CIFAR10
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Cifar10 LT r20
0.44AUBC
6
Cifar10 LT r10
51.2AUBC
6
CIFAR10 to STL natural resolution shift (target)
53.418Accuracy
6
CIFAR10
96.22Accuracy
6
CIFAR10 Dirichlet α=0.2 (test)
55.44Accuracy
6
CIFAR10 (test)
71.12Test Accuracy (%)
6
CIFAR10 Noise ratio 0.6, Open ratio 1.0 (test)
94Accuracy
6
CIFAR10 Noise ratio 0.3, Open ratio 1.0 (test)
0.961Best Accuracy
6
CIFAR10 (test)
83.4Mean Accuracy
6
CIFAR10 (test)
44.97R@1
6
CIFAR10 standard (test)
3.74FID
6
CIFAR10 (test)
43.61Accuracy
6
CIFAR10 Corrupted (test)
75.4Test Accuracy
6
CIFAR10 250 labels (2%) (test)
95.4Accuracy
6
CIFAR10 40 labels (0.08%) (test)
94.6Accuracy
6
CIFAR10-LT gamma=200, 1125 labels (10%) (test)
0.714Accuracy
6
CIFAR10-LT gamma=100, 1244 labels (10%) (test)
79.2Accuracy
6
CIFAR10 permutation invariant (test)
0.466Test Error (%)
6
CIFAR10
23.53Unlearning Runtime (s)
5
CIFAR10-DVS
80.66Aauth Success Rate
5
CIFAR10 first batch (train)
1,000Iterations
5
CIFAR10 (val)
0.95BVS
5
CIFAR10
100Mask Score
5
CIFAR10 Non-IID 2-Class partition (test)
61Accuracy
5
CIFAR10 unlabeled 2 (train)
91.17Accuracy
5