Table 4 RBF-SVM on MNIST. Mean of three runs, the bold values are the best results

From: A dynamic fractional generalized deterministic annealing for rapid convergence in deep learning optimization

Method

Val. Acc. (%)

Train time (s)

γ chosen

SMO + grid–search

97.4

124

0.060

DF–GDA (ours)

98.7

37

0.057 ± 0.004

Back to article page