CIFAR-10
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
CIFAR-10 (test)
5.77Error after Pruning
12
CIFAR-10 (test)
92.48Test Accuracy (Symmetric, η=0.2)
12
CIFAR-10S Dir(0.1) 50C
90.49Accuracy
12
CIFAR-10 (test)
95.2Clean Accuracy
12
CIFAR-10 Laplace5 noise 1.0 (test)
50.58Accuracy
12
CIFAR-10 Laplace1 noise 1.0 (test)
61.54Accuracy
12
CIFAR-10 Gaussian2 Noise (test)
60.34Accuracy
12
CIFAR-10 1v6
9.3Error (Actual)
12
CIFAR-10 (test)
0.949Baseline Accuracy
12
CIFAR-10 Flip2 noise 1.0 (test)
90.93Top-1 Acc
12
CIFAR-10 (test)
92.68Classification Accuracy (Cl-Acc)
12
CIFAR-10 250 labels (test)
4.73Test Error
12
CIFAR-10 LT (test)
13.9Error Rate
12
CIFAR-10
38.42FPR95 (SVHN)
12
CIFAR-10 long-tailed (test)
12.03FID
12
Long-tailed CIFAR-10 Imbalance factor 0.01 (test)
82.4Top-1 Accuracy
12
CIFAR-10 (epsilon=2/255) 1.0 (test)
28.07Clean Error (%)
12
Imbalanced CIFAR-10 (long-tailed, ratio 10) (val)
10.62Top-1 Error
12
Imbalanced CIFAR-10 long-tailed, ratio 100 (val)
20.18Top-1 Error
12
CIFAR-10 2009 (test)
96.53Accuracy
12
CIFAR-10N w
86.5Accuracy
11
CIFAR-10 (test)
94.45Test Accuracy
11
CIFAR-10 (test)
100Retain Accuracy
11
CIFAR-10 l_inf threat model, epsilon=8/255 1.0 (test)
95.18Standard Accuracy
11
CIFAR-10 (test)
100Utility Score
11