Fig. 4From: A parallel and distributed stochastic gradient descent implementation using commodity clustersDepicts a box plot comparing the training time vs. node count for the dataset size 1:1Back to article page