CIFAR10
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
CIFAR10-LT Blobs OOD synthetic noise (test)
99.12AUC
4
CIFAR10-LT synthetic Rademacher noise (test)
99.78AUC
4
CIFAR10-LT Gaussian OOD synthetic noise (test)
99.83AUC
4
CIFAR10 unlabeled 1 (train)
0.016Absolute Estimation Error
4
CIFAR10 Split (2 tasks)
88.77Avg Accuracy (A_T)
4
CIFAR10 SOLS with Bernoulli shift (test)
0.163Error Rate
4
CIFAR10 32x32 (test)
63.94FID
4
CIFAR10-LT gamma=200, 3365 labels (30%) (test)
74.1Accuracy
4
CIFAR10C gradually changing (test)
0.104Avg Error (%)
4
CIFAR10 6 random classes ID vs 4 random classes OOD (test)
0.97AUROC
4
CIFAR10-5 SSCL New Class
95.78Task 1 Accuracy
4
CIFAR10 RobustBench
4.29Mean Robust Accuracy Improvement
4
Cifar10 (test)
74.3Min Test Accuracy
4
C+50 (CIFAR10 with 50 unseen classes) (test)
74.6F1 Score
4
CIFAR10 C+10 10 unseen classes (test)
79.38F1 Score
4
CIFAR10 1,000 labels (train)
0.182Error Rate
4
CIFAR10 Pixel 32x32x3 (test)
3.67Wall-clock Speedup over DDPM
3
Cifar10 imb
85.2F-Accuracy (2%)
3
CIFAR10 2 modal
86Gap Statistic
3
CIFAR10 2 modal
0.86Gap
3
CIFAR10
91.73Clean Accuracy
3
CIFAR10 (test)
93.8Accuracy (1-shot)
3
CIFAR10 (test)
82.28PA
3
LT-CIFAR10 imb. 100 (test)
0.267Error Rate @ 80% TPR
3
CIFAR10.1
96Rejection Rate
3