List your company
ResearchFormerly WizWand

Image Classification on Flowers (Accuracy)

99.2AccuracyStableRep
6.32830.43954.5578.661Sep 15, 2022Apr 16, 2023Nov 16, 2023Jun 17, 2024Jan 16, 2025Aug 18, 2025Mar 20, 2026
Updated 7mo ago

Evaluation results

MethodLinks
StableRepPre-training dataset=RedCaps, Pre-training data source=Syn, Evaluation protocol=Few-shot2023.0699.2
CLIPPre-training dataset=RedCaps, Pre-training data source=Real, Evaluation protocol=Few-shot2023.0699
DINO-vMFBackbone=ViT-Base/16, Pre-training=ImageNet-1K, Evaluation Protocol=Fine-tuning2024.0598.9
iBOTBackbone=ViT-Base/16, Pre-training=ImageNet-1K, Evaluation Protocol=Fine-tuning2024.0598.9
CapPa L/14MAP head=true, Grain=Fine, Frozen representation=true, backbone=ViT-L/142023.0698.9
DINOBackbone=ViT-Base/16, Pre-training=ImageNet-1K, Evaluation Protocol=Fine-tuning2024.0598.8
iBOT-vMFBackbone=ViT-Base/16, Pre-training=ImageNet-1K, Evaluation Protocol=Fine-tuning2024.0598.8
IsoCLIPIntra-modal=✓, Classifier=NCM, Backbone=ViT-B/16-open2026.0398.8
Image-ImageIntra-modal=✓, Classifier=NCM, Backbone=ViT-B/16-open2026.0398.6
Random (DeIT)Backbone=ViT-Base/16, Pre-training=ImageNet-1K, Evaluation Protocol=Fine-tuning2024.0598.4
DeiT-B#Params (M)=86.02025.1298.4
GrowTAS-S#Params (M)=22.42025.1298.21
StableRepPre-training Data Source=Synthetic, Evaluation Protocol=5-way, 5-shot2023.0698.2
StableRepPre-training dataset=CC12M, Pre-training data source=Syn, Evaluation protocol=Few-shot2023.0698.2
CLIPPre-training dataset=RedCaps, Pre-training data source=Syn, Evaluation protocol=Few-shot2023.0698.2
BeITBackbone=ViT-Base/16, Pre-training=ImageNet-1K, Evaluation Protocol=Fine-tuning2024.0598
AutoFormer-S#Params (M)=22.92025.1297.9
ViTAE-S#Params (M)=24.02025.1297.8
PreNAS-S#Params (M)=23.02025.1297.6
CLIPPre-training Data Source=Real, Evaluation Protocol=5-way, 5-shot2023.0697.6
CLIPPre-training dataset=CC12M, Pre-training data source=Real, Evaluation protocol=Few-shot2023.0697.6
DearKD-S#Params (M)=22.02025.1297.4
CapPaMAP head=true, Grain=Fine, Frozen representation=true2023.0697
CLIPPre-training Data Source=Synthetic, Evaluation Protocol=5-way, 5-shot2023.0696.9
CLIPPre-training dataset=CC12M, Pre-training data source=Syn, Evaluation protocol=Few-shot2023.0696.9
CLIP* L/14MAP head=true, Grain=Fine, Frozen representation=true, backbone=ViT-L/142023.0696.6
FLAVAImg-text pairs=70M, Evaluation protocol=Linear probing2022.0996.4
Partial-LoRAParameter Count=0.32M2025.1295.21
VeRAParameter Count=0.10M2025.1294.91
SimCLRPre-training dataset=RedCaps, Pre-training data source=Real, Evaluation protocol=Few-shot2023.0694.8
Partial-AdaLoRAParameter Count=0.26M2025.1294.62
AdaLoRAParameter Count=0.67M2025.1294.36
LoRA+Parameter Count=0.67M2025.1294.35
SimCLRPre-training dataset=RedCaps, Pre-training data source=Syn, Evaluation protocol=Few-shot2023.0694.3
MMFTModel=ViT, Method=Ours2026.0194.09
CapMAP head=true, Grain=Fine, Frozen representation=true2023.0694
FLYPModel=ViT, Method=FLYP2026.0193.77
DoRAParameter Count=0.77M2025.1293.64
LoRAParameter Count=0.67M2025.1293.18
CLIP-ViT-B/16Img-text pairs=400M, Evaluation protocol=Linear probing2022.0993.1
FLAVAImg-text pairs=14M, Evaluation protocol=Linear probing2022.0992.7
SimCLRPre-training Data Source=Real, Evaluation Protocol=5-way, 5-shot2023.0692.1
SimCLRPre-training dataset=CC12M, Pre-training data source=Real, Evaluation protocol=Few-shot2023.0692.1
Full-FTProjection Type=Baseline, Number of Parameters=7.00M2024.1192.05
FTProjection Type=Both-Proj, Number of Parameters=5.33M2024.1192
SimCLRPre-training Data Source=Synthetic, Evaluation Protocol=5-way, 5-shot2023.0692
SimCLRPre-training dataset=CC12M, Pre-training data source=Syn, Evaluation protocol=Few-shot2023.0692
FTProjection Type=Out-Proj, Number of Parameters=1.77M2024.1191.45
FTProjection Type=In-Proj, Number of Parameters=3.56M2024.1191.02
FTModel=ViT, Method=FT2026.0190.25
OmniVLImg-text pairs=14M*, Evaluation protocol=Linear probing2022.0989.8
ViT-L/16#Params (M)=307.02025.1289.7
ViT-B/16#Params (M)=86.02025.1289.5
CLIP* (16k)MAP head=true, Grain=Fine, Frozen representation=true2023.0689.5
BLIPImg-text pairs=14M, Evaluation protocol=Linear probing2022.0988.4
ProDiaLProjection Type=Both-Proj, Number of Parameters=0.65M2024.1188
LoRAProjection Type=Both-Proj, Number of Parameters=0.61M2024.1187.32
StrongProjection Type=Baseline, Number of Parameters=0.94M2024.1186.76
DoRAProjection Type=Both-Proj, Number of Parameters=0.65M2024.1186.6
ALBEFImg-text pairs=14M, Evaluation protocol=Linear probing2022.0986.5
ProDiaLProjection Type=In-Proj, Number of Parameters=0.33M2024.1185.97
DoRAProjection Type=In-Proj, Number of Parameters=0.41M2024.1185.95
LoRAProjection Type=In-Proj, Number of Parameters=0.39M2024.1185.71
CLIP* (8k)MAP head=true, Grain=Fine, Frozen representation=true2023.0685.1
ZSModel=MLLM, Method=ZS2026.0184.68
MMFTModel=RN50, Method=Ours2026.0182.88
FLYPModel=RN50, Method=FLYP2026.0181.92
ProDiaLProjection Type=Out-Proj, Number of Parameters=0.25M2024.1181.9
DoRAProjection Type=Out-Proj, Number of Parameters=0.26M2024.1180.97
LoRAProjection Type=Out-Proj, Number of Parameters=0.24M2024.1180.6
BitFitProjection Type=Baseline, Number of Parameters=0.06M2024.1178.84
OpenAI-CLIP VIT-LN=400M, Zero-shot=true2024.0978.7
DINOv2-ARLN=20M, Zero-shot=true2024.0978.4
LiT L16LN=112M, Zero-shot=true2024.0976.3
DINOv2-MpNetN=20M, Zero-shot=true2024.0975.8
Image-TextIntra-modal=✗, Classifier=Zero-Shot, Backbone=ViT-B/16-open2026.0375.3
LAION-CLIP VIT-LN=400M, Zero-shot=true2024.0973
FTModel=RN50, Method=FT2026.0169.29
METER-CLIP-B/16Img-text pairs=400M+4M, Evaluation protocol=Linear probing2022.0967.1
Linear ProbingProjection Type=Baseline, Number of Parameters=0.02M2024.1159.9
ZSModel=ViT, Method=ZS2026.0156.42
ZSModel=RN50, Method=ZS2026.0156.25
SynerNetZero-shot=true2026.0149.2
ZSNLZero-shot=true2026.0148.3
TransCLIPZero-shot=true2026.0112.7
OpenCLIPZero-shot=true2026.019.9