Image Classification on Stanford Cars (test)
96.13AccuracyALIGN
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| ALIGNBackbone=EfficientNet-L2, Resolution=289/3602021.02 | 96.13 | — | — | — | — | — | — | — | — | — | |
| SR-GNN (Full Model)Param (M)=30.9, GFLOPS (B)=9.8, Infer. time per img. (ms)=5.02022.09 | 96.1 | — | — | — | — | — | 2.2 | — | — | — | |
| State of the artresolution=448x4482021.07 | 96 | — | — | — | — | — | — | — | — | — | |
| SAM-finalBackbone=EfficientNet-L2, Fine-tuning strategy=With SAM optimization2021.02 | 95.96 | — | — | — | — | — | — | — | — | — | |
| SaSPA + CutMixFT Strategy=No fine-tuning, Backbone=ViT-B/16@3842024.06 | 95.37 | — | — | — | — | — | — | — | — | — | |
| SaSPA + CutMixBackbone=ViT-B/16, Resolution=384x384, FT Strategy=No fine-tuning2024.06 | 95.37 | — | — | — | — | — | — | — | — | — | |
| Ensemblemode=ensemble2021.07 | 95.35 | — | — | — | — | — | — | — | — | — | |
| SaSPA + CutMixFT Strategy=No fine-tuning, Backbone=ResNet50@4482024.06 | 95.34 | — | — | — | — | — | — | — | — | — | |
| SaSPA + CutMixBackbone=ResNet50, Resolution=448x448, FT Strategy=No fine-tuning2024.06 | 95.34 | — | — | — | — | — | — | — | — | — | |
| SaSPAFT Strategy=No fine-tuning, Backbone=ResNet50@4482024.06 | 95.29 | — | — | — | — | — | — | — | — | — | |
| SaSPABackbone=ResNet50, Resolution=448x448, FT Strategy=No fine-tuning2024.06 | 95.29 | — | — | — | — | — | — | — | — | — | |
| Diff-Mix + CutMixFT Strategy=TI+DB, Backbone=ViT-B/16@3842024.06 | 95.26 | — | — | — | — | — | — | — | — | — | |
| Diff-Mix + CutMixBackbone=ViT-B/16, Resolution=384x384, FT Strategy=Textual Inversion + DreamBooth2024.06 | 95.26 | — | — | — | — | — | — | — | — | — | |
| Diff-Mix + CutMixFT Strategy=TI+DB, Backbone=ResNet50@4482024.06 | 95.21 | — | — | — | — | — | — | — | — | — | |
| Diff-Mix + CutMixBackbone=ResNet50, Resolution=448x448, FT Strategy=Textual Inversion + DreamBooth2024.06 | 95.21 | — | — | — | — | — | — | — | — | — | |
| Diff-MixFT Strategy=TI+DB, Backbone=ResNet50@4482024.06 | 95.12 | — | — | — | — | — | — | — | — | — | |
| SaSPAFT Strategy=No fine-tuning, Backbone=ViT-B/16@3842024.06 | 95.12 | — | — | — | — | — | — | — | — | — | |
| Diff-MixBackbone=ResNet50, Resolution=448x448, FT Strategy=Textual Inversion + DreamBooth2024.06 | 95.12 | — | — | — | — | — | — | — | — | — | |
| SaSPABackbone=ViT-B/16, Resolution=384x384, FT Strategy=No fine-tuning2024.06 | 95.12 | — | — | — | — | — | — | — | — | — | |
| EfficientNetV2-LParams=121M, Pre-trained=ImageNet ILSVRC2012, Evaluation Protocol=finetuned2021.04 | 95.1 | — | — | — | — | — | — | — | — | — | |
| Diff-MixFT Strategy=TI+DB, Backbone=ViT-B/16@3842024.06 | 95.09 | — | — | — | — | — | — | — | — | — | |
| Diff-MixBackbone=ViT-B/16, Resolution=384x384, FT Strategy=Textual Inversion + DreamBooth2024.06 | 95.09 | — | — | — | — | — | — | — | — | — | |
| SAM-baselineBackbone=EfficientNet-L2, Fine-tuning strategy=Without SAM optimization2021.02 | 95.07 | — | — | — | — | — | — | — | — | — | |
| ELOPEBackbone CNN=ResNet-1012019.11 | 95 | — | — | — | — | — | — | — | — | — | |
| MixupBackbone=ViT-B/16, Resolution=384x384, FT Strategy=Standard2024.06 | 94.98 | — | — | — | — | — | — | — | — | — | |
| CMNBackbone=ResNet-502020.12 | 94.9 | — | — | — | — | — | — | — | — | — | |
| Non-binary Transfer Learning (Optimal)reinitialized layers=1 new layer, high learning rate=0.025, low learning rate=0.01, low layers count=8, regularization=none2021.07 | 94.86 | — | — | — | — | — | — | — | — | — | |
| CutMixFT Strategy=-, Backbone=ViT-B/16@3842024.06 | 94.83 | — | — | — | — | — | — | — | — | — | |
| CutMixBackbone=ViT-B/16, Resolution=384x384, FT Strategy=Standard2024.06 | 94.83 | — | — | — | — | — | — | — | — | — | |
| Cubuk et al.2018.11 | 94.8 | — | — | — | — | — | — | — | — | — | |
| GPipeBackbone CNN=AmoebaNet-B2019.11 | 94.8 | — | — | — | — | — | — | — | — | — | |
| AutoAugmBackbone CNN=Inception-v42019.11 | 94.8 | — | — | — | — | — | — | — | — | — | |
| CutMixFT Strategy=-, Backbone=ResNet50@4482024.06 | 94.73 | — | — | — | — | — | — | — | — | — | |
| CutMixBackbone=ResNet50, Resolution=448x448, FT Strategy=Standard2024.06 | 94.73 | — | — | — | — | — | — | — | — | — | |
| GPipeParams=556M, Pre-trained=ImageNet ILSVRC2012, Evaluation Protocol=finetuned2021.04 | 94.7 | — | — | — | — | — | — | — | — | — | |
| EfficientNet-B7Params=66M, Pre-trained=ImageNet ILSVRC2012, Evaluation Protocol=finetuned2021.04 | 94.7 | — | — | — | — | — | — | — | — | — | |
| EfficientNet-B7Architecture=EfficientNet-B72021.06 | 94.7 | — | — | — | — | — | — | — | — | — | |
| DA-fusionBackbone=ResNet50, Resolution=448x448, FT Strategy=Textual Inversion2024.06 | 94.69 | — | — | — | — | — | — | — | — | — | |
| Real-filteringBackbone=ViT-B/16, Resolution=384x384, FT Strategy=No fine-tuning2024.06 | 94.66 | — | — | — | — | — | — | — | — | — | |
| Real-guidanceBackbone=ViT-B/16, Resolution=384x384, FT Strategy=No fine-tuning2024.06 | 94.65 | — | — | — | — | — | — | — | — | — | |
| GPipe (AmoebaNet-B (18, 512))Architecture=AmoebaNet-B (18, 512), Resolution=480x480, Protocol=Fine-tuned, Number of fine-tuning runs=5, Crop=Single-crop2018.11 | 94.6 | — | — | — | — | — | — | — | — | — | |
| Cross-X1-Stage=true, Sep. Init.=false, Backbone=ResNet-50, Multi-crop=false2019.09 | 94.6 | — | — | — | — | — | — | — | — | — | |
| Cross-XBackbone=ResNet-502020.12 | 94.6 | — | — | — | — | — | — | — | — | — | |
| ACNetBackbone=ResNet-502020.12 | 94.6 | — | — | — | — | — | — | — | — | — | |
| ACNetBase Model=ResNet502021.01 | 94.6 | — | — | — | — | — | — | — | — | — | |
| EfficientNetV2-MParams=55M, Pre-trained=ImageNet ILSVRC2012, Evaluation Protocol=finetuned2021.04 | 94.6 | — | — | — | — | — | — | — | — | — | |
| Real-filteringBackbone=ResNet50, Resolution=448x448, FT Strategy=No fine-tuning2024.06 | 94.59 | — | — | — | — | — | — | — | — | — | |
| L2 (Default)learning rate=0.025, regularization=L22021.07 | 94.59 | — | — | — | — | — | — | — | — | — | |
| Real-guidanceBackbone=ResNet50, Resolution=448x448, FT Strategy=No fine-tuning2024.06 | 94.55 | — | — | — | — | — | — | — | — | — | |
| BaselineBackbone=ResNet50, Resolution=448x448, FT Strategy=Standard2024.06 | 94.54 | — | — | — | — | — | — | — | — | — | |
| DA-fusionBackbone=ViT-B/16, Resolution=384x384, FT Strategy=Textual Inversion2024.06 | 94.53 | — | — | — | — | — | — | — | — | — | |
| Cross-X1-Stage=true, Sep. Init.=false, Backbone=SENet-50, Multi-crop=false2019.09 | 94.5 | — | — | — | — | — | — | — | — | — | |
| DCLBackbone CNN=ResNet-502019.11 | 94.5 | — | — | — | — | — | — | — | — | — | |
| ELOPEBackbone CNN=ResNet-502019.11 | 94.5 | — | — | — | — | — | — | — | — | — | |
| CIN1-Stage=true, Backbone=ResNet-1012020.03 | 94.5 | — | — | — | — | — | — | — | — | — | |
| DCLBackbone=ResNet-502020.12 | 94.5 | — | — | — | — | — | — | — | — | — | |
| LIOBackbone=ResNet-502020.12 | 94.5 | — | — | — | — | — | — | — | — | — | |
| DCLBase Model=ResNet502021.01 | 94.5 | — | — | — | — | — | — | — | — | — | |
| MixupBackbone=ResNet50, Resolution=448x448, FT Strategy=Standard2024.06 | 94.49 | — | — | — | — | — | — | — | — | — | |
| L2SP (Default)learning rate=0.025, regularization=L2SP2021.07 | 94.42 | — | — | — | — | — | — | — | — | — | |
| OursBase Model=ResNet502021.01 | 94.41 | — | — | — | — | — | — | — | — | — | |
| MoCo V2 Anisotropictraining_mode=SSL, texture_suppression=anisotropic2020.11 | 94.3 | — | — | — | — | — | — | — | — | — | |
| BaselineBackbone=ViT-B/16, Resolution=384x384, FT Strategy=Standard2024.06 | 94.21 | — | — | — | — | — | — | — | — | — | |
| DLABackbone CNN=DLA-X-60-C2019.11 | 94.1 | — | — | — | — | — | — | — | — | — | |
| ISEBackbone CNN=ResNet-502019.11 | 94.1 | — | — | — | — | — | — | — | — | — | |
| CIN1-Stage=true, Backbone=ResNet-502020.03 | 94.1 | — | — | — | — | — | — | — | — | — | |
| NTS-Net1-Stage=true, Sep. Init.=false, Multi-crop=true2019.09 | 93.9 | — | — | — | — | — | — | — | — | — | |
| MaxEnt-CNN1-Stage=true, Sep. Init.=false, Multi-crop=false2019.09 | 93.9 | — | — | — | — | — | — | — | — | — | |
| NTS-NetBackbone CNN=ResNet-502019.11 | 93.9 | — | — | — | — | — | — | — | — | — | |
| NTS1-Stage=false2020.03 | 93.9 | — | — | — | — | — | — | — | — | — | |
| NTS-NetBackbone=ResNet-502020.12 | 93.9 | — | — | — | — | — | — | — | — | — | |
| NTSBase Model=ResNet502021.01 | 93.9 | — | — | — | — | — | — | — | — | — | |
| NTS-NetInfer. time per img. (ms)=35.02022.09 | 93.9 | — | — | — | — | — | — | — | — | — | |
| DFL-CNNBase Model=VGG-16, Scale=2-scale2016.11 | 93.8 | — | — | — | — | — | — | — | — | — | |
| TASNBackbone=1x Resnet-502019.03 | 93.8 | — | — | — | — | — | — | — | — | — | |
| DFB-CNN1-Stage=true, Sep. Init.=true, Multi-crop=false2019.09 | 93.8 | — | — | — | — | — | — | — | — | — | |
| DFL-CNNBackbone CNN=VGG-162019.11 | 93.8 | — | — | — | — | — | — | — | — | — | |
| DFL-CNN1-Stage=false2020.03 | 93.8 | — | — | — | — | — | — | — | — | — | |
| TASNBackbone=ResNet-502020.12 | 93.8 | — | — | — | — | — | — | — | — | — | |
| TASNBase Model=ResNet502021.01 | 93.8 | — | — | — | — | — | — | — | — | — | |
| EfficientNetV2-SParams=24M, Pre-trained=ImageNet ILSVRC2012, Evaluation Protocol=finetuned2021.04 | 93.8 | — | — | — | — | — | — | — | — | — | |
| NTSNBackbone=3xResnet-502019.03 | 93.7 | — | — | — | — | — | — | — | — | — | |
| HBP1-Stage=true2020.03 | 93.7 | — | — | — | — | — | — | — | — | — | |
| XCiT-L24/16Architecture=XCiT-L24/16, distillation=true2021.06 | 93.7 | — | — | — | — | — | — | — | — | — | |
| WiSE-FTalpha=optimal2021.09 | 93.6 | — | — | — | — | — | — | — | — | — | |
| GMMBackbone CNN=VGG-192019.11 | 93.5 | — | — | — | — | — | — | — | — | — | |
| PC-ResNet-50Backbone=ResNet-50, Pairwise Confusion=true2017.05 | 93.43 | — | — | — | — | — | — | — | — | — | |
| Spatial RNNBackbone CNN=M-Net/D-Net2019.11 | 93.4 | — | — | — | — | — | — | — | — | — | |
| XCiT-M24/16Architecture=XCiT-M24/16, distillation=true2021.06 | 93.4 | — | — | — | — | — | — | — | — | — | |
| CBAMBase Model=ResNet502021.01 | 93.35 | — | — | — | — | — | — | — | — | — | |
| OursBase Model=VGG192021.01 | 93.32 | — | — | — | — | — | — | — | — | — | |
| DFL-CNNBase Model=VGG-16, Scale=1-scale2016.11 | 93.3 | — | — | — | — | — | — | — | — | — | |
| iSQRT-COVBackbone CNN=ResNet-1012019.11 | 93.3 | — | — | — | — | — | — | — | — | — | |
| DFL-CNNBackbone=VGG-162020.12 | 93.3 | — | — | — | — | — | — | — | — | — | |
| DeiT-B-384Params=86M, Pre-trained=ImageNet ILSVRC2012, Evaluation Protocol=finetuned2021.04 | 93.3 | — | — | — | — | — | — | — | — | — | |
| WiSE-FTalpha=0.52021.09 | 93.3 | — | — | — | — | — | — | — | — | — | |
| TASNBackbone=3xVGG-192019.03 | 93.2 | — | — | — | — | — | — | — | — | — | |
| DFLBase Model=ResNet502021.01 | 93.11 | — | — | — | — | — | — | — | — | — | |
| DFL-CNNBase Model=ResNet-50, Scale=1-scale2016.11 | 93.1 | — | — | — | — | — | — | — | — | — | |
| FCAN1-Stage=false, Sep. Init.=true, Multi-crop=true2019.09 | 93.1 | — | — | — | — | — | — | — | — | — |