List your company
ResearchFormerly WizWand

Image Classification on Stanford Cars (Adversarial Robustness)

68.91Top-1 Accuracy (Clean)CLBP
8.350824.072939.79555.5171Mar 28, 2026Apr 6, 2026Apr 16, 2026Apr 26, 2026May 5, 2026May 15, 2026May 25, 2026
Updated 5mo ago

Evaluation results

MethodLinks
CLBPShots=16, Training Epochs=10, Training Attack (PGD)=2-step (1/255), Evaluation Attack (PGD-100)=100-step (1/255)2026.0568.9156.24—
AdvMaPLeShots=16, Training Epochs=10, Training Attack (PGD)=2-step (1/255), Evaluation Attack (PGD-100)=100-step (1/255)2026.0556.1717.57—
AdvVLPShots=16, Training Epochs=10, Training Attack (PGD)=2-step (1/255), Evaluation Attack (PGD-100)=100-step (1/255)2026.055617.47—
FAPShots=16, Training Epochs=10, Training Attack (PGD)=2-step (1/255), Evaluation Attack (PGD-100)=100-step (1/255)2026.0554.2319.23—
CLIPBackbone=ViT-B/32, Evaluation Protocol=Zero-shot2026.0354.1——
WISE-FTBackbone=ViT-B/32, Pre-training Dataset=ImageNet, Evaluation Protocol=Zero-shot2026.0353.180—
GRACEBackbone=ViT-B/32, Pre-training Dataset=ImageNet, Evaluation Protocol=Zero-shot2026.0352.11.9—
CLIPYear=20212026.0452.07——
CLIPBackbone=ViT-B/32, Evaluation Protocol=Zero-shot2026.0351.470—
OursModel Configuration=5 trees2026.0448.23——
OursModel Configuration=1 tree2026.0448.11——
TPGMBackbone=ViT-B/32, Pre-training Dataset=ImageNet, Evaluation Protocol=Zero-shot2026.0347.690—
AoSYear=20252026.0446.42——
SPDBackbone=ViT-B/32, Pre-training Dataset=ImageNet, Evaluation Protocol=Zero-shot2026.0346.310—
AGFTBackbone=ViT-B/32, Evaluation Protocol=Zero-shot, Training Strategy=Adversarially fine-tuned2026.0345.69——
Vanilla FTBackbone=ViT-B/32, Pre-training Dataset=ImageNet, Evaluation Protocol=Zero-shot2026.0345.440—
FAREYear=20242026.0445.06——
GLADIATORBackbone=ViT-B/32, Evaluation Protocol=Zero-shot, Training Strategy=Adversarially fine-tuned2026.0343.75——
PMG-AFTBackbone=ViT-B/32, Evaluation Protocol=Zero-shot, Training Strategy=Adversarially fine-tuned2026.0342.21——
PMG-FTYear=20242026.0440.49——
TeCoABackbone=ViT-B/32, Evaluation Protocol=Zero-shot, Training Strategy=Adversarially fine-tuned2026.0333.88——
FAREBackbone=ViT-B/32, Pre-training Dataset=ImageNet, Evaluation Protocol=Zero-shot2026.0333.871.3—
FLYPBackbone=ViT-B/32, Pre-training Dataset=ImageNet, Evaluation Protocol=Zero-shot2026.0333.720—
PMG-AFTBackbone=ViT-B/32, Pre-training Dataset=ImageNet, Evaluation Protocol=Zero-shot2026.0329.52—
TGA-ZSRBackbone=ViT-B/32, Evaluation Protocol=Zero-shot, Training Strategy=Adversarially fine-tuned2026.0329.34——
LAATBackbone=ViT-B/32, Pre-training Dataset=ImageNet, Evaluation Protocol=Zero-shot2026.0328.81.6—
TeCoAYear=20232026.0426.59——
AdvVPShots=16, Training Epochs=10, Training Attack (PGD)=2-step (1/255), Evaluation Attack (PGD-100)=100-step (1/255)2026.0514.833.57—
TeCoABackbone=ViT-B/32, Pre-training Dataset=ImageNet, Evaluation Protocol=Zero-shot2026.0310.681.5—
A-TPTBackbone=ViT-B/162026.05—39.2—
A-TPTBackbone=ViT-B/322026.05—31.8—
CLIPBackbone=ViT-B/162026.05—0—
CLIPBackbone=ViT-B/322026.05—0—
GPT-4V2024.03——58.3
MTABackbone=ViT-B/162026.05—18.5—
MTABackbone=ViT-B/322026.05—26.4—
R-TPTBackbone=ViT-B/162026.05—34.7—
R-TPTBackbone=ViT-B/322026.05—28.4—
RARBase Model=LLaVA1.52024.03——72.6
RARBase Model=Intern-IXC22024.03——65.4
RARBase Model=Qwen-VL2024.03——81.6
TPT-EnsembleBackbone=ViT-B/162026.05—26—
TPT-EnsembleBackbone=ViT-B/322026.05—25.9—
TTCBackbone=ViT-B/162026.05—2.9—
TTCBackbone=ViT-B/322026.05—2.3—