Open in new window / Try shogun cloud
--- Log opened Tue Mar 19 00:00:14 2013
-!- zxtx [zv@conference/pycon/x-rnauhrcjrxankcvr] has quit [Ping timeout: 260 seconds]00:07
-!- zxtx [zv@conference/pycon/x-xhgwzhjusmxtgavv] has joined #shogun00:08
shogun-buildbot_build #366 of rpm1 - libshogun is complete: Failure [failed test]  Build details are at  blamelist: Soeren Sonnenburg <>00:09
-!- zxtx [zv@conference/pycon/x-xhgwzhjusmxtgavv] has quit [Ping timeout: 255 seconds]00:27
-!- zxtx [zv@conference/pycon/x-ujcvrjcchgjuzfif] has joined #shogun00:46
-!- shogun-notifier- [] has quit [Quit: transmission timeout]01:05
-!- zxtx [zv@conference/pycon/x-ujcvrjcchgjuzfif] has quit [Ping timeout: 240 seconds]01:19
-!- zxtx [zv@conference/pycon/x-hlwmyxstkbabctqs] has joined #shogun01:25
-!- zxtx [zv@conference/pycon/x-hlwmyxstkbabctqs] has quit [Ping timeout: 252 seconds]02:01
-!- sumit_ [73f91219@gateway/web/freenode/ip.] has joined #shogun04:53
-!- sumit_ [73f91219@gateway/web/freenode/ip.] has quit [Ping timeout: 245 seconds]05:16
@sonney2kshogun-buildbot_, force build 'rpm1 - libshogun'05:48
shogun-buildbot_build forced [ETA 5m02s]05:48
shogun-buildbot_I'll give a shout when the build finishes05:48
shogun-buildbot_build #367 of rpm1 - libshogun is complete: Success [build successful]  Build details are at
-!- hoijui [] has joined #shogun07:41
-!- zxtx [] has joined #shogun08:05
-!- blackburn [~blackburn@] has quit [Quit: Leaving.]08:47
-!- blackburn [] has joined #shogun09:14
sonne|workblackburn: look
sonne|workwe've got some press coverage09:45
-!- blackburn [] has quit [Ping timeout: 264 seconds]09:45
-!- blackburn [] has joined #shogun09:46
-!- n4nd0 [] has joined #shogun09:53
sonne|workblackburn: ok again10:09
sonne|workblackburn, n4nd0 look we've got some press coverage
n4nd0sonne|work: awesome10:09
blackburnsonne|work: heh cool10:10
n4nd0sonne|work: you are famous10:10
sonne|worknow if we had fancy applications on the website (d3js ...) then these would be on...10:10
n4nd0sonne|work: blackburn and I are getting something with d3js in tapkee10:13
n4nd0we could put it to in shogun if you like it10:13
sonne|workbtw they also announced our workshop...
blackburnsonne|work: n4nd0 is going to be a d3 expert soon as he prepares d3.js demo for tapkee :)10:13
blackburnsonne|work: preparing DR idea description10:13
n4nd0I didn't get anything from Christoph yet :S10:13
blackburnwho is Christoph?10:14
n4nd0I will write him again soon telling him about what I think it could be done in a project and ask him if he would be interested10:14
n4nd0blackburn: Christoph Lampert10:14
sonne|workn4nd0: please do it today10:14
sonne|worktime is running...10:14
n4nd0I contacted him and asked him if he would like to mentor a project in Structured Learning10:14
n4nd0sonne|work: ok10:15
blackburnn4nd0: ah I see10:15
n4nd0blackburn: he has done nice things in computer vision with structured learning10:15
sonne|workn4nd0: maybe it helps if you CC me...10:15
n4nd0sonne|work: good idea10:15
n4nd0blackburn: with CRFs and PGMs in general10:15
n4nd0this is a very nice tutorial10:16
sonne|workblackburn: btw I am extremely happy that we managed to get a new release out before gsoc :)10:18
sonne|workthis and that change in development model / tests should make the whole process go much smoother10:19
blackburnsonne|work: yes that's cool10:23
sonne|workblackburn: I've had a phone chat with alex yesterday10:24
sonne|workhe is working out the 2 days post the shogun workshop10:24
sonne|workit is likely to work out well.. so we can probably announce things end of the week10:25
blackburnsonne|work: alright, cool!10:25
n4nd0SWIG question10:34
n4nd0I am making an instance of RealMatrixFeatures in python10:34
n4nd0and giving it to a StructuredModel10:34
n4nd0the member of the StructuredModel is of class Features, parent of RealMatrixFeatures10:35
n4nd0I am trying to call get_features_vector() but it doesn't find it in Features - it doesn't exist there10:35
n4nd0but it does exist in RealMatrixFeatures10:35
n4nd0I am guessing that it either requires some SWIG code or an obtain_from_generic method to cast in python Features to RealMatrixFeatures10:36
blackburnsonne|work: I shared a doc with ideas - please send my idea around when you have time10:37
sonne|workn4nd0: exactly10:38
sonne|workblackburn: just do this directly on the website...10:38
blackburnsonne|work: lets keep that in doc yet and publish in a batch mode10:38
n4nd0sonne|work: what is prefereable, obtain_from_generic or swig code?10:39
blackburnthe thing that worries me is that I'd like to get my idea done by myself LOL10:39
n4nd0blackburn: with the info that Georg wrote we can probably write in there the description10:39
blackburnn4nd0: yes I'll put it in10:39
n4nd0blackburn: haha :) it is a nice description10:40
sonne|workblackburn: that should not worry you - to the contrary. that will help massively - you might even join the student in coding stuff...10:41
blackburnsonne|work: prerequisites: C++, metagenomics10:43
blackburnsonne|work: alright 2 ideas ready10:54
sonne|workblackburn: put them live and then we can send around another call!10:55
blackburnsonne|work: can I put some CI stuff for tapkee on one of our buildbots?10:58
blackburnsonne|work: btw I learned how to use gcov so will put it for shogun soon too10:59
blackburnsonne|work: tapkee offers me a possibility to play as it compiles just in a few minutes, not hour :D11:00
sonne|workblackburn: what is wrong with travis for tapkee then?11:00
blackburnsonne|work: I can't put anything out of travis11:00
sonne|workwhat not? only issue why we don't put everything on travis is that shogun's stuff takes too long11:01
blackburnsonne|work: how can I push generated lcov output somewhere?11:02
blackburnsonne|work: I generate stuff (valgrind output, etc), how can I get it back from travis?11:04
sonne|workno idea but how would you get it back from shogun's bb's?11:05
blackburnsonne|work: I can store some pass there may be11:07
blackburnokay I'll try to do that with travis then11:07
sonne|workblackburn: you mean ssh key and we rsync stuff? or put it in some folder that can be seen from the web?11:07
blackburnsonne|work: yeah may be ssh11:07
blackburnsonne|work: other way could be to update gh_pages folder11:08
blackburnsonne|work: webpage on github11:09
blackburnsonne|work: we could have page if we had gh_pages branch11:09
blackburnor shogun-toolbox.github.com11:10
blackburnsonne|work: just like that
blackburndjango editor drives me mad11:11
blackburnfscking wysiwyg11:12
sonne|workblackburn: it is not django editor11:13
blackburn sonne|work I put them live11:13
sonne|workit is tinymce11:13
sonne|workbern4rd added it for convenience11:13
blackburnsonne|work: I see11:14
blackburnsonne|work: anyway it works like windows11:14
blackburnin the worst mean of these words11:14
-!- hoijui [] has quit [Ping timeout: 255 seconds]11:19
n4nd0blackburn: it fails pretty bad lot of times11:20
n4nd0if happens to me sometimes that I click on "save and continue" and errors start popping up11:20
-!- hoijui_ [] has joined #shogun11:20
sonne|workn4nd0: which errors?11:23
n4nd0sonne|work: it couldn't be save, try again later11:24
n4nd0I ended up not using save and continue11:24
sonne|workI mean I don't mind to return to using just a raw text form11:24
sonne|workn4nd0: when did that occur and what do you have to do for that?11:24
n4nd0sonne|work: it has occurred when I have added some info in the webpage, for the workshop program for example that I remember now11:25
sonne|workn4nd0: speeking of which please add the date to the schedule :)11:26
n4nd0sonne|work: hehe sure11:27
sonne|workn4nd0: and if the problem occurs again tell it to us11:29
n4nd0sonne|work: error occurred :)11:30
sonne|workn4nd0: so what do you do exactly?11:30
n4nd0I edited on workshop2013_program "in July" to "on July 12th"11:30
n4nd0save and continue11:30
n4nd0that worked fine11:30
sonne|workand then?11:30
n4nd0although I was directed to the same page but with poor formatting11:31
n4nd0not the normal Django admin style page11:31
n4nd0then I clicked save11:31
n4nd0An internal server error occurred. Please try again later.11:31
sonne|workbut that is not the editor then11:32
sonne|workI wonder why we didn't receive an email with the exception11:32
n4nd0sonne|work: this is what I mean with page with poor formatting
n4nd0sonne|work: anyway the modification is fine11:33
sonne|workn4nd0: I am not so worried about the formatting (IIRC reloading would fix this) but more about the internal server error...11:36
blackburnsonne|work: how to say 'thanks, just in time' in german? :D11:57
blackburnI do not trust google translate11:57
sonne|workgerade noch rechtzeitig11:57
sonne|workbut this sounds a bit negative11:58
blackburnsonne|work: what about positive?11:58
blackburnI am sorry to distract you - got curious to answer andreas in german11:58
sonne|workno idea11:58
blackburnalright ;)11:59
n4nd0blackburn: andreas from scikits?12:05
blackburnn4nd0: no, Ziehe12:06
blackburnnot mueller :)12:06
sonne|workblackburn: ha pushing helps :)12:06
blackburnsonne|work: pushing?12:07
sonne|workblackburn: I've email andreas just an hour ago or so12:07
blackburnahh :)12:07
blackburnthey forgot I guess12:07
blackburnsonne|work: let me kick some asses too :D12:07
blackburnsonne|work: 3 ideas done12:18
-!- heiko [] has joined #shogun12:21
blackburnheiko: could you please fill in your gsoc idea?12:22
blackburnI shared a doc12:22
heikoblackburn: yes I saw it12:24
heikoI will do it soon12:25
heikoblackburn: and I will also help on the application12:25
heikoblackburn: can you update me on the deadlines?12:25
blackburnheiko: 29 march12:25
blackburnheiko: so 10 days, pretty enough12:25
heikoblackburn: ok, I will add some stuff tonight and continue to update this week12:26
heikoblackburn: what do you think on a project on GP classification?12:26
heiko(in case nobody wants to do the large sparse Gaussian stuff)12:27
heikocould do: binary GP with laplace approx and EP, same for multiclass, unit-testing the existing stuff   ....and thats already a nice little project12:28
blackburnheiko: mentor?12:32
heikoblackburn:  me12:32
heikoblackburn: I implemented most of this stuff end of last year, not all though, but its all covered by the GP book12:32
blackburnheiko: I never heard about these crazy letters so that's just a mentor availability12:33
heikoblackburn: also, there is no multiclass GP implementation out there that is easy to use12:33
heikothe book matlab toolbox doesnt contain multiclass12:33
heikoEP=expectation propagation12:33
blackburnGP, EP, laplace, go figure :D12:33
heikovariational optimisation12:33
blackburnheiko: jk12:33
heikohaha :)12:33
heikoGP=Grand prix12:34
heikomulticlass GP = mens/woman Grand prix12:34
blackburnheiko: ooh that's interesting!12:34
blackburnheiko: it is out of law to mix up men and women here though!12:34
heikoin Russia? :)12:35
blackburnheiko: well not really out of law but pretty pressing12:35
blackburnheiko: I'd kill myself if I was a gay in russia :D12:35
heikoblackburn: yeah, you dont like gay guys12:36
n4nd0any idea why trying to asign an numpy array of float64 to an SGVector<float64> fails in python_modular??12:36
heikon4nd0: it might be the form12:36
blackburnheiko: that's because we are barbarians living in 1600 :)12:36
n4nd0the error in particular is *** TypeError: in method 'ResultSet_psi_pred_set', argument 2 of type 'shogun::SGVector< float64_t > *'12:36
heikon4nd0: you  need [ [row 1] [row 2]]12:36
blackburnn4nd0: assign?12:36
n4nd0blackburn: =12:36
blackburnn4nd0: that won't work12:36
n4nd0heiko: it is a vector, just one row, or?12:36
heikoblackburn:  Germany were also barbarians12:36
heikon4nd0: ah sorry12:37
blackburnheiko: yes originally :D12:37
n4nd0blackburn: mmm why not? typemaps do not handle that?12:37
n4nd0heiko: no problem :)12:37
heikon4nd0: I had this problem quite often12:37
blackburnn4nd0: no, they handle only methods12:37
n4nd0blackburn: aham I see12:37
heikon4nd0: sometimes the type has to be set explicitly with dtype="float64"12:37
heikoI made this experience12:37
n4nd0heiko: yep, I tried doing that12:37
blackburnheiko: I am afraid assigning won't work with types12:37
heikoI am doing that all the type12:38
blackburnheiko: I think you misunderstood something! :)12:38
blackburnheiko: if x is sgvector12:38
blackburnx = numpy.array([3,4.0])12:38
blackburnthat won't work AFAIK12:38
heikoah I see12:38
blackburnI mean it would work but x now is an array12:38
heikoblackburn: I usually understand things wrong :D12:39
blackburnheiko: set_pizza(numpy.array([3.0,4.0]))12:39
n4nd0blackburn: I probably need to put a setter or the like to make it work then12:39
blackburnheiko: you meant that?12:39
heikoblackburn:  what I meant is I pass numpy arrays to methods that require an sgvector all the time12:39
n4nd0heiko: yeah12:39
blackburnn4nd0: yes12:39
n4nd0I thought also that meant you could do =12:39
n4nd0but it seems not :)12:39
blackburnn4nd0: you just assign it to new thing12:40
heikon4nd0: ah I see, so you have a sgvector and want to replace its content so that somebody else implicitly does something different12:40
n4nd0blackburn: the thing is since this is a field or an struct12:40
n4nd0it doesn't let me do it directly12:40
n4nd0field of* an struct12:41
blackburnn4nd0: could you provide an example?12:41
heikoblackburn:  whats the desired field in the gsoc?12:42
n4nd0struct CResultSet { SGVector<float64_t> psi; }12:42
n4nd0from python I am doing12:42
blackburnheiko: field?12:42
blackburnheiko: ahh12:42
blackburnheiko: I meant desired skills or so12:42
n4nd0result_set.psi = method_that_gives_numpy_array()12:42
blackburnheiko: feel free to update it with something better12:42
blackburnn4nd0: no, it needs a setter12:42
heikoand the difference to prerequisites?12:43
n4nd0blackburn: all right, thanks!12:43
blackburnheiko: prereqs are must, desired is would be better - I meant that12:43
n4nd0blackburn: hey13:38
blackburnn4nd0: hey13:38
n4nd0what about a director class with a method returning a numpy array13:38
n4nd0where the return value of that method in C++ is defined to be an SGVector13:38
n4nd0it fails for some reason13:38
n4nd0TypeError: SWIG director type mismatch in output value of type 'shogun::SGVector< float64_t >'13:39
sonne|workn4nd0: code signature?14:00
n4nd0virtual SGVector< float64_t > get_joint_feature_vector(int32_t feat_idx, CStructuredData* y);14:02
n4nd0sonne|work: ^14:02
n4nd0I don't really get why the typemap doesn't handle this14:03
sonne|workn4nd0: so you get back a SGVector?14:03
sonne|workin python I mean?14:03
n4nd0sonne|work: yes, SGVector is the return value14:03
n4nd0sonne|work: yeah, I am doing a StructuredModel in python using directors14:03
n4nd0sonne|work: and I am writing this get_joint_feature_vector method in python14:04
sonne|workn4nd0: well then use vec.get_vector()14:04
n4nd0sonne|work: in python? vec would be the numpy array?14:04
n4nd0mmm no, I guess no14:05
sonne|workn4nd0: no vec would be SGVector14:05
sonne|workand then vec.get()14:05
sonne|workit will give you a numpy translated SGVector14:05
sonne|workn4nd0: but note that you can do things like vec +=1 IIRC14:05
n4nd0sonne|work: but how can I get and SGVector from python??14:06
n4nd0sonne|work: I mean, is it possible to create a new SGVector from python interface?14:09
sonne|worknot typically used but I've seen heiko abusing it :)14:10
heikosonne|work: yep :) but not necessary actually14:11
n4nd0heiko, sonne|work : and how would you create the SGVector from the numpy array?14:21
n4nd0I'd rather avoid setting element by element14:22
n4nd0but I don't manage to get anything to work using the constructor14:22
sonne|workheiko: btw :)14:22
heikon4nd0: I dont understand, you can pase numpy arrays to functions that accept SGVectors right?14:24
sonne|workn4nd0: looks like we don't have a convenience constructor for that yet - you could create a vector of appropriate size, zero it and += a numpy array to it14:24
heikosonne|work: haha14:25
heikogreat :)14:25
n4nd0heiko: for some reason the SGVector constructor that accepts another SGVector constructor doesn't work for this14:25
sonne|workn4nd0: well it needs no const whatever crap14:26
n4nd0sonne|work: ok - out of curiosity, why the SGVector constructor that gets another SGVector doesn't work in this case?14:26
heikon4nd0: weird14:26
n4nd0sonne|work: aaah because of the const14:26
heikosonne|work: they also  mentioned the c-base workshop!14:26
sonne|workn4nd0: you need SGVector(SGVector vec);14:26
sonne|workthen it would work14:26
sonne|workheiko: yeah cool isn't it?14:26
heikosonne|work: indeed cool14:27
heikosonne|work: did they mention us before?14:28
sonne|workheiko: no idea - I accidentally stumbled upon this14:28
heikosonne|work: blackburn, I added my two project suggestions, feel free to feedback (will also continue to improve)14:29
heikotime for lunch!14:29
-!- FSCV [~FSCV@] has joined #shogun14:34
n4nd0sonne|work: isn't weird that some_SGVector = some_numpy_array14:34
n4nd0doesn't work while it does work with +=14:34
blackburnheiko: cool! thanks14:34
sonne|workn4nd0: just write the operator!14:35
n4nd0sonne|work: yeah yeah, I am doing it like that, I am just wondering :D14:35
blackburnheiko: haha what are you waiting for, an einstein?14:35
sonne|workn4nd0: and the constructor too please :)14:36
blackburnn4nd0: no he meant add operator= and it could work14:36
n4nd0getting lost here :S14:36
blackburnI am curious where can we find such students haha14:37
sonne|workyes exactly n4nd0 please write the constructor and the operator=14:38
blackburnPrerequisites: C++, (sparse) numerical linear algebra (positive definiteness, trace, Cholesky decomposition, Eigenvalues,iterative methods for linear systems (conjugate gradients), preconditioning), Basic probability theory: ( Gaussian distributions and manipulating them, expected value and variance), Graph colourings14:38
n4nd0blackburn: lol yeah, maybe some of those are more things the student will work with than actual reqs14:39
n4nd0at least IMO14:39
sonne|workblackburn: would you qualify?14:43
blackburnsonne|work: I am afraid no14:43
sonne|workscary isn't it?14:43
blackburnsonne|work: yeah so heiko needs someone more qualified than me and probably you :D14:44
n4nd0the scary I meant14:44
blackburnsonne|work: well preconditioners and conj gradient are most specific - most know what it is but I can't say I deeply understand that14:45
sonne|workblackburn: yeah an each set of algorithms in shogun is so specialized that there likely is just 1 shogun dev who *really* knows about it14:45
blackburnsonne|work: yeah, I don't know about string kernels at all and you probably don't know how these tapkee impls work14:46
blackburnthat's normal though I think14:46
sonne|workand when ICA enters shogun then we have something only andreas and 1 student know about14:47
blackburnsonne|work: I was scaried with his idea14:47
sonne|workit is somehow crucial that we keep having experts for each ML domain14:47
blackburnsonne|work: it is just some domain people avoid :D14:48
sonne|workand for sure tests to ensure that this stuff does sth14:48
blackburnsonne|work: not really a lot of references for that14:48
sonne|workblackburn: ?14:48
blackburnsonne|work: ICA14:48
sonne|workandreas' abstract was perfect14:48
sonne|workwikipedia then14:48
blackburnsonne|work: no I mean in machine learning ICA is not well acknowledged14:48
blackburnand BSS problem at all14:49
sonne|workbut in signal processing14:49
sonne|workand even ICA has its own conference14:49
blackburnsonne|work: it is close to matrix factorization and etc14:51
blackburnsonne|work: so igor's ideas are close to14:52
blackburnsonne|work: I don't like any separation - math is the same14:52
blackburnso that's quite nice we can unify things here in shogun14:53
sonne|workthats true14:54
sonne|workafter all that is what is shogun for14:55
n4nd0my SSVM learning for grid CRFs is crap :D14:55
n4nd0sonne|work: I am implementing this in python using the DirectorStructuredModel14:57
n4nd0but the results are bad - I have to debug it14:57
n4nd0sonne|work: what I mean graphically,
n4nd0top row is ground truth and bottom prediction15:06
sonne|workn4nd0: not totally random though :)15:06
n4nd0sonne|work: hehe, I have (at least) a bug for sure15:06
-!- n4nd0 [] has quit [Quit: leaving]15:24
-!- shogun-notifier- [] has joined #shogun15:59
shogun-notifier-shogun: iglesias :master * 8f1cfdd / src/shogun/structure/ (2 files):
shogun-notifier-shogun: some fixes for DirectorStructuredModel to make it work with PrimalMosekSOSVM15:59
shogun-notifier-shogun: remove general delta_loss and get_joint_feature_vector that are in StructuredModel15:59
shogun-notifier-shogun: add init_opt and check_training_setup (used by PrimalMosekSOSVM)15:59
shogun-notifier-shogun: iglesias :master * 06df817 / src/shogun/structure/PrimalMosekSOSVM.cpp:
shogun-notifier-shogun: add debug messages in PrimalMosekSOSVM15:59
shogun-notifier-shogun: iglesias :master * 6072d89 / src/shogun/features/ (3 files):
shogun-notifier-shogun: add obtain_from_generic method in MatrixFeatures15:59
shogun-notifier-shogun: iglesias :master * 2fddcb7 / src/shogun/structure/StructuredModel.cpp,src/shogun/structure/StructuredModel.h:
shogun-notifier-shogun: move implementation of CResultSet methods from header to cpp15:59
shogun-notifier-shogun: Soeren Sonnenburg :master * 323cf50 / src/shogun/ (8 files):
shogun-notifier-shogun: Merge pull request #945 from iglesias/structure15:59
shogun-notifier-shogun: Minor fixes for structured learning15:59
-!- blackburn [] has quit [Quit: Leaving.]16:07
-!- flxb [] has joined #shogun16:21
shogun-buildbot_build #952 of deb3 - modular_interfaces is complete: Failure [failed git]  Build details are at  blamelist: Soeren Sonnenburg <>, iglesias <>16:24
-!- hoijui_ [] has quit [Ping timeout: 264 seconds]16:49
heikon4nd0. write tests ! :)16:57
sonne|workheiko: that was just doc updates error message fixes :)17:04
heikosonne|work: I meant the bugs on the results17:05
sonne|workheiko: and n4nd0 ran away :-)17:05
sonne|workahh heh17:05
heikosonne|work: hehe :) as blackburn said17:05
heikoI walk around with a gun ;D17:05
sonne|workheiko: US would be the place to go then :D17:08
heikosonne|work: uuh, maybe I better relax again in this case ;)17:09
-!- zxtx [] has quit [Ping timeout: 245 seconds]17:21
shogun-buildbot_build #953 of deb3 - modular_interfaces is complete: Success [build successful]  Build details are at
-!- n4nd0 [] has joined #shogun17:37
-!- n4nd0 [] has quit [Client Quit]17:39
-!- hoijui_ [~hoijui@] has joined #shogun17:40
-!- hoijui_ [~hoijui@] has quit [Ping timeout: 260 seconds]18:21
shogun-notifier-shogun: Soeren Sonnenburg :master * 3630c69 / src/shogun/kernel/Kernel.cpp:
shogun-notifier-shogun: handle missing case in kernel18:33
-!- zxtx [zv@conference/pycon/x-ngldgkqwgolzbklc] has joined #shogun19:18
shogun-notifier-shogun: Soeren Sonnenburg :master * 5e619ed / tests/integration/python_modular/ (2 files):
shogun-notifier-shogun: blacklist a few failing integration tests and exit with exit code 1 if a test fails19:19
-!- zxtx [zv@conference/pycon/x-ngldgkqwgolzbklc] has quit [Ping timeout: 258 seconds]19:44
-!- zxtx [zv@conference/pycon/x-xjwwzbcndhdisryg] has joined #shogun19:45
-!- blackburn [~blackburn@] has joined #shogun19:52
heikoblackburn: I found out how to store matrix factor for later use19:54
blackburnheiko: insight me!19:54
heikox = L.triangularView<Lower>().solve(b);19:54
heikox = L.triangularView<Lower>().adjoint().solve(x); 19:54
heikowhere L is the factor19:54
blackburnheiko: ohh nice to know!19:55
heikosimply do this "double backslash" solve that you might know from matlab19:55
heikoquite cool19:55
blackburnheiko: no matlab is a foreign thing for me :)19:55
blackburnheiko: that's quite cool!19:55
heikoblackburn: you do often19:56
heikoinv(A)*b equals to L\L'\b19:56
heikobut whatever19:56
heikosame idea19:56
heikotrianular solver19:56
heikocool stuff this library19:56
heikoand also actively maintained :)19:57
blackburnheiko: hah much more better than blas thing19:57
heikothats a good trick if one has to solve multiple systems with the same matrix19:57
heikoone can simply store the cholesky19:57
heikoand then its very fast19:57
heikoblackburn: yes much better19:57
heikoI will put it into the GP project: replace all lapack/blas calls by eigen :)19:58
@sonney2kapologies guys19:58
blackburnsonney2k: ?19:58
-!- zxtx [zv@conference/pycon/x-xjwwzbcndhdisryg] has quit [Ping timeout: 264 seconds]19:58
blackburnheiko: eigen3 is well tested also - you may like it19:59
heikoblackburn: I know, it passes the blas testsuite :)19:59
heikoblackburn:  thats crucial for these things19:59
heikoimagine your solver returns nonsense :)20:00
heikowill never find that bug20:00
blackburnfriends presented me a toy helicopter I am playing around20:00
heikothats a good ressource20:00
blackburnheiko: yes, though sometimes I think they lack some doc20:00
heikoindeed, I found the thing in the forum20:01
blackburnheiko: btw I think we should use stackoverflow as that kind of forum20:01
heikohow can we do that?20:02
blackburnheiko: just subscribe to shogun tag at stackoverflow20:03
wikingheiko: yo20:03
heikoblackburn: good idea, we could also  mention on the website20:03
heikowiking: hi20:04
heikowas just about to leave :)20:04
wikingah ok20:04
wikingtalk to u then anoher time20:04
heikowiking: I will be here tonight or if not tomorrow20:04
heikobut its already 7 and still at work, should go home :)20:04
heikosee you later all20:05
-!- heiko [] has left #shogun []20:05
shogun-notifier-shogun: Soeren Sonnenburg :master * d3b6183 / / (14 files):
shogun-notifier-shogun: rename Bmrm_return_value_T -> BmrmStatistics20:07
shogun-notifier-shogun: Soeren Sonnenburg :master * 81a8ffb / src/Makefile,src/Makefile.template,src/configure:
shogun-notifier-shogun: make failed examples actually fail things20:07
@sonney2kblackburn, now you know20:07
@sonney2klots of fixing needed to get this stuff operational again20:08
shogun-buildbot_build #955 of deb3 - modular_interfaces is complete: Failure [failed test python_modular]  Build details are at  blamelist: Soeren Sonnenburg <>20:10
shogun-buildbot_build #661 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at  blamelist: Soeren Sonnenburg <>20:14
-!- travis-ci [] has joined #shogun20:17
travis-ci[travis-ci] it's Soeren Sonnenburg's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun:
-!- travis-ci [] has left #shogun []20:17
shogun-buildbot_build #374 of rpm1 - libshogun is complete: Failure [failed test]  Build details are at  blamelist: Soeren Sonnenburg <>20:25
-!- n4nd0 [] has joined #shogun20:26
shogun-buildbot_build #662 of cyg1 - libshogun is complete: Success [build successful]  Build details are at
shogun-buildbot_build #956 of deb3 - modular_interfaces is complete: Failure [failed test python_modular]  Build details are at  blamelist: Soeren Sonnenburg <>20:55
shogun-buildbot_build #957 of deb3 - modular_interfaces is complete: Failure [failed test libshogun]  Build details are at  blamelist: Soeren Sonnenburg <>21:10
-!- zxtx [zv@conference/pycon/x-jnkmsxrsyydeitia] has joined #shogun21:20
-!- hoijui_ [] has joined #shogun21:39
-!- zxtx [zv@conference/pycon/x-jnkmsxrsyydeitia] has quit [Ping timeout: 252 seconds]22:19
-!- zxtx [zv@conference/pycon/x-gjvzfaqdorndenme] has joined #shogun22:24
-!- zxtx [zv@conference/pycon/x-gjvzfaqdorndenme] has quit [Ping timeout: 258 seconds]22:53
-!- n4nd0 [] has quit [Ping timeout: 245 seconds]23:05
-!- n4nd0 [] has joined #shogun23:05
-!- hoijui_ [] has quit [Quit: Leaving]23:07
-!- shogun-notifier- [] has quit [Quit: transmission timeout]23:07
-!- n4nd0 [] has quit [Quit: leaving]23:21
-!- zxtx [zv@conference/pycon/x-mkrdomcndvqaktzk] has joined #shogun23:27
--- Log closed Wed Mar 20 00:00:15 2013