Fig. 14From: Sequence-to-sequence modeling for graph representation learningGrowth of training time versus number of graphsBack to article page