CIFAR-10
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
CIFAR-10 (test)
82.1Clean Accuracy
7
CIFAR-10 (test)
94Top-1 Accuracy (BS=64)
7
CIFAR-10 Dirichlet alpha=0.3 (test)
46.57Global Accuracy
7
CIFAR-10 S4 search space (test)
2.59Error Rate
7
CIFAR-10 (test)
1Model Retention
7
CIFAR-10-C Corruption average (test)
62.56AUROC
7
Augmented CIFAR-10 Implicit Neural Representations (test)
56.95Accuracy
7
CIFAR-10 Dirichlet 0.5 (test)
79.47Personalized Accuracy
7
CIFAR-10 Instance-Dependent Noise 30% (test)
88.6Accuracy
7
CIFAR-10 (test)
85.7Test Accuracy (Dir μ=0.6)
7
CIFAR-10.2 (test)
0.122RMSE (%)
7
CIFAR-10.1 (test)
0.186RMSE (%)
7
CIFAR-10 50% symmetric noise
94.5Accuracy
7
CIFAR-10 η = 0.4 (class-conditional noise) (test)
70.63Accuracy
7
CIFAR-10 In and Out 1.0 (test)
0.928AUROC
7
CIFAR-10 ID-TEST DEEPNETS-1M (test)
0.885Avg Accuracy
7
CIFAR-10 10K-10K samples Protocol (test)
11.4FID
7
CIFAR-10 (test)
0.21Runtime (s)
6
CIFAR-10 @5 (5 sequential splits)
95.35Accuracy
6
CIFAR-10
92.73Classification Accuracy (%)
6
CIFAR-10
80.3Accuracy
6
CIFAR-10 WaNet trigger (test)
7ASR
6
CIFAR-10 (D)
64ASR
6
CIFAR-10 (LS K=1)
72.56Top-1 Accuracy
6
CIFAR-10N (Aggregate (A))
97.1Accuracy (ACC)
6