Digits
Benchmarks
TaskDatasetSOTA resultTrendResults
Digits (train)
0.113Training Loss
12
Digits (UCI)
0.001MSE
12
Digits MNIST
99.47Average Accuracy
12
Digits
73.1Regularity
11
Digits
0.577Fidelity (R2 Score)
11
Digits
100Stability
11
Digits
95.15Accuracy
11
Digits
88.1F1 Score
10
Digits
0.023Absolute Estimation Error
10
Digits (test)
93.6Accuracy
9
Digits
98.52Accuracy (MNIST)
9
digits
70.78F-measure
9
DIGITS subsampled to 100 (train)
82R500
9
Digits-Four (test)
92.8Accuracy (sv->mt)
8
Digits alpha=0.1 (test)
86.87Accuracy (MNIST Domain)
7
Digits
0.836ARI
7
Digits 5-9 (R2 protocol)
89.98Accuracy
7
Digits 0-4 D0–4-D4 protocol
86.11Accuracy
7
Digits 0-4 (D0–4-R6 protocol)
91.59Accuracy
7
Digits 10 classes scikit-learn bundled (test)
96.44Accuracy
6
Digits alpha=0.01
74.95Accuracy (MNIST)
6
Digits off-sphere (held-out split)
98.2Accuracy
6
Digits + Related External Data MNIST, SVHN, MNIST-M, USPS, KMNIST (test)
99.38Accuracy (MNIST)
6
Digits MNIST SVHN MNIST-M USPS (test)
99.23Accuracy (MNIST)
6
Digits
0.02Runtime (s)
6