Open in new window / Try shogun cloud
--- Log opened Thu Sep 22 00:00:15 2011
-!- blackburn [~blackburn@188.168.5.118] has quit [Quit: Leaving.]00:44
-!- Brian [4b930b45@gateway/web/freenode/ip.75.147.11.69] has joined #shogun05:44
BrianHi05:44
-!- Brian is now known as Guest6394305:44
-!- Guest63943 [4b930b45@gateway/web/freenode/ip.75.147.11.69] has quit [Ping timeout: 252 seconds]06:15
-!- sonne|work [~sonnenbu@194.78.35.195] has joined #shogun10:09
-!- sonne|work [~sonnenbu@194.78.35.195] has left #shogun []13:45
-!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has joined #shogun14:12
-!- sonne|work [~sonnenbu@194.78.35.195] has joined #shogun14:25
CIA-3shogun: Soeren Sonnenburg master * rb73b2c1 / (7 files in 5 dirs):17:34
CIA-3shogun: add some helper functions like clear vector, fill vector etc to SGVector17:34
CIA-3shogun: This unfortunately causes conflicts w/ modelselection that using17:34
CIA-3shogun: SGVector<void>. Until ms is resolved to use sth else it will remain broken. - http://git.io/Lnpp6g17:34
CIA-3shogun: Soeren Sonnenburg master * rdd4e99f / (10 files): add new .cs examples from daniel - http://git.io/qok8NQ17:34
shogun-buildbotbuild #247 of cmdline_static is complete: Failure [failed test_1]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cmdline_static/builds/247  blamelist: sonne@debian.org17:44
shogun-buildbotbuild #227 of python_static is complete: Failure [failed test_1]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/python_static/builds/227  blamelist: sonne@debian.org17:44
shogun-buildbotbuild #235 of r_static is complete: Failure [failed test_1]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/r_static/builds/235  blamelist: sonne@debian.org17:44
shogun-buildbotbuild #232 of octave_static is complete: Failure [failed test_1]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/octave_static/builds/232  blamelist: sonne@debian.org17:44
-!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has quit [Quit: Page closed]19:30
-!- blackburn [~blackburn@188.168.4.56] has joined #shogun21:16
blackburnsonney2k: heh did you broke modelselection?21:45
@sonney2kyes21:45
@sonney2kit is using SGVector<void>21:45
@sonney2kthat clashes with any of the new functions like zero()21:46
@sonney2kfill()21:46
@sonney2ketc21:46
@sonney2kso MS has to be done differently21:46
@sonney2knot exactly sure how though - need to talk/chat to heiko21:47
blackburnwill you fix it?21:47
@sonney2kI am not sure I can21:47
blackburnI don't have a clue about his impls21:47
blackburnsonney2k: what is the next release date?21:48
blackburnI'd say december is perfect21:48
@sonney2kblackburn, whenever we have sth ready21:53
@sonney2kcurrently I am only moving stuff from .h -> .cpp21:53
@sonney2kand will hopefully be able to remove Array*21:53
@sonney2k(not easy)21:54
blackburnI did superlu integration and continuing with auto-dim-reduction techniques21:54
@sonney2kthen I would wish to have all the platforms, osx, cygwin, linux working and being integration tested in the buildbot21:54
@sonney2kand hopefully heiko's migration stuff and nicer modelselection ...21:55
@sonney2kand other bugfixes of course21:55
blackburnsonney2k: what do you think about CUDA, etc?21:56
@sonney2kopencl you mean?21:57
@sonney2kfor which algorithms?21:57
blackburnyeah techs of this kind21:57
blackburnno idea21:57
@sonney2kI played around with this some time ago - it didn't help for strings21:58
blackburnI will be involved with some cuda svms21:58
@sonney2kit might for other simpler applications21:58
blackburnprobably21:58
blackburnwell I will be do images21:58
@sonney2kthere is already one which is OSS21:58
blackburnOSS?21:58
blackburnwell nor me or you do string kernel learning right now21:59
blackburnno chance I will do with next two years :)21:59
blackburnbut I will be working with images22:00
@sonney2khttp://code.google.com/p/multisvm/22:00
blackburnand why not to create branch of gpu accelerated shogun22:00
blackburnsure but we could have many more things22:01
blackburnnot only svm22:01
blackburnit would be impressive to use any advantages of machines22:02
blackburnno idea where to get so much time to get it done :D22:03
@sonney2kand the hardware that uses it...22:03
@sonney2kI don't even have any serious graphics adapter22:03
@sonney2kso *I* cannot do this22:03
blackburnme too22:03
@sonney2kI don't mind it being done22:04
@sonney2kthough22:04
blackburnokay we will see22:04
blackburnI guess if I will need some cuda svms I will do it standalone first22:04
blackburnhmm I won't do svms I guess22:04
@sonney2kso I would continue to parallelize other code for multicore before attempting this22:04
blackburnbut may be something useful22:04
@sonney2kheh22:05
blackburnand then merge if possibel22:05
--- Log closed Fri Sep 23 00:00:18 2011