List your company
ResearchFormerly WizWand

Classification on Food101

96.2Top-1 AccuracyFlorence-CoSwin-H
16.74437.3725878.628Nov 25, 2020Oct 28, 2021Oct 1, 2022Sep 4, 2023Aug 6, 2024Jul 10, 2025Jun 13, 2026
Updated 3mo ago

Evaluation results

MethodLinks
Florence-CoSwin-HResolution=384px, Backbone=CoSwin-H, Evaluation Protocol=Linear Probing2021.1196.2——
CLIP-ViT-L/14Resolution=336px, Evaluation Protocol=Linear Probing2021.1195.9——
Food-R12026.0695.21——
Florence-CoSwin-HBackbone=CoSwin-H, Resolution=384px, Evaluation Protocol=Zero-shot2021.1195.1——
CLIP-ResNet-50x64Evaluation Protocol=Linear Probing2021.1194.8——
RoDE2026.0694.02——
FoodLMM2026.0693.93——
CLIP-ViT-L/14Backbone=ViT-L/14, Resolution=336px, Evaluation Protocol=Zero-shot2021.1193.8——
GrafitBackbone=RegNetY-8.0GF, Pre-training=ImageNet-1k, Resolution=384x384, Number of Parameters=39M, Classifier=MLP2020.1193.7——
GrafitBackbone=RegNetY-8.0GF, Parameters=39M, Resolution=384x3842020.1193.7——
CAE*Supervised=false, Self-supervised=true, Backbone=ViT-B2022.0293.32——
MAESupervised=false, Self-supervised=true, Backbone=ViT-B2022.0293.19——
EfficientNet-B7Pre-training=ImageNet-1k, Resolution=600x600, Number of Parameters=64M2020.1193——
EfficientNet-B7Parameters=64M, Resolution=600x6002020.1193——
SigLIP2Data size=10B, Zero-shot=true2026.0392.8——
PEcoreData size=5B, Zero-shot=true2026.0392.5——
FLIP-ViT-L/14Backbone=ViT-L/14, Evaluation Protocol=Zero-shot2021.1192.2——
EfficientNet-L2Resolution=800px, Evaluation Protocol=Linear Probing2021.1192——
DeiTSupervised=true, Self-supervised=false, Backbone=ViT-B2022.0291.81——
CLIP-ResNet-50x64Backbone=ResNet-50x64, Evaluation Protocol=Zero-shot2021.1191.8——
DINOSupervised=false, Self-supervised=true, Backbone=ViT-B2022.0291.67——
SigLIPData size=10B, Zero-shot=true2026.0391.6——
PRENet2026.0691.13——
ITOPre-training Dataset=DataComp-1B, Backbone=ViT-B/16, Training Epochs=10 epochs, Zero-shot protocol=EVA-CLIP2026.0390.8——
CLIPPre-training Dataset=DataComp-1B, Backbone=ViT-B/16, Training Epochs=10 epochs, Zero-shot protocol=EVA-CLIP2026.0390.2——
LSA+OnZetaBackbone=ViT-B/162026.0689.85——
ITO sub2Pre-training Dataset=DataComp-1B, Backbone=ViT-B/16, Training Epochs=10 epochs, Zero-shot protocol=EVA-CLIP2026.0389.7——
LSABackbone=ViT-B/162026.0689.51——
GrafitBackbone=ResNet-50, Pre-training=ImageNet, Resolution=224x224, Evaluation Protocol=Single center crop, Classifier=MLP2020.1189.5——
OnZetaBackbone=ViT-B/162026.0689.07——
LlipData size=2.5B, Zero-shot=true2026.0389——
BaselineBackbone=ResNet-50, Pre-training=ImageNet, Resolution=224x224, Evaluation Protocol=Single center crop, Classifier=Linear FC2020.1188.9——
ClusterFit+Backbone=ResNet-50, Pre-training=ImageNet, Resolution=224x224, Evaluation Protocol=Single center crop, Classifier=Linear FC2020.1188.9——
CLIP baselineBackbone=ViT-B/162026.0688.86——
SNCA+Backbone=ResNet-50, Pre-training=ImageNet, Resolution=224x224, Evaluation Protocol=Single center crop, Classifier=MLP2020.1188.8——
Grafit FCBackbone=ResNet-50, Pre-training=ImageNet, Resolution=224x224, Evaluation Protocol=Single center crop, Classifier=Linear FC2020.1188.7——
MetaCLIPData size=2.5B, Zero-shot=true2026.0388.3——
Noun SubmanifoldBackbone=CLIP ViT-L-14, zero-shot=true2023.0587.7——
POS PGABackbone=CLIP ViT-L-14, zero-shot=true2023.0587.5——
ViT-L/16Resolution=384px, Evaluation Protocol=Linear Probing2021.1187.4——
CLIPBackbone=CLIP ViT-L-14, zero-shot=true2023.0587.4——
POS PCABackbone=CLIP ViT-L-14, zero-shot=true2023.0587.4——
TPTBackbone=ViT-B/162026.0686.97——
OpenCLIPData size=2B, Zero-shot=true2026.0386.2——
SimCLRv2Backbone=ResNet-152x3, Evaluation Protocol=Linear Probing2021.1183.6——
Random Init.Supervised=false, Self-supervised=false, Backbone=ViT-B2022.0282.77——
LSA+OnZetaBackbone=ResNet-502026.0682.01——
LSABackbone=ResNet-502026.0681.58——
OnZetaBackbone=ResNet-502026.0680.85——
CLIP baselineBackbone=ResNet-502026.0680.59——
RényiCLProtocol=linear evaluation2022.0878——
NNCLRBackbone=ResNet-50, Pre-training=ImageNet2021.0476.7——
NNCLRBackbone=ResNet-50, Pre-training Dataset=ImageNet-1K, Pre-training Epochs=8002022.0376.7——
NNCLRProtocol=linear evaluation2022.0876.7——
CaCoBackbone=ResNet-50, Pre-training Dataset=ImageNet-1K, Pre-training Epochs=8002022.0375.4——
DreamLIPData size=30M, Zero-shot=true2026.0375.4——
BYOLBackbone=ResNet-50, Pre-training=ImageNet2021.0475.3——
BYOLBackbone=ResNet-50, Pre-training Dataset=ImageNet-1K, Pre-training Epochs=8002022.0375.3——
BYOLProtocol=linear evaluation2022.0875.3——
TPTBackbone=ResNet-502026.0674.89——
GoldiCLIPData size=30M, Zero-shot=true2026.0374——
COSMOSData size=30M, Zero-shot=true2026.0373.9——
AdCo +Backbone=ResNet-50, Pre-training Dataset=ImageNet-1K, Pre-training Epochs=8002022.0373.8——
SimCLRBackbone=ResNet-50, Pre-training=ImageNet2021.0472.8——
SimCLRBackbone=ResNet-50, Pre-training Dataset=ImageNet-1K, Pre-training Epochs=8002022.0372.8——
FLAIRData size=30M, Zero-shot=true2026.0372.5——
Sup.-INBackbone=ResNet-50, Pre-training=ImageNet2021.0472.3——
SupervisedProtocol=linear evaluation2022.0872.3——
MoCo v2+Backbone=ResNet-50, Pre-training Dataset=ImageNet-1K, Pre-training Epochs=8002022.0371.5——
SimCLRProtocol=linear evaluation2022.0868.4——
SigLIPData size=30M, Zero-shot=true2026.0364.2——
CLIPData size=30M, Zero-shot=true2026.0361.3——
CLIP-PGSImage Encoder=ViT-B/16, p=0.3, Protocol=Zero-shot2025.0346.5——
CLIP-PGSImage Encoder=ViT-B/16, p=0.5, Protocol=Zero-shot2025.0342.8——
CLIPImage Encoder=ViT-B/16, Protocol=Zero-shot2025.0342.3——
E-CLIPImage Encoder=ViT-B/16, Protocol=Zero-shot2025.0342.1——
A-CLIPImage Encoder=ViT-B/16, Protocol=Zero-shot2025.0341.8——
FLIPImage Encoder=ViT-B/16, Protocol=Zero-shot2025.0339.9——
CLIP-PGSImage Encoder=ViT-S/16, p=0.3, Protocol=Zero-shot2025.0339.1——
CLIP-PGSImage Encoder=ViT-S/16, p=0.5, Protocol=Zero-shot2025.0338.7——
ITO sub2Backbone=ViT-B/16, Pre-training Dataset=CC3M-recap, Evaluation Protocol=Zero-shot2026.0323.1——
ITO sub3Backbone=ViT-B/16, Pre-training Dataset=CC3M-recap, Evaluation Protocol=Zero-shot2026.0323.1——
FLAIRBackbone=ViT-B/16, Pre-training Dataset=CC3M-recap, Evaluation Protocol=Zero-shot2026.0319.8——
BiFTAProtocol=Zero-shot, Backbone=RN-1012026.01—84.02—
BiFTAZero-shot=true, Backbone=Average over five CLIP variants2026.01—87.18—
BiFTABackbone=ViT-B/32, Evaluation Protocol=Zero-shot2026.01—86.43—
BiFTABackbone=RN-50, Zero-shot evaluation protocol=true2026.01—81.39—
BYOLArchitecture=ResNet-50, Pre-training Dataset=ImageNet, Evaluation Protocol=Linear evaluation2020.06—75.3—
BYOLArchitecture=ResNet-50, Pre-training Dataset=ImageNet, Evaluation Protocol=Fine-tuned2020.06—88.5—
C-TPTBackbone=ResNet50, Optim.=true2026.07—74.83—
CaBMBottleneck type=Strictly text bottleneck, Protocol=text-only classifier2026.07—93.68—
CLIPBackbone=ViT-B/16, Pre-trained on=IT35M, Mode=Zero-shot2022.10—61.2—
CLIPProtocol=Zero-shot, Backbone=RN-1012026.01—82.44—
CLIPBackbone=ViT-B/32, Evaluation Protocol=Zero-shot2026.01—82.6—
CLIPBackbone=RN-50, Zero-shot evaluation protocol=true2026.01—78.62—
CLIPShots=0, Venue=ICML’222025.12—77.3—
CLIPWay=5, Shot=52026.03—86—
CLIPBackbone=ResNet50, Optim.=false2026.07—73.94—
CLIP-DProtocol=Zero-shot, Backbone=RN-1012026.01—83.25—
CLIP-DBackbone=ViT-B/32, Evaluation Protocol=Zero-shot2026.01—84.12—