Open in new window / Try shogun cloud
--- Log opened Thu Jul 19 00:00:17 2012
-!- heiko [] has left #shogun []00:09
wikingsonney2k: noup01:05
wikingsonney2k: just got stuck with one part of the code01:05
-!- romi_ [] has quit [Remote host closed the connection]01:16
-!- puffin444 [62e3926e@gateway/web/freenode/ip.] has quit [Quit: Page closed]01:16
blackburnwiking: here?01:21
wikingblackburn: yes01:21
blackburnwiking: I need some kind of help with QP task01:21
blackburnproblem is \min_s \sum_i (s_i - s*_i)01:22
blackburns.t. \sum_i s_i = k01:22
blackburn0 <= s_i <= 101:22
wikingu need a solver for that?01:22
blackburnI am confused01:22
blackburnis it a QP task?01:22
blackburnargh ()^201:23
blackburnnot ()01:23
wikingah ok01:23
wiking\min_s \sum_i (s_i - s*_i)^201:23
blackburnQP task is x^T H x + f x01:23
blackburnso what is H and f here?01:23
blackburnif it is a qp01:23
wikingu mean u wanna solve this with libqp?01:23
blackburnyeah fits perfectly exepct H part01:24
blackburnI actually even have mosek calling code for that01:24
wikingfor which?01:25
blackburnfor solution of this problem01:25
wikingaaah and you wanna have it instead with libqp?01:26
blackburnokay actually it is (x-y)*I*(x-y)^T, right?01:26
blackburnyeah would be cool to avoid mosek01:26
wikingonly question is what is s*_i?01:28
blackburnit is given01:28
wikingah cool01:28
wikingso you only have s as a vector01:28
blackburneverything is known01:28
wikingand another s* vector01:28
blackburnexcept s01:28
wikingand wanna minimize your s*01:28
wikingi mean s01:28
wikingok get is01:29
wikingget it01:29
blackburnthe other option is to use nlopt but I am confused whether it can be considered as qp01:29
wikingx = s01:30
wikingx^T = s*01:30
wikingand define Q that it creates s-s*01:30
wikingonly question here with the ^2 part01:31
wikingbut the problem is that in qp definition01:31
wikingyou have actually xHx^t01:32
wikingso that x is the same01:32
blackburnhmm wait01:33
blackburnwhat if x = s - s*01:33
blackburnthen H = identity01:33
blackburn\sum_i x_i = k - \sum_i s*_i01:34
wikingu can do that01:34
wikingH is positive semi-def in that way01:34
blackburns*_i < x_i < 1+s*_i01:34
blackburndoes it make sense to you?01:34
wikingit was 0 < s_i < 101:35
wikingtry it01:36
blackburnlooks legit01:36
blackburnwiking: funny thing is that problem size is usually <100 here01:44
blackburnbut I just don't know how to work with mosek :D01:45
wikingi dont approve mosek01:54
blackburnbirds are like sponges for oil :D01:57
-!- blackburn [~blackburn@] has left #shogun []03:04
-!- ckwidmer [45c9b18c@gateway/web/freenode/ip.] has joined #shogun03:46
-!- ckwidmer [45c9b18c@gateway/web/freenode/ip.] has quit [Client Quit]03:47
shogun-buildbotbuild #33 of nightly_default is complete: Success [build successful]  Build details are at
shogun-buildbotbuild #23 of nightly_all is complete: Success [build successful]  Build details are at
-!- splade [~chatzilla@] has joined #shogun06:33
-!- gsomix [~gsomix@] has quit [Ping timeout: 248 seconds]07:29
-!- pluskid [~pluskid@] has joined #shogun08:12
-!- gsomix [] has joined #shogun08:46
-!- splade [~chatzilla@] has quit [Ping timeout: 252 seconds]08:48
-!- uricamic [~uricamic@2001:718:2:1634:7de6:7891:7fa4:5d68] has joined #shogun08:48
-!- splade [~splade@] has joined #shogun08:48
-!- gsomix [] has quit [Read error: Connection reset by peer]08:53
-!- gsomix [] has joined #shogun08:54
gsomixgood morning08:54
gsomixdroopy, nya08:54
-!- gsomix_ [] has joined #shogun09:06
-!- gsomix [] has quit [Read error: Connection reset by peer]09:06
-!- gsomix_ is now known as gsomix09:32
-!- pluskid [~pluskid@] has quit [Ping timeout: 248 seconds]09:40
-!- pluskid [~pluskid@] has joined #shogun09:40
-!- pluskid [~pluskid@] has quit [Ping timeout: 276 seconds]10:15
-!- pluskid [~pluskid@] has joined #shogun10:35
-!- blackburn [~blackburn@] has joined #shogun11:29
-!- gsomix [] has quit [Remote host closed the connection]11:31
blackburnuricamic: hey, around?12:08
-!- splade [~splade@] has left #shogun []12:18
uricamicblackburn: yep12:39
blackburnuricamic: I have some struggles with libqp12:39
uricamicgo on12:40
blackburnI am solving following problem12:40
blackburn\min_s \sum_i (s_i - s*_i)^2,      s.t.     \sum_i s_i = k,    0 <= s_i <= 112:40
blackburns* and k are given12:40
blackburnso I introduced x = s - s*12:42
blackburnthen the problem became12:42
blackburn\min_x \sum_i x_i^2       s.t.     \sum_i x_i = k - \sum_i s*_i12:42
blackburn0-s*_i <= x <= 1-s*_i12:43
blackburnuricamic: does it look ok?12:43
blackburnI am unsure such thing is allowed here12:43
uricamicit seems to be ok12:43
uricamicbut I have to look on it more carefully12:44
blackburnuricamic: main problem is that I got solution not in bounds with gsmo for some reason12:44
uricamicunfortunately Vojta is not here, so I can't ask him, but I will try to look on it more carefully12:46
blackburnit sets bounds12:46
blackburnargh bad example12:46
blackburnit looks like solution stays at initial point here :D12:46
blackburnuricamic: does it depend on initial guess?12:47
uricamicnot sure, what does the return value says?12:47
blackburn1 iteration, KKT12:47
blackburnobjective = 0.012:47
blackburnlb = -0.512:49
uricamicI mean the exitflag12:49
blackburnub = 0.512:49
blackburnexitflag = 412:49
blackburnb = 1.012:49
blackburnbut solution is x = [0,0]12:49
uricamicI see, I am not sure if that exitflag is good12:50
uricamicbut I have been just using libqp_splx only so far12:50
blackburnuricamic: well it is the only non-emergency code here12:51
blackburnargh no wrong12:51
blackburn-1 is no memory available12:51
blackburn0 is max iterations code12:51
blackburnand 4 is KKT satisfied code12:51
uricamicI see, so it should be ok, but it is strange though12:52
uricamicin gsmo the exitflag cannot be 1, 2 or 3, tight?12:53
blackburnyeah it seems so12:53
uricamicin that case 4 should be ok12:53
blackburnI set initial x to [10,10]12:54
blackburnand it did nothing with it actually12:54
blackburnI think there is some illformed vector here12:55
uricamicI am looking on the webpage of libqp now, and it seems that your constraints are not ok for it now12:55
uricamicthere is just x^T a = b12:56
blackburnyeah but what is wrong with it?12:56
blackburnI set a to ones12:56
uricamicbut in your task you said that constraints are sum_i x_i = k - sum_i s^*_i12:57
blackburnbut s* is given12:57
uricamicI see, ok12:58
blackburnso I just b = k - sum_i s*_i12:58
uricamicok, in that case it should be ok12:58
uricamichave you tried without using initial solution?12:58
blackburnhow to do that?12:59
blackburnI just tried x = [1000 , 1000] and libqp made it a solution.. strange12:59
uricamicit depends also how have you set the TolKKT parameter13:00
blackburnin doc here13:00
blackburnI set it to 1e-613:00
uricamicand of course on the initial solution provided13:00
blackburnit is stated here that x 'must be feasible'13:00
blackburndoes it mean initial solution should satisfy bound?13:01
uricamicsure, I guess13:03
blackburnshould it satisfy a'x=b then?13:03
blackburnuh that is another optimization problem then :D13:05
droopyblackburn, ^_^13:05
blackburnuricamic: but seriously - how to do that initial guess13:07
-!- yoo [575b08cb@gateway/web/freenode/ip.] has joined #shogun13:07
-!- yoo [575b08cb@gateway/web/freenode/ip.] has quit [Client Quit]13:07
blackburnhmmm apparently upper bound would work in that case it seems13:08
blackburnanyway, thanks uricamic13:10
uricamicno problem blackburn13:11
CIA-18shogun: Sergey Lisitsyn master * r759fdb7 / (17 files in 4 dirs): Refactored MALSAR-based multitask algorithms and added clustered multitask logistic regression -
-!- blackburn [~blackburn@] has left #shogun []13:22
wikinguricamic: here?14:34
-!- pluskid [~pluskid@] has quit [Ping timeout: 248 seconds]15:07
-!- pluskid [~pluskid@] has joined #shogun15:08
-!- ermutarra [55578d64@gateway/web/freenode/ip.] has joined #shogun15:12
-!- ermutarra [55578d64@gateway/web/freenode/ip.] has left #shogun []15:14
-!- pluskid [~pluskid@] has quit [Read error: Connection reset by peer]15:19
-!- uricamic [~uricamic@2001:718:2:1634:7de6:7891:7fa4:5d68] has quit [Quit: Leaving.]16:16
-!- heiko [] has joined #shogun16:17
-!- gsomix [~gsomix@] has joined #shogun17:04
gsomixgood evening17:04
heikosonney2k, are you around?18:03
-!- os252 [] has joined #shogun18:36
-!- heiko [] has quit [Quit: Leaving.]19:04
-!- blackburn [~blackburn@] has joined #shogun19:34
-!- blackburn [~blackburn@] has quit [Quit: Leaving.]19:54
-!- puffin444 [a6f986d8@gateway/web/freenode/ip.] has joined #shogun20:23
puffin444hello os25220:23
os252hi puffin44420:24
puffin444how are things?20:24
os252ok, still little time due to family matters. Sorry for having been difficult to reach.20:25
os252How are things on your end?20:25
puffin444that's okay20:25
os252Fortunately you are on top of everything, so I am not really needed ;-)20:26
puffin444good finally done with the gradient stuff20:26
os252Nice,so including all the ARD variants ?20:26
puffin444now working on some vis. demos and fitc20:26
puffin444yes the ard works nicely20:26
os252its separate kernels?20:27
puffin444there are ard linear and gaussian kernels20:27
os252yeah, I guess that's all for nwo.20:27
os252I was wondering whether it's worth porting the matern kernel.20:27
puffin444the weights can be learned thtough gradient search20:27
os252It is a bit rougher and not quite as smooth as Gaussian.20:27
os252Statisticians like it...20:27
os252and has some good properties.20:28
os252But I'd put somewhere further down on the list - there are probably more exciting things than implementing exotic kernels :-).20:28
puffin444interesting. never heard of that kernel20:28
os252The periodic kernels are also sort of cool.20:28
os252If you want to get inspiration:20:29
os252(4.31) is a cool kernel.20:29
os252It basically is like a squared exponential with prior knowledge that the function you are regressing is periodic.20:30
os252It\s cool to do spectral decomposition form a just a handful of points;you can optimize the hyper parameter that corresponds to the period length and read of what's happening.20:30
puffin444hmm... I was hoping to implement fitc/laplacian first if thats okay20:31
os252Could be means to get good demos - but you really decide what you like, there is infinite freedom to put some cool demos together.20:31
-!- blackburn [~blackburn@] has joined #shogun20:32
os252Totally. I just thought I through up some ideas for demos that illustrate what you can do with GPs and cannot easily with SVMs etc.20:32
puffin444that way adding exotic kernels/likelihoods would be an easy task20:32
puffin444yes i am looking at gpml for inspirarion for the denos20:32
os252Agree - and as I already pushed for the ARD stuff, let's defer that one for later. But take a look at that stuff.20:33
puffin444i am hoping to do something similar via matploylib20:33
os252What have you got so far?20:33
puffin444frankly i dont have demos yet, been looking at fitc20:34
os252Yes, I understand - more fun to do.20:35
puffin444but i plan to have at least some simple demos by the rnd of the week20:35
os252I have some old code for sausage plots in matplotlib20:35
os252some super simple simulation, but it's similar to things in gpml20:35
os252is here20:35
os252As suggested before. I'd vote for simple demos now and more complex things at the end of all this.20:36
os252Have you read the unfitting sparse GP approximations paer?20:36
puffin444yes these might be quite useful20:36
os252I think it will be very helpful to get an efficient implementation of a general framework for sparse approximations (including fits).20:37
puffin444yes i have looked at it briefly a fee times20:37
puffin444few* lol20:37
os252It may make sense to discuss this if you like.20:38
os252The approximations are certainly non-trivial.20:38
os252and take a bit of time to get into.20:38
puffin444there was a paper i found that compared many sparse methods, but i think its safest now to look at fitc20:38
os252yes, there are several comparisons. FITc is definitely high up in the ranking.20:39
puffin444as i can compare it directly to gpml20:39
os252Yeah, it's brill that they have this all implemented :-)20:40
puffin444lots of potential math error implementing this20:40
os252But let me know if you want to discuss. The initialization of the pseudo points etc. can be a bit of a tricky business.20:40
os252Another question is how to handle the pseudo points.20:40
puffin444yes its wonderful they have it in gpml20:40
os252Effectively they are parameters like everything else in the covariance.20:41
os252The question is how to best incorporate this into your framework.20:41
os252Definitely. That's why C++ is not the best prototyping language (at least for me).20:41
-!- puffin444_ [a6f986d8@gateway/web/freenode/ip.] has joined #shogun20:43
puffin444_yes it is something that i have been thinking about20:44
-!- puffin444 [a6f986d8@gateway/web/freenode/ip.] has quit [Ping timeout: 245 seconds]20:45
os252B.t.w Sergey has switched shogun to Eigen, right?20:45
os252You could consider coding up the maths for this stuff in Eigen.20:45
os252It will be a lot less pain than native blas/lapack calls.20:45
puffin444_yes that would be wonderful20:45
os252I am not aware of the latest state of the play but this: suggests it's in there yet.20:46
blackburnos252: moreover I am already using it for my algorithms20:46
os252I think Eigen is superb in particular if you are porting code from matlaby/python like scripting languages. You can almost copy & paste thins.20:46
os252blackburn: are the libraries copied into shogun?20:47
os252or is there an external dependency?20:47
blackburnos252: no, we rely on dependency20:47
blackburnit can be a problem though20:47
os252is it optional or a requirement, i.e. can core code use Eigen ?20:47
puffin444_yes this might make things muxh rasier20:47
blackburnwhat is core?20:47
os252Well, not sure. GP support of course ;-)20:48
puffin444_well already gp only works with lapack installed20:48
blackburnI don't actually know whether we can make it mandatory20:49
puffin444_is there a performance difference with eigen20:49
blackburnyes, and performance of coding in particular20:49
os252Yeah, numerically it's bit faster than black/lapack, coding is orders of magnitude ;-)20:50
blackburnpuffin444_: it uses SSE stuff if you know what it is20:50
puffin444_oh wow, so do we keep lapack just because it runs on more pkayforms?20:50
os252I think they really tackled the last bottle necks. You can even now put Intel MKL underneath to get multi-core support for large matrices (which Eigen does not have yet).20:50
blackburnpuffin444_: we still have lapack code..20:51
os252but Eigen has lapak interfaces. Blackburn: have you explored whether those are sufficient for Shogun?20:51
blackburnos252: problem is that they are provided only in standalone packages20:52
blackburnso if you install it on say debian/ubuntu20:52
blackburnyou get no blas/lapack headers provided20:52
os252but otherwise they work?20:52
os252If yes: I am really wondering why not copy all of Eigen into shogun and kick all the lapas/black dependencies.20:52
blackburnlet me check20:53
os252Linalg is sufficiently "central" to justify to few kilobyte of third party code I guess.20:53
blackburnI am not big fan of integrating code actually20:53
blackburneigen3 is being developed pretty fast still20:54
os252It surely is a tradeoff - I completely understand.20:54
blackburnlapack and blas haven't changed for years20:55
os252But either way: if you add stuff to shogun that depends on Eigen I see now reason why puffin444 should code up sophisticated PG linalg in lapack.20:55
blackburnno just guard it with HAVE_EIGEN320:55
blackburnit is not a problem actually20:56
os252The fitc stuff is pretty complicated and it will take a lot of time to code this up in lapak otherwise.20:56
os252puffin444: have you used Eigen before? Keen on going down that route?20:57
puffin444_yes especially the derivatives i found out that  i have my work cut out for me20:57
blackburnos252: puffin444_: see shogun/lib/malsar folder - a few examples of eigen3 code are here20:57
os252Sorry. I was not aware that it's already sorted and ok to use Eigen in shogun; otherwise would have pointed this out earlier. Surely all the linalg you need is pretty messy...20:58
puffin444_Yes but it is workable20:59
os252I'd suggest to then use Eigen for the GP stuff. You can also mix and match as you see fit.20:59
puffin444_will do21:01
puffin444_os252, anything more you would like to discuss?21:02
blackburnactually you can gain pretty much with eigen321:02
blackburne.g. afaik lapack has no LDL'21:03
blackburnI remember you were using cholesky somewhere21:03
-!- cwidmer [8ca3fe9d@gateway/web/freenode/ip.] has joined #shogun21:03
blackburnoh famous cwidmer is in21:03
puffin444_yes actually everwhere :)21:03
blackburnpuffin444_: LDLT is something more stable IIRC21:04
blackburnworks for semi-definite21:04
os252Yes, Eigen has LDLT which is quite nice. It's also quite fast.21:05
os252I have some basic GP code for a very applied project in Eigen; I'd be happy to just send it. You could get some inspiration of Eigen for this purpose.21:05
puffin444_sure that would be great21:07
os252I'll just put it into an email and you'll see what you can use.21:07
os252ok, anything else we should discuss now?21:12
os252Definitely ping when you need input on the sparse stuff.21:12
puffin444_nothing else on my mind21:12
puffin444_will do21:12
os252Great. Then happy coding!21:13
puffin444_ok! :)21:13
-!- os252 [] has left #shogun []21:15
-!- os252 [] has joined #shogun21:15
@sonney2kwiking, so what is the problem / where are you stuck?21:24
@sonney2kgsomix, how is it going?21:24
gsomixsonney2k, working and reading numpy code.21:24
wikingsonney2k: so solver21:24
wikingsonney2k: need to do some redesigning there to be able to handle both primal mosek and the bmrm solver21:24
wikingsonney2k: had a chat today with uricamic about it, and we've agreed on some changes21:25
@sonney2kwiking, but you know what to do?21:26
wikingnow i'm trying it out21:26
@sonney2kok then21:27
@sonney2kgsomix, still about the slicing / dictionary?21:27
gsomixsonney2k, yes, this numpy code is not well commented.21:28
gsomixso, I'm doing experiments (as physicist hehe)21:28
-!- emrecelikten [~emre@] has joined #shogun21:32
-!- puffin444_ [a6f986d8@gateway/web/freenode/ip.] has left #shogun []21:37
@sonney2kgsomix, but can't you determine the types of the PyObjects21:37
@sonney2kwiking, is your PR ready to be merged or will it break things?21:38
@sonney2kI meant will it still compile and other (unrelated) stuff still work.21:39
wikingsonney2k: it compiled for me21:39
wikingso it shouldn't break anything21:39
@sonney2kwiking, ok then21:40
gsomixsonney2k, I can and I know that [a:b, c:d] in my_cool_features[bla, bla] is tuple21:40
@sonney2kwiking, btw could you figure out why the linking error occurs on *bsd21:40
wikingsonney2k: not yet21:40
@sonney2kgsomix, and the sub-elements?21:40
wikingsonney2k: that's a really weird thing21:40
wikingsonney2k: i think i have a conflict there... since most of the .so are compiled with cc and c++ (which is 4.2.x)21:41
gsomixsonney2k, hm, at know I can say that it is not array or sequence.21:41
wikingso that's why ldd shows that will depend on that other libstdc++21:41
@sonney2kwiking, hmmhh mixture of gcc 3.X and 4.X?21:46
wikingsonney2k: since we have to use 4.6 gcc on that machine because of that gcc bug21:50
wikingbut maybe i'm just doing something fishy there21:51
@sonney2kwiking, well can you explicitly force gcc 4.6 for the examples then?21:52
gsomixsonney2k, yes!21:53
@sonney2kgsomix, yes?!21:53
gsomixsub_elements are PySlices, huh.21:54
@sonney2kok then :)21:54
gsomixbut little deep in python docs.21:54
-!- rieck [] has quit [Quit: ZNC -]21:54
wikingsonney2k: retrying again something21:55
wikinggmake[3]: Leaving directory `/usr/home/wiking/shogun-buildbot/bsd1_-_libshogun/build/examples/undocumented/libshogun'22:01
wikingError expanding embedded variable.22:01
-!- rieck [~rieck@] has joined #shogun22:01
wikingso sonney2k22:01
wikingi don't think that this patch did the job22:01
wikingit's still calling make :(22:02
@sonney2kwiking, please replace the test -n with test -z in check.sh22:04
@sonney2kin the make line22:04
CIA-18shogun: Soeren Sonnenburg master * r96a80ed / examples/undocumented/libshogun/ : use test -z not -n to use configured make -
wikingi've did this22:05
wiking LIBRARY_PATH="LD_LIBRARY_PATH" LIB_PATH="//usr/local/lib"         INC_PATH="//usr/local/include" LIBS="-Xlinker --no-undefined  -lshogun" LD_LIBRARY_PATH=//usr/local/lib INCLUDES="-I. -I..  -I/usr/local/include/json   -I/usr/local/include/libxml2 -I/usr/local/include   -I../libshogun -I.." MAKE="gmake" ./check.sh22:05
wikingin i did an echo $MAke22:05
wikingi mean echo $MAKE22:05
wikingand it's make and not gmake22:05
wikingwhich is very weird22:05
@sonney2kwiking, well either git pull again or change the test -n "$MAKE" line22:09
@sonney2kto test -z22:09
wikingsonney2k: i'll wait now22:10
wikingsince it's building22:10
wikingand i was using the build dir of the bot to do the tests22:10
wiking3 more mins22:12
wikingsonney2k: is there a way to do gmake -j4 ?22:16
wikingor soemthing like this22:16
wikingsonney2k: ok so the problem is now the following22:17
wikingok let me check it for 100% before i start telling b.s.22:18
wikinghuh so first of all22:19
wikingnot the right c++ is used for building the libshogun examples22:19
@sonney2kwiking, set MAKEFLAGS="-j4"22:19
wikingsonney2k: i mean that u should change that in the buildbot22:20
wikingsince there's really free cpu there in the machine22:20
wikingin examples/undocumented/libshogun/Makefile22:20
wikingwhich will not be ok, since i've used c++46 for building the libshogun22:21
wikingso that should be set as well by ./configure script not just when building libshogun itself22:21
wikingbut the examples as well22:21
wikingbut even then i'm seeing some weird shit22:21
wikingLD_LIBRARY_PATH=/usr/home/wiking/shogun-buildbot/bsd1_-_libshogun/build/src/shogun ./basic_minimal22:22
wikingthis give me22:22
wiking/usr/lib/ version GLIBCXX_3.4.11 required by /usr/home/wiking/shogun-buildbot/bsd1_-_libshogun/build/src/shogun/
wikingi'm trying to find info about this22:23
wikingok i think i see wtf22:24
wikingjust a sec i'll try to experiment with some extra compiler flags22:25
wikingok i think the problem occures here when it's linking the libshogun.so22:28
wikingthat it doesn't link it to the right
-!- os252 [] has left #shogun []22:29
gsomix>> You probably do not want to use this function. | in python docs :)22:30
wikingwoah i've just seen in a printed book a screenshot of netscape navigator :DDD22:30
wikinggsomix: then u most defintely should use it :)22:31
droopycool!?! ;>22:31
gsomixdroopy, what do you think about higgs boson?22:32
droopyi'll re-pm you on this serv if pm is actually 'private'!?!22:32
* gsomix sighs22:33
CIA-18shogun: Soeren Sonnenburg master * rfa6ce12 / (2 files in 2 dirs): pass COMP_CPP through for compiling libshogun examples -
@sonney2kwiking, another attempt...22:39
wikingsonney2k: okey! will check22:39
wikingor wait till it builds22:39
@sonney2kyeah better that then22:40
@sonney2kwiking, btw do you have some nice graphics produced w/ the help of shogun?22:40
@sonney2kI am trying to create some kind of gallery22:40
@sonney2killustrating various applications/use cases/benchmarks whatever for the new website22:41
wikingsonney2k: will do it for you22:41
blackburnI have one idea but I need someone blender skilled :D22:41
droopygrin ^_^22:41
wikingsonney2k: i'll create some screenshots for the latent svm22:41
wikingsonney2k: but we can even for fun make a webservice :)22:42
@sonney2kblackburn, heh22:42
@sonney2kwiking, you mean with security audited code like shogun's :D22:42
wikingsonney2k: let say there's a cat regonizer :) and people can just upload any images to see if it locates the cat (if there's one) in the pic22:42
@sonney2krieck, woudl cry :D22:42
wikingsonney2k: DDDD22:42
wikingsonney2k: :>22:42
wikingwell then just screenshots22:42
@sonney2kwiking, well ok that is certainly safe22:43
@sonney2kwell to some extend22:43
wikingblackburn: :>>>>>22:43
@sonney2k(bugs in image parsing lib)22:43
@sonney2kI think such examples are really highly needed22:43
@sonney2kwhat shogun can do is too abstract for everyone...22:44
@sonney2kexcept the authors :D22:44
wikingok i think i'll need to add this flag to the linker22:44
blackburnI still convinced shogun is used by ten guys22:44
wikingbut i'll test this now22:44
wikingblackburn: :DDD22:44
blackburnworst thing is that  we have outdated version in ubuntu/debian22:45
wikingblackburn: dude i think the train set is down! :D22:50
blackburnwiking: what do you mean?22:50
wikingit downloaded :)))22:50
wikingthat 135 gigs22:50
blackburnokay now i have no idea what to do with it22:51
blackburnI mean RealFeatures(traindat) wouldn't work here LOL22:51
@sonney2kblackburn, no just I use shogun - you all only work for me *eg*22:55
blackburnsonney2k: yeah I tend to believe ;)22:55
wikingblackburn: well now i'll upload the whole thing into the hdfs on the cluster22:59
wikingand then we can run on it feature extractors22:59
wikingyou've said you'll create a c based hog extractor23:00
blackburnaham I have to hurry up with hogs23:00
wikingi'll port that retina or what the fuck was that descriptr :)23:00
blackburnyeah FREAK23:00
blackburnI didn't call you freak :D23:01
wikingthat's the one23:01
wikingsonney2k: hmmm did that run well with other builders?23:02
wikingi mean libshogun build this time23:02
wikingsonney2k: almost... ;)23:07
shogun-buildbotbuild #176 of deb3 - modular_interfaces is complete: Failure [failed test libshogun]  Build details are at  blamelist: Soeren Sonnenburg <>23:07
wikingerror in gmake basic_minimal CC=c++46 LIBRARY_PATH=LD_LIBRARY_PATH LIB_PATH=//usr/home/wiking/shogun-buildbot/bsd1_-_libshogun/build/src/build/lib INC_PATH=//usr/home/wiking/shogun-buildbot/bsd1_-_libshogun/build/src/build/include LIBS=-Xlinker --no-undefined  -lshogun23:07
wikingerror in ./basic_minimal23:07
wikinggmake: unrecognized option `--no-undefined'23:07
wikingsonney2k: oh ok it's the same for you apparently23:08
shogun-buildbotbuild #152 of deb2 - static_interfaces is complete: Failure [failed test libshogun]  Build details are at  blamelist: Soeren Sonnenburg <>23:08
@sonney2kwiking, grmpf23:15
CIA-18shogun: Soeren Sonnenburg master * r2442926 / examples/undocumented/libshogun/ : escape flags -
shogun-buildbotbuild #153 of deb2 - static_interfaces is complete: Success [build successful]  Build details are at
blackburnwiking: hmm it looks like my substitution we discussed last night was wrong23:24
blackburnsonney2k: are you busy with that building stuff?23:27
blackburnwiking: probably I do not cover negative and positive cases when I transform s.t. part23:30
gsomixsonney2k, ta dam!23:33
rieck /msg *status loadmod colloquy23:37
wikingsonney2k: ok23:53
shogun-buildbotbuild #177 of deb3 - modular_interfaces is complete: Success [build successful]  Build details are at
wikingsonney2k: i have the solution... but i really don't know how to add this into the configure script23:54
wikingso basically i would need to give this compiler flag: -Wl,-rpath,/usr/local/lib/gcc4623:54
wikingi'll try to see how's this solved usually23:55
wikingwith the configure script23:55
wikingbecause now for some reason when c++46 is doing the linking23:56
wikingit uses the standard /usr/lib/ path for the libgcc and libstdc++ libs23:56
wikingand that'll break of course23:56
wikingsince it's for c++ (which is 4.2..)23:56
wikingthis is the only thing left to be solved and then we can enable libshogun tests on freebsd23:57
gsomixgood night guys23:59
@sonney2kwiking, yeah but that is wrong23:59
@sonney2kgsomix, tada?23:59
gsomixsonney2k, ta dam.23:59
@sonney2kgsomix, cool23:59
@sonney2kPR'd alreadY?23:59
gsomixsonney2k, PR will be tomorrow.23:59
--- Log closed Fri Jul 20 00:00:05 2012