List your company
ResearchFormerly WizWand

Image Classification on Stylized-ImageNet

56.2Top-1 AccuracySIN
2.95216.77630.644.424Nov 21, 2019Sep 1, 2020Jun 14, 2021Mar 27, 2022Jan 6, 2023Oct 19, 2023Jul 31, 2024
Updated 4mo ago

Evaluation results

MethodLinks
SINBackbone=ResNet-502020.1056.2
MAE-H + DAT2022.0932.77
LLEarch=ViT-H/14, train data=IN-1k, base_framework=MAE2022.1231.15
LLEarch=ViT-H/14, train data=IN-1k, base_framework=MAE, augmentation=Edge Aug2022.1230.94
LLEarch=ViT-L/16, train data=IN-1k, base_framework=MAE2022.1229.52
LLEarch=ViT-L/16, train data=IN-1k, base_framework=MAE, augmentation=Edge Aug2022.1229.32
AdvPropBackbone=EfficientNet-B7, Resolution=224x2242019.1126.6
MAE-H2022.0926.36
AdvPropBackbone=EfficientNet-B6, Resolution=224x2242019.1125.9
VOLO-D5HAT=true, Resolution=224x2242022.0425.9
LLEarch=ViT-B/16, train data=IN-1k, base_framework=MAE2022.1224.9
DATBackbone=ViT-B/16, Pre-training=ImageNet-21K, Fine-tuning=ImageNet-1K, Input Resolution=224x2242023.1124.8
ViT-BHAT=true, Resolution=224x2242022.0424.7
LLEarch=ViT-B/16, train data=IN-1k, base_framework=MAE, augmentation=Edge Aug2022.1224.65
VOLO-D5HAT=false, Resolution=224x2242022.0424.6
LLEarch=ViT-B/16, train data=IG-3.6B, base_framework=SWAG (FT)2022.1224.45
AdvPropBackbone=EfficientNet-B5, Resolution=224x2242019.1124.4
LLEarch=ViT-B/16, train data=IG-3.6B, base_framework=SWAG (FT), augmentation=Edge Aug2022.1224.07
DADBackbone=ViT-B/16, Pre-training=ImageNet-21K, Fine-tuning=ImageNet-1K, Input Resolution=224x2242023.1123.4
AugReg-ViT + DAT2022.0923.13
AdvPropBackbone=EfficientNet-B4, Resolution=224x2242019.1122.5
Dr. ViTBackbone=ViT-B, Discrete Only=true2021.1122.19
DeepAugment + Augmix + DAT2022.0921.89
EfficientNet-B7Resolution=224x2242019.1121.8
ViT-BHAT=false, Resolution=224x2242022.0421.8
AdvPropBackbone=EfficientNet-B3, Resolution=224x2242019.1121.4
DeepAugment + Augmix2022.0921.23
DeepAugment + AugMixBackbone=ResNet-502020.1021.2
VOLO-D1HAT=true, Resolution=224x2242022.0421.2
EfficientNet-B6Resolution=224x2242019.1120.9
EfficientNet-B5Resolution=224x2242019.1120.8
EfficientNet-B4Resolution=224x2242019.1120.2
ViT-B/16 + PyramidATBackbone=ViT-B/16, Resolution=512x512, Pre-training=ImageNet-21K, Fine-tuning=ImageNet-1K, Adversarial Training=PyramidAT2021.1119.84
Dr. ViTPre-trained=ImageNet-21K, Fine-tuned on=ImageNet, Resolution=384x384, Discrete representation=discrete only2021.1119.69
Dr. ViTBackbone=ViT-B, Discrete Only=false2021.1119.38
DrViT2022.0919.38
VOLO-D1HAT=false, Resolution=224x2242022.0419.2
Dr. ViTBackbone=ViT-S, Discrete Only=true2021.1118.44
ViT-SHAT=true, Resolution=224x2242022.0418.1
Dr. ViTPre-trained=ImageNet-21K, Fine-tuned on=ImageNet, Resolution=384x384, Discrete representation=true2021.1118.05
AdvPropBackbone=EfficientNet-B2, Resolution=224x2242019.1117.8
EfficientNet-B3Resolution=224x2242019.1117.8
ViTBackbone=ViT-B/16, Pre-training=ImageNet-21K, Fine-tuning=ImageNet-1K, Input Resolution=224x2242023.1117.7
ViT-B/16 + PixelATBackbone=ViT-B/16, Resolution=512x512, Pre-training=ImageNet-21K, Fine-tuning=ImageNet-1K, Adversarial Training=PixelAT2021.1117.66
Shape-Texture Debiased TrainingBackbone=ResNet-502020.1017.4
EfficientNet-B2Resolution=224x2242019.1116.8
AdvPropBackbone=EfficientNet-B1, Resolution=224x2242019.1116.7
AugReg-ViT2022.0916.62
Swin-BHAT=true, Resolution=224x2242022.0416.4
Dr. ViTPre-trained=ImageNet-21K, Fine-tuned on=ImageNet, Resolution=512x512, Discrete representation=true2021.1116.02
ViT-SHAT=false, Resolution=224x2242022.0415.8
Swin-BHAT=false, Resolution=224x2242022.0415.7
LLEarch=ResNet-50, train data=IN-1k2022.1215.25
Swin-SHAT=true, Resolution=224x2242022.0415.1
EfficientNet-B1Resolution=224x2242019.1115
LLEarch=ResNet-50, train data=IN-1k, augmentation=Edge Aug2022.1214.72
DeepAugment2022.0914.68
AdvPropBackbone=EfficientNet-B0, Resolution=224x2242019.1114.6
CNFABackbone=Resnet-182024.0714.4
Dr. ViTBackbone=ViT-S, Discrete Only=false2021.1114.22
Swin-SHAT=false, Resolution=224x2242022.0414.2
Dr. ViTBackbone=ViT-B, Resolution=384x3842021.1114.06
Swin-THAT=true, Resolution=224x2242022.0413.8
ViT-BPre-trained=ImageNet-21K, Fine-tuned on=ImageNet, Resolution=384x384, Discrete representation=false2021.1113.59
Dr. ViTBackbone=ViT-Ti, Discrete Only=true2021.1113.44
ViT-B/16Backbone=ViT-B/16, Resolution=512x512, Pre-training=ImageNet-21K, Fine-tuning=ImageNet-1K, Adversarial Training=None2021.1113.44
CFABackbone=Resnet-182024.0713.3
EfficientNet-B0Resolution=224x2242019.1113.1
Swin-THAT=false, Resolution=224x2242022.0413.1
Dr. ViTBackbone=MLPMixer, Discrete Only=false2021.1113.05
NCLBackbone=Resnet-182024.0712.7
ViT-THAT=true, Resolution=224x2242022.0412.5
CLBackbone=Resnet-182024.0712.3
ResNet50 + DAT2022.0910.8
ViT-SBackbone=ViT-S2021.1110.39
ViT-BBackbone=ViT-B2021.1110.39
AdaBNTraining Augmentations=false2021.0310.2
ViT-THAT=false, Resolution=224x2242022.049.5
Dr. ViTBackbone=MLPMixer, Discrete Only=true2021.118.98
Dr. ViTBackbone=ViT-B, Data Augmentation=Dosovitskiy et al., 2020, Discrete Only=true2021.118.94
AdaBNTraining Augmentations=true2021.038.4
ResNet-50Resolution=224x2242019.118
ViT-BBackbone=ViT-B, Resolution=384x3842021.117.58
Hybrid ViT-BBackbone=Hybrid ViT-B2021.117.5
Hybrid ViT-SBackbone=Hybrid ViT-S2021.117.42
ResNet-50Backbone=ResNet-502020.107.4
ResNet502022.097.38
MLPMixerBackbone=MLPMixer2021.117.03
ResNet-50Backbone=ResNet-502021.116.56
ViT-TiBackbone=ViT-Ti2021.116.17
ViT-B VanillaBackbone=ViT-B, Data Augmentation=Dosovitskiy et al., 20202021.115.94
ViT2022.095.82
CutMix + MoExBackbone=ResNet-502020.105