Tiny-ImageNet
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Tiny ImageNet 50% known 50% novel classes (test)
58.3Known Accuracy
5
Tiny ImageNet Long-Tailed, p=100 (test)
0.427Accuracy
5
Tiny-ImageNet Symmetric Noise 0.2 (test)
0.829Accuracy (Best)
5
Tiny ImageNet 40% Mixed noise: A-20, S-20
37.9Accuracy
5
Tiny ImageNet 40% Mixed noise: A-30, S-10
39.7Accuracy
5
Tiny ImageNet 40% Asymmetric noise
37.6Accuracy
5
Tiny-ImageNet Clean AA Aggregate standard (test)
36.55Accuracy
4
Tiny-ImageNet Clean standard (test)
54.15Accuracy
4
Tiny-ImageNet Forget 10 Class
99.68Accuracy (Forgetting)
4
Tiny-ImageNet
64.18Top-1 Accuracy
4
Tiny-ImageNet (Non-IID 1)
52.62Accuracy
4
Tiny-ImageNet LT
39Clean Accuracy
4
Tiny ImageNet-C noise-free (test)
38.66Relative Accuracy
4
Tiny-ImageNet Symmetric Noise 0.5 (test)
71.07Accuracy
4
Tiny-ImageNet Cross-Device, 100 clients
1Average TPR
3
Tiny ImageNet-KR (test)
95.55Df (Forgetting Difference)
3
Tiny-ImageNet-C v1 (test)
46.48Accuracy (Severity 1)
3
Tiny-ImageNet 20-class 32x32
25.2Accuracy (Batch 1)
3
Tiny-ImageNet (All)
79.35Performance Score 1
2
Tiny-ImageNet Unseen
10.9Accuracy
2
Tiny-ImageNet 200 (test)
87.2mAP
2
Tiny ImageNet 200 (train)
227.647Training Delta R
2
Tiny-ImageNet Split 100 tasks
—Average Forgetting
0
Split Tiny-ImageNet 100 tasks (test)
—Accuracy
0
Tiny-ImageNet LT
—Primary metric
0