Few-shot Image Classification on miniImageNet (test)
99.07AccuracyMPA
Evaluation results
| Method | Links | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| MPAVenue=None, Pre-trained=true, Way=5-way, Shot=5-shot2026.02 | 99.07 | — | — | — | — | — | — | — | — | — | — | — | |
| CAMLVenue=ICLR-24, Pre-trained=true, Way=5-way, Shot=5-shot2026.02 | 98.6 | — | — | — | — | — | — | — | — | — | — | — | |
| FD-AlignNumber of ways=5, Number of shots=52023.10 | 98.52 | — | — | — | — | — | — | — | — | — | — | — | |
| WiSE-FTNumber of ways=5, Number of shots=52023.10 | 98.44 | — | — | — | — | — | — | — | — | — | — | — | |
| SPMVenue=AAAI-24, Pre-trained=true, Way=5-way, Shot=5-shot2026.02 | 98.3 | — | — | — | — | — | — | — | — | — | — | — | |
| CLIPNumber of ways=5, Number of shots=52023.10 | 97.46 | — | — | — | — | — | — | — | — | — | — | — | |
| FD-AlignNumber of ways=5, Number of shots=12023.10 | 95.04 | — | — | — | — | — | — | — | — | — | — | — | |
| WiSE-FTNumber of ways=5, Number of shots=12023.10 | 93.55 | — | — | — | — | — | — | — | — | — | — | — | |
| HCTransformersVenue=CVPR-22, Pre-trained=false, Way=5-way, Shot=5-shot2026.02 | 89.19 | — | — | — | — | — | — | — | — | — | — | — | |
| CLIPNumber of ways=5, Number of shots=12023.10 | 88.21 | — | — | — | — | — | — | — | — | — | — | — | |
| CPEAVenue=CVPR-23, Pre-trained=false, Way=5-way, Shot=5-shot2026.02 | 87.26 | — | — | — | — | — | — | — | — | — | — | — | |
| SemFew-TransVenue=CVPR-24, Pre-trained=true, Way=5-way, Shot=5-shot2026.02 | 86.49 | — | — | — | — | — | — | — | — | — | — | — | |
| VDLF-Netshot=5-shot2026.04 | 86.33 | — | — | — | — | — | — | — | — | — | — | — | |
| Diff-ResNetVenue=TPAMI-24, Pre-trained=false, Way=5-way, Shot=5-shot2026.02 | 83.86 | — | — | — | — | — | — | — | — | — | — | — | |
| SP-CLIPVenue=CVPR-23, Pre-trained=true, Way=5-way, Shot=5-shot2026.02 | 83.42 | — | — | — | — | — | — | — | — | — | — | — | |
| TEENWay=5-way, Shot=5-shot2023.12 | 83.11 | — | — | — | — | — | — | — | — | — | — | — | |
| LR with DCWay=5-way, Shot=5-shot2023.12 | 82.88 | — | — | — | — | — | — | — | — | — | — | — | |
| NegCosineWay=5-way, Shot=5-shot2023.12 | 80.94 | — | — | — | — | — | — | — | — | — | — | — | |
| Prototypical Networksshot=5-shot2026.04 | 80.09 | — | — | — | — | — | — | — | — | — | — | — | |
| Meta-UAFSVenue=IJCAI-21, Pre-trained=false, Way=5-way, Shot=5-shot2026.02 | 79.99 | — | — | — | — | — | — | — | — | — | — | — | |
| SEVProVenue=IJCAI-24, Pre-trained=true, Way=5-way, Shot=5-shot2026.02 | 78.88 | — | — | — | — | — | — | — | — | — | — | — | |
| MetaOptNetArchitecture=ResNet12, Shots=52023.04 | 78.63 | — | — | — | — | — | — | — | — | — | — | — | |
| MetaBaselineArchitecture=WRN 28-10, Shots=52023.04 | 78.17 | — | — | — | — | — | — | — | — | — | — | — | |
| TADAMArchitecture=ResNet12, Shots=52023.04 | 76.7 | — | — | — | — | — | — | — | — | — | — | — | |
| Sparse and Shared Feature ActivationsArchitecture=ResNet12, Shots=52023.04 | 76.6 | — | — | — | — | — | — | — | — | — | — | — | |
| MAMLway=5, shot=50, Learning Paradigm=Supervised meta-learning2023.03 | 75.54 | — | — | — | — | — | — | — | — | — | — | — | |
| ProtoNetWay=5-way, Shot=5-shot2023.12 | 73.68 | — | — | — | — | — | — | — | — | — | — | — | |
| PsCoway=5, shot=50, Learning Paradigm=Unsupervised meta-learning2023.03 | 73.5 | — | — | — | — | — | — | — | — | — | — | — | |
| PsCoway=5, shot=502023.03 | 73.5 | — | — | — | — | — | — | — | — | — | — | — | |
| Meta-SVEBMway=5, shot=50, Learning Paradigm=Unsupervised meta-learning2023.03 | 72.28 | — | — | — | — | — | — | — | — | — | — | — | |
| PsCoway=5, shot=20, Learning Paradigm=Unsupervised meta-learning2023.03 | 72.22 | — | — | — | — | — | — | — | — | — | — | — | |
| PsCoway=5, shot=202023.03 | 72.22 | — | — | — | — | — | — | — | — | — | — | — | |
| FedFSL-CFRDVenue=AAAI-25, Pre-trained=false, Way=5-way, Shot=5-shot2026.02 | 72.15 | — | — | — | — | — | — | — | — | — | — | — | |
| ProtoNetsway=5, shot=50, Learning Paradigm=Supervised meta-learning2023.03 | 72.04 | — | — | — | — | — | — | — | — | — | — | — | |
| MAMLway=5, shot=20, Learning Paradigm=Supervised meta-learning2023.03 | 71.03 | — | — | — | — | — | — | — | — | — | — | — | |
| VDLF-Netshot=1-shot2026.04 | 70.72 | — | — | — | — | — | — | — | — | — | — | — | |
| ProtoNetsway=5, shot=20, Learning Paradigm=Supervised meta-learning2023.03 | 70.05 | — | — | — | — | — | — | — | — | — | — | — | |
| Lachapelle et alArchitecture=ResNet12, Shots=52023.04 | 70.01 | — | — | — | — | — | — | — | — | — | — | — | |
| CACTUS-MAMLway=5, shot=50, Learning Paradigm=Unsupervised meta-learning2023.03 | 69.64 | — | — | — | — | — | — | — | — | — | — | — | |
| LASIUM-MAMLway=5, shot=50, Learning Paradigm=Unsupervised meta-learning2023.03 | 69.13 | — | — | — | — | — | — | — | — | — | — | — | |
| Matching Networksshot=5-shot2026.04 | 68.87 | — | — | — | — | — | — | — | — | — | — | — | |
| LR with DCWay=5-way, Shot=1-shot2023.12 | 68.57 | — | — | — | — | — | — | — | — | — | — | — | |
| Meta-GMVAEway=5, shot=50, Learning Paradigm=Unsupervised meta-learning2023.03 | 68.26 | — | — | — | — | — | — | — | — | — | — | — | |
| Prototypical NetArchitecture=Conv64(x4), Shots=52023.04 | 68.2 | — | — | — | — | — | — | — | — | — | — | — | |
| UMTRAway=5, shot=50, Learning Paradigm=Unsupervised meta-learning2023.03 | 67.15 | — | — | — | — | — | — | — | — | — | — | — | |
| Meta-SVEBMway=5, shot=20, Learning Paradigm=Unsupervised meta-learning2023.03 | 67.07 | — | — | — | — | — | — | — | — | — | — | — | |
| TEENWay=5-way, Shot=1-shot2023.12 | 65.7 | — | — | — | — | — | — | — | — | — | — | — | |
| LASIUM-MAMLway=5, shot=20, Learning Paradigm=Unsupervised meta-learning2023.03 | 65.17 | — | — | — | — | — | — | — | — | — | — | — | |
| LASIUM-ProtoNetsway=5, shot=50, Learning Paradigm=Unsupervised meta-learning2023.03 | 64.89 | — | — | — | — | — | — | — | — | — | — | — | |
| SimCLRway=5, shot=50, Learning Paradigm=Self-supervised learning2023.03 | 64.85 | — | — | — | — | — | — | — | — | — | — | — | |
| SimCLRway=5, shot=502023.03 | 64.85 | — | — | — | — | — | — | — | — | — | — | — | |
| SwAVway=5, shot=50, Learning Paradigm=Self-supervised learning2023.03 | 64.52 | — | — | — | — | — | — | — | — | — | — | — | |
| SwAVway=5, shot=502023.03 | 64.52 | — | — | — | — | — | — | — | — | — | — | — | |
| CACTUS-MAMLway=5, shot=20, Learning Paradigm=Unsupervised meta-learning2023.03 | 63.84 | — | — | — | — | — | — | — | — | — | — | — | |
| CACTUS-ProtoNetsway=5, shot=50, Learning Paradigm=Unsupervised meta-learning2023.03 | 63.55 | — | — | — | — | — | — | — | — | — | — | — | |
| MoCo v2way=5, shot=50, Learning Paradigm=Self-supervised learning2023.03 | 63.45 | — | — | — | — | — | — | — | — | — | — | — | |
| MoCo v2way=5, shot=502023.03 | 63.45 | — | — | — | — | — | — | — | — | — | — | — | |
| PsCoway=5, shot=5, Learning Paradigm=Unsupervised meta-learning2023.03 | 63.26 | — | — | — | — | — | — | — | — | — | — | — | |
| PsCoway=5, shot=52023.03 | 63.26 | — | — | — | — | — | — | — | — | — | — | — | |
| Meta-GMVAEway=5, shot=20, Learning Paradigm=Unsupervised meta-learning2023.03 | 63.14 | — | — | — | — | — | — | — | — | — | — | — | |
| MAMLArchitecture=Conv32(x4), Shots=52023.04 | 63.11 | — | — | — | — | — | — | — | — | — | — | — | |
| MetaOptNetArchitecture=ResNet12, Shots=12023.04 | 62.64 | — | — | — | — | — | — | — | — | — | — | — | |
| NegCosineWay=5-way, Shot=1-shot2023.12 | 62.33 | — | — | — | — | — | — | — | — | — | — | — | |
| ProtoNetsway=5, shot=5, Learning Paradigm=Supervised meta-learning2023.03 | 62.29 | — | — | — | — | — | — | — | — | — | — | — | |
| MAMLway=5, shot=5, Learning Paradigm=Supervised meta-learning2023.03 | 62.13 | — | — | — | — | — | — | — | — | — | — | — | |
| Matching Networksshot=1-shot2026.04 | 61.87 | — | — | — | — | — | — | — | — | — | — | — | |
| SimCLRway=5, shot=20, Learning Paradigm=Self-supervised learning2023.03 | 61.83 | — | — | — | — | — | — | — | — | — | — | — | |
| SimCLRway=5, shot=202023.03 | 61.83 | — | — | — | — | — | — | — | — | — | — | — | |
| CACTUS-ProtoNetsway=5, shot=20, Learning Paradigm=Unsupervised meta-learning2023.03 | 61.54 | — | — | — | — | — | — | — | — | — | — | — | |
| SwAVway=5, shot=20, Learning Paradigm=Self-supervised learning2023.03 | 61.36 | — | — | — | — | — | — | — | — | — | — | — | |
| SwAVway=5, shot=202023.03 | 61.36 | — | — | — | — | — | — | — | — | — | — | — | |
| UMTRAway=5, shot=20, Learning Paradigm=Unsupervised meta-learning2023.03 | 61.11 | — | — | — | — | — | — | — | — | — | — | — | |
| LASIUM-ProtoNetsway=5, shot=20, Learning Paradigm=Unsupervised meta-learning2023.03 | 61.09 | — | — | — | — | — | — | — | — | — | — | — | |
| MoCo v2way=5, shot=20, Learning Paradigm=Self-supervised learning2023.03 | 60.23 | — | — | — | — | — | — | — | — | — | — | — | |
| MoCo v2way=5, shot=202023.03 | 60.23 | — | — | — | — | — | — | — | — | — | — | — | |
| Sparse and Shared Feature ActivationsArchitecture=ResNet12, Shots=12023.04 | 60.1 | — | — | — | — | — | — | — | — | — | — | — | |
| Training from Scratchway=5, shot=50, Learning Paradigm=Supervised learning2023.03 | 59.63 | — | — | — | — | — | — | — | — | — | — | — | |
| MetaBaselineArchitecture=WRN 28-10, Shots=12023.04 | 59.62 | — | — | — | — | — | — | — | — | — | — | — | |
| TADAMArchitecture=ResNet12, Shots=12023.04 | 58.5 | — | — | — | — | — | — | — | — | — | — | — | |
| Meta-SVEBMway=5, shot=5, Learning Paradigm=Unsupervised meta-learning2023.03 | 58.03 | — | — | — | — | — | — | — | — | — | — | — | |
| Prototypical Networksshot=1-shot2026.04 | 56.8 | — | — | — | — | — | — | — | — | — | — | — | |
| Meta-GMVAEway=5, shot=5, Learning Paradigm=Unsupervised meta-learning2023.03 | 55.73 | — | — | — | — | — | — | — | — | — | — | — | |
| LASIUM-MAMLway=5, shot=5, Learning Paradigm=Unsupervised meta-learning2023.03 | 54.56 | — | — | — | — | — | — | — | — | — | — | — | |
| Lachapelle et alArchitecture=ResNet12, Shots=12023.04 | 54.22 | — | — | — | — | — | — | — | — | — | — | — | |
| ProtoNetWay=5-way, Shot=1-shot2023.12 | 54.16 | — | — | — | — | — | — | — | — | — | — | — | |
| CACTUS-MAMLway=5, shot=5, Learning Paradigm=Unsupervised meta-learning2023.03 | 53.97 | — | — | — | — | — | — | — | — | — | — | — | |
| CACTUS-ProtoNetsway=5, shot=5, Learning Paradigm=Unsupervised meta-learning2023.03 | 53.36 | — | — | — | — | — | — | — | — | — | — | — | |
| LASIUM-ProtoNetsway=5, shot=5, Learning Paradigm=Unsupervised meta-learning2023.03 | 52.53 | — | — | — | — | — | — | — | — | — | — | — | |
| SimCLRway=5, shot=5, Learning Paradigm=Self-supervised learning2023.03 | 52.5 | — | — | — | — | — | — | — | — | — | — | — | |
| SimCLRway=5, shot=52023.03 | 52.5 | — | — | — | — | — | — | — | — | — | — | — | |
| SwAVway=5, shot=5, Learning Paradigm=Self-supervised learning2023.03 | 52.41 | — | — | — | — | — | — | — | — | — | — | — | |
| SwAVway=5, shot=52023.03 | 52.41 | — | — | — | — | — | — | — | — | — | — | — | |
| Training from Scratchway=5, shot=20, Learning Paradigm=Supervised learning2023.03 | 51.53 | — | — | — | — | — | — | — | — | — | — | — | |
| MoCo v2way=5, shot=5, Learning Paradigm=Self-supervised learning2023.03 | 50.94 | — | — | — | — | — | — | — | — | — | — | — | |
| MoCo v2way=5, shot=52023.03 | 50.94 | — | — | — | — | — | — | — | — | — | — | — | |
| UMTRAway=5, shot=5, Learning Paradigm=Unsupervised meta-learning2023.03 | 50.73 | — | — | — | — | — | — | — | — | — | — | — | |
| Prototypical NetArchitecture=Conv64(x4), Shots=12023.04 | 49.42 | — | — | — | — | — | — | — | — | — | — | — | |
| MAMLArchitecture=Conv32(x4), Shots=12023.04 | 48.7 | — | — | — | — | — | — | — | — | — | — | — | |
| MAMLway=5, shot=1, Learning Paradigm=Supervised meta-learning2023.03 | 46.81 | — | — | — | — | — | — | — | — | — | — | — | |
| PsCoway=5, shot=1, Learning Paradigm=Unsupervised meta-learning2023.03 | 46.7 | — | — | — | — | — | — | — | — | — | — | — |