List your company
ResearchFormerly WizWand

Image Classification on CIFAR10 Corrupted (test)

86.5AccuracySTAMP + Hi-Vec
59.56466.55773.5580.543Aug 11, 2025
Updated 7mo ago

Evaluation results

MethodLinks
STAMP + Hi-VecBackbone=ResNet-182025.0886.569.977.3
STAMPBackbone=ResNet-182025.088569.476.4
SAR + Hi-VecBackbone=ResNet-182025.0884.174.979.2
Tent + Hi-VecBackbone=ResNet-182025.0883.97579.2
SoTTABackbone=ResNet-182025.0882.762.771.3
SARBackbone=ResNet-182025.0882.36673.2
CoTTABackbone=ResNet-182025.0881.865.472.5
EATABackbone=ResNet-182025.0881.567.974.1
TentBackbone=ResNet-182025.0880.165.972.3
BN StatsBackbone=ResNet-182025.0879.466.972.6
RoTTABackbone=ResNet-182025.0879.462.469.8
MRLBackbone=ResNet-182025.0871.363.866.4
OWTTTBackbone=ResNet-182025.0865.963.564.4
SourceBackbone=ResNet-182025.0860.661.560