List your company
ResearchFormerly WizWand

Image Classification on CIFAR-100 N (test)

83.17Accuracy (Noisy)FE + NI-ERM
54.393261.864169.33576.8059Feb 28, 2022Aug 9, 2022Jan 19, 2023Jul 1, 2023Dec 10, 2023May 21, 2024Oct 31, 2024
Updated 8mo ago

Evaluation results

MethodLinks
FE + NI-ERMBackbone=DINOv2, Classifier=Logistic Regression, Data Augmentation=false2024.1083.17——
ProMix2024.1073.79——
PLS2024.1073.25——
DivideMix2024.1071.13——
DivideMixBackbone=ResNet-34, Number of independent runs=32023.0571.13——
ILL2024.1068.07——
ILLBackbone=ResNet-34, Number of independent runs=32023.0568.07—78.53
SOP+Backbone=PreActResNet182022.0267.81—78.91
SOPBackbone=ResNet-34, Number of independent runs=32023.0567.81—78.91
ELR+Backbone=ResNet342022.0266.72—78.57
ELRBackbone=ResNet-34, Number of independent runs=32023.0566.72—78.57
Co-teachingBackbone=ResNet342022.0260.37—73.46
Co-teachingBackbone=ResNet-34, Number of independent runs=32023.0560.37—73.46
ForwardBackbone=ResNet342022.0257.01—76.18
ForwardBackbone=ResNet-34, Number of independent runs=32023.0557.01—76.18
CORES*Backbone=ResNet342022.0255.72—73.87
CORESBackbone=ResNet-34, Number of independent runs=32023.0555.72—73.87
CEBackbone=ResNet342022.0255.5—76.7
CEBackbone=ResNet-34, Number of independent runs=32023.0555.5—76.7
CEBackbone=CLIP, Adaptation=FFT2024.09—72.41—
Deep kNN2021.11—52.71—
DEFTBackbone=CLIP, Adaptation=FFT2024.09—79.04—
ELRBackbone=CLIP, Adaptation=FFT2024.09—72.83—
FasTEN2021.11—63.53—
GLC2021.11—54.72—
GMMBackbone=CLIP, Adaptation=FFT2024.09—76.06—
L2RW2021.11—57.54—
LongReMixBackbone=CLIP, Adaptation=FFT2024.09—73.94—
MLaC2021.11—45.95—
MLoC2021.11—61.49—
MSLC2021.11—63.51—
MW-Net2021.11—61.12—
ProMixBackbone=CLIP, Adaptation=FFT2024.09—75.97—
RoLTBackbone=CLIP, Adaptation=FFT2024.09—75.91—
SCEBackbone=CLIP, Adaptation=FFT2024.09—72.52—
UNICONBackbone=CLIP, Adaptation=FFT2024.09—77.68—