4

I am training ANN via package neuralnet in R and the problem is that it uses only 1 core in the training process. I've installed package parallel and I've tried to use something like

nn<-doParallel|parallel and so...(neuralnet(formula,.....)

but I am not wel skilled in these issues. Any advices to how to let all the cores work during the training process to make it faster? If I use

detectCores()

It sees all 4 cores.

Bury
  • 527
  • 2
  • 5
  • 15
  • I realize this is pretty old, but if you're still around and looking for a solution, you might find it here http://www.r-bloggers.com/grnn-and-pnn/ I'm very interested in this subject and have been trying to figure it out as well, to no avail. good luck! – Liz Young Jul 23 '15 at 20:30
  • Refer to my code here to see if it helps https://stackoverflow.com/questions/34861114/r-neuralnet-package-too-slow-for-millions-of-records – Nikhil Gupta Feb 21 '19 at 18:28

0 Answers0