List your company
ResearchFormerly WizWand

Fine-grained Classification on Stanford Dogs 20% symmetric noise (test)

81.4Best AccuracyDivideMix + SNSCL
63.002467.778772.55577.3313Mar 4, 2023
Updated 8mo ago

Evaluation results

MethodLinks
DivideMix + SNSCLSNSCL=true2023.0381.481.16
DivideMixSNSCL=false2023.0379.2277.86
MLC + SNSCLSNSCL=true2023.0378.9278.56
SYM + SNSCLSNSCL=true2023.0377.5577.24
MW-Net + SNSCLSNSCL=true2023.0377.4977.08
Label Smooth + SNSCLSNSCL=true2023.0376.8576.12
Cross-Entropy + SNSCLSNSCL=true2023.0376.3375.83
Conf. Penalty + SNSCLSNSCL=true2023.0376.1475.73
GCE + SNSCLSNSCL=true2023.0375.9974.56
JoCoR + SNSCLSNSCL=true2023.0375.7974.99
Co-teaching + SNSCLSNSCL=true2023.0374.1873.09
MLCSNSCL=false2023.0374.0870.51
Label SmoothSNSCL=false2023.0373.5164.42
Conf. PenaltySNSCL=false2023.0373.2266.89
Cross-EntropySNSCL=false2023.0373.0163.82
MW-NetSNSCL=false2023.0371.9969.2
SYMSNSCL=false2023.0369.262.13
GCESNSCL=false2023.0366.9666.93
JoCoRSNSCL=false2023.0366.9460.81
Co-teachingSNSCL=false2023.0363.7158.43