Image Classification on Caltech101 (test)
99.12AccuracySDFed
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| SDFedBackbone=ViT-B162026.02 | 99.12 | — | — | — | — | — | |
| UOPPBackbone=ViT-B162026.02 | 98.24 | — | — | — | — | — | |
| ViT-L/16 + BackgroundArchitecture=ViT-L/16, Epochs=2, Spinal FC=false, Background Class=true2023.05 | 98.02 | — | — | — | — | — | |
| ViT-L/16 + (Spinal FC & Background)Architecture=ViT-L/16, Epochs=2, Spinal FC=true, Background Class=true2023.05 | 97.69 | — | — | — | — | — | |
| GPT-FLBackbone=ViT-B162026.02 | 97.58 | — | — | — | — | — | |
| ViT-L/16 + Spinal FCArchitecture=ViT-L/16, Epochs=2, Spinal FC=true, Background Class=false2023.05 | 97.31 | — | — | — | — | — | |
| Deformable ProtoPNetBackbone=Inception V3, Explanation=Single-Scale2024.04 | 97.22 | — | — | — | — | — | |
| ST-ProtoPNetBackbone=ConvNeXt-tiny, Explanation=Single-Scale2024.04 | 97.17 | — | — | — | — | — | |
| ViT-L/16Architecture=ViT-L/16, Epochs=2, Spinal FC=false, Background Class=false2023.05 | 97.02 | — | — | — | — | — | |
| FedOTPBackbone=ViT-B162026.02 | 97.02 | — | — | — | — | — | |
| ST-ProtoPNetBackbone=Inception V3, Explanation=Single-Scale2024.04 | 96.99 | — | — | — | — | — | |
| PIP-NetBackbone=ConvNeXt-tiny, Explanation=Single-Scale2024.04 | 96.61 | — | — | — | — | — | |
| BaselineBackbone=ConvNeXt-tiny, Explanation=N/A2024.04 | 96.56 | — | — | — | — | — | |
| BaselineBackbone=Inception V3, Explanation=N/A2024.04 | 96.42 | — | — | — | — | — | |
| SDFedBackbone=ResNet502026.02 | 96.08 | — | — | — | — | — | |
| DSPTNoise Type=PAIR, Noise Rate=20%2026.05 | 96.06 | — | — | — | — | — | |
| ST-ProtoPNetBackbone=ResNet50, Explanation=Single-Scale2024.04 | 95.95 | — | — | — | — | — | |
| MCPNetBackbone=ConvNeXt-tiny, Explanation=Multi-Scale2024.04 | 95.95 | — | — | — | — | — | |
| FR-ResNet18-EntropyBackbone=ResNet18, Acquisition Strategy=Entropy, #Samples=60002022.10 | 95.9 | — | — | — | — | — | |
| DSPTNoise Type=PAIR, Noise Rate=30%2026.05 | 95.85 | — | — | — | — | — | |
| FR-ResNet18-EntropyBackbone=ResNet18, Acquisition Strategy=Entropy, #Samples=40002022.10 | 95.83 | — | — | — | — | — | |
| MCPNetBackbone=Inception V3, Explanation=Multi-Scale2024.04 | 95.76 | — | — | — | — | — | |
| FR-ResNet18-EntropyBackbone=ResNet18, Acquisition Strategy=Entropy, #Samples=50002022.10 | 95.7 | — | — | — | — | — | |
| HEART-PFLalpha=0.12026.03 | 95.67 | — | — | — | — | — | |
| ResNet18-core-setBackbone=ResNet18, Acquisition Strategy=core-set, #Samples=50002022.10 | 95.56 | — | — | — | — | — | |
| NLPromptNoise Type=PAIR, Noise Rate=20%2026.05 | 95.56 | — | — | — | — | — | |
| ResNet18-EntropyBackbone=ResNet18, Acquisition Strategy=Entropy, #Samples=60002022.10 | 95.54 | — | — | — | — | — | |
| ResNet18-EntropyBackbone=ResNet18, Acquisition Strategy=Entropy, #Samples=40002022.10 | 95.5 | — | — | — | — | — | |
| ResNet18-core-setBackbone=ResNet18, Acquisition Strategy=core-set, #Samples=60002022.10 | 95.48 | — | — | — | — | — | |
| ResNet18-EntropyBackbone=ResNet18, Acquisition Strategy=Entropy, #Samples=50002022.10 | 95.42 | — | — | — | — | — | |
| FR-ResNet18-EntropyBackbone=ResNet18, Acquisition Strategy=Entropy, #Samples=30002022.10 | 95.38 | — | — | — | — | — | |
| NLPromptNoise Type=SYM, Noise Rate=40%2026.05 | 95.36 | — | — | — | — | — | |
| ResNet18-core-setBackbone=ResNet18, Acquisition Strategy=core-set, #Samples=40002022.10 | 95.34 | — | — | — | — | — | |
| DSPTNoise Type=SYM, Noise Rate=60%2026.05 | 95.31 | — | — | — | — | — | |
| ResNet18-EntropyBackbone=ResNet18, Acquisition Strategy=Entropy, #Samples=30002022.10 | 95.24 | — | — | — | — | — | |
| NLPromptNoise Type=PAIR, Noise Rate=30%2026.05 | 95.21 | — | — | — | — | — | |
| DSPTNoise Type=SYM, Noise Rate=40%2026.05 | 95.07 | — | — | — | — | — | |
| ResNet18-core-setBackbone=ResNet18, Acquisition Strategy=core-set, #Samples=30002022.10 | 95.04 | — | — | — | — | — | |
| UOPPBackbone=ResNet502026.02 | 94.78 | — | — | — | — | — | |
| PowderBackbone=ViT-B162026.02 | 94.75 | — | — | — | — | — | |
| Supervised-INEvaluation protocol=Linear evaluation, Backbone=ResNet-50, Pre-trained=ImageNet2022.01 | 94.5 | — | — | — | — | — | |
| HEART-PFLalpha=0.32026.03 | 94.49 | — | — | — | — | — | |
| NLPromptNoise Type=SYM, Noise Rate=60%2026.05 | 94.42 | — | — | — | — | — | |
| HEART-PFLalpha=0.52026.03 | 94.35 | — | — | — | — | — | |
| CoCoOp2023.09 | 94.3 | — | — | — | — | — | |
| DePTBaseline=KgCoOp2023.09 | 94.23 | — | — | — | — | — | |
| BaselineBackbone=ResNet50, Explanation=N/A2024.04 | 94.21 | — | — | — | — | — | |
| BYOLEvaluation protocol=Linear evaluation, Backbone=ResNet-50, Pre-trained=ImageNet2022.01 | 94.2 | — | — | — | — | — | |
| LLALAcquisition Strategy=LLAL, #Samples=60002022.10 | 94.18 | — | — | — | — | — | |
| KgCoOp2023.09 | 94.17 | — | — | — | — | — | |
| DePTBaseline=CoCoOp2023.09 | 94.1 | — | — | — | — | — | |
| LLALAcquisition Strategy=LLAL, #Samples=50002022.10 | 94.08 | — | — | — | — | — | |
| LLALAcquisition Strategy=LLAL, #Samples=30002022.10 | 94.02 | — | — | — | — | — | |
| DSPTNoise Type=SYM, Noise Rate=80%2026.05 | 94.01 | — | — | — | — | — | |
| CoOp2023.09 | 93.97 | — | — | — | — | — | |
| LLALAcquisition Strategy=LLAL, #Samples=40002022.10 | 93.89 | — | — | — | — | — | |
| Deformable ProtoPNetBackbone=ResNet50, Explanation=Single-Scale2024.04 | 93.88 | — | — | — | — | — | |
| MCPNetBackbone=ResNet50, Explanation=Multi-Scale2024.04 | 93.88 | — | — | — | — | — | |
| FedTPGnum_prompts=8, alpha=0.32024.09 | 93.88 | — | — | — | — | — | |
| BYOLEvaluation protocol=Fine-tuned, Backbone=ResNet-50, Pre-trained=ImageNet2022.01 | 93.8 | — | — | — | — | — | |
| DePTBaseline=PromptSRC2023.09 | 93.8 | — | — | — | — | — | |
| PromptSRC2023.09 | 93.77 | — | — | — | — | — | |
| DSPTNoise Type=PAIR, Noise Rate=40%2026.05 | 93.71 | — | — | — | — | — | |
| Deformable ProtoPNetBackbone=ConvNeXt-tiny, Explanation=Single-Scale2024.04 | 93.65 | — | — | — | — | — | |
| NLPromptNoise Type=PAIR, Noise Rate=40%2026.05 | 93.63 | — | — | — | — | — | |
| PromptFolionum_prompts=8, alpha=0.32024.09 | 93.59 | — | — | — | — | — | |
| PromptFL+FedProxnum_prompts=8, alpha=0.3, federated_optimizer=FedProx2024.09 | 93.57 | — | — | — | — | — | |
| ResNet18-core-setBackbone=ResNet18, Acquisition Strategy=core-set, #Samples=20002022.10 | 93.5 | — | — | — | — | — | |
| PromptFLnum_prompts=8, alpha=0.32024.09 | 93.47 | — | — | — | — | — | |
| GPT-FLBackbone=ResNet502026.02 | 93.45 | — | — | — | — | — | |
| FR-ResNet18-EntropyBackbone=ResNet18, Acquisition Strategy=Entropy, #Samples=20002022.10 | 93.3 | — | — | — | — | — | |
| Supervised-INEvaluation protocol=Fine-tuned, Backbone=ResNet-50, Pre-trained=ImageNet2022.01 | 93.3 | — | — | — | — | — | |
| DePTBaseline=CoOp2023.09 | 93.3 | — | — | — | — | — | |
| ResNet18-EntropyBackbone=ResNet18, Acquisition Strategy=Entropy, #Samples=20002022.10 | 93.27 | — | — | — | — | — | |
| ReLICv2Evaluation protocol=Fine-tuned, Backbone=ResNet-50, Pre-trained=ImageNet2022.01 | 93.2 | — | — | — | — | — | |
| FedPeralpha=0.12026.03 | 93.11 | — | — | — | — | — | |
| MaPLe2023.09 | 92.97 | — | — | — | — | — | |
| SmoothingNoise Type=SYM, Noise Rate=40%2026.05 | 92.82 | — | — | — | — | — | |
| ReLICv2Evaluation protocol=Linear evaluation, Backbone=ResNet-50, Pre-trained=ImageNet2022.01 | 92.8 | — | — | — | — | — | |
| FedAvg-peralpha=0.52026.03 | 92.68 | — | — | — | — | — | |
| FedProx-peralpha=0.52026.03 | 92.63 | — | — | — | — | — | |
| DePTBaseline=MaPLe2023.09 | 92.53 | — | — | — | — | — | |
| Dittoalpha=0.52026.03 | 92.51 | — | — | — | — | — | |
| FedALAalpha=0.52026.03 | 92.42 | — | — | — | — | — | |
| FedOTPBackbone=ResNet502026.02 | 92.15 | — | — | — | — | — | |
| SimCLREvaluation protocol=Fine-tuned, Backbone=ResNet-50, Pre-trained=ImageNet2022.01 | 92.1 | — | — | — | — | — | |
| FedALAalpha=0.32026.03 | 92.02 | — | — | — | — | — | |
| FedBABUalpha=0.52026.03 | 92.01 | — | — | — | — | — | |
| Dittoalpha=0.32026.03 | 91.92 | — | — | — | — | — | |
| FedBABUalpha=0.32026.03 | 91.9 | — | — | — | — | — | |
| FedAvg-peralpha=0.32026.03 | 91.89 | — | — | — | — | — | |
| LLALAcquisition Strategy=LLAL, #Samples=20002022.10 | 91.88 | — | — | — | — | — | |
| NLPromptNoise Type=SYM, Noise Rate=80%2026.05 | 91.85 | — | — | — | — | — | |
| FedProx-peralpha=0.32026.03 | 91.73 | — | — | — | — | — | |
| PromptFL+FedPernum_prompts=8, alpha=0.3, federated_strategy=FedPer2024.09 | 91.7 | — | — | — | — | — | |
| VPT2023.09 | 91.67 | — | — | — | — | — | |
| LG-FedAvgalpha=0.12026.03 | 91.53 | — | — | — | — | — | |
| FedBABUalpha=0.12026.03 | 91.5 | — | — | — | — | — | |
| PromptFL+FedAMPnum_prompts=8, alpha=0.3, federated_strategy=FedAMP2024.09 | 91.38 | — | — | — | — | — | |
| NNCLREvaluation protocol=Linear evaluation, Backbone=ResNet-50, Pre-trained=ImageNet2022.01 | 91.3 | — | — | — | — | — |