Image Classification on miniImageNet
98.87AccuracyMPA
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MPAMarker=-, PT (Pre-trained)=true, Protocol=5-way 1-shot2026.02 | 98.87 | 95.6 | |
| MLVLMMarker=AAAI-25, PT (Pre-trained)=true, Protocol=5-way 1-shot2026.02 | 98.24 | — | |
| CAMLMarker=ICLR-24, PT (Pre-trained)=true, Protocol=5-way 1-shot2026.02 | 96.2 | — | |
| SPMMarker=AAAI-24, PT (Pre-trained)=true, Protocol=5-way 1-shot2026.02 | 93.7 | 83.31 | |
| iSICEBackbone=ConvNext-T2023.04 | 85.1 | — | |
| Precision ΩBackbone=ConvNext-T2023.04 | 83.6 | — | |
| iSQRT-COVBackbone=ConvNext-T2023.04 | 82 | — | |
| ECER-FSLMarker=AAAI-25, PT (Pre-trained)=true, Protocol=5-way 1-shot2026.02 | 81.14 | 76.58 | |
| Geometric Mean Lossshots=52025.01 | 81.13 | — | |
| iSICEBackbone=ResNeXt-1012023.04 | 81 | — | |
| RFSIC-simpleshots=52025.01 | 79.64 | — | |
| CANshots=52025.01 | 79.44 | — | |
| PN Lossshots=52025.01 | 79.13 | — | |
| SemFew-TransMarker=CVPR-24, PT (Pre-trained)=true, Protocol=5-way 1-shot2026.02 | 78.94 | 74.98 | |
| NCA Lossshots=52025.01 | 78.93 | — | |
| iSICEBackbone=VGG-162023.04 | 78.7 | — | |
| iSICEBackbone=ResNet-1012023.04 | 78 | — | |
| Precision ΩBackbone=ResNeXt-1012023.04 | 77.6 | — | |
| iSICEBackbone=VGG-192023.04 | 77.1 | — | |
| iSQRT-COVBackbone=VGG-162023.04 | 76.2 | — | |
| iSQRT-COVBackbone=ResNeXt-1012023.04 | 76.1 | — | |
| DN4shots=52025.01 | 75.66 | — | |
| iSQRT-COVBackbone=VGG-192023.04 | 75.4 | — | |
| HCTransformersMarker=CVPR-22, PT (Pre-trained)=false, Protocol=5-way 1-shot2026.02 | 74.62 | 70.31 | |
| Precision ΩBackbone=VGG-162023.04 | 74 | — | |
| Fully SupervisedFully Supervised=true2024.11 | 73.95 | — | |
| iSQRT-COVBackbone=ResNet-1012023.04 | 73.9 | — | |
| Precision ΩBackbone=VGG-192023.04 | 73.8 | — | |
| Diff-ResNetMarker=TPAMI-24, PT (Pre-trained)=false, Protocol=5-way 1-shot2026.02 | 73.47 | — | |
| Transductive CNAPSBackbone=ResNet-18, N-way=5, K-shot=52022.10 | 73.1 | — | |
| Precision ΩBackbone=ResNet-1012023.04 | 73 | — | |
| Meta-Baselineshots=52025.01 | 72.96 | — | |
| MetaQDABackbone=Conv4, N-way=5, K-shot=52022.10 | 72.64 | — | |
| SAMPTransfer*Backbone=Conv4b, N-way=5, K-shot=52022.10 | 72.52 | — | |
| SP-CLIPMarker=CVPR-23, PT (Pre-trained)=true, Protocol=5-way 1-shot2026.02 | 72.31 | 70.26 | |
| iSICEBackbone=ResNet-502023.04 | 72 | — | |
| CPEAMarker=ICCV-23, PT (Pre-trained)=false, Protocol=5-way 1-shot2026.02 | 71.97 | 68.49 | |
| SEVProMarker=IJCAI-24, PT (Pre-trained)=true, Protocol=5-way 1-shot2026.02 | 71.81 | — | |
| FEATBackbone=Conv4, N-way=5, K-shot=52022.10 | 71.61 | — | |
| Simple CNAPSBackbone=ResNet-18, N-way=5, K-shot=52022.10 | 70.8 | — | |
| iSQRT-COVBackbone=ResNet-502023.04 | 70.7 | — | |
| L2P-AHILBag Size=162024.11 | 70.26 | — | |
| iSICEBackbone=Swin-T2023.04 | 69.1 | — | |
| iSICEBackbone=Swin-B2023.04 | 68.4 | — | |
| SAMPTransferBackbone=Conv4, N-way=5, K-shot=52022.10 | 68.33 | — | |
| XOResNet44Time-steps=42026.05 | 68 | — | |
| iSQRT-COVBackbone=Swin-T2023.04 | 67.7 | — | |
| ROTBag Size=162024.11 | 67.02 | — | |
| SimpleShotBackbone=Conv4, N-way=5, K-shot=52022.10 | 66.92 | — | |
| TPN-semiShots=5-shot, Distraction=None2018.05 | 66.42 | — | |
| Precision ΩBackbone=Swin-B2023.04 | 66.4 | — | |
| ProtoNetBackbone=Conv4, N-way=5, K-shot=52022.10 | 66.33 | — | |
| CleanEvaluation Protocol=SL2024.02 | 66.2 | — | |
| SNN-ResNet-12Time-steps=16, Evaluation Protocol=5-way 5-shot2026.05 | 65.61 | — | |
| Precision ΩBackbone=ResNet-502023.04 | 65.6 | — | |
| Precision ΩBackbone=Swin-T2023.04 | 65.6 | — | |
| Geometric Mean Lossshots=12025.01 | 65.51 | — | |
| Revisiting UMLBackbone=Conv4, N-way=5, K-shot=52022.10 | 65.33 | — | |
| TPN-semiShots=5-shot, Distraction=With Distraction (w/D)2018.05 | 64.95 | — | |
| iSQRT-COVBackbone=Swin-B2023.04 | 64.9 | — | |
| C3LRBackbone=Conv4, N-way=5, K-shot=52022.10 | 64.81 | — | |
| Soft k-MeansShots=5-shot, Distraction=None2018.05 | 64.59 | — | |
| DLLPBag Size=162024.11 | 64.53 | — | |
| Masked Soft k-MeansShots=5-shot, Distraction=None2018.05 | 64.39 | — | |
| MMCBackbone=Conv4, N-way=5, K-shot=52022.10 | 64.39 | — | |
| Meta-UAFSMarker=IJCAI-21, PT (Pre-trained)=false, Protocol=5-way 1-shot2026.02 | 64.22 | 62.36 | |
| LLP-VATBag Size=162024.11 | 64.17 | — | |
| CANshots=12025.01 | 63.85 | — | |
| Soft k-MeansShots=5-shot, Distraction=With Distraction (w/D)2018.05 | 63.55 | — | |
| Meta-Baselineshots=12025.01 | 63.17 | — | |
| CSSL-FSL_Mini64Backbone=Conv4, N-way=5, K-shot=52022.10 | 63.13 | — | |
| Soft k-Means+ClusterShots=5-shot, Distraction=None2018.05 | 63.08 | — | |
| Pre+LinearBackbone=Conv4, N-way=5, K-shot=52022.10 | 63.01 | — | |
| ProtoTransferBackbone=Conv4, N-way=5, K-shot=52022.10 | 62.99 | — | |
| Masked Soft k-MeansShots=5-shot, Distraction=With Distraction (w/D)2018.05 | 62.96 | — | |
| NCA Lossshots=12025.01 | 62.68 | — | |
| PN Lossshots=12025.01 | 62.42 | — | |
| MAMLBackbone=Conv4, N-way=5, K-shot=52022.10 | 62.13 | — | |
| RFSIC-simpleshots=12025.01 | 62.02 | — | |
| Soft k-Means+ClusterShots=5-shot, Distraction=With Distraction (w/D)2018.05 | 61.27 | — | |
| DN4shots=12025.01 | 61.23 | — | |
| SAMPTransfer*Backbone=Conv4b, N-way=5, K-shot=12022.10 | 61.02 | — | |
| Matching networks2026.05 | 60 | — | |
| L2P-AHILBag Size=322024.11 | 59.81 | — | |
| CleanEvaluation Protocol=MoCo2024.02 | 57.6 | — | |
| OA ResNet44Time-steps=42026.05 | 57.59 | — | |
| FedFSL-CFRDMarker=AAAI-25, PT (Pre-trained)=false, Protocol=5-way 1-shot2026.02 | 56.66 | — | |
| MetaQDABackbone=Conv4, N-way=5, K-shot=12022.10 | 56.41 | — | |
| SAMPTransferBackbone=Conv4, N-way=5, K-shot=12022.10 | 55.75 | — | |
| Meta-GMVAEBackbone=Conv4, N-way=5, K-shot=52022.10 | 55.73 | — | |
| Transductive CNAPSBackbone=ResNet-18, N-way=5, K-shot=12022.10 | 55.6 | — | |
| ULDA-ProtoNetBackbone=Conv4, N-way=5, K-shot=52022.10 | 55.41 | — | |
| U-SoSN+ArLBackbone=Conv4, N-way=5, K-shot=52022.10 | 55.39 | — | |
| U-MISoBackbone=Conv4, N-way=5, K-shot=52022.10 | 55.38 | — | |
| DLLPBag Size=322024.11 | 55.37 | — | |
| CleanEvaluation Protocol=SimCLR2024.02 | 55.3 | — | |
| FEATBackbone=Conv4, N-way=5, K-shot=12022.10 | 55.15 | — | |
| CUMCABackbone=Conv4, N-way=5, K-shot=52022.10 | 54.55 | — | |
| CleanEvaluation Protocol=SimSiam2024.02 | 54.5 | — | |
| ULDA-MetaNetBackbone=Conv4, N-way=5, K-shot=52022.10 | 54.49 | — |