Tiny-ImageNet
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Tiny-ImageNet 200
53.98BadNets CA
9
Tiny ImageNet C
26.26Accuracy
9
Tiny-ImageNet Pathological 10
261Accuracy@40
9
Tiny-ImageNet-C Severity 1-5
0.4Accuracy
9
Tiny ImageNet 64x64 (val)
27.5FID
8
Tiny ImageNet 5000-sample half seed-42 (test)
1.6ECE
8
Tiny-ImageNet Pathological 20
258Top-30 Accuracy
8
Tiny-ImageNet Dir(beta) (test)
80.24Accuracy (beta=0.5)
8
Tiny-ImageNet Full-Dataset setup (test)
54.9Accuracy
7
Tiny-ImageNet Held-out Clients (test)
90.6Accuracy
7
Tiny-ImageNet Dirichlet 0.3
212Accuracy@14
7
Tiny ImageNet 2,000 labels (test)
15.4Error Rate
7
Tiny ImageNet Dirichlet α=0.2 (test)
23.83Accuracy
6
Tiny-ImageNet-200
34.53FGSM Error Rate
6
Tiny-ImageNet-200 T200-B100-S20 (test)
34.85Final Accuracy
6
Tiny-ImageNet-200 T200-B100-S10 (test)
37.99Final Accuracy
6
Tiny-ImageNet-200 T200-B100-S5 (test)
38.1Final Accuracy
6
Tiny ImageNet 64x64x3 (train)
72.1Model size overhead (MB)
6
Tiny-ImageNet (test val)
24.23FID
6
Tiny Imagenet standard (val)
41.7kNN Accuracy
5
Tiny-ImageNet alpha = 0.4
50.97Top-1 Accuracy
5
Tiny-ImageNet alpha = 0.1
43.79Top-1 Accuracy
5
Tiny-ImageNet 200 (test)
55.89Accuracy
5
Tiny-ImageNet 5% v1 (test)
72.6C-Acc (BadNet)
5
Tiny-ImageNet Instance-dependent Noise 0.2 (test)
83.37Accuracy (Best)
5