2011/10/7  <[email protected]>:
> We wrestled with exactly this issue for decision trees, so its clear now that 
> a general solution would be very beneficial to scikit-learn.

For liblinear it might be a bit complicated since both the C++ code
and the cython wrapper would have to be rewritten to generate the two
versions (single and double precision) using some kid of preprocessor
at built time. Furthermore that would make upgrading from one version
of liblinear to another more complicate since we would have propagate
the same source level modifications each time.

For SGDClassifier, this is a single pure cython file that we fully
control. Does anybody knows if there is an easy way to script cython
to generate 2 versions of the function, one for single precision input
and the other for double precision? Or will we have to use an adhoc
source file transformation scheme?

-- 
Olivier
http://twitter.com/ogrisel - http://github.com/ogrisel

------------------------------------------------------------------------------
All of the data generated in your IT infrastructure is seriously valuable.
Why? It contains a definitive record of application performance, security
threats, fraudulent activity, and more. Splunk takes this data and makes
sense of it. IT sense. And common sense.
http://p.sf.net/sfu/splunk-d2dcopy2
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general

Reply via email to