Non-Negative Matrix Factorization (NMF) using CUDA

Wrapper package for the nmfgpu library, which implements several Non-negative Matrix Factorization (NMF) algorithms for CUDA platforms. By using the acceleration of GPGPU computing, the NMF can be used for real-world problems inside the R environment. All CUDA devices starting with Kepler architecture are supported by the library.


Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.

install.packages("nmfgpu4R") by Sven Koitka, 2 years ago

Report a bug at

Browse source code at

Authors: Sven Koitka [aut, cre, cph] , Christoph M. Friedrich [ctb]

Documentation:   PDF Manual  

GPL-3 | file LICENSE license

Imports Rcpp, Matrix, SparseM, stats, stringr, tools, utils

Suggests gdata

Linking to Rcpp

System requirements: CUDA >= v7.0, Nvidia GPU (e.g. GeForce or Tesla) with compute capability >= 3.0 (Kepler)

See at CRAN