List your company
ResearchFormerly WizWand

Image Classification on CIFAR-100 Imbalance Ratio LT-50 (test)

90.2AccuracyLIFT
41.205653.925366.64579.3647Apr 7, 2021Sep 3, 2021Jan 30, 2022Jun 28, 2022Nov 24, 2022Apr 22, 2023Sep 18, 2023
Updated 8mo ago

Evaluation results

MethodLinks
LIFTBackbone=ViT-B/16, Learnable Params.=0.10M, Epochs=10, Training Protocol=Fine-tuning pre-trained model from ImageNet-21K2023.0990.2
LPTBackbone=ViT-B/16, Learnable Params.=1.01M, Epochs=40+40, Training Protocol=Fine-tuning pre-trained model from ImageNet-21K2023.0990
LIFT w/ TTEBackbone=ViT-B/16, Learnable Params.=0.10M, Epochs=10, Training Protocol=Fine-tuning pre-trained model2023.0983.1
LIFTBackbone=ViT-B/16, Learnable Params.=0.10M, Epochs=10, Training Protocol=Fine-tuning pre-trained model2023.0982
Meta-SuperDisco2023.0358.3
Alshammari et al.Venue=CVPR 222023.0357.7
Samuel et al.Venue=ICCV 212023.0357.6
DRO-LTBackbone=ResNet-32, Radius Config=Learned ε2021.0457.57
DRO-LTBackbone=ResNet-32, Radius Config=ε/√n2021.0457.2
BS + CMOBackbone=ResNet-32, Training Epochs=400, Augmentation=AutoAugment [9]2021.1256.7
Zhu et al.Venue=CVPR 222023.0356.6
BCLBackbone=ResNet-32, Learnable Params.=0.46M, Epochs=200, Training Protocol=Training from scratch2023.0956.6
Cui et al.Venue=ICCV 212023.0356
PaCoBackbone=ResNet-32, Training Epochs=400, Augmentation=AutoAugment [9]2021.1256
PaCoBackbone=ResNet-32, Learnable Params.=0.46M, Epochs=400, Training Protocol=Training from scratch2023.0956
DRO-LTBackbone=ResNet-32, Radius Config=Shared ε2021.0455.32
BSBackbone=ResNet-32, Training Epochs=400, Augmentation=AutoAugment [9]2021.1254.2
BSBackbone=ResNet-32, Learnable Params.=0.46M, Epochs=400, Training Protocol=Training from scratch2023.0954.2
DRO-LTBackbone=ResNet-32, Radius Config=ε = 0 (ERM)2021.0452.31
MiSLASBackbone=ResNet-32, Learnable Params.=0.46M, Epochs=200+10, Training Protocol=Training from scratch2023.0952.3
Wang et al.Venue=ICLR 212023.0351.7
Zhong et al.Venue=CVPR 212023.0351.5
DiVEBackbone=ResNet-32, Learnable Params.=0.46M, Epochs=200, Training Protocol=Training from scratch2023.0951.1
CBD + TailCalibXType=Ours2021.1150.9
Li et al.Venue=CVPR 212023.0350.5
GBMType=Loss or Re-weighting2021.1150.3
CosineCE + TailCalibXType=Ours2021.1149.8
CBDType=Distillation, re-implementation=true2021.1149.19
CosineCE + TailCalibType=Ours2021.1149.18
Domain AdaptationType=Loss or Re-weighting2021.1148.53
MODALSType=Generation, re-implementation=true2021.1147.57
Park et al.Venue=ICCV 212023.0347.4
TSCBackbone=ResNet-32, Imbalance Ratio=502021.1147.4
CRTType=Decouple, re-implementation=true2021.1147.26
SSPBackbone=ResNet-322021.0447.11
BBNType=Decouple2021.1147.02
BBNBackbone=ResNet-32, Learnable Params.=0.46M, Epochs=200, Training Protocol=Training from scratch2023.0947
smDRAGONBackbone=ResNet-322021.0446.85
cRTBackbone=ResNet-32, Reproduced=true2021.0446.83
L2RWType=Loss or Re-weighting2021.1146.83
τ-normBackbone=ResNet-32, Reproduced=true2021.0446.74
LDAMBackbone=ResNet-32, Learnable Params.=0.46M, Epochs=200, Training Protocol=Training from scratch2023.0946.6
KCL†Backbone=ResNet-32, Imbalance Ratio=502021.1146.3
LDAM-DRWBackbone=ResNet-32, Imbalance Ratio=502021.1146.2
CosineCEType=Baseline, re-implementation=true2021.1145.72
Meta Weight NetType=Loss or Re-weighting2021.1145.66
Class-Balanced FocalType=Loss or Re-weighting2021.1145.32
CB-FocalBackbone=ResNet-32, Imbalance Ratio=502021.1145.2
LDAMBackbone=ResNet-32, Imbalance Ratio=502021.1145
mixupType=Generation2021.1144.99
LDAM LossBackbone=ResNet-322021.0444.97
CEType=Baseline, re-implementation=true2021.1144.88
CE-DRWBackbone=ResNet-32, Imbalance Ratio=502021.1144.7
CB-CEBackbone=ResNet-32, Imbalance Ratio=502021.1144.6
CE-DRSBackbone=ResNet-32, Imbalance Ratio=502021.1144.5
Focal LossBackbone=ResNet-322021.0444.32
Focal LossType=Loss or Re-weighting2021.1144.32
FocalBackbone=ResNet-32, Imbalance Ratio=502021.1144.3
SSLBackbone=ResNet-32, Reproduced=true2021.0444.02
CEBackbone=ResNet-32, Imbalance Ratio=502021.1143.9
CEBackbone=ResNet-32, Reproduced=true2021.0443.85
Manifold mixupType=Generation2021.1143.09