List your company
ResearchFormerly WizWand

Image Classification on Flowers102 (Accuracy)

99.6AccuracyLoRA (r = 32)
8.308832.009455.7179.4106May 16, 2025Jul 16, 2025Sep 16, 2025Nov 17, 2025Jan 17, 2026Mar 20, 2026May 21, 2026
Updated 5mo ago

Evaluation results

MethodLinks
LoRA (r = 32)Backbone=DINOv2-ViT-B/14, Params (M)=4.722026.0499.6
RandLoRA (r = 10)Backbone=DINOv2-ViT-B/14, Params (M)=3.592026.0499.6
VeRA (r = 256)Backbone=DINOv2-ViT-B/14, Params (M)=0.092026.0499.3
GiVA (r = 32, SecSV-B)Backbone=DINOv2-ViT-B/14, Params (M)=0.082026.0499.3
GiVA (r = 32, TopSV-B)Backbone=DINOv2-ViT-B/14, Params (M)=0.082026.0499.2
GiVA (r = 32, Rand-B)Backbone=DINOv2-ViT-B/14, Params (M)=0.082026.0499.2
Full Fine-TuningBackbone=DINOv2-ViT-B/14, Params (M)=86.732026.0499.1
VeRA (r = 256)Backbone=CLIP-ViT-L/14, Params (M)=0.232026.0498.7
GiVA (r = 32, Rand-B)Backbone=CLIP-ViT-L/14, Params (M)=0.202026.0498.7
GiVA (r = 32, SecSV-B)Backbone=CLIP-ViT-L/14, Params (M)=0.202026.0498.5
GiVA (r = 32, TopSV-B)Backbone=CLIP-ViT-L/14, Params (M)=0.202026.0498.4
Full Fine-TuningBackbone=CLIP-ViT-L/14, Params (M)=303.282026.0498.3
LoRA (r = 32)Backbone=CLIP-ViT-L/14, Params (M)=12.582026.0498.3
RandLoRA (r = 10)Backbone=CLIP-ViT-L/14, Params (M)=12.782026.0498.2
EBMmode=Fine-tuning2025.1098.2
SCANShot=162025.0597.61
PLOTmode=Fine-tuning2025.1097.6
PromptSRCmode=Fine-tuning2025.1097.6
CoOpmode=Fine-tuning2025.1097.1
SCANShot=82025.0597.03
MaPLemode=Fine-tuning2025.1097
SimNLShot=162025.0596.51
TaskResShot=162025.0596.03
SimNLShot=82025.0595.23
Tip-Adapter-FShot=162025.0595.01
SCANShot=42025.0594.74
TaskResShot=82025.0594.73
CLIP-AdapterShot=162025.0593.91
SimNLShot=42025.0592.04
Tip-Adapter-FShot=82025.0591.85
CLIP-AdapterShot=82025.0591.79
SCANShot=22025.0591.03
CoOpShot=82025.0590.84
TaskResShot=42025.0590.2
Tip-Adapter-FShot=42025.0589.02
SimNLShot=22025.0588.43
CoCoOpmode=Fine-tuning2025.1087.8
CLIP-AdapterShot=42025.0587.3
TaskResShot=22025.0586.57
SCANShot=12025.0586.47
CoOpShot=42025.0585.81
CoOpShot=162025.0584.39
SimNLShot=12025.0584.17
Tip-Adapter-FShot=22025.0582.5
CLIP-AdapterShot=22025.0581.77
Tip-Adapter-FShot=12025.0580.13
Co-CoOpSource Dataset=ImageNet, Evaluation Protocol=Cross-dataset transfer (zero-shot)2025.1179.41
TaskResShot=12025.0579.17
CoCoOpShot=162025.0579.14
AXISArchitecture=ViT-B-322025.0878.37
CoOpShot=22025.0577.47
CoCoOpShot=82025.0575.88
BEiT-B/16Alignment train data=IN21K repr. + 1 img-capt. pair per class, Zero-shot=true2026.0573.74
CLIP-AdapterShot=12025.0573.43
CoCoOpShot=42025.0572.72
MMRL++Backbone=ViT-B/16, Training Shots=16-shot2026.0572.57
CLIP-SVD2025.0972.45
MaPLe2025.0972.23
MMRLBackbone=ViT-B/16, Training Shots=16-shot2026.0572.22
A3B2Backbone=ViT-B/16, Training Shots=16-shot2026.0572.15
Co-CoOp2025.0971.88
TCPBackbone=ViT-B/16, Training Shots=16-shot2026.0571.46
CLIP2025.0971.3
MaPLeSource Dataset=ImageNet, Evaluation Protocol=Cross-dataset transfer (zero-shot)2025.1171.3
KgCoOpBackbone=ViT-B/16, Training Shots=16-shot2026.0571.23
CoCoOpShot=22025.0570.7
CLIPmode=Fine-tuning2025.1070.7
CoOpBackbone=ViT-B/16, Training Shots=16-shot2026.0570.66
PromptSRCSource Dataset=ImageNet, Evaluation Protocol=Cross-dataset transfer (zero-shot)2025.1170.52
MaPLeBackbone=ViT-B/16, Training Shots=16-shot2026.0570.28
aTLASArchitecture=ViT-B-322025.0870.24
KgCoOp2025.0970.01
CoOp2025.0968.71
MMABackbone=ViT-B/16, Training Shots=16-shot2026.0568.52
CoOpShot=12025.0567.9
ProGrad2025.0967.87
CoCoOpShot=12025.0567.52
CLIP ViT-B/16Alignment train data=WIT400M, Zero-shot=true2026.0565.96
CLIP2026.0465.62
CLIPShot=02025.0565.46
BEiT-B/16Alignment train data=IN21K repr., Zero-shot=true2026.0562.63
FARE2026.0461.94
AdvFLYPfull2026.0457.67
PMG-AFT2026.0455.23
TGA-ZSR2026.0452.92
AdvFLYP2026.0452.5
TeCoA2026.0451.63
BEiT-B/16Alignment train data=IN1K repr., Zero-shot=true2026.0511.82