12
$\begingroup$

I've had p = Predict[x -> y] running for hours and I have now idea if I should kill it or not.

I'm looking for some rough bounds for an estimate of how long Predict[] or Classify[] will take to train depending on the size of the input and the method used:

  • "LinearRegression"
  • "NearestNeighbors"
  • "RandomForest"
  • "NeuralNetwork"

Would be really helpful to monitor the loss function like you can in other ML systems.

$\endgroup$
13
$\begingroup$

This is on our list of planned functionality.

For now, you can estimate the run time by training on a smaller set first, and then extrapolating.

$\endgroup$
  • $\begingroup$ In 10.3 or 11? what would a rough time estimate be? Also the run time would be problem specific, can you show an example with code maybe? $\endgroup$ – M.R. Jul 29 '15 at 20:18

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.