List your company
ResearchFormerly WizWand

Image Classification on Imbalanced CIFAR-10 long-tailed, ratio 10 (test)

88.85Top-1 AccuracyMETA (CB-CE)
83.244484.699786.15587.6103Jun 22, 2020Jul 21, 2020Aug 19, 2020Sep 17, 2020Oct 16, 2020Nov 14, 2020Dec 13, 2020
Updated 8mo ago

Evaluation results

MethodLinks
META (CB-CE)Backbone=ResNet32, Loss Function=CB-CE2020.1288.85—
ABSGD (CB-Focal)Backbone=ResNet32, Loss Function=CB-Focal2020.1288.76—
ABSGD (CB-CE)Backbone=ResNet32, Loss Function=CB-CE2020.1288.57—
ABSGD (CB-CE)Backbone=ResNet32, Loss Function=CB-CE2020.1288.42—
ABSGD (CB-LDAM)Backbone=ResNet32, Loss Function=CB-LDAM2020.1288.4—
META (CB-Focal)Backbone=ResNet32, Loss Function=CB-Focal2020.1288.37—
BBNBackbone=ResNet-322020.0688.3100
ABSGD (CB-LDAM)Backbone=ResNet32, Loss Function=CB-LDAM2020.1288.27—
SGD (CB-LDAM)Backbone=ResNet32, Loss Function=CB-LDAM2020.1288.12—
ELF(LDAM)+DRWBackbone=ResNet-322020.0688-19.6
SGD (CB-LDAM)Backbone=ResNet32, Loss Function=CB-LDAM2020.1287.81—
CE+DRWBackbone=ResNet-322020.0687.60
LDAM+DRWBackbone=ResNet-322020.0687.60
ELF(CE)+DRWBackbone=ResNet-322020.0687.6-26.4
META (CB-LDAM)Backbone=ResNet32, Loss Function=CB-LDAM2020.1287.4—
Focal+DRWBackbone=ResNet-322020.0687.30
SGD (CB-Focal)Backbone=ResNet32, Loss Function=CB-Focal2020.1287.1—
MixupBackbone=ResNet-322020.0687.10
Manifold MixupBackbone=ResNet-322020.06870
ResamplingBackbone=ResNet32, Loss Function=CE2020.1286.99—
SGD (CB-CE)Backbone=ResNet32, Loss Function=CB-CE2020.1286.77—
FocalBackbone=ResNet-322020.0686.70
SGD (LDAM)Backbone=ResNet32, Loss Function=LDAM2020.1286.69—
CEBackbone=ResNet-322020.0686.40
ABSGD (CB-Focal)Backbone=ResNet32, Loss Function=CB-Focal2020.1285.9—
SGD (LDAM)Backbone=ResNet32, Loss Function=LDAM2020.1285—
ResamplingBackbone=ResNet32, Loss Function=CE2020.1284.59—
SGD (CB-CE)Backbone=ResNet32, Loss Function=CB-CE2020.1283.8—
SGD (CB-Focal)Backbone=ResNet32, Loss Function=CB-Focal2020.1283.46—