List your company
ResearchFormerly WizWand

Image Classification on CIFAR-10 (Symmetry/Asymmetry Accuracy)

94.37Accuracy (Test, symm 20%)CAR
85.904488.102290.392.4978Aug 18, 2021
Updated 8mo ago

Evaluation results

MethodLinks
CARBackbone=ResNet34, Learning rate scheduler=cosine annealing, Trials=32021.0894.3793.4990.5680.9892.09
SATBackbone=ResNet34, Learning rate scheduler=cosine annealing, Trials=32021.0894.1492.6489.2378.5891.43
DACBackbone=ResNet34, Trials=32021.0892.9190.7186.374.8491.54
Joint OptBackbone=ResNet34, Trials=32021.0892.2590.7986.8769.1691.21
ELRBackbone=ResNet34, Learning rate scheduler=cosine annealing, Trials=32021.0892.1291.4388.8780.6990.35
SELFBackbone=ResNet34, Trials=32021.0891.1390.386.463.5988.32
GCEBackbone=ResNet34, Trials=32021.0889.8387.1382.5464.0785.21
SLBackbone=ResNet34, Trials=32021.0889.8387.1382.8168.1281.55
Forward TBackbone=ResNet34, Trials=32021.0887.9983.2574.9654.6486.44
NLNLBackbone=ResNet34, Trials=32021.0887.481.2568.339.5481.53
Cross EntropyBackbone=ResNet34, Trials=32021.0886.9881.8874.1453.8285.54
BootstrapBackbone=ResNet34, Trials=32021.0886.2382.2375.1254.1285.11