Open in new window / Try shogun cloud
--- Log opened Fri Jun 29 00:00:17 2012
CIA-18shogun: Soeren Sonnenburg master * r11eba43 / src/shogun/optimization/lbfgs/lbfgs.h : drop doxygen mainpage from liblbfgs (clashes with shogun mainpage) -
@sonney2kblackburn, finally it is there:
@sonney2kbut look at it
@sonney2klots of new stuff is not documented...00:23
blackburna lot of work00:24
blackburnI like work :D00:24
-!- n4nd0 [] has quit [Ping timeout: 240 seconds]00:25
@sonney2kblackburn, so then blackburn work for us :D00:25
alexlovesdataI had a long chat with my PhDstudent Viktor00:48
-!- alexlovesdata [55b21f1e@gateway/web/freenode/ip.] has quit [Quit: Page closed]00:49
-!- zxtx [] has quit [Ping timeout: 246 seconds]00:56
-!- blackburn [~blackburn@] has quit [Ping timeout: 248 seconds]02:17
-!- naywhaya1e [] has joined #shogun03:45
-!- naywhayare [] has quit [Ping timeout: 264 seconds]03:45
-!- zxtx [~zv@] has joined #shogun03:59
-!- romi_ [~mizobe@] has quit [Ping timeout: 252 seconds]04:45
-!- romi_ [~mizobe@] has joined #shogun04:48
-!- romi_ [~mizobe@] has quit [Ping timeout: 245 seconds]04:53
-!- romi_ [~mizobe@] has joined #shogun04:55
-!- romi_ [~mizobe@] has quit [Ping timeout: 265 seconds]05:19
-!- romi_ [~mizobe@] has joined #shogun05:20
-!- zxtx [~zv@] has quit [Ping timeout: 246 seconds]05:43
-!- sks [~sks@] has joined #shogun07:01
-!- gsomix [~gsomix@] has joined #shogun07:17
-!- romi_ [~mizobe@] has quit [Ping timeout: 264 seconds]07:56
-!- romi_ [~mizobe@] has joined #shogun07:56
-!- romi_ [~mizobe@] has quit [Ping timeout: 248 seconds]08:05
-!- sks [~sks@] has quit [Ping timeout: 248 seconds]08:42
-!- uricamic [~uricamic@2001:718:2:1634:a834:e48f:6ac0:beac] has joined #shogun08:50
shogun-buildbotbuild #1 of nightly_all is complete: Success [build successful]  Build details are at
CIA-18shogun: Soeren Sonnenburg master * r2660a96 / (2 files): fix compile error when lapack is not available -
-!- n4nd0 [] has joined #shogun09:52
shogun-buildbotbuild #2 of nightly_none is complete: Success [build successful]  Build details are at
-!- shogun-buildbot [] has quit [Remote host closed the connection]10:16
-!- shogun-buildbot [] has joined #shogun10:16
-!- zxtx [] has joined #shogun10:33
-!- nickon [] has joined #shogun10:33
-!- romi_ [~mizobe@] has joined #shogun10:38
-!- zxtx [] has quit [Ping timeout: 246 seconds]10:40
-!- blackburn [~blackburn@] has joined #shogun10:45
-!- romi_ [~mizobe@] has quit [Ping timeout: 248 seconds]10:46
-!- n4nd0 [] has quit [Ping timeout: 252 seconds]11:18
-!- zxtx [] has joined #shogun11:25
-!- blackburn [~blackburn@] has quit [Read error: Operation timed out]11:44
-!- romi_ [~mizobe@] has joined #shogun11:58
-!- heiko [] has joined #shogun12:00
-!- romi_ [~mizobe@] has quit [Ping timeout: 250 seconds]12:02
-!- romi_ [~mizobe@] has joined #shogun12:07
-!- romi_ [~mizobe@] has quit [Ping timeout: 246 seconds]12:12
CIA-18shogun: puffin444 master * rc1e2e27 / (11 files in 5 dirs): Added Kernel Scaling in CInferenceMethod. Fixed some warnings. Fixed -
CIA-18shogun: Heiko Strathmann master * r047feef / (11 files in 5 dirs): Merge pull request #609 from puffin444/master -
-!- n4nd0 [] has joined #shogun12:31
-!- romi_ [~mizobe@] has joined #shogun12:45
-!- blackburn [6da9d60d@gateway/web/freenode/ip.] has joined #shogun12:50
-!- pluskid [~pluskid@] has joined #shogun12:55
-!- romi_ [~mizobe@] has quit [Ping timeout: 246 seconds]13:14
-!- blackburn [6da9d60d@gateway/web/freenode/ip.] has quit [Ping timeout: 245 seconds]13:22
-!- romi_ [~mizobe@] has joined #shogun13:22
-!- romi_ [~mizobe@] has quit [Ping timeout: 240 seconds]13:43
-!- romi_ [~mizobe@] has joined #shogun13:58
-!- uricamic [~uricamic@2001:718:2:1634:a834:e48f:6ac0:beac] has quit [Quit: Leaving.]13:59
-!- romi_ [~mizobe@] has quit [Ping timeout: 244 seconds]14:08
-!- romi_ [~mizobe@] has joined #shogun14:21
-!- gsomix [~gsomix@] has quit [Ping timeout: 255 seconds]14:24
-!- gsomix [~gsomix@] has joined #shogun14:28
-!- romi_ [~mizobe@] has quit [Ping timeout: 246 seconds]14:34
gsomixsonney2k, around?14:39
-!- romi_ [~mizobe@] has joined #shogun14:45
-!- nickon [] has quit [Quit: ( :: NoNameScript 4.22 :: )]14:47
gsomixI just want to ask, what preprocessors I should "to direct"? Dense, Sparse, String, Dim reduction?14:48
-!- romi_ [~mizobe@] has quit [Ping timeout: 248 seconds]14:57
-!- gsomix [~gsomix@] has quit [Quit: Ex-Chat]15:05
-!- gsomix [~gsomix@] has joined #shogun15:05
-!- romi_ [~mizobe@] has joined #shogun15:10
-!- pluskid [~pluskid@] has quit [Ping timeout: 252 seconds]15:10
-!- pluskid [~pluskid@] has joined #shogun15:10
-!- alexlovesdata_ [c25faeba@gateway/web/freenode/ip.] has joined #shogun15:23
-!- romi_ [~mizobe@] has quit [Read error: Connection timed out]15:30
-!- romi_ [~mizobe@] has joined #shogun16:00
-!- n4nd0 [] has quit [Quit: leaving]16:06
-!- nicococo [] has joined #shogun16:29
-!- nicococo [] has left #shogun []16:29
@sonney2kgsomix, did you finish dotfeatures yet?16:53
@sonney2kotherwise just dense preprocs16:53
-!- blackburn [~blackburn@] has joined #shogun16:58
heikosonney2k, how can I activate debug messages from python?17:00
heikothanks blackburn17:04
heikoblackburn, are you still there?17:13
heikoI have a weird thing17:13
heikoI have a feature object17:14
heikosay CFeatures* p17:14
heikoand I call p->get_num_vectors()17:14
heikowhy is that possible?17:14
heikothe method is pure17:14
heikobut it works17:14
blackburnbut what is the actual type?17:14
heikothe method is overwritten there17:15
blackburnso what is the problem?17:15
heikobut why can I call it from the CFeatures interface?17:15
heikoI thought that wasnt possible17:15
blackburnbecause it is virtual17:15
heikobut its abstract17:15
heikoI thought the compiler forbids that, in fact I have experienced that a few times17:15
heikoand the linker complaining about undefined references17:16
blackburnhmm strange I've never experienced problem with pure virtual methods17:16
heikommh perhaps I have kept that wrong in memory17:17
blackburnit looks fine to me - you just declare it in base class and put impl in derived ones17:18
heikoI mean in C++ you always have that problem that a class cannot call its own pure virtual methods17:18
blackburnCSGObject::get_name is pure virtual as well17:18
heikoyes true17:18
heikoseems like, I have to read about that :D17:18
-!- pluskid [~pluskid@] has quit [Quit: Leaving]17:36
-!- romi_ [~mizobe@] has quit [Remote host closed the connection]18:00
-!- puffin444 [62e3926e@gateway/web/freenode/ip.] has joined #shogun18:07
@sonney2kblackburn, can you please have a look why all sparse features examples in static interfaces crash:
@sonney2kheiko, blackburn that would be the 'new' buildbot
@sonney2konly debian 1-3 now18:11
@sonney2kand nightly build18:11
heikosonney2k, nice!18:11
heikomuch more accessible :)18:11
@sonney2k is now the latest *released* documentation18:12
@sonney2kand nightly we get an updated doc18:12
@sonney2kcreated at18:12
heikoah thats also good18:12
@sonney2kit is run in the nightly_default process18:13
@sonney2kheiko, so now we only need a few more machines for osx, cygwin18:14
@sonney2kand potentially other linuxes18:14
heikoany ideas on the osx thing so far?18:14
heikoand cygwin?18:14
@sonney2kheiko, btw I am pretty impressed by the size of the build directory 1261504 kB for the whole build!18:15
heikohehe wow,18:15
heikobut that probably are the modular interfaces right?18:15
heikohow big is libshogun?18:15
blackburnsonney2k: ok will take a look18:16
@sonney2kheiko, well look at the buildbot :D18:16
@sonney2kheiko, I already have a cygwin machine18:16
@sonney2ktough to setup though but working on it18:16
@sonney2kfor osx  I have absolutely no idea yet :`-(18:19
@sonney2kso plan for now is to get cygwin running18:20
@sonney2kand email notifications18:20
heikoyeah step by step18:20
@sonney2kI always got email notifications btw18:20
@sonney2kso maybe I should at least put you: heiko, blackburn on the always recipient list for failures...18:21
-!- alexlovesdata_ [c25faeba@gateway/web/freenode/ip.] has quit [Ping timeout: 245 seconds]18:21
heikoyes you can do that18:21
blackburnI don't mind18:21
@sonney2kfor both of you your google addresses?18:22
@sonney2kwiking, any news on your PR?18:23
@sonney2kdidn't you want to send it by yesterday night?18:23
@sonney2kblackburn, all sparse stuff fails in modular too
heikosonney2k, I would like to add a little scripts to the python_modular/tools folder18:27
heikobut it is ignored by git18:27
heikoshould I add with -f?18:27
@sonney2kheiko, is it in .gitignores?18:28
@sonney2kheiko, what is its name?18:29
heikoId like to git add ../examples/undocumented/python_modular/tools/two_distributions_data.py18:29
heikoexamples/undocumented/python_modular/tools is in gitignore18:30
@sonney2kheiko, well it should not18:30
@sonney2kso better fix git ignore18:30
@sonney2kalright 30 C here18:30
@sonney2kI better go swimming before the thunderstorm comes18:31
@sonney2kshogun-buildbot, force build nightly_default18:32
shogun-buildbotbuild #5 forced18:32
shogun-buildbotI'll give a shout when the build finishes18:32
@sonney2kshogun-buildbot, help18:32
shogun-buildbotGet help on what? (try 'help <foo>', or 'commands' for a command list)18:32
@sonney2kshogun-buildbot, commands18:32
shogun-buildbotbuildbot commands: commands, dance, destroy, force, hello, help, last, list, mute, notify, source, status, stop, unmute, version, watch18:32
@sonney2kshogun-buildbot, list18:32
shogun-buildbottry 'list builders'18:32
@sonney2kshogun-buildbot, list builders18:32
shogun-buildbotConfigured builders: deb1 - libshogun deb2 - static_interfaces deb3 - modular_interfaces nightly_all nightly_default nightly_none18:32
@sonney2kshogun-buildbot, help watch18:32
shogun-buildbotUsage: watch <which> - announce the completion of an active build18:32
CIA-18shogun: Heiko Strathmann master * rccf6742 / (14 files in 6 dirs): Merge pull request #610 from karlnapf/master (+7 more commits...) -
-!- blackburn1 [~blackburn@] has joined #shogun18:42
puffin444Heiko, I have a question about the linear kernel18:42
heikopufin444, yes, hi18:42
puffin444It does not appear that there are any registered parameters for the LinearKernel. Is my observation correct?18:44
heikothe two should be registered18:45
heikothis is often forgotten :)18:45
puffin444Normal and normal_length?18:46
blackburn1too bad C++ have no annotations18:47
puffin444Do you think it makes sense to make them model selection parameters? I am not sure how a vector could  be modified in the framework.18:48
puffin444Or at least the normal vector18:48
heikono I think they shouldnt be available18:49
heikoAs far as I understand it they are only used in the optimisation case and set from outside anyway18:50
puffin444Okay. So I don't think I can add any kernel derivatives for the linear kernel then.18:50
heikobut thats natural isnt it?18:50
heikofor a linear kernel, what can you do? :)18:51
heikothey is nothing to vary18:51
puffin444I think so. GPML has a linear kernel with ARD that has kernel derivatives available18:51
heikolinear kernel with ARD?18:51
heikohow does that work?18:51
heikofor ARD you would need some kind of parameter for every dimension18:52
heikodo you have a link?18:52
heikoso I can read that?18:52
heikowhat are their kernel parameters?18:52
puffin444Yes its under covLINard.m at
puffin444a parameter for each dimension as you said.18:53
blackburn1two kernel statisticians walk into a bar18:53
heikopuffin44, I see, they bsically add a factor to each dimension18:54
heikoyou could add that to our linear kernel18:54
heikothe kernel is then k(x,y)= x1*y1*alpha1+x2*y2*alpha2 ....18:55
heikoblackburn and then? :)18:55
heikopuffin444, I like the idea18:55
puffin444How would this fit into the model selection framework? Wouldn't the dimensionality have to be set at construction?18:55
puffin444If the dimension of the inputs change then somehow the modelselection parameters would have to be deleted and created over again.18:56
blackburn1heiko: and then your conversation :)18:56
heikobut why should the dimension change when you use a linear kernel?18:57
heikoor do you mean when you register one parameter per dimension?18:58
puffin444I'm not sure. Suppose you allocated a linear kernel for some data set, use it, call cleanup(), and then init() if for some other dataset with different dimensions.18:58
puffin444one parameter per dimension if I want to integrate it into hyperparameter learning.18:58
heikoah, ok, I would say, just register a vector with the weights per dimension18:58
heikothen the parameter (also for modelselection) is *one* vector18:58
heikowould that be a problem in the lerning methods you wrote?18:59
puffin444This might complicate the kernel derivatives, as right now they are accessed by name.18:59
heikoyou would have to treat all elements of the vector as single parameter18:59
heikocant you just access the whole vector and then take only the desired index?19:00
heikommh the derivative of a linear kernel which is linear in its parameters19:00
heikowait, what would be the derivative of this kernel?19:01
puffin444The derivative would be respect to each dimension.19:02
puffin444K = -2*x(:,i)*x(:,i)';19:02
heikoI mean, the parameter vanishes19:03
puffin444The issue is that in order to abstractly access kernel parameters, I current take the total number of registered model selection parameters and use their name to get their derivative.19:03
shogun-buildbotHey! build nightly_default #5 is complete: Failure [failed test]19:04
shogun-buildbotBuild details are at
puffin444Thanks buildbot19:04
blackburn1shogun-buildbot: Hey!19:04
shogun-buildbotWhat you say!19:04
heikodamn that was my fault :)19:05
puffin444I am not sure how the indices of a vector parameter would fit it.19:05
puffin444The buildbot seems cheerful considering the build failed.19:05
heikopuffin444, mmh, hard to tell for me currently19:07
heikoI mean if you can access the vector itself19:07
heikoand you can get the derivative with respect to all these parameters, then you would be fine right?19:07
heikothe way they do it in the GP box is the same, they just compute the derivative wrt all dimensions19:08
heikoblackburn wtf ? :)19:09
heikohave you split?19:09
blackburn1heiko: transferring data from old notebook to new19:09
blackburn1so work on both now19:09
heikomake backups19:09
heikoI recently deleted all my data :)19:10
heikowhen doing exactly that19:10
blackburn1that should be funny19:10
puffin444One possible, but perhaps awkward way to do it is have the derivative function in the kernel return a list of matrices instead of one.19:10
puffin444Wait forget it that would mess up the GradientSelection stuff.19:11
blackburn1oh I like having 8 gb of ram19:11
heikopuffin444, no too much memory19:12
puffin444The problem is currently the framework treats each registered parameter as one dimension to be optimized. Access is by name to bypass possibly different indicies used for the same parameters in different data structures.19:12
heikoyes I see19:12
CIA-18shogun: Heiko Strathmann master * r11c6ed7 / examples/undocumented/libshogun/statistics_quadratic_time_mmd.cpp : corrected compile error -
CIA-18shogun: Heiko Strathmann master * rc54c124 / src/shogun/statistics/TestStatistic.h : added compute threshold dummy -
CIA-18shogun: Heiko Strathmann master * r029a375 / examples/undocumented/python_modular/ : minor fixes -
CIA-18shogun: Heiko Strathmann master * r5e9cd50 / (3 files in 3 dirs): Merge pull request #611 from karlnapf/master -
heikobut puffin444, when the kernel matrix is derived wrt to say first dimension parameter19:14
heikowhat is the difference when you derive wrt to second?19:15
heikono, this doesnt lead somewhere :(19:16
heikohow hard would it be to extend the framework to also work on vectors?19:16
heiko(and matrices) of parameters19:16
puffin444I really don't know.19:17
heikothis problem arises with any ard right?19:17
-!- blackburn [~blackburn@] has left #shogun []19:17
-!- blackburn1 is now known as blackburn19:18
-!- blackburn [~blackburn@] has left #shogun []19:18
puffin444Yes, and now that I look at Combined Kernel we have the same problem I think with computing derivatives.19:19
puffin444With the vector of subkernel weights19:20
heikoyes I see19:20
heikommh, we have to think about it some more19:20
heikoI would currently say to extend the proceure19:21
heikoyou currently get parameters by name, but since you can access the TParameter construct you also got type information19:21
heikoand using that you can make a distinction between CT_SCALAR, CT_VECTOR etc19:22
-!- blackburn [~blackburn@] has joined #shogun19:22
heikoand then iterator over all indices19:22
heikothen whats missing is how to get the derivative from the kernel19:22
heikosince we only know the index we are currently at19:22
heikothis could be solved by simply passing the index as parameter19:23
heikobut then theres the problem of having a unified interface for derivatives in general19:23
puffin444Yeah, I could overload the function by adding one that takes both a name and index.19:23
heikoyes that would work for vectors19:23
heikocould also do it for matrices19:24
puffin444I call one of them based on the TParameter Type.19:24
heikoand since these are the only container types shogun got19:24
heikothere are only three possibilites19:24
blackburnheiko: statistics_quadratic_time_mmd19:24
heikoblackburn, already fixed19:24
blackburnah forgot to pull sorry19:24
heikopuffin444, I think that would be nice19:25
blackburnheiko: statistics_linear_time_mmd19:25
heikoblackburn, damn :)19:25
puffin444Another issue is within the Sum and Combined Kernels. Suppose a Combined Kernel has a Gaussian Subkernel that has a width parameter.19:25
puffin444Forget it. If I pass the name to the subkernels it shouldn't be an issue.19:26
CIA-18shogun: Heiko Strathmann master * r11ba3d7 / examples/undocumented/python_modular/ : corrected compile error -
CIA-18shogun: Heiko Strathmann master * r263031f / examples/undocumented/libshogun/statistics_linear_time_mmd.cpp : another example fix -
CIA-18shogun: Heiko Strathmann master * r166b510 / examples/undocumented/python_modular/ : Wrong file. -
CIA-18shogun: Heiko Strathmann master * rdb38a5d / examples/undocumented/libshogun/statistics_linear_time_mmd.cpp : Merge pull request #612 from karlnapf/master -
heikopuffin444, ok19:28
heikopuffin444, well this addition by vectors and matrices would make the framework even nicer :)19:29
heikoand the problem with the number of dimensions (dont know how many parameters to register) is resolved on the fly19:29
puffin444This will take more time than expected though.19:30
heikoshogun-buildbot faster!19:30
heikopuffin444, yes, these framework things always take ages. I still remember the horror of these parameter trees last year :)19:30
-!- blackburn [~blackburn@] has quit [Quit: Leaving.]19:41
-!- blackburn [~blackburn@] has joined #shogun19:42
heikoblackburn, is there a corresponding thing to #ifdef HAVE_LAPACK in python_modular19:46
blackburnhmm no probably no19:47
heikohow to proceed with that? tests will fail if i use the methods19:47
blackburnI am curious how do we handle LDA then19:47
blackburnheiko: any idea? :D19:53
heikoblackburn, not really :D19:53
CIA-18shogun: Heiko Strathmann master * re9b94ce / (5 files in 3 dirs): Merge pull request #613 from karlnapf/master (+5 more commits...) -
heikoI just added the example19:59
heikoblackburn I have an idea19:59
heikowe could add a method to shogun class19:59
heikoboolean have_lapack()19:59
heikowhich then returns true/false depending on #ifdef stuff19:59
blackburnwell you may handle this via import19:59
heikothat would work19:59
heikothen in python you can ask that20:00
blackburnif import fails you have no your stuff available20:00
heikobut still the example fails20:00
blackburnno, why20:00
heikooh can one check for failing imports in pythong?20:00
blackburnjust catch exception and do nothing20:00
blackburnyeah try except20:00
heikobut in my case only one method is defined or not20:01
heikoimport would work20:01
blackburnheiko: you can check if method is here20:02
heiko"mymethod" in dir(dyn)20:02
heikonice I will add that20:03
CIA-18shogun: Heiko Strathmann master * r42dd334 / examples/undocumented/python_modular/ : added check for method availability to avoid errors when lapack isnt installed -
CIA-18shogun: Heiko Strathmann master * r9962e1d / examples/undocumented/python_modular/ : Merge pull request #614 from karlnapf/master -
heikoblackburn, nice works20:07
heikoI will leave now, see you later20:07
blackburnheiko: see you20:07
puffin444hey Heiko20:12
-!- blackburn [~blackburn@] has quit [Read error: Connection reset by peer]20:31
-!- blackburn [~blackburn@] has joined #shogun20:35
-!- blackburn1 [~blackburn@] has joined #shogun21:09
-!- blackburn [~blackburn@] has quit [Ping timeout: 265 seconds]21:09
-!- heiko [] has quit [Ping timeout: 264 seconds]21:19
-!- gsomix [~gsomix@] has quit [Quit: Ex-Chat]21:47
-!- gsomix [~gsomix@] has joined #shogun21:47
-!- blackburn1 [~blackburn@] has quit [Quit: Leaving.]22:08
-!- blackburn [~blackburn@] has joined #shogun22:34
-!- puffin444 [62e3926e@gateway/web/freenode/ip.] has quit [Quit: Page closed]22:34
@sonney2kwiking, could you please give me an update?22:38
blackburnsonney2k: hey have 5 mins?22:46
blackburngsomix asked me about add_to_dense_sgvec22:47
blackburnI am a little confused with it as well22:47
@sonney2kgsomix, why don't you ask me?22:52
@sonney2kblackburn, sure22:52
blackburnsonney2k: okay we both are little confused with SGVector parameter22:53
gsomixsonney2k, blackburn is closer to me... :)22:53
blackburnwe copy memory here you know22:53
blackburnhow can it work at all?22:53
@sonney2kblackburn, I don't understand the question22:53
blackburnsonney2k: add_to_sgvec has SGVector parameter22:53
blackburnlol nevermind22:54
@sonney2kproblem is you cannot do any operations with pointers22:54
@sonney2kso you need some sgvector interface22:54
@sonney2kgsomix, I see22:54
blackburnI mixed up what is being added and where22:54
blackburngsomix: add_to_sgvec adds given vector (you map it from numpy)22:54
@sonney2kso the idea is that this sg vector based function has to be overloaded22:54
blackburnto ith one22:54
blackburnsonney2k: I got it - that's ok22:55
@sonney2kand the sg_vector based function will be called from the standard one22:55
blackburnfor some reason I had idea that something modifies given vector22:55
blackburnso there would be a problem22:55
@sonney2kgsomix, btw the other thing on the plan once all this is finished are the typemaps for python for easier model selection interface22:55
blackburnsonney2k: I actually think we should stop at some point22:56
blackburnnot with >5 of director classes22:56
blackburnonce example is here it is easy to continue22:56
blackburnso for example I'd rather let say n4nd0 do it by given example - he just knows his domain22:57
blackburnit takes some time for gsomix to get into stuff he modifies22:57
blackburnsonney2k: wait I was wrong22:58
blackburnI am still confused22:58
blackburnsonney2k: you input vec2 (SGVector)22:58
blackburnand want to modify it22:58
blackburnhow can it go back if we typemap w/ copying22:58
gsomixsonney2k, ok.23:02
@sonney2kblackburn, it will be copied, then modified from python and then the modified thing is copied again and goes back to the add_to_dense_vec23:02
blackburnsonney2k: uhh how does it work?23:02
@sonney2kblackburn, as you would expect23:03
blackburndoes swig maps it back and forth?23:03
@sonney2kvia the in / out typemaps23:03
blackburnI thought out typemaps are only for return types23:03
@sonney2kone can even define new director in / out typemaps btw23:03
@sonney2kthey are23:03
@sonney2kbut you forget that this thing is going to python23:03
@sonney2kand somehow has to go back to C23:03
@sonney2kso C -> python -> C23:04
blackburnokay then I expect director dot features a little slow23:04
@sonney2kand arrows use the director stuff23:04
@sonney2ksame with kernel computation btw23:04
gsomixHhmm, and what about vec2 (SGVector)? Actually it's not numpy array, just pointer. I think there is a little problem, because vec2 is <Swig Object of type 'shogun::SGVector< double > *' at 0x1d069d0>.23:05
@sonney2kwe used some get_vector there - well only once23:05
@sonney2kgsomix, which function are you talking about?23:05
@sonney2kwell it is a SGVector23:06
@sonney2kwhat you can do is use its operations23:06
@sonney2klike +=23:06
@sonney2klike in the example23:07
gsomixBut I thought he should to be transformed into an array.23:07
gsomix*numpy array23:07
@sonney2khmmhh you are right23:08
gsomixAnd this is numpy array. After +=, I mean.23:08
@sonney2kmaybe it is not after all23:08
@sonney2kI wonder if this only happens when we use directorin / out typemaps23:08
gsomixsonney2k, should I try use directorin/out in typemaps?23:11
@sonney2kgsomix, we don't need them for Dotfeatures it seems23:11
@sonney2kit is sufficient to have the += operator overloaded23:11
@sonney2khmmhh strong thunderstorm outside...23:11
gsomixI'm afraid of thunderstorms... last was with a hurricane23:13
@sonney2knah not this one23:13
@sonney2kgsomix, you might need to overload *= etc operators to get other dot feature operations to work23:14
gsomixsonney2k, All right. I need to think a little and sleep.23:21
gsomixsonney2k, see you tomorrow, ok? =___=23:21
gsomixgood night23:21
-!- gsomix [~gsomix@] has quit [Ping timeout: 264 seconds]23:40
-!- n4nd0 [] has joined #shogun23:44
n4nd0sonney2k: hi there! do you have a moment for a question?23:45
-!- alexlovesdata [55b21452@gateway/web/freenode/ip.] has joined #shogun23:49
n4nd0hey alexlovesdata, how are you doing?23:50
alexlovesdatahey nando23:50
alexlovesdatapreparing to try out somethin for medical data ... but in one hour23:50
alexlovesdatacurrently yet busy ;)23:50
n4nd0alexlovesdata: can I ask you something about hmsvm? something about theory, nothing fancy I think23:52
alexlovesdataye, but maybe later23:53
n4nd0all right23:53
alexlovesdataI am still helping a friend with an insurance23:53
-!- blackburn [~blackburn@] has quit [Quit: Leaving.]23:58
--- Log closed Sat Jun 30 00:00:17 2012