CIFAR
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
CIFAR100
83.65Base Accuracy
15
CIFAR-10 (test)
10.81FID
15
CIFAR10 (ID) vs CIFAR100 (OOD)
92.5AUROC
15
CIFAR-10-P (test)
2.11mFP
15
CIFAR-10
1.85Running Time (sec)
15
CIFAR10
0.62Loss Series Score
15
CIFAR-10 8-bit color (test)
2.49Bits per Dimension
15
CIFAR-10 BadNet DenseNet-161 (test)
90.17C-Acc
15
CIFAR-100 20 classes
89.4AUROC
15
CIFAR-10 6 classes
97.3AUROC
15
CIFAR-100
70.72Top-1 Acc (p=0.0)
15
CIFAR-10
93.23Top-1 Acc (p=0.0)
15
CIFAR10 (test)
91.6AUC
15
CIFAR-100 (test)
35.6Worst Subgroup Acc (1st)
15
CIFAR-10 (test)
95.02Baseline Accuracy
15
CIFAR10-DVS
85.69Accuracy
15
CIFAR-10 6/4 known-unknown class split
97.2Accuracy
15
CIFAR-100 standard (train test)
74.14Accuracy
15
CIFAR-10 25 labels per class 32x32
4.39Error Rate
15
CIFAR-10 Long-Tailed lambda=150 (test)
0.202Error Rate
15
CIFAR-100-LT Imbalance Factor 50
58.1Accuracy
15
CIFAR10 (test)
50.91Clean Error Rate
15
CIFAR-10 (test)
73.2Test Accuracy
15
CIFAR-10 50 images per class
74.5Test Accuracy
15
CIFAR-10 50 images/class (test)
74.5Accuracy
15