1
$\begingroup$

I was planning to build a Neural Network model and create a list of the same model with the elements of the list being different Initializations of the same model. Is it possible to Train the entire list consisting of say 5 Neural Network models, simultaneously, either by using parallelization technique, or by mapping the NetTrain function on the list? Any help will be highly appreciated, as this will save a lot of time for my project.

Improve this question
$\endgroup$
2
  • 1
    $\begingroup$ I don't believe it makes a lot of sense, because NetTrain automatically uses all available cores on the current machine (if using TargetDevice->CPU) anyways. Because of this the time difference would be negligible just training them serially anyways. (If you have the option of a bunch of different kernels running on different machines or something, disregard this message) $\endgroup$
    – Carl Lange
    Commented May 21, 2021 at 16:35
  • $\begingroup$ Are you asking if WL has anything similar to TensorFlow distributed training? $\endgroup$
    – Rohit Namjoshi
    Commented May 21, 2021 at 23:15

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.