Image Classification on CIFAR10
97.9AccuracyMVT
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| MVTMLLM=MMICL, Backbone=ViT-L2025.12 | 97.9 | — | — | — | |
| MVT + FTMLLM=MMICL, Backbone=ViT-L2025.12 | 96.7 | — | — | — | |
| MVT + FTMLLM=Otter, Backbone=ViT-L2025.12 | 96.6 | — | — | — | |
| TimeSCLBackbone=ResNet502025.12 | 95.9 | — | — | — | |
| SCS-SupConBackbone=ResNet502025.12 | 95.9 | — | — | — | |
| PaCoBackbone=ResNet502025.12 | 95.8 | — | — | — | |
| CSTCNBackbone=ResNet502025.12 | 95.8 | — | — | — | |
| DACLBackbone=ResNet502025.12 | 95.8 | — | — | — | |
| PCLBackbone=ResNet502025.12 | 95.7 | — | — | — | |
| FNCLBackbone=ResNet502025.12 | 95.7 | — | — | — | |
| CS-SupCon w. ov.Backbone=ResNet502025.12 | 95.7 | — | — | — | |
| SupConBackbone=ResNet502025.12 | 95.6 | — | — | — | |
| SelfConBackbone=ResNet502025.12 | 95.6 | — | — | — | |
| Circle LossBackbone=ResNet502025.12 | 95.6 | — | — | — | |
| CSA-RSICBackbone=ResNet502025.12 | 95.6 | — | — | — | |
| CLIPMLLM=None, Backbone=ViT-L2025.12 | 95.6 | — | — | — | |
| CS-SupConBackbone=ResNet502025.12 | 95.4 | — | — | — | |
| VeRAParameter Count=0.10M2025.12 | 95.23 | — | — | — | |
| BYOLBackbone=ResNet502025.12 | 95 | — | — | — | |
| BaselineBackbone=ResNet502025.12 | 94.9 | — | — | — | |
| DoRAParameter Count=0.77M2025.12 | 94.78 | — | — | — | |
| AdaLoRAParameter Count=0.67M2025.12 | 94.72 | — | — | — | |
| MVTMLLM=Otter, Backbone=ViT-L2025.12 | 94.7 | — | — | — | |
| LoRAParameter Count=0.67M2025.12 | 94.5 | — | — | — | |
| Partial-LoRAParameter Count=0.17M2025.12 | 94.35 | — | — | — | |
| MD-SNNModel=ResNet18, Precision (W / U)=8 / 8, Timestep=42025.12 | 94.27 | — | — | — | |
| LoRA+Parameter Count=0.67M2025.12 | 94.22 | — | — | — | |
| Partial-AdaLoRAParameter Count=0.15M2025.12 | 94.11 | — | — | — | |
| MD-SNNModel=ResNet18, Precision (W / U)=4 / 4, Timestep=42025.12 | 94.05 | — | — | — | |
| SimCLRBackbone=ResNet502025.12 | 93.6 | — | — | — | |
| MINTModel=ResNet19, Precision (W / U)=4 / 4, Timestep=42025.12 | 91.45 | — | — | — | |
| MINTModel=ResNet19, Precision (W / U)=8 / 8, Timestep=42025.12 | 91.36 | — | — | — | |
| ADMM-QuantModel=7Conv+3FC, Precision (W / U)=4 / 32, Timestep=82025.12 | 89.4 | — | — | — | |
| ST-QuantModel=VGG9, Precision (W / U)=5 / 32, Timestep=252025.12 | 88.6 | — | — | — | |
| SpikeSimModel=VGG9, Precision (W / U)=8 / 32, Timestep=52025.12 | 88.11 | — | — | — | |
| STBP-QuantModel=AlexNet, Precision (W / U)=8 / 14, Timestep=82025.12 | 86.65 | — | — | — | |
| STBP-QuantModel=AlexNet, Precision (W / U)=4 / 10, Timestep=82025.12 | 84.09 | — | — | — | |
| Test-Time Dynamic Model SelectionModel Pair=EfficientNet-ViT, Trials=3002025.12 | 75.9 | — | — | — | |
| BPtraining_data_fraction=100%, Architecture=AlexNet2026.01 | 75.82 | — | — | — | |
| PC-Iavgtraining_data_fraction=100%, Architecture=AlexNet2026.01 | 75.41 | — | — | — | |
| PC-Iavgtraining_data_fraction=50%, Architecture=AlexNet2026.01 | 71.37 | — | — | — | |
| BPtraining_data_fraction=50%, Architecture=AlexNet2026.01 | 70.67 | — | — | — | |
| Test-Time Dynamic Model SelectionModel Pair=AlexNet-GoogleNet, Trials=3002025.12 | 70.6 | — | — | — | |
| Test-Time Dynamic Model SelectionModel Pair=AlexNet-ViT, Trials=3002025.12 | 69.5 | — | — | — | |
| Test-Time Dynamic Model SelectionModel Pair=EfficientNet-ViT, Trials=7002025.12 | 68 | — | — | — | |
| Test-Time Dynamic Model SelectionModel Pair=EfficientNet-ViT, Trials=10002025.12 | 67.8 | — | — | — | |
| Best Individual ModelModel Pair=EfficientNet-ViT, Trials=3002025.12 | 67.7 | — | — | — | |
| PC-Ifwtraining_data_fraction=100%, Architecture=AlexNet2026.01 | 66.72 | — | — | — | |
| Best Individual ModelModel Pair=EfficientNet-ViT, Trials=7002025.12 | 66.4 | — | — | — | |
| Best Individual ModelModel Pair=EfficientNet-ViT, Trials=10002025.12 | 66.4 | — | — | — | |
| Test-Time Dynamic Model SelectionModel Pair=AlexNet-ViT, Trials=7002025.12 | 66.2 | — | — | — | |
| PC-Iavgtraining_data_fraction=25%, Architecture=AlexNet2026.01 | 66.13 | — | — | — | |
| Test-Time Dynamic Model SelectionModel Pair=AlexNet-ViT, Trials=10002025.12 | 65.9 | — | — | — | |
| Best Individual ModelModel Pair=AlexNet-ViT, Trials=7002025.12 | 64.8 | — | — | — | |
| BPtraining_data_fraction=25%, Architecture=AlexNet2026.01 | 64.41 | — | — | — | |
| PC-Ifwtraining_data_fraction=50%, Architecture=AlexNet2026.01 | 64.25 | — | — | — | |
| PC-Ifwtraining_data_fraction=25%, Architecture=AlexNet2026.01 | 63.83 | — | — | — | |
| Best Individual ModelModel Pair=AlexNet-GoogleNet, Trials=3002025.12 | 62.7 | — | — | — | |
| Best Individual ModelModel Pair=AlexNet-ViT, Trials=3002025.12 | 62.4 | — | — | — | |
| Best Individual ModelModel Pair=AlexNet-ViT, Trials=10002025.12 | 62.4 | — | — | — | |
| Test-Time Dynamic Model SelectionModel Pair=EfficientNet-GoogleNet, Trials=3002025.12 | 59 | — | — | — | |
| Test-Time Dynamic Model SelectionModel Pair=AlexNet-GoogleNet, Trials=10002025.12 | 58.4 | — | — | — | |
| Best Individual ModelModel Pair=AlexNet-GoogleNet, Trials=7002025.12 | 57.7 | — | — | — | |
| Test-Time Dynamic Model SelectionModel Pair=AlexNet-GoogleNet, Trials=7002025.12 | 57.5 | — | — | — | |
| Test-Time Dynamic Model SelectionModel Pair=EfficientNet-GoogleNet, Trials=10002025.12 | 57.3 | — | — | — | |
| Best Individual ModelModel Pair=AlexNet-GoogleNet, Trials=10002025.12 | 56.8 | — | — | — | |
| Test-Time Dynamic Model SelectionModel Pair=EfficientNet-GoogleNet, Trials=7002025.12 | 55.8 | — | — | — | |
| Best Individual ModelModel Pair=EfficientNet-GoogleNet, Trials=3002025.12 | 54.8 | — | — | — | |
| Best Individual ModelModel Pair=EfficientNet-GoogleNet, Trials=10002025.12 | 54.8 | — | — | — | |
| Best Individual ModelModel Pair=EfficientNet-GoogleNet, Trials=7002025.12 | 53.6 | — | — | — | |
| ADA-NetBackbone=Conv-Large2019.05 | — | 10.3 | — | — | |
| ADA-NetBackbone=Conv-Large, Translation range=4, ZCA whitening=false2019.05 | — | 8.72 | — | — | |
| ADA-Net+Backbone=Conv-Large, Combined with=VAT+Ent2019.05 | — | 10.09 | — | — | |
| BaseCNNParams(M)=0.52019.07 | — | 8.3 | — | — | |
| Best result of others2012.02 | — | 18.5 | — | — | |
| CAMNet2Params(M)=1.22019.07 | — | 7.59 | — | — | |
| CAMNet3Params(M)=2.02019.07 | — | 7.02 | — | — | |
| CAMNet4Params(M)=3.02019.07 | — | 7.06 | — | — | |
| Cr-Stitch2Params(M)=12019.07 | — | 8.22 | — | — | |
| FASTalpha=0.52026.01 | — | — | — | 35.14 | |
| FASTalpha=1002026.01 | — | — | — | 45.07 | |
| FedAvgalpha=0.52026.01 | — | — | — | 43.93 | |
| FedAvgalpha=1002026.01 | — | — | — | 52.05 | |
| FedGHalpha=0.52026.01 | — | — | — | 43.38 | |
| FedGHalpha=1002026.01 | — | — | — | 52.94 | |
| FedNovaalpha=0.52026.01 | — | — | — | 43.72 | |
| FedNovaalpha=1002026.01 | — | — | — | 51.73 | |
| FedProtoalpha=0.52026.01 | — | — | — | 44.33 | |
| FedProtoalpha=1002026.01 | — | — | — | 56.05 | |
| FedProxalpha=0.52026.01 | — | — | — | 42.7 | |
| FedProxalpha=1002026.01 | — | — | — | 50.37 | |
| II ModelBackbone=Conv-Large2019.05 | — | 12.36 | — | — | |
| MA-DNNBackbone=Conv-Large2019.05 | — | 11.91 | — | — | |
| MCDNN2012.02 | — | 11.21 | 39 | — | |
| Mean TeacherBackbone=Conv-Large2019.05 | — | 12.31 | — | — | |
| Mean Teacher+fastSWABackbone=Conv-Large, Translation range=4, ZCA whitening=false2019.05 | — | 9.05 | — | — | |
| MultiCNN3Params(M)=1.52019.07 | — | 9.71 | — | — | |
| RefProtoFLalpha=0.52026.01 | — | — | — | 45.51 | |
| RefProtoFLalpha=1002026.01 | — | — | — | 56.36 | |
| SaaSBackbone=Conv-Large2019.05 | — | 13.22 | — | — |