=k=22022.07 | 97.04 | — | — | — | — | |
| SimCLREvaluation protocol=Fine-tuned, Pre-training=ImageNet, Backbone=ResNet-502020.02 | 97 | — | — | — | — | |
| MobileNetV3Throughput (imgs/s)=75602023.05 | 97 | — | — | — | — | |
| MaPLeShot=16, Time (s)=14070, Mem. (M)=30612024.12 | 97 | — | — | — | — | |
| CeiT-TFLOPs=1.2G, Resolution=224x224, Pre-training=ImageNet2021.03 | 96.9 | — | — | — | — | |
| LogMETop-k=k=22022.07 | 96.87 | — | — | — | — | |
| NASNet-A-MThroughput (imgs/s)=26232023.05 | 96.8 | — | — | — | — | |
| MobileNetV1Throughput (imgs/s)=85432023.05 | 96.7 | — | — | — | — | |
| ViT-B/16PEFT=-, Params(%)=86M2026.03 | 96.7 | — | — | — | — | |
| MobileNetV2Throughput (imgs/s)=65342023.05 | 96.6 | — | — | — | — | |
| Skip TuningShot=8, Time (s)=508, Mem. (M)=5282024.12 | 96.6 | — | — | — | — | |
| ProGradShots=162025.10 | 96.6 | — | — | — | — | |
| BEITBackbone=ViT-S/162022.03 | 96.4 | — | — | — | — | |
| ProSRCShot=8, Time (s)=7636, Mem. (M)=33732024.12 | 96.27 | — | — | — | — | |
| TIP-AdapterShots=162025.10 | 96.2 | — | — | — | — | |
| BYOLArch=R502021.04 | 96.1 | — | — | — | — | |
| BYOLBackbone=ResNet-50, Pre-training Dataset=ImageNet-1K, Pre-training Epochs=8002022.03 | 96.1 | — | — | — | — | |
| CaCoBackbone=ResNet-50, Pre-training Dataset=ImageNet-1K, Pre-training Epochs=8002022.03 | 96.1 | — | — | — | — | |
| CLIPBackbone=ResNet-50, Evaluation Protocol=Linear probe2021.06 | 96.1 | — | — | — | — | |
| UniFGVCLearning Paradigm=Training-free, Method Strategy=Reasoning-based2026.03 | 95.84 | — | — | — | — | |
| NNCLRArch=ViT-B/82021.04 | 95.8 | — | — | — | — | |
| MaPLeShot=8, Time (s)=7060, Mem. (M)=30612024.12 | 95.8 | — | — | — | — | |
| DISSLEvaluation Protocol=Linear probe, Backbone=ResNet-50, Pre-training Dataset=ImageNet, Multi-crop Augmentation=2 x 160 + 4 x 962022.09 | 95.3 | — | — | — | — | |
| NNCLRArch=R502021.04 | 95.1 | — | — | — | — | |
| NNCLRBackbone=ResNet-50, Pre-training Dataset=ImageNet-1K, Pre-training Epochs=8002022.03 | 95.1 | — | — | — | — | |
| CRISPPEFT=-, Params(%)=21M2026.03 | 95.1 | — | — | — | — | |
| SimCLRProtocol=Linear evaluation, Backbone=ResNet-50 (4x), Pre-training=ImageNet2020.02 | 95 | — | — | — | — | |
| LaCLIPArchitecture=ViT-B/16, Pre-training Data=CC3M, Protocol=5-way 5-shot2023.05 | 95 | — | — | — | — | |
| CLIP-LoRA + FMAShots=42025.10 | 95 | — | — | — | — |