List your company
ResearchFormerly WizWand

Error Detection on EuroSAT

98.1AuROCOurs
40.504855.457470.4185.3626May 25, 2024Sep 18, 2024Jan 13, 2025May 10, 2025Sep 4, 2025Dec 30, 2025Apr 26, 2026
Updated 6mo ago

Evaluation results

MethodLinks
OursBackbone=SigLIP ViT-B/162025.1198.198.032.93
OursBackbone=CLIP ViT-B/162025.1197.6496.163.35
Ours-DBackbone=CLIP ViT-B/16, Covariance=Diagonal2025.1197.3295.917.39
Ours-DBackbone=SigLIP ViT-B/16, Covariance=Diagonal2025.1196.9396.9512.67
Ours-DBackbone=CLIP ViT-B/32, Covariance=Diagonal2025.1196.5692.5724.82
OursBackbone=CLIP ViT-B/322025.1196.3794.1416.39
DynProto2026.0495.33—18.28
PCME++Backbone=CLIP ViT-B/322025.1189.3199.842.57
RMDSBackbone=ViT-B/16, Evaluation Protocol=Fully Fine-tuned2024.0585.29——
ASHBackbone=ViT-B/16, Evaluation Protocol=Fully Fine-tuned2024.0584.23——
ReactBackbone=ViT-B/16, Evaluation Protocol=Fully Fine-tuned2024.0584.2——
ODINBackbone=ViT-B/16, Evaluation Protocol=Fully Fine-tuned2024.0583.74——
VIMBackbone=ViT-B/16, Evaluation Protocol=Fully Fine-tuned2024.0582.34——
MSPBackbone=ViT-B/16, Evaluation Protocol=Fully Fine-tuned2024.0581.97——
MaxLogitBackbone=ViT-B/16, Evaluation Protocol=16-shot Prompt Learning2024.0580.02——
CLS-MBackbone=ViT-B/16, Evaluation Protocol=16-shot Prompt Learning2024.0579.52——
GradNormBackbone=ViT-B/16, Evaluation Protocol=Fully Fine-tuned2024.0579.02——
EnergyBackbone=ViT-B/16, Evaluation Protocol=16-shot Prompt Learning2024.0578.87——
CLS-EBackbone=ViT-B/16, Evaluation Protocol=16-shot Prompt Learning2024.0578.58——
MDSBackbone=ViT-B/16, Evaluation Protocol=Fully Fine-tuned2024.0577.76——
TempScalingBackbone=CLIP ViT-B/162025.1176.2273.781.28
MaxSoftmaxBackbone=CLIP ViT-B/162025.1176.2173.6981.15
KNNBackbone=ViT-B/16, Evaluation Protocol=Fully Fine-tuned2024.0576.14——
TrustVLM*Backbone=SigLIP ViT-B/16, Extra features=DinoV22025.1174.9792.1674.2
MaxSoftmaxBackbone=SigLIP ViT-B/162025.1174.9169.2480.3
NegLabel2026.0474.46—64.89
EntropyBackbone=SigLIP ViT-B/162025.1174.4169.0482.51
TempScalingBackbone=SigLIP ViT-B/162025.1174.468.6280.85
TrustVLM*Backbone=CLIP ViT-B/16, Extra features=DinoV22025.1173.7291.8473.66
EntropyBackbone=CLIP ViT-B/162025.1171.9767.9983.73
BayesVLMBackbone=SigLIP ViT-B/162025.1170.5462.5282.52
NegPromptBackbone=ViT-B/16, Evaluation Protocol=16-shot Prompt Learning2024.0569.98——
LoCoOp (GL-MCM)Backbone=ViT-B/16, Evaluation Protocol=16-shot Prompt Learning2024.0569.08——
ZeroBackbone=CLIP ViT-B/162025.1166.7361.8187.23
ZeroBackbone=SigLIP ViT-B/162025.1166.6355.7987.95
MaxSoftmaxBackbone=CLIP ViT-B/322025.1163.2551.6287.97
TempScalingBackbone=CLIP ViT-B/322025.1163.2551.6287.95
ProLIPBackbone=CLIP ViT-B/162025.1162.4260.6489.44
BayesVLMBackbone=CLIP ViT-B/322025.1161.4641.4990.56
EntropyBackbone=CLIP ViT-B/322025.1161.0350.2892.07
MaxCosineBackbone=SigLIP ViT-B/162025.1160.4754.298.93
MaxCosineBackbone=CLIP ViT-B/162025.1159.2448.1887.98
CLIPNBackbone=ViT-B/16, Evaluation Protocol=Zero-shot CLIP2024.0557.56——
AdaNegBackbone=ViT-B/16, Evaluation Protocol=Zero-shot CLIP2024.0553.05——
NegLabelBackbone=ViT-B/16, Evaluation Protocol=Zero-shot CLIP2024.0553.02——
MCMBackbone=ViT-B/16, Evaluation Protocol=Zero-shot CLIP2024.0550.97——
MaxCosineBackbone=CLIP ViT-B/322025.1145.7235.1497.84
ProbVLMBackbone=CLIP ViT-B/322025.1142.7230.2395.53