List your company
ResearchFormerly WizWand

Image Classification on Clothing1M (test)

82.1AccuracySOMNet
74.466476.448278.4380.4118Sep 13, 2016Apr 14, 2018Nov 14, 2019Jun 15, 2021Jan 15, 2023Aug 16, 2024Mar 18, 2026
Updated 6mo ago

Evaluation results

MethodLinks
SOMNetBackbone=ResNet-502019.0682.1——
50-ResNet (Fine-tuned)loss=cross-entropy, init=#6, training=50k2016.0980.38——
Nonedata=50k clean, pretrained=#3 model2017.1180.38——
Robust Self-Training with Closed-loop Label CorrectionExtra data=Meta-knowledge2026.0380.23——
Robust Self-Training with Closed-loop Label CorrectionExtra data=Meta-knowledge, Clean data percentage=50%2026.0379.92——
Nonedata=50k clean, pretrained=#6 model2017.1179.9——
CurriculumNetBackbone=ResNet-50, Source=Re-run2019.0679.8——
EMLCExtra Data=true, k=12023.0579.35——
EMLCExtra data=Meta-knowledge, Self-supervised learning=true2026.0379.35——
AlexNetloss=cross-entropy, init=#1, training=1M, 50k2016.0978.24——
Robust Self-Training with Closed-loop Label CorrectionExtra data=Meta-knowledge, Clean data percentage=10%2026.0378.17——
EMLCExtra Data=true, k=52023.0578.16——
FasTENClean set usage=true, Loss Correction=true, Backbone=ResNet-50, Pre-trained=ImageNet2021.1177.83——
FasTENExtra Data=true2023.0577.83——
FasTENExtra data=Direct training on clean instances2026.0377.83——
PUDistillBackbone=ResNet-50, Pre-trained=ImageNet2021.0577.7——
L2B-Naive2022.0277.5——
DMLPExtra data=Meta-knowledge, Self-supervised learning=true2026.0377.31——
DMLPExtra data=Meta-knowledge, Self-supervised learning=true, Clean data percentage=50%2026.0377.3——
FasTEN w/o LCClean set usage=true, Loss Correction=false, Backbone=ResNet-50, Pre-trained=ImageNet2021.1177.07——
Self LearningExtra Data=true2023.0576.44——
AlexNetloss=cross-entropy, init=#1, training=1M, 50k2016.0976.22——
MSLGExtra Data=true2023.0576.02——
MSLG2022.0276——
MLC2022.0275.8——
MLaCClean set usage=true, Backbone=ResNet-50, Pre-trained=ImageNet2021.1175.78——
MLCExtra Data=true2023.0575.78——
LRA-diffusionConditioned on=CC* feature encoder2023.0575.7——
SANMbase_method=DivideMix2023.0275.63——
SANMExtra Data=false2023.0575.63——
SANM2023.0575.63——
DMLPExtra data=Meta-knowledge, Self-supervised learning=true, Clean data percentage=10%2026.0375.5——
CC2023.0575.4——
MFRW2022.1175.35——
CC*Backbone=ResNet50, Pre-processing=Sample selection via CC [44]2023.0575.32——
DivideMix+SNSCL2022.0275.3——
EPL2023.0575.21——
CLEANnoise structure=agnostic2018.0575.2——
Knockoffs-SPR2023.0175.2——
50-ResNetloss=cross-entropy, init=ImageNet, training=50k2016.0975.19——
Nonedata=50k clean, pretrained=ImageNet2017.1175.19——
PGDF2021.1275.19——
VRIBackbone=ResNet-50, Pre-trained=ImageNet2026.0375.19——
DM-AugDescstrategy=WS-SAW2021.0375.11——
DM-AugDesc-WS-SAW2021.1275.11——
AugDescClean set usage=false, Backbone=ResNet-50, Pre-trained=ImageNet2021.1175.11——
AugDesc2023.0275.11——
AugDescExtra Data=false2023.0575.11——
AugDesc2022.1175.11——
AugDesc2022.0275.1——
Nested+Co-teachingBackbone=ResNet-50, Balanced subset or loss=true2022.0675——
PESEnsemble=true2021.0674.99——
UNICONBackbone=ResNet-502022.0374.98——
UNICON2023.0574.98——
WarPIBackbone=ResNet-50, Pre-trained=ImageNet2026.0374.98——
ProMixBackbone=ResNet-50, Pre-trained=ImageNet2022.0774.94——
CoDisBackbone=ResNet-50, Pre-trained=ImageNet2026.0374.92——
Nested-CoTeachingtraining_labels=noisy labels only2022.0974.9——
Nested2023.0274.9——
RRL2023.0274.9——
ULCBackbone=ResNet-50, Pre-trained=ImageNet2022.0774.9——
Nested2022.0274.9——
RRL2022.0274.9——
CLIPCleanerIntegrated module=DivideMix2024.0874.87——
MixFixBackbone=PreResNet-18, Zero-shot=false2024.0874.87——
C2D2024.0874.84——
C2DBackbone=PreResNet-18, Zero-shot=false2024.0874.84——
SSR+2024.0874.83——
kMEIDTM (+DivideMix)Integration=DivideMix, Implementation=Based on authors' code (*)2022.0674.82——
ELR+Backbone=ResNet-50, Pre-trained=ImageNet2020.0674.81——
ELR+2021.0374.81——
ELR+Initializer=ImageNet2021.0474.81——
ELR+Ensemble=true2021.0674.81——
ELR+2021.1274.81——
ELRBackbone=ResNet-502022.0374.81——
ELR+Backbone=ResNet-50, Pre-trained=ImageNet2022.0774.81——
ELR+Extra Data=false2023.0574.81——
ELR2023.0574.81——
ELR+2024.0874.81——
ELR+Backbone=ResNet-50, Pre-training=ImageNet, Mixup=true, Two Networks=true, Weight Averaging=true2021.0874.81——
ELRBackbone=ResNet-50, Pre-trained=ImageNet2026.0374.81——
ELR+Backbone=ResNet-50, Balanced subset or loss=true2022.0674.8——
DivideMixBackbone=ResNet-50, Balanced subset or loss=true2022.0674.8——
RRLBackbone=ResNet-502023.0374.8——
TCLBackbone=ResNet-502023.0374.8——
DivideMix2022.0274.8——
TCL2024.0874.8——
DivideMix2020.0274.76——
DivideMixBackbone=ResNet-50, Pre-trained=ImageNet2020.0674.76——
DivideMix2021.0374.76——
DivideMixInitializer=ImageNet2021.0474.76——
DivideMixBackbone=ResNet-50, Pre-trained=ImageNet2021.0574.76——
DivideMixEnsemble=true2021.0674.76——
DivideMix2021.1274.76——
DividemixSetting=CL2022.0474.76——
DMixBackbone=ResNet-502022.0374.76——
DivideMixClean set usage=false, Backbone=ResNet-50, Pre-trained=ImageNet2021.1174.76——
DivideMixPre-training=ImageNet2021.0374.76——
DivideMixtraining_labels=noisy labels only2022.0974.76——
DivideMix2021.0374.76——