MNIST
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
MNIST-10 (test)
92.9Test Accuracy
19
MNIST (test)
89.15Accuracy
18
MNIST-10C
84.12Accuracy
18
MNIST (train)
0.028Cross-Entropy Loss
18
MNIST
95.8NMI
18
MNIST 10,000 samples (test)
4.2Test Error
18
mnist
100Algorithm Success Rate
18
MNIST
84.1Average Accuracy
18
MNIST
0.48Running Time
18
MNIST 30% noise
0.05MSE
18
MNIST 5% noise
0.039MSE
18
MNIST uniform label noise η=0.1 (7-fold CV)
97.74Accuracy
18
MNIST (test)
0.028Test Cross-Entropy
18
MNIST 100 clients, highly heterog. (test)
96Test Accuracy
18
MNIST (test)
95.8FGSM Accuracy
18
MNIST
87AER
18
MNIST (test)
98.6Accuracy
18
MNIST
95.4Top-1 Accuracy
18
MNIST-scale ideal downsampling (test)
99.14Acc
18
MNIST
1.7Classification Error (Bernoulli)
18
Split MNIST S (test)
99.83Task-Averaged Accuracy
18
Mnist-r
-0.793Test Squared Error
18
MNIST (test)
55Execution Time (s)
18
MNIST full
640Execution Time (s)
18
MNIST unpermuted pixel-by-pixel (test)
99.2Accuracy (Test)
18