Yes, and we had a number of problems with it in the past (strange errors,
unacceptable latency...)
For the moment we'll probably split the topology at the prediction bolt and
use a queuing system to stack/unstack instances ; hence we'll be able to
write the prediction code in Python.
Also, after hacking for a week or so with Spark / MLLib, I'm quite in love
with Scala and Spark for distributed processing but frankly MLLib is still
really rough and eons away from sklearn in terms of API maturity, tooling
and experimental agility. I guess the core devs could borrow some
inspiration from sklearn :)
2014-07-05 13:36 GMT+02:00 Olivier Grisel <[email protected]>:
> Actually it's just using the ShellBolt.
>
> --
> Olivier
>
>
> ------------------------------------------------------------------------------
> Open source business process management suite built on Java and Eclipse
> Turn processes into business applications with Bonita BPM Community Edition
> Quickly connect people, data, and systems into organized workflows
> Winner of BOSSIE, CODIE, OW2 and Gartner awards
> http://p.sf.net/sfu/Bonitasoft
> _______________________________________________
> Scikit-learn-general mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>
------------------------------------------------------------------------------
Open source business process management suite built on Java and Eclipse
Turn processes into business applications with Bonita BPM Community Edition
Quickly connect people, data, and systems into organized workflows
Winner of BOSSIE, CODIE, OW2 and Gartner awards
http://p.sf.net/sfu/Bonitasoft
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
