On Tue, Oct 30, 2012 at 01:38:48PM +0000, Andreas Mueller wrote: > I tried working on the neural network a bit more and I think it would be > good to use > blas3 calls there. Unfortunately, these are not included in the blas > that sklearn ships. > This problem already came up for the euclidean distance speedup PR.
And it's not resolved, as the merge of this branch ended up in a fascio with a random segfault at import that still hasn't been solved. I spent a while on it and could solve it. I'd say that the current situation is that we don't have a clean way of using blas 3. I think that the most promising step is this direction would be to stabilize the eucliean distance speed up branch (a new pull request should probably be created) and try to get it to be reliable. I don't know out of the top of my head how to get the last part done (having it reliable). I think that it's just a 'simple matter of programming', but last time I tried, I fail to achieve it in the time lapse that I head. Gael ------------------------------------------------------------------------------ Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_sfd2d_oct _______________________________________________ Scikit-learn-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
