On 10/18/2013 04:23 AM, Lars Buitinck wrote: > 2013/10/18 Stephane Clinchant <stephane.clinch...@gmail.com>: >> normWHT = np.trace(np.dot(np.dot(H.T, np.dot(W.T, W)), H)) >> MemoryError >> >> Do you have any explanations/motivations for changing the implementations of >> NMF ? > I'm not sure why this was done, but it's already fixed in master. The > current NMF code scales quite gracefully to large sparse matrices, as > long as the number of components is kept small (and the tolerance > increased a bit). > Some variants (I think sparse input?) were completely broken before, which prompted a rewrite. I thought the 0.14.1 had two implementations for the different cases but apparently that was not the case? Vladn knows more ;)
------------------------------------------------------------------------------ October Webinars: Code for Performance Free Intel webinars can help you accelerate application performance. Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from the latest Intel processors and coprocessors. See abstracts and register > http://pubads.g.doubleclick.net/gampad/clk?id=60135031&iu=/4140/ostg.clktrk _______________________________________________ Scikit-learn-general mailing list Scikit-learn-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
