List your company
ResearchFormerly WizWand

Image Classification on CIFAR-10 4k labels

3.89Error RateMeta Pseudo Labels
3.39086.760410.1313.4996Nov 21, 2020Apr 23, 2021Sep 24, 2021Feb 25, 2022Jul 28, 2022Dec 29, 2022Jun 1, 2023
Updated 8mo ago

Evaluation results

MethodLinks
Meta Pseudo LabelsBackbone=Wide ResNet-28-22023.063.89
Doubly Robust Self-TrainingBackbone=Wide ResNet-28-22023.063.93
FixMatchBackbone=Wide ResNet-28-22023.064.31
UDABackbone=Wide ResNet-28-22023.064.32
ReMixMatchBackbone=Wide ResNet-28-22023.064.72
MixMatchBackbone=Wide ResNet-28-22023.064.95
SWABackbone=Wide ResNet-28-22023.065
EnAETBackbone=Wide ResNet-28-22023.065.35
MixMatchParameter Amount=36.5 M2020.115.53
SHOT-VAEParameter Amount=36.5 M2020.116.11
Mean TeacherParameter Amount=36.5 M2020.116.28
ICTBackbone=Wide ResNet-28-22023.067.66
SHOT-VAEParameter Amount=1.5 M2020.118.51
Mean TeacherBackbone=Wide ResNet-28-22023.069.19
CT-GANParameter Amount=1.5 M2020.1110.62
MixupParameter Amount=1.5 M2020.1110.71
LPParameter Amount=1.5 M2020.1111.82
LGA + VATBackbone=Wide ResNet-28-22023.0612.06
Pi-ModelParameter Amount=36.5 M2020.1112.16
VATParameter Amount=1.5 M2020.1113.13
Mean TeacherParameter Amount=1.5 M2020.1115.87
Pseudo-LabelingBackbone=Wide ResNet-28-22023.0616.09
Pi-ModelParameter Amount=1.5 M2020.1116.37