List your company
ResearchFormerly WizWand

Image Classification on CIFAR-100 80% symmetric noise (test)

73.8AccuracyCUFIT
3.621.82540.0558.275Dec 2, 2021Jun 2, 2022Dec 1, 2022Jun 1, 2023Nov 30, 2023May 30, 2024Nov 29, 2024
Updated 8mo ago

Evaluation results

MethodLinks
CUFITBackbone=DINOv2, Noise rate=0.82024.1173.8
PGDFEpoch selection=best2021.1266.7
Co-teachingBackbone=DINOv2, Noise rate=0.82024.1166.7
PGDFEpoch selection=last2021.1265.9
JoCorBackbone=DINOv2, Noise rate=0.82024.1164.2
CoDisBackbone=DINOv2, Noise rate=0.82024.1163.7
LongReMixEpoch selection=best2021.1262.9
LongReMixEpoch selection=last2021.1262.3
DM-AugDesc-WS-SAWEpoch selection=best2021.1261.8
DM-AugDesc-WS-SAWEpoch selection=last2021.1261.6
ELR+Epoch selection=best2021.1260.8
DivideMixEpoch selection=best2021.1260.2
DivideMixEpoch selection=last2021.1259.6
Linear probingBackbone=DINOv2, Noise rate=0.82024.1159.6
M-correctionEpoch selection=best2021.1248.2
M-correctionEpoch selection=last2021.1247.6
Meta-LearningEpoch selection=best2021.1242.4
Meta-LearningEpoch selection=last2021.1240.1
MixupEpoch selection=best2021.1230.8
ReinBackbone=DINOv2, Noise rate=0.82024.1125.6
Cross-EntropyEpoch selection=best2021.1219.9
MixupEpoch selection=last2021.1217.6
Cross-EntropyEpoch selection=last2021.128.8
Full-trainingBackbone=DINOv2, Noise rate=0.82024.116.3