List your company
ResearchFormerly WizWand

Image Classification on CIFAR-100-LT Imbalance 200

48.32Top-1 AccuracyDifficulty-Net + LAS
39.292841.636443.9846.3236Sep 7, 2022
Updated 8mo ago

Evaluation results

MethodLinks
Difficulty-Net + LASTraining setting=decoupled learning, Backbone=ResNet-322022.0948.32
Difficulty-Net + LWSTraining setting=decoupled learning, Backbone=ResNet-322022.0947.91
Difficulty-Net + Balanced SoftmaxTraining setting=e2e training, Backbone=ResNet-322022.0947.53
Difficulty-Net + cRTTraining setting=decoupled learning, Backbone=ResNet-322022.0947.45
PaCo + Balanced SoftmaxTraining setting=e2e training, Backbone=ResNet-322022.0946.72
BALMSTraining setting=decoupled learning, Backbone=ResNet-322022.0946.12
MWN + LASTraining setting=decoupled learning, Backbone=ResNet-322022.0945.04
LASTraining setting=decoupled learning, Backbone=ResNet-322022.0944.87
Difficulty-NetTraining setting=e2e training, Backbone=ResNet-322022.0944.8
MWN + LWSTraining setting=decoupled learning, Backbone=ResNet-322022.0944.71
MWN + cRTTraining setting=decoupled learning, Backbone=ResNet-322022.0944.56
LWSTraining setting=decoupled learning, Backbone=ResNet-322022.0944.42
cRTTraining setting=decoupled learning, Backbone=ResNet-322022.0944.05
EQLTraining setting=e2e training, Backbone=ResNet-322022.0943.46
PaCoTraining setting=e2e training, Backbone=ResNet-322022.0943.09
LDAMTraining setting=e2e training, Backbone=ResNet-322022.0941.42
CB-DATraining setting=e2e training, Backbone=ResNet-322022.0940.89
CDB-CETraining setting=e2e training, Backbone=ResNet-322022.0940.42
MWNTraining setting=e2e training, Backbone=ResNet-322022.0940.25
Class-BalancedTraining setting=e2e training, Backbone=ResNet-322022.0939.95
Focal LossTraining setting=e2e training, Backbone=ResNet-322022.0939.64