Open in new window / Try shogun cloud
--- Log opened Tue Jan 28 00:00:13 2014
-!- thoralf [] has quit [Ping timeout: 272 seconds]00:49
-!- zxtx [] has quit [Ping timeout: 260 seconds]06:29
-!- zxtx [] has joined #shogun07:12
-!- lisitsyn [~lisitsyn@] has quit [Read error: Operation timed out]09:14
-!- benibadman [~benibadma@] has joined #shogun09:15
-!- benibadm_ [~benibadma@] has joined #shogun09:41
-!- benibadman [~benibadma@] has quit [Ping timeout: 245 seconds]09:45
-!- lisitsyn [] has joined #shogun10:00
@wikingsonney2k: lisitsyn: have u seen yesterday the conversation with naywhayare10:02
@wikingabout their benchmarking system?10:02
lisitsynwiking: no let me read logs10:03
@wikinglisitsyn: kk10:03
lisitsynwiking: oh it is pretty big ;)10:10
lisitsyntl;dr :D10:10
lisitsynwiking: jk, what is the important point here?10:10
@wikinglisitsyn: basically that they've developed some sort benchmarking system10:11
@wikingand i've suggested that they should just put that into a github repo10:11
@wikingso that we can start hacking it as well10:11
lisitsynwiking: yeah that's opensource :D10:11
lisitsynwiking: we should definitely share libcovertree10:11
@wikingyeah sure but now with that svn repo10:11
lisitsynwhich does not exist yet10:11
@wikingit's not so easy to hack...10:12
lisitsynbut jl is a kung fu guy of unsupported code10:12
@wikingi mean contribute10:12
lisitsynyeah github is much better in any sense10:12
@wikinglibcovertree ?10:12
lisitsynwiking: :D10:12
lisitsynwiking: sorry that's another point10:12
lisitsynwiking: so I meant10:13
lisitsynjohn langford and his covertree10:13
lisitsynwhich naywhayare hates a bit10:13
lisitsynand actually I tried THRICE to patch it10:13
lisitsynwiking: yeah10:13
lisitsynwiking: we have his code in our repo as well10:13
lisitsynbut naywhayare claims it is buggy10:14
lisitsynwiking: some shogun/lib/ I guess10:14
lisitsynwiking: yeah that too10:16
@wikingah somewhere else as well? :D10:17
lisitsynand this one10:17
@wikingah src/shogun/lib/JLCoverTree.h10:17
@wikingwhy twice? :D10:17
lisitsynwiking: they are a bit different somehow10:17
lisitsynbut that's to be solved as well10:18
@wikingwell ok that's another story10:18
lisitsynwiking: what I meant is10:18
lisitsynwe both need fast covertree10:18
lisitsynI mean shogun and mlpack10:18
lisitsynso I'd push push push some shared thing10:19
@wikingsounds reasonable10:19
lisitsynit would be stupid to have two implementations of covertrees10:19
@wikingwell u know10:19
@wikingthere's like millions of reimplementations :)10:19
@wikingin shogun ::PP10:19
lisitsynyeah but no good covertree10:19
-!- Saurabh7 [~Saurabh7@] has joined #shogun11:47
-!- votjakovr [~votjakovr@] has joined #shogun12:12
-!- Saurabh7 [~Saurabh7@] has quit [Quit: Leaving]12:18
-!- HeikoS1 [] has joined #shogun12:31
votjakovrHeikoS1: hi! i'm glad to see you :) how is it going with kernel->clone() issue?12:50
HeikoS1votjakovr: hi! good to see you too12:50
HeikoS1ah yeah, I will do that soonish12:51
HeikoS1currently some other things are very pressing regarding next gsoc12:51
HeikoS1votjakovr: how is it going for you=?12:51
HeikoS1votjakovr: thanks for you patch btw!12:52
HeikoS1votjakovr: good work as usual, ... speaking of how are the string examples going? did you find any time for thoise?12:53
votjakovrHeikoS1: Oh, yeah :) Btw I'm trying to use GPs with preprocessed features and find a bug in parameters hashing framework12:53
HeikoS1votjakovr: ah bugs bugs ;)12:54
HeikoS1cool, preprcessed features would be nice12:55
HeikoS1I am currently working on categorical data12:55
HeikoS1with diffusion kernels12:55
HeikoS1votjakovr: and btw also using lots of your GPC code for the experiments, it works really nicely12:56
votjakovrHeikoS1: cool, but i'd like to speed up them12:56
HeikoS1votjakovr: that would be very helpful indeed12:56
HeikoS1votjakovr: did you think about this shallow copying of the kernel?12:57
votjakovrHeikoS1: nope, i'm waiting when you'll try to fix kernel->clone() issue :)12:59
HeikoS1votjakovr: I see13:00
HeikoS1so I will do that next13:00
votjakovrHeikoS1: thanks :)13:00
HeikoS1votjakovr: btw what are your plans for the summer?13:04
votjakovrHeikoS1: i still don't have plans13:07
HeikoS1votjakovr: are you interested in doing another gsoc?13:07
votjakovrHeikoS1: i'm not a student unfortunately13:08
HeikoS1votjakovr: ah too bad13:08
HeikoS1votjakovr: do you want to stay in university or more into proper business world?13:09
HeikoS1do you even want to stay in Russia?13:09
HeikoS1votjakovr: sorry, questions over questions!13:09
HeikoS1votjakovr: in any case, you should stay involved in shogun! :)13:09
votjakovrHeikoS1: i'm pretty flexible to change my location, but now i'm in Russia13:12
HeikoS1ever thought about doing msc or phd in machine learning?13:13
votjakovrHeikoS1: actually nope :)13:13
votjakovrHeikoS1: but it would be cool, i think13:13
HeikoS1votjakovr: so here in europe at least, that would be a quite good match to your skills, just to let you know13:14
HeikoS1I am sure many people would appreciate that13:14
votjakovrHeikoS1: Yeah. Btw I think my patch may be merged13:45
lisitsynvotjakovr: still in saint p? ;)14:23
votjakovrlisitsyn: yeah14:26
-!- zxtx [] has quit [Ping timeout: 260 seconds]16:28
-!- lisitsyn [] has quit [Quit: Leaving.]16:29
-!- shogun-notifier- [] has joined #shogun16:54
shogun-notifier-shogun: Roman Votyakov :develop * 0e35508 / / (7 files):
shogun-notifier-shogun: various minor fixes in GPs framework16:54
shogun-notifier-shogun: Heiko Strathmann :develop * e94c133 / / (7 files):
shogun-notifier-shogun: Merge pull request #1829 from votjakovr/feature/gp_refactoring16:54
shogun-notifier-shogun: Various minor fixes in GPs framework16:54
-!- zxtx [] has joined #shogun17:14
shogun-buildbotbuild #47 of deb4 - python3 is complete: Failure [failed test python modular]  Build details are at  blamelist: Heiko Strathmann <>, Roman Votyakov <>17:26
shogun-buildbotbuild #410 of precise - libshogun is complete: Failure [failed test]  Build details are at  blamelist: Heiko Strathmann <>, Roman Votyakov <>17:38
-!- votjakovr [~votjakovr@] has quit [Quit: WeeChat 0.4.0]17:47
-!- travis-ci [] has joined #shogun17:53
travis-ci[travis-ci] it's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun:
-!- travis-ci [] has left #shogun []17:53
-!- benibadm_ [~benibadma@] has quit [Ping timeout: 252 seconds]17:58
-!- lisitsyn [~lisitsyn@] has joined #shogun18:18
-!- thoralf [] has joined #shogun19:18
-!- HeikoS1 [] has quit [Ping timeout: 272 seconds]19:31
-!- shogun-notifier- [] has quit [Quit: transmission timeout]19:55
-!- zxtx [] has quit [Ping timeout: 272 seconds]20:32
-!- thoralf [] has quit [Read error: Operation timed out]20:50
-!- thoralf [] has joined #shogun21:26
-!- zxtx [] has joined #shogun22:27
-!- FSCV [~FSCV@] has joined #shogun22:39
thoralfwiking: I found a case NCBM gets stuck due to numerical issues... and I discovered a few places where NaNs occur and result in non-terminating QP solver.  (e.g. in the line search)22:43
thoralfInterested in this or are you already over it? ;)22:43
@wikingjust send the issue23:27
@wikingand i'll give it a look23:28
thoralfCool.  On my way.23:36
thoralfThe numerical stuff is at least indirectly related to the problem.23:52
thoralfHot-fixing only a few of the points turns out to solve the problem at least for my example.23:53
thoralfBut what makes me more thoughtful is that your code creates something that cannot be solved by our QP solver.23:54
@wikingwhat do u mean?23:56
@wikingor what exactly do u mean about 'creates something that cannot be solved by our QP solver.'?23:56
thoralfThe qp solver does not converge.23:56
thoralfIt simply got stuck (not the first time though)23:56
@wikingyou mean ncbm gets stuck23:57
thoralfProbably NaNs in the OP.23:57
@wikingwill try to check23:57
@wikinghave a new qp solver idea23:57
@wikingbut no time to do it :<23:57
thoralfVarying parameters in the problem yields *either* to non-converging qp solver or forever iterating ncbm.23:58
@wikingbtw is your problem concave?23:58
@wikingi.e. non-convex23:58
thoralfIt's a toy example.  Maybe.23:58
thoralfOr, probably.23:58
thoralfNever thought of, but good point.23:59
@wikingsince ncbm should be used mostly for non-convex problems23:59
@wikingi mean that's what it's for actually23:59
@wikingalthough that shouldn't mean that it cannot solve a standard convex problem23:59
--- Log closed Wed Jan 29 00:00:08 2014