Zero-Shot Learning on CUB
90.1Top-1 AccuracyRLVC
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| RLVCBackbone=ViT2026.03 | 90.1 | — | — | — | — | |
| ZeroDiff++Venue=Ours, Backbone=Res101, Fine-tune features=Author's fine-tune features2026.02 | 87.7 | — | — | — | — | |
| ZeroDiffVenue=ICLR25, Backbone=Res101, Fine-tune features=Author's fine-tune features2026.02 | 87.5 | — | — | — | — | |
| VADSVenue=CVPR24, Backbone=ViT, Fine-tune features=Other fine-tune features2026.02 | 86.8 | — | — | — | — | |
| VADSBackbone=ViT, Method Category=Generative2026.03 | 86.8 | — | — | — | — | |
| TFVAEGANVenue=ECCV20, Backbone=Res101, Fine-tune features=Author's fine-tune features2026.02 | 85.8 | — | — | — | — | |
| TF-VAEGANSetting=Fine-tuned Transductive (FT-TR)2020.03 | 85.1 | — | — | — | — | |
| f-VAEGANVenue=CVPR19, Backbone=Res101, Fine-tune features=Author's fine-tune features2026.02 | 85.1 | — | — | — | — | |
| SDVAEVenue=ICCV21, Backbone=Res101, Fine-tune features=Author's fine-tune features2026.02 | 85.1 | — | — | — | — | |
| CEGANVenue=CVPR21, Backbone=Res101, Fine-tune features=Author's fine-tune features2026.02 | 84.6 | — | — | — | — | |
| f-CLSWGANVenue=CVPR18, Backbone=Res101, Fine-tune features=Author's fine-tune features2026.02 | 84.5 | — | — | — | — | |
| DFCAFlowVenue=TCSVT23, Backbone=Res101, Fine-tune features=Author's fine-tune features2026.02 | 83.9 | — | — | — | — | |
| f-VAEGAN-D2Learning Setting=Transductive, Fine-tuned=true2019.03 | 82.6 | — | — | — | — | |
| f-VAEGANSetting=Fine-tuned Transductive (FT-TR)2020.03 | 82.6 | — | — | — | — | |
| ReZSLVenue=TIP23, Backbone=Res101, Fine-tune features=None2026.02 | 80.9 | — | — | — | — | |
| ZeroMambaBackbone=VMamba, Method Category=Embedding2026.03 | 80.8 | — | — | — | — | |
| VSPCNBackbone=ViT, Method Category=Embedding2026.03 | 80.6 | — | — | — | — | |
| AENetBackbone=ViT, Method Category=Embedding2026.03 | 80.3 | — | — | — | — | |
| ZSLViTFeature Type=ViT2024.04 | 78.9 | — | — | — | — | |
| ZSLViTVenue=CVPR24, Backbone=ViT, Fine-tune features=None2026.02 | 78.9 | — | — | — | — | |
| ZSLViTBackbone=ViT, Method Category=Embedding2026.03 | 78.9 | — | — | — | — | |
| PSVMA++Venue=TPAMI25, Backbone=ViT, Fine-tune features=None2026.02 | 78.8 | — | — | — | — | |
| PSVMA+Backbone=ViT, Method Category=Embedding2026.03 | 78.8 | — | — | — | — | |
| MSDN++Category=Embedding-based Methods, Attention-based=true, Backbone=ResNet1012026.03 | 78.5 | — | — | — | — | |
| TransZero++Venue=TPAMI22, Backbone=Res101, Fine-tune features=None2026.02 | 78.3 | — | — | — | — | |
| TransZero++Backbone=ResNet, Method Category=Embedding2026.03 | 78.3 | — | — | — | — | |
| GEM-ZSL2022.07 | 77.8 | — | — | — | — | |
| GEM-ZSLVenue=CVPR'22, Feature Type=CNN, Attention=true2024.04 | 77.8 | — | — | — | — | |
| CE-GZSL2021.03 | 77.5 | — | — | — | — | |
| CE-GZSL2022.07 | 77.5 | — | — | — | — | |
| CE-GZSLBackbone=ResNet, Method Category=Generative2026.03 | 77.5 | — | — | — | — | |
| TransZero2022.07 | 76.8 | — | — | — | — | |
| TransZeroVenue=AAAI'22, Feature Type=CNN, Attention=true2024.04 | 76.8 | — | — | — | — | |
| TransZeroCategory=Embedding-based Methods, Attention-based=true, Backbone=ResNet1012026.03 | 76.8 | — | — | — | — | |
| DiffusionZSLVenue=ICCV23, Backbone=Res101, Fine-tune features=Author's fine-tune features2026.02 | 76.7 | — | — | — | — | |
| HASBackbone=ResNet, Method Category=Embedding2026.03 | 76.5 | — | — | — | — | |
| RGEN2022.07 | 76.1 | — | — | — | — | |
| MSDN2022.07 | 76.1 | — | — | — | — | |
| MSDNVenue=CVPR'22, Feature Type=CNN, Attention=true2024.04 | 76.1 | — | — | — | — | |
| MSDNCategory=Embedding-based Methods, Attention-based=true, Backbone=ResNet1012026.03 | 76.1 | — | — | — | — | |
| GNDANCategory=Embedding-based Methods, Attention-based=true, Backbone=ResNet1012026.03 | 75.1 | — | — | — | — | |
| SC-EGGBackbone=ResNet, Method Category=Generative2026.03 | 75.1 | — | — | — | — | |
| APNLearning Paradigm=Non-generative, ZoomInMod=true2022.04 | 75 | — | — | — | — | |
| APNCategory=Embedding-based Methods, Attention-based=true, Backbone=ResNet1012026.03 | 75 | — | — | — | — | |
| TF-VAEGANSetting=Transductive (TR)2020.03 | 74.7 | — | — | — | — | |
| APN+TF-VAEGANLearning Paradigm=Generative, Backbone Features=APN features2022.04 | 74.7 | — | — | — | — | |
| ViFRBackbone=ResNet, Method Category=Generative2026.03 | 74.5 | — | — | — | — | |
| TF-VAEGANSetting=Fine-tuned Inductive (FT-IN)2020.03 | 74.3 | — | — | — | — | |
| TF-VAEGAN*Learning Paradigm=Generative, Backbone Features=ResNet101 (finetuned)2022.04 | 74.3 | — | — | — | — | |
| SBAR-TSetting=Fine-tuned Transductive (FT-TR)2020.03 | 74 | — | — | — | — | |
| APN+f-VAEGAN-D2*Learning Paradigm=Generative, Backbone Features=APN features2022.04 | 73.9 | — | — | — | — | |
| APN+ABPLearning Paradigm=Generative, Backbone Features=APN features2022.04 | 73.2 | — | — | — | — | |
| f-VAEGAN-D2Learning Setting=Inductive, Fine-tuned=true2019.03 | 72.9 | — | — | — | — | |
| f-VAEGANSetting=Fine-tuned Inductive (FT-IN)2020.03 | 72.9 | — | — | — | — | |
| f-VAEGAN-D2*Learning Paradigm=Generative, Backbone Features=ResNet101 (finetuned)2022.04 | 72.9 | — | — | — | — | |
| ICCEBackbone=ResNet, Method Category=Generative2026.03 | 72.7 | — | — | — | — | |
| DUET2022.07 | 72.3 | — | — | — | — | |
| DUETVenue=AAAI'23, Feature Type=ViT2024.04 | 72.3 | — | — | — | — | |
| DUETBackbone=ViT, Method Category=Embedding2026.03 | 72.3 | — | — | — | — | |
| UE-finetuneLearning Setting=Transductive, Fine-tuned=true2019.03 | 72.1 | — | — | — | — | |
| UE-finetuneSetting=Fine-tuned Transductive (FT-TR)2020.03 | 72.1 | — | — | — | — | |
| APN+Learning Paradigm=Non-generative, ZoomInMod=false2022.04 | 72 | — | — | — | — | |
| APNNon-generative=true2022.07 | 72 | — | — | — | — | |
| APNVenue=NeurIPS'20, Feature Type=CNN, Attention=true2024.04 | 72 | — | — | — | — | |
| APNVenue=NeurIPS20, Backbone=Res101, Fine-tune features=None2026.02 | 72 | — | — | — | — | |
| ARENLearning Paradigm=Non-generative2022.04 | 71.8 | — | — | — | — | |
| ARENVenue=CVPR'19, Feature Type=CNN, Attention=true2024.04 | 71.8 | — | — | — | — | |
| ARENCategory=Embedding-based Methods, Attention-based=true, Backbone=ResNet1012026.03 | 71.8 | — | — | — | — | |
| f-VAEGAN-D2Learning Setting=Transductive, Fine-tuned=false2019.03 | 71.1 | — | — | — | — | |
| f-VAEGANSetting=Transductive (TR)2020.03 | 71.1 | — | — | — | — | |
| SGMALearning Paradigm=Non-generative2022.04 | 71 | — | — | — | — | |
| APN+CVCLearning Paradigm=Generative, Backbone Features=APN features2022.04 | 71 | — | — | — | — | |
| SGMACategory=Embedding-based Methods, Attention-based=true, Backbone=ResNet1012026.03 | 71 | — | — | — | — | |
| ABP*Learning Paradigm=Generative, Backbone Features=ResNet101 (finetuned)2022.04 | 70.7 | — | — | — | — | |
| CVC*Learning Paradigm=Generative, Backbone Features=ResNet101 (finetuned)2022.04 | 70 | — | — | — | — | |
| Composer2022.07 | 69.4 | — | — | — | — | |
| ComposerVenue=NeurIPS'20, Feature Type=CNN, Attention=true2024.04 | 69.4 | — | — | — | — | |
| ComposerCategory=Generative Methods, Attention-based=true, Backbone=ResNet1012026.03 | 69.4 | — | — | — | — | |
| ViFRCategory=Generative Methods, Backbone=ResNet1012026.03 | 69.1 | — | — | — | — | |
| LFGAA2021.03 | 67.6 | — | — | — | — | |
| LFGAA+HybridLearning Paradigm=Non-generative2022.04 | 67.6 | — | — | — | — | |
| LFGAACategory=Embedding-based Methods, Attention-based=true, Backbone=ResNet1012026.03 | 67.6 | — | — | — | — | |
| ZeroNASBackbone=ResNet, Method Category=Generative2026.03 | 66.4 | — | — | — | — | |
| DAZLE2022.07 | 66 | — | — | — | — | |
| DAZLEVenue=CVPR'20, Feature Type=CNN, Attention=true2024.04 | 66 | — | — | — | — | |
| DAZLECategory=Embedding-based Methods, Attention-based=true, Backbone=ResNet1012026.03 | 66 | — | — | — | — | |
| TF-VAEGANSetting=Inductive (IN)2020.03 | 64.9 | — | — | — | — | |
| TF-VAEGAN2021.03 | 64.9 | — | — | — | — | |
| TF-VAEGANGenerative=true2022.07 | 64.9 | — | — | — | — | |
| TF-VAEGANVenue=ECCV'20, Feature Type=CNN2024.04 | 64.9 | — | — | — | — | |
| SBAR-ISetting=Fine-tuned Inductive (FT-IN)2020.03 | 63.9 | — | — | — | — | |
| HSVA2022.07 | 62.8 | — | — | — | — | |
| HSVAVenue=NeurIPS'21, Feature Type=CNN2024.04 | 62.8 | — | — | — | — | |
| HSVACategory=Common Space Learning, Backbone=ResNet1012026.03 | 62.8 | — | — | — | — | |
| DLFZRL2021.03 | 61.8 | — | — | — | — | |
| ALEFeature Generator=f-CLSWGAN2017.12 | 61.5 | — | — | — | — | |
| f-VAEGAN-D2Learning Setting=Inductive, Fine-tuned=false2019.03 | 61 | — | — | — | — | |
| f-VAEGANSetting=Inductive (IN)2020.03 | 61 | — | — | — | — | |
| f-VAEGAN-D22021.03 | 61 | — | — | — | — | |
| f-VAEGANVenue=CVPR'19, Feature Type=CNN2024.04 | 61 | — | — | — | — |