CIFAR
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
CIFAR-10
94.73BA
18
CIFAR-100
26.6Top-1 Error Rate
18
CIFAR-10-LT Imbalance Ratio 50 (test)
88.5Top-1 Accuracy
18
CIFAR-100 Sequential
80.22Accuracy
18
Sequential CIFAR-10
96.32Accuracy
18
CIFAR-10 Sequential
83.97Accuracy
18
S-CIFAR-100 (test)
5.75Forgetting
18
S-Cifar-10 (test)
1.24Forgetting
18
S-Cifar-100 (test)
19.29Forgetting
18
S-Cifar-10 (test)
8.23Forgetting Rate
18
S-Cifar-100 streaming (test)
8.9Forgetting Rate (FRG)
18
S-Cifar-10 streaming (test)
0.201Forgetting Rate (FRG)
18
Binarized CIFAR
82.7Accuracy
18
CIFAR-FS (test)
32.1Worst Accuracy
18
CIFAR-10
96.74Accuracy
18
CIFAR-100 imbalanced version (test)
0.625AUPRC
18
CIFAR-100 instance-dependent noise (IDN) (test)
68.82Acc (η=0.2)
18
CIFAR-10 instance-dependent noise (IDN) (test)
90.31Accuracy (η=0.2)
18
CIFAR10-MC
55.2Type NMI
17
CIFAR-10 OpenOOD v1.5 (test)
91.34AUROC
17
CIFAR-100 10 task
70.59Avg Incremental Acc
17
CIFAR100
88.8Accuracy
17
cifar10
126,492,578.88Clustering Inertia
17
CIFAR-100 (test)
64.8Accuracy (k=80)
17
CIFAR-10
1.339Loss @ Epoch 30
17