MNIST
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Static MNIST (test)
-79.905Log Likelihood
3
MNIST OOD (test)
0.513RMS Calibration Error
3
MNIST (Target (mix of Group 1 and 2))
65Accuracy
3
MNIST (Target (Group 3))
55Accuracy
3
MNIST Target (Group 2)
69.9Accuracy
3
MNIST Target (Group 1)
80Accuracy
3
MNIST far-OOD ID: CIFAR-10 OpenOOD (test)
44.5FPR@95
3
MNIST 5-layer MLPs (test)
0.871Kendall's Tau
3
MNIST 3-layer MLPs (test)
0.876Kendall's Tau
3
MNIST-Even (test)
96.37Accuracy
3
MNIST (test)
0.048MSE
3
MNIST (test)
0.019MSE
3
psMNIST permuted sequential (test)
97.3Test Accuracy
3
MNIST 28x28 (test)
78.01NLL (nats)
3
mnist (test)
99.91Accuracy
3
MNIST 8M
77.4Triplet Centroid Accuracy
3
MNIST 8M
96.7Trustworthiness
3
MNIST original (test)
0.95Accuracy
3
MNIST
4.23Avg Execution Time (ms/query)
3
MNIST (inlier) vs. FGSM (outlier) (test)
81.7ROC AUC
3
Shuffled MNIST 100 tasks
85.4Accuracy
3
Split MNIST
97.92Accuracy
3
MNIST
11.56Sampling Time (s)
3
MNIST (train)
-3,483.94Negative Log-Likelihood
3
MNIST Clustered (test)
85.2Accuracy
3