List your company
ResearchFormerly WizWand

Image Classification on CIFAR-100 10000 labels (test)

16.24Error RateSelf-Tuning+HCR
14.841224.283133.72543.1669Nov 21, 2019Dec 21, 2020Jan 22, 2022Feb 23, 2023Mar 26, 2024Apr 27, 2025May 29, 2026
Updated 5mo ago

Evaluation results

MethodLinks
Self-Tuning+HCRNetwork=EfficientNet-B2, Number of Parameters=9.43M2022.0616.24
Self-TuningNetwork=EfficientNet-B2, Number of Parameters=9.43M2022.0617.57
FixMatchNetwork=EfficientNet-B2, Number of Parameters=9.43M2022.0621.69
Fine-TuningNetwork=EfficientNet-B2, Number of Parameters=9.43M2022.0621.74
FixMatch + IAM-Dlabeled samples=100002026.0521.99
Co-TuningNetwork=EfficientNet-B2, Number of Parameters=9.43M2022.0622.22
FixMatch + CMW-Net2022.0222.55
FixMatch2022.0222.6
FixMatchlabeled samples=100002026.0522.93
ReMixMatchNetwork=WRN-28-8, Number of Parameters=11.76M2022.0623.03
FixMatchNetwork=WRN-28-8, Number of Parameters=11.76M2022.0623.18
UDANetwork=WRN-28-8, Number of Parameters=11.76M2022.0624.5
UDA2022.0224.5
EnAETLabels=100002019.1126.93
MixMatchNetwork=WRN-28-8, Number of Parameters=11.76M2022.0628.31
MixMatch2022.0228.31
Mean TeacherNetwork=WRN-28-8, Number of Parameters=11.76M2022.0635.83
Mean Teacher2022.0235.83
Pseudo-LabelingNetwork=WRN-28-8, Number of Parameters=11.76M2022.0636.21
Pseudo-Labeling2022.0236.21
Π-ModelNetwork=WRN-28-8, Number of Parameters=11.76M2022.0637.88
Π-Model2022.0237.88
Temporal ensemblingLabels=100002019.1138.65
Pi-ModelLabels=100002019.1139.19
IAM-Dlabeled samples=100002026.0540.02
SAMlabeled samples=100002026.0543.3
SGDlabeled samples=100002026.0545.94
Supervised-onlyLabels=100002019.1151.21