MNIST
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
MNIST (test)
85Test Accuracy
6
MNIST (train)
0.001Loss
6
MNIST 20k samples
0.002MSE
5
MNIST
14Inferred K (K_hat)
5
MNIST INR
0.005Runtime (s)
5
MNIST (test)
0.097Cross-Entropy Loss
5
MNIST (train)
0.033Cross-entropy
5
MNIST permuted (val)
98.6Accuracy
5
MNIST (test)
0.4Error Rate
5
MNIST
1,000Iterations
5
MNIST
149Iterations
5
MNIST
14Iterations
5
MNIST (train)
95.55T5
5
MNIST (test)
97.79Accuracy (\epsilon=1)
5
MNIST 1000 samples (test)
19Number of Clusters
5
MNIST (test)
403Iterations
5
MNIST 2k
-0.23MLP Metric
5
MNIST CKKS (encrypted)
1,864.19Total Exp. Time (s)
5
MNIST
97.6Plaintext Accuracy
5
MNIST INR
97.2Accuracy
5
MNIST
97.21Accuracy
5
MNIST 10,000 images (val)
0.018L1 Loss
5
MNIST (Hard)
94.95Token Accuracy
5
MNIST (Hard)
94.37Token Accuracy
5
MNIST Hard
88.51Token Accuracy
5