7 Commits

Author	SHA1	Message	Date
Danny Sullivan	e5ae879ea0	clean up and average can now be set to an int indicating the number of iterations before averaging	2014-10-03 15:44:15 +02:00
Danny Sullivan	7b6cb91459	adding add_average method and solving iteration bug by replacing definition of t	2014-10-03 09:51:16 +02:00
Danny Sullivan	a5852c67e8	changing averaging to use sparse trick	2014-09-12 18:25:18 +02:00
Lars Buitinck	b8e1a6c357	ENH SGD Cython improvements ... * nogil declarations where possible (but no with nogil blocks, since the performance benefit of those is not clear) * use BLAS where appropriate * replaced DOUBLE by double, INTEGER and int32 by int/np.intc, since that's the type used by scipy.sparse * explicit checks that dataset dims don't exceed INT_MAX * replaced an np.all by an explicit loop (npymath library for C99 isfinite)	2013-09-07 18:24:36 +02:00
Peter Prettenhofer	63c7478c23	un-done NormedWeightVector factorization; performance decrease on RCV1 is neglectable.	2012-03-18 20:55:25 +01:00
Peter Prettenhofer	9b4604614d	factored out norm updates and moved them to a dedicated subclass	2012-03-07 21:18:08 +01:00
Peter Prettenhofer	7fcb2dbb3e	moved WeightVector und SequentialDataset into seperate modules.	2012-03-07 20:09:05 +01:00