[mlpack-git] [mlpack] mlpack KMeans class much slower than armadillo kmeans() (#514)

[Ryan Curtin]

mlpack's k-means does not use OpenMP at this point in time.  So yes, what you're seeing is normal.  But do keep in mind that mlpack's k-means implementation is far more flexible, allowing you to use five or six different algorithms---one of which will probably be faster than the default O(kN) algorithm.  (Try it;  for your dataset I'd expect Hamerly's algorithm or the single-tree Pelleg-Moore algorithm to be fastest.)

---
Reply to this email directly or view it on GitHub:
https://github.com/mlpack/mlpack/issues/514#issuecomment-178551026
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack-git/attachments/20160202/8bfdf8fb/attachment.html>