Galois Notes
============

Inputs are in /net/faraday/inputs/leml


Compilation 
===========

0. A compiled binary is included in codes/
1. If you want to compile your self, you need to change 
   the ``BLASFLAG'' in codes/Makefile manually.
2. type ``make'' to make the binary train-ml


Data sets
=========
	bibtex: a smaller dataset
	eurlex: a medium dataset
	wiki: a large dataset

Example Usage
=============

./codes/train-ml -k 50 smat-data/bibtex/


The current routines for sparse_matrix times dense_matrix
=================================================
Please see the last two functions in codes/smat.cpp
	void smat_x_dmat(const smat_t &X, const double* W, const size_t k, double *H);
	void smat_x_dmat(const double a, const smat_t &X, const double* W, const size_t k, const double *H0, double *H);
