List your company
ResearchFormerly WizWand

Image Classification on Oxford-Flowers

99.7Top-1 AccuracyARC*
57.99668.82379.6590.477Mar 30, 2020Apr 13, 2021Apr 27, 2022May 12, 2023May 25, 2024Jun 8, 2025Jun 23, 2026
Updated 4mo ago

Evaluation results

MethodLinks
ARC*Backbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.25, Data Augmentation=Advanced (cutmix, mixup, label smoothing)2023.1099.7
SSF*Backbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.39, Data Augmentation=Advanced (cutmix, mixup, label smoothing)2023.1099.6
ARCattBackbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.222023.1099.4
ARCBackbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.252023.1099.3
LoRABackbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.442023.1099.2
TResNet-LSpeed (img/sec)=500, Input resolution=368, Precision=Mixed Precision V100 GPU2020.0399.1
VPT-DeepBackbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.852023.1099
EfficientNet-B7Speed (img/sec)=70, Input resolution=600, Precision=Mixed Precision V100 GPU2020.0398.8
DeiT-Bvariant=22022.0498.8
DearKD-BStudent Network=DeiT-B2022.0498.8
EfficientNet-B7# Params=64.0M, # FLOPs=37.2B, Resolution=600, Pre-training=ImageNet, Fine-tuning=true2021.0798.8
TNT-S# Params=23.8M, # FLOPs=17.3B, Resolution=384, Pre-training=ImageNet, Fine-tuning=true2021.0798.8
Full fine-tuningBackbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=85.982023.1098.8
BiasBackbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.282023.1098.8
AlphaNet-A6Backbone=AlphaNet-A6, FLOPs=709 M, Pre-training=ImageNet, Fine-tuning=true2021.0298.7
CMT-S# Params=25.1M, # FLOPs=4.04B, Pre-training=ImageNet, Fine-tuning=true2021.0798.7
CeiT-S# Params=24.2M, # FLOPs=12.9B, Resolution=384, Pre-training=ImageNet, Fine-tuning=true2021.0798.6
EffiNet-B52022.0498.5
Inception-v4# Params=41.1M, # FLOPs=16.1B, Pre-training=ImageNet, Fine-tuning=true2021.0798.5
AdapterBackbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.412023.1098.5
SSFBackbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.39, Data Augmentation=Basic2023.1098.5
DeiT-Bvariant=12022.0498.4
DeiT-B# Params=85.8M, # FLOPs=17.6B, Pre-training=ImageNet, Fine-tuning=true2021.0798.4
VPT-ShallowBackbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.252023.1098.4
DINO + SEREMode=Fine-tuning2022.0698
Linear probingBackbone=ViT-B/16, Pre-trained=ImageNet-21k, Params(M)=0.182023.1097.9
EfficientNet-B1Backbone=EfficientNet-B1, FLOPs=700 M, Pre-training=ImageNet, Fine-tuning=true2021.0297.8
DINOMode=Fine-tuning2022.0697.8
AlphaNet-A0Backbone=AlphaNet-A0, FLOPs=203 M, Pre-training=ImageNet, Fine-tuning=true2021.0297.7
DearKD-SStudent Network=DeiT-S2022.0497.4
ResNet-152# Params=58.1M, # FLOPs=11.3B, Pre-training=ImageNet, Fine-tuning=true2021.0797.4
EfficientNet-B0Backbone=EfficientNet-B0, FLOPs=390 M, Pre-training=ImageNet, Fine-tuning=true2021.0297.2
DearKD-TiStudent Network=DeiT-Ti2022.0495.1
AttrVREvaluation Protocol=few-shot, Backbone=ViT-B/162026.0693.1
BioMedVREvaluation Protocol=few-shot, Backbone=ViT-B/162026.0693
Gated Prompt TuningBackbone=ViT-S, Pre-training strategy=MoCo v3, # of prompt tokens=482023.0691.14
ViT-L/162022.0489.7
VPT-DEEPBackbone=ViT-S, Pre-training strategy=MoCo v3, # of prompt tokens=482023.0689.53
ViT-B/162022.0489.5
ViT-B/16# Params=85.8M, # FLOPs=17.6B, Resolution=384, Pre-training=ImageNet, Fine-tuning=true2021.0789.5
ViT-L/322022.0486.4
AREvaluation Protocol=few-shot, Backbone=ViT-B/162026.0686.2
ViT-B/322022.0485.4
VPT-SHALLOWBackbone=ViT-S, Pre-training strategy=MoCo v3, # of prompt tokens=482023.0684.65
VPEvaluation Protocol=few-shot, Backbone=ViT-B/162026.0683.3
DMAEModel Scale=L6H62026.0183.18
SWEETModel Scale=L6H62026.0183.12
WAVEModel Scale=L6H62026.0180.26
WT-SelectModel Scale=L6H62026.0176.65
Iso. PruningModel Scale=L6H62026.0175.77
AWTTrain=false2024.0775.07
Gated Prompt TuningBackbone=ViT-L, Pre-training strategy=MAE, # of prompt tokens=482023.0674.71
MMRL + MetaTPTadaptation=MetaTPT2025.1274.22
SuS-X-SDTrain=false2024.0773.81
MMRL + TPTadaptation=TPT2025.1272.89
MMRLadaptation=none2025.1272.77
POMPTrain=true2024.0772.72
MaPLe + MetaTPTadaptation=MetaTPT2025.1272.7
PromptAlignTrain=true2024.0772.39
MaPLe + PromptAlignadaptation=PromptAlign2025.1272.39
MaPLe + TPTadaptation=TPT2025.1272.37
WaffleCLIPTrain=false2024.0772.35
MaPLeTrain=true2024.0772.23
MaPLeadaptation=none2025.1272.23
BioMedVREvaluation Protocol=zero-shot, Backbone=ViT-B/162026.0672.2
CoCoOpTrain=true2024.0771.88
Self-TPT-vTrain=true2024.0771.79
ProVP-RefTrain=true2024.0771.62
CLIP-ViT-B/16Layers=-2026.0371.38
CuPLTrain=false2024.0771.3
TCA (Baseline)Layers=3, 6, 9, KeepRate=0.72026.0371.17
VisDescTrain=false2024.0770.85
Col-LnLayers=3, 6, 9, KeepRate=0.72026.0370.85
CLIP + MetaTPTadaptation=MetaTPT2025.1270.77
DiffTPTTrain=true2024.0770.1
TCA (Baseline)Layers=0, 3, 6, KeepRate=0.72026.0369.59
CoOp + MetaTPTadaptation=MetaTPT2025.1269.14
PLOT++Train=true2024.0769.1
TPTTrain=true2024.0768.98
CLIP + TPTadaptation=TPT2025.1268.98
CoOpTrain=true2024.0768.71
CoOpadaptation=none2025.1268.71
Col-LnLayers=0, 3, 6, KeepRate=0.72026.0368.53
CoOp + TPTadaptation=TPT2025.1268.48
CLIPTrain=false2024.0767.44
CLIPadaptation=none2025.1267.44
VPT-DEEPBackbone=ViT-L, Pre-training strategy=MAE, # of prompt tokens=482023.0665.44
VPT-SHALLOWBackbone=ViT-L, Pre-training strategy=MAE, # of prompt tokens=482023.0662.77
CLIPEvaluation Protocol=zero-shot, Backbone=ViT-B/162026.0659.6