Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

STL10

Benchmarks

Task NameDataset NameSOTA ResultTrend
Image ClassificationSTL10
Accuracy99.6
103
Image ClassificationSTL10-LT (gamma_l = 10) (test)
Accuracy98.7
65
Image ClassificationSTL10 (test)
Error Rate (%)4.82
53
Image ClassificationSTL10 gamma_l = 20 long-tail (test)
Accuracy98.9
49
Image ClassificationSTL10-LT (test)
Accuracy84.8
36
Image ClassificationSTL10 (val)
Accuracy98
33
Backdoor Attack in Dataset CondensationSTL10
CTA72.92
31
Semi-supervised Long-tailed Image ClassificationSTL10 LT
Accuracy79
24
Image ClassificationSTL10 Dp: ImageNet (test)
BA82.94
22
Zero-shot ClassificationSTL10
Top-1 Clean Acc97.6
21
Debugging Effectiveness EstimationSTL10
ATPF88.67
21
Image ClassificationSTL10
Testing Accuracy97.16
18
ClassificationSTL10 Noise (test)
Accuracy (Original)98.58
18
ClassificationSTL10 Rotation (test)
Accuracy (Original Data)98.62
18
ClusteringSTL10 small-scale (val)
NMI99.2
17
Image ClusteringSTL10 (labelled)
Accuracy59.8
17
Image ClassificationSTL10 (Nl=150, M=100k, γl=10) long-tailed (test)
Accuracy71.3
16
Attribution Adversarial SensitivitySTL10 96x96 (test)
VGrad0.91
15
Image ClassificationSTL10 original (test)
Accuracy98.31
15
Image ClassificationSTL10 43 (test)
Accuracy77.45
15
Image GenerationSTL10
Inception Score10.41
15
Attention-based attribution sensitivitySTL10 96×96 (test)
AttRaw Score0.32
12
Backdoor Attack on Dataset CondensationSTL10
CTA72.4
12
Generative ModelingSTL10 96x96 (test)
FID77.24
12
Model Stealing AttackSTL10
Agreement94.2
11
Showing 25 of 61 rows