Image Classification on Cars
96.4AccuracyKL
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| KLSelection Method=KL, Adaptation Strategy=Full, Training Epochs=4e, Backbone Architecture=Resnet-50-v22020.09 | 96.4 | — | — | — | — | — | |
| EPNSelection Method=EPN, Adaptation Strategy=Full, Training Epochs=2e, Backbone Architecture=Resnet-50-v22020.09 | 96.2 | — | — | — | — | — | |
| KNNSelection Method=KNN, Adaptation Strategy=Full, Training Epochs=4e, Backbone Architecture=Resnet-50-v22020.09 | 96.1 | — | — | — | — | — | |
| KLSelection Method=KL, Adaptation Strategy=Full, Training Epochs=2e, Backbone Architecture=Resnet-50-v22020.09 | 96.1 | — | — | — | — | — | |
| KNNSelection Method=KNN, Adaptation Strategy=Full, Training Epochs=2e, Backbone Architecture=Resnet-50-v22020.09 | 96 | — | — | — | — | — | |
| EPNSelection Method=EPN, Adaptation Strategy=Full, Training Epochs=4e, Backbone Architecture=Resnet-50-v22020.09 | 96 | — | — | — | — | — | |
| KNNSelection Method=KNN, Adaptation Strategy=Adapters, Training Epochs=4e, Backbone Architecture=Resnet-50-v22020.09 | 95.9 | — | — | — | — | — | |
| KLSelection Method=KL, Adaptation Strategy=Adapters, Training Epochs=4e, Backbone Architecture=Resnet-50-v22020.09 | 95.9 | — | — | — | — | — | |
| EPNSelection Method=EPN, Adaptation Strategy=Adapters, Training Epochs=4e, Backbone Architecture=Resnet-50-v22020.09 | 95.8 | — | — | — | — | — | |
| Dom-Ad (Am-B)Backbone Architecture=AmoebaNet-B2020.09 | 95.8 | — | — | — | — | — | |
| CapPa L/14MAP head=true, Grain=Fine, Frozen representation=true, backbone=ViT-L/142023.06 | 95.8 | — | — | — | — | — | |
| Dom-Ad (In-v3)Backbone Architecture=Inception-v32020.09 | 95.7 | — | — | — | — | — | |
| BaselineBackbone Architecture=Resnet-50-v22020.09 | 95.6 | — | — | — | — | — | |
| θA fine-tuneNumber of calibration samples (N)=-2026.05 | 95.22 | — | — | — | — | — | |
| HIVEFoundation Model=SigLIP2026.03 | 95.09 | — | — | — | — | — | |
| BaseFoundation Model=SigLIP2026.03 | 94.86 | — | — | — | — | — | |
| SAFoundation Model=SigLIP2026.03 | 94.85 | — | — | — | — | — | |
| iSICEBackbone=ResNeXt-1012023.04 | 94.5 | — | — | — | — | — | |
| iSICEBackbone=VGG-162023.04 | 94 | — | — | — | — | — | |
| CLIP* L/14MAP head=true, Grain=Fine, Frozen representation=true, backbone=ViT-L/142023.06 | 94 | — | — | — | — | — | |
| iSICEBackbone=VGG-192023.04 | 93.9 | — | — | — | — | — | |
| iSICEBackbone=ResNet-1012023.04 | 93.6 | — | — | — | — | — | |
| iSICEBackbone=ResNet-502023.04 | 93.5 | — | — | — | — | — | |
| CapMAP head=true, Grain=Fine, Frozen representation=true2023.06 | 93.4 | — | — | — | — | — | |
| iSICEBackbone=Swin-B2023.04 | 93.3 | — | — | — | — | — | |
| CapPaMAP head=true, Grain=Fine, Frozen representation=true2023.06 | 93.3 | — | — | — | — | — | |
| iSICEBackbone=ConvNext-T2023.04 | 93.1 | — | — | — | — | — | |
| Precision ΩBackbone=Swin-B2023.04 | 93.1 | — | — | — | — | — | |
| ISO-CTS/LARVBackbone=ViT-L/142026.02 | 93 | — | 0.8 | — | — | — | |
| fine-tunedBackbone=ViT-L/142026.02 | 92.8 | — | — | — | — | — | |
| CLIPPre-training Data Source=Synthetic, Evaluation Protocol=5-way, 5-shot2023.06 | 92.5 | — | — | — | — | — | |
| CLIPPre-training dataset=CC12M, Pre-training data source=Syn, Evaluation protocol=Few-shot2023.06 | 92.5 | — | — | — | — | — | |
| θB fine-tuneNumber of calibration samples (N)=-2026.05 | 92.5 | — | — | — | — | — | |
| Spot-tuneBackbone=ResNet-50, Param Count=7x (7x)2022.03 | 92.4 | — | — | — | — | — | |
| KPBackbone=VGG-162023.04 | 92.4 | — | — | — | — | — | |
| iSQRT-COVBackbone=ResNeXt-1012023.04 | 92.4 | — | — | — | — | — | |
| ISO-C/LARVBackbone=ViT-L/142026.02 | 92.4 | — | 0.9 | — | — | — | |
| PaRaMSModel Status=Protected Standalone (θ̂_def)2025.11 | 92.39 | — | — | — | — | — | |
| iSQRT-COVBackbone=ResNet-1012023.04 | 92.3 | — | — | — | — | — | |
| Precision ΩBackbone=VGG-192023.04 | 92.2 | — | — | — | — | — | |
| Precision ΩBackbone=ConvNext-T2023.04 | 92.2 | — | — | — | — | — | |
| BA^2Backbone=ResNet-50, Param Count=3.8x (1.71x)2022.03 | 92.14 | — | — | — | — | — | |
| Supervised-INArchitecture=ResNet-50, Pre-training Dataset=ImageNet, Evaluation Protocol=Fine-tuned2020.06 | 92.1 | — | — | — | — | — | |
| iSQRT-COVBackbone=ResNet-502023.04 | 92.1 | — | — | — | — | — | |
| FNCEvaluation Protocol=Fine-tuned, Backbone=ResNet, Pre-training Dataset=ImageNet2020.11 | 92 | — | — | — | — | — | |
| Improved BCNNBackbone=VGG-162023.04 | 92 | — | — | — | — | — | |
| Precision ΩBackbone=VGG-162023.04 | 92 | — | — | — | — | — | |
| Precision ΩBackbone=ResNet-502023.04 | 92 | — | — | — | — | — | |
| iSQRT-COVBackbone=Swin-B2023.04 | 92 | — | — | — | — | — | |
| Fine-TuningBackbone=ResNet-50, Param Count=6x2022.03 | 91.89 | — | — | — | — | — | |
| ScratchTraining mode=Training from Scratch, FLOPs (x 1e18)=12.92023.10 | 91.89 | — | — | — | — | — | |
| bert2BERTTraining mode=Training from the Pretrained Model: M(12,384) → M(12,768), FLOPs (x 1e18)=4.6, Ratio (Saving)=64.4%2023.10 | 91.88 | — | — | — | — | — | |
| MangoTraining mode=Training from the Pretrained Model: M(12,384) → M(12,768), FLOPs (x 1e18)=3.0, Ratio (Saving)=76.4%2023.10 | 91.83 | — | — | — | — | — | |
| LIGOTraining mode=Training from the Pretrained Model: M(12,384) → M(12,768), FLOPs (x 1e18)=5.7, Ratio (Saving)=55.7%2023.10 | 91.82 | — | — | — | — | — | |
| SimCLR v2Evaluation Protocol=Fine-tuned, Backbone=ResNet, Pre-training Dataset=ImageNet2020.11 | 91.8 | — | — | — | — | — | |
| StableRepPre-training Data Source=Synthetic, Evaluation Protocol=5-way, 5-shot2023.06 | 91.8 | — | — | — | — | — | |
| StableRepPre-training dataset=CC12M, Pre-training data source=Syn, Evaluation protocol=Few-shot2023.06 | 91.8 | — | — | — | — | — | |
| StackBERTTraining mode=Training from Scratch, FLOPs (x 1e18)=11.3, Ratio (Saving)=12.6%2023.10 | 91.71 | — | — | — | — | — | |
| HIHCABackbone=VGG-162023.04 | 91.7 | — | — | — | — | — | |
| DeepCOVBackbone=VGG-162023.04 | 91.7 | — | — | — | — | — | |
| BYOLArchitecture=ResNet-50, Pre-training Dataset=ImageNet, Evaluation Protocol=Fine-tuned2020.06 | 91.6 | — | — | — | — | — | |
| BYOLEvaluation Protocol=Fine-tuned, Backbone=ResNet, Pre-training Dataset=ImageNet2020.11 | 91.6 | — | — | — | — | — | |
| DeepKSPDBackbone=VGG-162023.04 | 91.6 | — | — | — | — | — | |
| CLIP* (16k)MAP head=true, Grain=Fine, Frozen representation=true2023.06 | 91.6 | — | — | — | — | — | |
| WTPBBackbone=ResNet-50, Param Count=6x (2.25x)2022.03 | 91.5 | — | — | — | — | — | |
| SimCLR (repro)Architecture=ResNet-50, Pre-training Dataset=ImageNet, Evaluation Protocol=Fine-tuned2020.06 | 91.4 | — | — | — | — | — | |
| Random initArchitecture=ResNet-50, Pre-training Dataset=ImageNet, Evaluation Protocol=Fine-tuned2020.06 | 91.4 | — | — | — | — | — | |
| iSQRT-COVBackbone=VGG-192023.04 | 91.4 | — | — | — | — | — | |
| Precision ΩBackbone=ResNet-1012023.04 | 91.4 | — | — | — | — | — | |
| IsoCLIPIntra-modal=✓, Classifier=NCM, Backbone=ViT-B/16-open2026.03 | 91.4 | — | — | — | — | — | |
| SimCLRArchitecture=ResNet-50, Pre-training Dataset=ImageNet, Evaluation Protocol=Fine-tuned2020.06 | 91.3 | — | — | — | — | — | |
| SimCLR v1Evaluation Protocol=Fine-tuned, Backbone=ResNet, Pre-training Dataset=ImageNet2020.11 | 91.3 | — | — | — | — | — | |
| iSICEBackbone=Swin-T2023.04 | 91.3 | — | — | — | — | — | |
| CBPBackbone=VGG-162023.04 | 91.2 | — | — | — | — | — | |
| iSQRT-COVBackbone=VGG-162023.04 | 91.2 | — | — | — | — | — | |
| BaseFoundation Model=CLIP2026.03 | 91.17 | — | — | — | — | — | |
| HIVEFoundation Model=CLIP2026.03 | 91.15 | — | — | — | — | — | |
| KPBackbone=ResNet-502023.04 | 91.1 | — | — | — | — | — | |
| SAFoundation Model=CLIP2026.03 | 91.04 | — | — | — | — | — | |
| TSV-M/LARVBackbone=ViT-L/142026.02 | 91 | — | 1.2 | — | — | — | |
| LRBPBackbone=VGG-162023.04 | 90.9 | — | — | — | — | — | |
| CLIPPre-training Data Source=Real, Evaluation Protocol=5-way, 5-shot2023.06 | 90.9 | — | — | — | — | — | |
| CLIPPre-training dataset=CC12M, Pre-training data source=Real, Evaluation protocol=Few-shot2023.06 | 90.9 | — | — | — | — | — | |
| CLIP* (8k)MAP head=true, Grain=Fine, Frozen representation=true2023.06 | 90.8 | — | — | — | — | — | |
| FLIPPre-training Data=LAION-400M, Backbone=ViT-L/14, Evaluation Protocol=Zero-shot classification2024.07 | 90.7 | — | — | — | — | — | |
| BCNNBackbone=VGG-162023.04 | 90.6 | — | — | — | — | — | |
| Image-ImageIntra-modal=✓, Classifier=NCM, Backbone=ViT-B/16-open2026.03 | 90.6 | — | — | — | — | — | |
| GalLoP+OursShots=42026.05 | 90.6 | — | — | — | — | — | |
| Precision ΩBackbone=Swin-T2023.04 | 90.5 | — | — | — | — | — | |
| DeiT-TGFLOPS=1.262023.05 | 90.3 | — | — | — | — | — | |
| ADAProtection Status=None (theta_merge)2025.11 | 90.3 | — | — | — | — | — | |
| MLCDPre-training Data=LAION-400M, Backbone=ViT-L/14, Evaluation Protocol=Zero-shot classification2024.07 | 90.1 | — | — | — | — | — | |
| CLIPPre-training dataset=RedCaps, Pre-training data source=Syn, Evaluation protocol=Few-shot2023.06 | 90.1 | — | — | — | — | — | |
| Precision ΩBackbone=ResNeXt-1012023.04 | 89.9 | — | — | — | — | — | |
| MPN-COVBackbone=VGG-162023.04 | 89.8 | — | — | — | — | — | |
| Image-TextIntra-modal=✗, Classifier=Zero-Shot, Backbone=ViT-B/16-open2026.03 | 89.8 | — | — | — | — | — | |
| TAPSBackbone=ResNet-50, Param Count=4.12x2022.03 | 89.76 | — | — | — | — | — | |
| iSQRT-COVBackbone=Swin-T2023.04 | 89.7 | — | — | — | — | — | |
| PiggybackBackbone=ResNet-50, Param Count=6x (2.25x)2022.03 | 89.62 | — | — | — | — | — | |
| LAION-CLIP ViT-LN=400M, Zero-shot=true2024.09 | 89.6 | — | — | — | — | — |