Few-shot classification on tieredImageNet (test)
98.79AccuracyMPA
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| MPAVenue=None, Pre-trained=true, Way=5-way, Shot=5-shot2026.02 | 98.79 | — | — | — | — | — | — | — | — | |
| CAMLVenue=ICLR-24, Pre-trained=true, Way=5-way, Shot=5-shot2026.02 | 98.1 | — | — | — | — | — | — | — | — | |
| SPMVenue=AAAI-24, Pre-trained=true, Way=5-way, Shot=5-shot2026.02 | 96.2 | — | — | — | — | — | — | — | — | |
| HCTransformersVenue=CVPR-22, Pre-trained=false, Way=5-way, Shot=5-shot2026.02 | 91.72 | — | — | — | — | — | — | — | — | |
| CPEAVenue=CVPR-23, Pre-trained=false, Way=5-way, Shot=5-shot2026.02 | 90.12 | — | — | — | — | — | — | — | — | |
| FewTUREBackbone=Swin-T, Way=5-way, Shot=5-shot, Method Category=Visual-Based2023.11 | 89.96 | — | — | — | — | — | — | — | — | |
| SemFew-TransBackbone=Swin-T, Way=5-way, Shot=5-shot, Method Category=Semantic-Based2023.11 | 89.89 | — | — | — | — | — | — | — | — | |
| SemFew-TransVenue=CVPR-24, Pre-trained=true, Way=5-way, Shot=5-shot2026.02 | 89.89 | — | — | — | — | — | — | — | — | |
| S2M2 + TCPRWay=5-way, Shot=5-shot2022.10 | 88.89 | — | — | — | — | — | — | — | — | |
| SP-CLIPBackbone=Visformer-T, Way=5-way, Shot=5-shot, Method Category=Semantic-Based2023.11 | 88.55 | — | — | — | — | — | — | — | — | |
| SP-CLIPVenue=CVPR-23, Pre-trained=true, Way=5-way, Shot=5-shot2026.02 | 88.55 | — | — | — | — | — | — | — | — | |
| S2M2Way=5-way, Shot=5-shot2022.10 | 87.61 | — | — | — | — | — | — | — | — | |
| FGFLBackbone=ResNet-12, Way=5-way, Shot=5-shot, Method Category=Visual-Based2023.11 | 87.21 | — | — | — | — | — | — | — | — | |
| AM3-BERTBackbone=ResNet-12, Way=5-way, Shot=5-shot, Method Category=Semantic-Based2023.11 | 87.2 | — | — | — | — | — | — | — | — | |
| Diff-ResNetVenue=TPAMI-24, Pre-trained=false, Way=5-way, Shot=5-shot2026.02 | 87.1 | — | — | — | — | — | — | — | — | |
| Inv-EquWay=5-way, Shot=5-shot2022.10 | 87.08 | — | — | — | — | — | — | — | — | |
| BD-CSPNSetting=Transductive, Backbone=WRN-28-102019.11 | 86.92 | — | — | — | — | — | — | — | — | |
| SUNBackbone=ViT-S, Way=5-way, Shot=5-shot, Method Category=Visual-Based2023.11 | 86.74 | — | — | — | — | — | — | — | — | |
| MCLBackbone=ResNet-12, Shot=5-shot, Feature Extraction Category=Dense features (PyramidGrid)2021.06 | 86.29 | — | — | — | — | — | — | — | — | |
| MCL-KatzBackbone=ResNet-12, Shot=5-shot, Feature Extraction Category=Dense features (PyramidGrid)2021.06 | 86.21 | — | — | — | — | — | — | — | — | |
| SemFewBackbone=ResNet-12, Way=5-way, Shot=5-shot, Method Category=Semantic-Based2023.11 | 85.88 | — | — | — | — | — | — | — | — | |
| SVAE-ProtoBackbone=ResNet-12, Way=5-way, Shot=5-shot, Method Category=Semantic-Based2023.11 | 85.77 | — | — | — | — | — | — | — | — | |
| CSEIWay=5-way, Shot=5-shot2022.10 | 85.72 | — | — | — | — | — | — | — | — | |
| Transductive Fine-TuningSetting=Transductive, Backbone=WRN-28-102019.11 | 85.5 | — | — | — | — | — | — | — | — | |
| Meta-AdaMBackbone=ResNet-12, Way=5-way, Shot=5-shot, Method Category=Visual-Based2023.11 | 85.24 | — | — | — | — | — | — | — | — | |
| LSTSetting=Semi-Supervised, Backbone=ResNet-122019.11 | 85.2 | — | — | — | — | — | — | — | — | |
| DeepEMDWay=5-way, Shot=5-shot2022.10 | 85.01 | — | — | — | — | — | — | — | — | |
| FEATBackbone=ResNet-12, Way=5-way, Shot=5-shot, Method Category=Visual-Based2023.11 | 84.79 | — | — | — | — | — | — | — | — | |
| RFSBackbone=ResNet-12, Way=5-way, Shot=5-shot, Method Category=Visual-Based2023.11 | 84.41 | — | — | — | — | — | — | — | — | |
| Meta-UAFSVenue=IJCAI-21, Pre-trained=false, Way=5-way, Shot=5-shot2026.02 | 84.33 | — | — | — | — | — | — | — | — | |
| CTMBackbone=ResNet-18, Way=5-way, Shot=5-shot, Method Category=Visual-Based2023.11 | 84.28 | — | — | — | — | — | — | — | — | |
| CANBackbone=ResNet-12, Way=5-way, Shot=5-shot, Method Category=Visual-Based2023.11 | 84.23 | — | — | — | — | — | — | — | — | |
| CANWay=5-way, Shot=5-shot2022.10 | 84.08 | — | — | — | — | — | — | — | — | |
| SEVProVenue=IJCAI-24, Pre-trained=true, Way=5-way, Shot=5-shot2026.02 | 84.04 | — | — | — | — | — | — | — | — | |
| ProtoNetBackbone=ResNet-12, Way=5-way, Shot=5-shot, Method Category=Visual-Based2023.11 | 84.03 | — | — | — | — | — | — | — | — | |
| ProtoNet+MCLBackbone=ResNet-12, Shot=5-shot, Feature Extraction Category=Global features2021.06 | 83.84 | — | — | — | — | — | — | — | — | |
| ProtoNetBackbone=ResNet-12, Shot=5-shot, Feature Extraction Category=Global features2021.06 | 83.46 | — | — | — | — | — | — | — | — | |
| Meta-BaselineBackbone=ResNet-12, Way=5-way, Shot=5-shot, Method Category=Visual-Based2023.11 | 83.29 | — | — | — | — | — | — | — | — | |
| MC2shots=5-shot, feature extraction=center, meta-training protocol=meta-train + meta-validation, backbone=fixed features from [2]2019.02 | 82.61 | — | — | — | — | — | — | — | — | |
| AM3Backbone=ResNet-12, Way=5-way, Shot=5-shot, Method Category=Semantic-Based2023.11 | 82.58 | — | — | — | — | — | — | — | — | |
| SemFew-TransBackbone=Swin-T, Way=5-way, Shot=1-shot, Method Category=Semantic-Based2023.11 | 82.37 | — | — | — | — | — | — | — | — | |
| MC2shots=5-shot, feature extraction=center, backbone=fixed features from [2]2019.02 | 82.21 | — | — | — | — | — | — | — | — | |
| TADAMBackbone=ResNet-12, Way=5-way, Shot=5-shot, Method Category=Visual-Based2023.11 | 81.92 | — | — | — | — | — | — | — | — | |
| MetaOptNet-SVMshots=5-shot, 15-shot meta-training=true, meta-training protocol=meta-train + meta-validation, backbone=ResNet-122019.02 | 81.75 | — | — | — | — | — | — | — | — | |
| MetaOptWay=5-way, Shot=5-shot2022.10 | 81.75 | — | — | — | — | — | — | — | — | |
| MetaOptNet-SVMSetting=Inductive, Backbone=ResNet-122019.11 | 81.56 | — | — | — | — | — | — | — | — | |
| LEOshots=5-shot, feature extraction=center, backbone=fixed features from [2]2019.02 | 81.44 | — | — | — | — | — | — | — | — | |
| LEOSetting=Inductive, Backbone=WRN-28-10, training_details=Training set and validation set used for training2019.11 | 81.44 | — | — | — | — | — | — | — | — | |
| LEOWay=5-way, Shot=5-shot2022.10 | 81.44 | — | — | — | — | — | — | — | — | |
| MatchNetBackbone=ResNet-12, Way=5-way, Shot=5-shot, Method Category=Visual-Based2023.11 | 80.6 | — | — | — | — | — | — | — | — | |
| RelationNet+MCLBackbone=ResNet-12, Shot=5-shot, Feature Extraction Category=Global features2021.06 | 80.27 | — | — | — | — | — | — | — | — | |
| TapNetBackbone=ResNet-12, Number of ways=5-way, Number of shots=5-shot2019.05 | 80.26 | — | — | — | — | — | — | — | — | |
| EGNN+TransductionTransduction=Yes, Setting=5-Way 5-Shot2019.05 | 80.15 | — | — | — | — | — | — | — | — | |
| EGNN+TransductionTransduction=Yes, Way=5-way, Shot=5-shot2019.05 | 80.15 | — | — | — | — | — | — | — | — | |
| EGNNSetting=Transductive, Backbone=ConvNet-2562019.11 | 80.15 | — | — | — | — | — | — | — | — | |
| SemFewBackbone=ResNet-12, Way=5-way, Shot=1-shot, Method Category=Semantic-Based2023.11 | 78.96 | — | — | — | — | — | — | — | — | |
| FedFSL-CFRDVenue=AAAI-25, Pre-trained=false, Way=5-way, Shot=5-shot2026.02 | 78.95 | — | — | — | — | — | — | — | — | |
| BD-CSPNSetting=Transductive, Backbone=WRN-28-102019.11 | 78.74 | — | — | — | — | — | — | — | — | |
| RelationNet+Backbone=ResNet-12, Shot=5-shot, Feature Extraction Category=Global features2021.06 | 78.41 | — | — | — | — | — | — | — | — | |
| SP-CLIPBackbone=Visformer-T, Way=5-way, Shot=1-shot, Method Category=Semantic-Based2023.11 | 78.03 | — | — | — | — | — | — | — | — | |
| LSTSetting=Semi-Supervised, Backbone=ResNet-122019.11 | 77.7 | — | — | — | — | — | — | — | — | |
| S2M2 + TCPRWay=5-way, Shot=1-shot2022.10 | 77.67 | — | — | — | — | — | — | — | — | |
| LwoFSetting=Inductive, Backbone=ConvNet-128, implementation=Results by current paper's implementation2019.11 | 77.24 | — | — | — | — | — | — | — | — | |
| AM3-BERTBackbone=ResNet-12, Way=5-way, Shot=1-shot, Method Category=Semantic-Based2023.11 | 77.03 | — | — | — | — | — | — | — | — | |
| SVAE-ProtoBackbone=ResNet-12, Way=5-way, Shot=1-shot, Method Category=Semantic-Based2023.11 | 76.98 | — | — | — | — | — | — | — | — | |
| FewTUREBackbone=Swin-T, Way=5-way, Shot=1-shot, Method Category=Visual-Based2023.11 | 76.32 | — | — | — | — | — | — | — | — | |
| S2M2Way=5-way, Shot=1-shot2022.10 | 74.87 | — | — | — | — | — | — | — | — | |
| KTNBackbone=Conv-128, Way=5-way, Shot=1-shot, Method Category=Semantic-Based2023.11 | 74.16 | — | — | — | — | — | — | — | — | |
| Transductive Fine-TuningSetting=Transductive, Backbone=WRN-28-102019.11 | 73.34 | — | — | — | — | — | — | — | — | |
| TPN (Higher Shot)Transduction=Yes, Training Strategy=Higher Shot, N-way=5-way, Shots=5-shot2018.05 | 73.3 | — | — | — | — | — | — | — | — | |
| Transductive Propagation NetsBackbone=ResNet-12, Number of ways=5-way, Number of shots=5-shot2019.05 | 73.3 | — | — | — | — | — | — | — | — | |
| TPNSetting=Semi-Supervised, Backbone=ConvNet-642019.11 | 73.3 | — | — | — | — | — | — | — | — | |
| FGFLBackbone=ResNet-12, Way=5-way, Shot=1-shot, Method Category=Visual-Based2023.11 | 73.21 | — | — | — | — | — | — | — | — | |
| SUNBackbone=ViT-S, Way=5-way, Shot=1-shot, Method Category=Visual-Based2023.11 | 72.99 | — | — | — | — | — | — | — | — | |
| TPNTransduction=Yes, N-way=5-way, Shots=5-shot2018.05 | 72.85 | — | — | — | — | — | — | — | — | |
| PROTO NET (Higher Way)Transduction=No, Training Strategy=Higher Way, N-way=5-way, Shots=5-shot2018.05 | 72.69 | — | — | — | — | — | — | — | — | |
| Prototypical NetsBackbone=ResNet-12, Number of ways=5-way, Number of shots=5-shot2019.05 | 72.69 | — | — | — | — | — | — | — | — | |
| Prototypical NetworksSetting=Inductive, Backbone=ConvNet-64, training_details=Training set and validation set used for training2019.11 | 72.69 | — | — | — | — | — | — | — | — | |
| TPNTransduction=Yes, Setting=5-Way 5-Shot2019.05 | 72.58 | — | — | — | — | — | — | — | — | |
| TPNTransduction=Yes, Way=5-way, Shot=5-shot2019.05 | 72.58 | — | — | — | — | — | — | — | — | |
| CSEIWay=5-way, Shot=1-shot2022.10 | 72.57 | — | — | — | — | — | — | — | — | |
| MAMLBackbone=ResNet-12, Way=5-way, Shot=5-shot, Method Category=Visual-Based2023.11 | 72.41 | — | — | — | — | — | — | — | — | |
| Inv-EquWay=5-way, Shot=1-shot2022.10 | 72.21 | — | — | — | — | — | — | — | — | |
| Euclideannoise_proportion=0%, shots=5-shot, ways=5-way, noise_type=symmetric label swap2022.04 | 71.48 | — | — | — | — | — | — | — | — | |
| Oraclenoise_proportion=0%, shots=5-shot, ways=5-way, noise_type=symmetric label swap2022.04 | 71.42 | — | — | — | — | — | — | — | — | |
| Vanilla ProtoNetnoise_proportion=0%, shots=5-shot, ways=5-way, noise_type=symmetric label swap2022.04 | 71.36 | — | — | — | — | — | — | — | — | |
| RNNPnoise_proportion=0%, shots=5-shot, ways=5-way, noise_type=symmetric label swap2022.04 | 71.36 | — | — | — | — | — | — | — | — | |
| Relation NetSetting=Inductive, Backbone=ConvNet-2562019.11 | 71.32 | — | — | — | — | — | — | — | — | |
| RELATION NETTransduction=BN, N-way=5-way, Shots=5-shot2018.05 | 71.31 | — | — | — | — | — | — | — | — | |
| Relation NetTransduction=BN, Setting=5-Way 5-Shot2019.05 | 71.31 | — | — | — | — | — | — | — | — | |
| Relation NetTransduction=BN, Way=5-way, Shot=5-shot2019.05 | 71.31 | — | — | — | — | — | — | — | — | |
| Relation NetsBackbone=ResNet-12, Number of ways=5-way, Number of shots=5-shot2019.05 | 71.31 | — | — | — | — | — | — | — | — | |
| Mediannoise_proportion=0%, shots=5-shot, ways=5-way, noise_type=symmetric label swap2022.04 | 71.28 | — | — | — | — | — | — | — | — | |
| Absolutenoise_proportion=0%, shots=5-shot, ways=5-way, noise_type=symmetric label swap2022.04 | 71.27 | — | — | — | — | — | — | — | — | |
| Baseline++noise_proportion=0%, shots=5-shot, ways=5-way, noise_type=symmetric label swap2022.04 | 71.24 | — | — | — | — | — | — | — | — | |
| TraNFS-3noise_proportion=0%, shots=5-shot, ways=5-way, noise_type=symmetric label swap2022.04 | 71.17 | — | — | — | — | — | — | — | — | |
| Reptile + BNTransduction=BN, N-way=5-way, Shots=5-shot2018.05 | 71.03 | — | — | — | — | — | — | — | — | |
| Reptile + BNTransduction=BN, Setting=5-Way 5-Shot2019.05 | 71.03 | — | — | — | — | — | — | — | — | |
| Reptile + BNTransduction=BN, Way=5-way, Shot=5-shot2019.05 | 71.03 | — | — | — | — | — | — | — | — | |
| TPN-semiShots=5-shot, With Distraction=false2018.05 | 71.01 | — | — | — | — | — | — | — | — |