List your company

Image Classification on ImageNet-R (val)

96.5AccuracyCoCa
-2.09223.50449.174.696Jul 5, 2021Feb 17, 2022Oct 2, 2022May 17, 2023Dec 30, 2023Aug 13, 2024Mar 29, 2025

Evaluation results

MethodLinks
CoCaEvaluation Protocol=Zero-shot2023.0296.5
LiT-e/14Evaluation Protocol=Zero-shot, Backbone=ViT-e/142023.0296.1
LiT-22BEvaluation Protocol=Zero-shot, Backbone=ViT-22B/142023.0296
BASICEvaluation Protocol=Zero-shot2023.0295.7
LiT-g/14Evaluation Protocol=Zero-shot, Backbone=ViT-g/142023.0294.9
ALIGNEvaluation Protocol=Zero-shot2023.0292.2
CLIPEvaluation Protocol=Zero-shot2023.0288.9
ZEROBackbone=CLIP-ViT-B-16 + CLIP-ViT-L-142024.0587.21
RLCFBackbone=CLIP-ViT-B-16 + CLIP-ViT-L-14, Tuning=full image encoder tuning (Theta_i), Adaptation steps (t)=32024.0586.19
Zero-ShotBackbone=CLIP-ViT-B-16 + CLIP-ViT-L-142024.0585.4
RLCFBackbone=CLIP-ViT-B-16 + CLIP-ViT-L-14, Tuning=prompt tuning (Theta_etx), Adaptation steps (t)=12024.0583.35
PaLI-17BShot Setting=0-shot2022.0981.97
PaLI-15BShot Setting=0-shot2022.0981.21
ZEROBackbone=CLIP-ViT-B-16, Ensemble variant=true2024.0580.75
PaLI-3BShot Setting=0-shot2022.0980.15
ZEROBackbone=MaPLe2024.0579.74
PromptAlignBackbone=MaPLe2024.0579.33
TPTBackbone=MaPLe2024.0578.12
EnsembleBackbone=CLIP-ViT-B-162024.0577.65
ZEROBackbone=CLIP-ViT-B-162024.0577.22
TPTBackbone=CLIP-ViT-B-162024.0577.06
Zero-ShotBackbone=MaPLe2024.0576.98
Zero-ShotBackbone=CLIP-ViT-B-162024.0573.98
Zero-Shot baselineEvaluation Protocol=Zero-Shot, Label Distribution Estimation=None2024.0468.9
OTTEREvaluation Protocol=Zero-Shot, Label Distribution Estimation=BBSE+OT2024.0463.3
OTTEREvaluation Protocol=Linear Probing, Label Distribution Estimation=BBSE+OT2024.0457.1
STL (FAN-L-Hybrid)Params (M)=77.32024.0156.6
STL (FAN-B-Hybrid)Params (M)=50.92024.0155.3
Linear Probing baselineEvaluation Protocol=Linear Probing, Label Distribution Estimation=None2024.0454.9
FAN-L-HybridParams (M)=76.82024.0153.2
FAN-B-HybridParams (M)=50.42024.0152.9
STL (FAN-S-Hybrid)Params (M)=26.52024.0151.8
ConvNext-BParams (M)=88.62024.0151.3
FAN-S-HybridParams (M)=26.32024.0150.7
MSNArchitecture=ViT-B/16, Pre-training method=MSN2022.0450
ConvNext-SParams (M)=50.22024.0149.5
RVT-BParams (M)=91.82024.0148.7
MAE-ViT-BParams (M)=862024.0148.3
MAEArchitecture=ViT-B/16, Pre-training method=MAE2022.0448.3
RVT-SParams (M)=23.32024.0147.7
BBSE+PMEvaluation Protocol=Linear Probing, Label Distribution Estimation=BBSE+PM2024.0447.6
DAT-AugReg-ViTParams (M)=862024.0147.3
ConvNext-TParams (M)=28.62024.0147.2
LV-ViT-MParams (M)=562024.0147.2
Swin-SParams (M)=502024.0146.6
Swin-BParams (M)=87.82024.0146.6
LV-ViT-SParams (M)=262024.0145.8
XCIT-S12Params (M)=26.32024.0145.5
XCIT-S24Params (M)=47.72024.0145.5
CvT*-LS-21Params (M)=32.12021.0745
LSNet-BFLOPS=1.3G2025.0343.1
CvT-13Params (M)=202021.0742.9
CvT-21Params (M)=322021.0742.6
CvT*-LS-13Params (M)=20.32021.0742.6
DeiT-SParams (M)=22.12021.0741.9
PVTv2-B1FLOPS=2.1G2025.0341.8
Swin-TParams (M)=28.32024.0141.3
FasterNet-T2FLOPS=1.9G2025.0340.5
EdgeNeXt-SFLOPS=1.3G2025.0340.1
FastViT-T12FLOPS=1.4G2025.0339.9
LSNet-SFLOPS=0.5G2025.0339.4
LSNet-TFLOPS=0.3G2025.0338.5
UniRepLKNet-AFLOPS=0.6G2025.0337.9
PoolFormer-S12FLOPS=1.8G2025.0337.7
FastViT-T8FLOPS=0.7G2025.0336.8
ResNet50Architecture=ResNet50, Pre-training method=Supervised2022.0436.11
EfficientViT-M3FLOPS=0.3G2025.0336.1
ResNet-50Params (M)=25.62021.0735.3
PVTv2-B0FLOPS=0.6G2025.0334.2
StarNet-S1FLOPS=0.4G2025.0334.1
PVT-TinyFLOPS=1.9G2025.0333.9
EdgeNeXt-XSFLOPS=0.5G2025.0332.5
EdgeNeXt-XXSFLOPS=0.3G2025.0329.5
MoCo-v2 + Patch-base NSalpha=2, k=163842021.1029.45
MoCo-v2 + Patch-base NSalpha=2, k=655362021.1028.95
FasterNet-T0FLOPS=0.3G2025.0328.6
MoCo-v2 + Patch-base NSalpha=2, k=327682021.1028.5
MoCo-v2 + Patch-base NSalpha=3, k=655362021.1028.11
MoCo-v22021.1027.81
MoCo-v2 + MoCHiN=512, K=1024, s=5122021.1025.71
MoCo-v2 + MoCHiN=128, K=1024, s=5122021.1025.29
BBSE+PMEvaluation Protocol=Zero-Shot, Label Distribution Estimation=BBSE+PM2024.041.7