Open in new window / Try shogun cloud
--- Log opened Sun Mar 18 00:00:19 2012
blackburnwhat do you call generic?00:01
PhilTilletWell, in the user guide there's "The code provides support to include your own coding, decoding, and base classifiers"00:03
PhilTilletShould also be the case for the C++ version I suppose :)00:03
blackburnjust some well-designed thing00:04
blackburnit shouldn't be very templated or so00:04
PhilTilletokay, that's what I wanted to know00:05
PhilTilletI have done an overdose of templates :p00:05
blackburnwe have not much templates here00:06
blackburnonly Features stuff00:07
PhilTilletWell, ECOC seems complicated but interesting :)00:09
PhilTilletis there some Information Theory  involved?00:09
blackburnyes, somewhere deep inside00:10
blackburnmain challenge here is how to decompose these bits00:10
PhilTilletYes, constructing the matrix they call Mc00:11
blackburnbut with diff distances between codes it involves some inf. theory things00:11
blackburnI mean practical00:11
blackburni.e. when you have 10 classes with some complex dependencies00:11
PhilTilletI see00:12
blackburnhow many bits you would like to use? what will each bit discriminate? and etc00:12
PhilTilletI see, each bit kind of represents a feature in this class doesn't it?00:13
blackburnno, it represents one discriminative classifier00:14
PhilTilletoh, yes00:14
PhilTilletone-versus-all, one-versus one, etc00:15
blackburnpretty interesting thing is these schemes are just some concrete ECOC matrices00:16
PhilTilletI love concrete things ^^00:18
PhilTilletI was used to one-versus-all methods but ECOC seems really more powerful00:21
blackburnmore complex though00:21
@sonney2k'evening gentlemen!00:22
PhilTillet'evening sonney2k !00:22
blackburn????? ????00:22
blackburn;)00:22
PhilTillet?????00:23
blackburnsonney2k: what kind of insomnia brought you here?00:23
blackburnPhilTillet: my chars were more informative ;)00:23
PhilTilletindeed :)00:23
@sonney2kwell I can read PhilTillet's ones :)00:24
blackburnoh really? what did he write? ;)00:25
blackburnsonney2k: back to release00:25
@sonney2kgood evening00:26
@sonney2kblackburn, yeah so how are things?00:26
blackburnsonney2k: did you take any japanese? ;)00:26
@sonney2kany progress?00:26
@sonney2kyes00:26
blackburnsonney2k: I need heiko's help here, he would help later00:26
blackburnI hope tomorrow actually00:27
@sonney2kok so we won't make progress today?00:27
blackburnexactly00:27
blackburnsonney2k: but we've got TWO new ideas suggestions00:27
blackburnone author is here btw ;)00:28
@sonney2k2 even?00:28
blackburnyeah check ml00:28
@sonney2kI read about deep learning on the ml00:28
blackburnyes and opencl00:29
blackburnsonney2k: I have totally mixed up opinion on any NNs00:29
PhilTilletyes, i'm here :p00:30
blackburnwe are pretty tight with slots and no idea what could we abandon00:30
PhilTilletjust to be clear, my mail told that I had already implemented NN using OpenCL, but making OpenCL code for SVM is also an option00:30
@sonney2kyeah that is the problem - the opencl idea is nice though00:31
blackburnopencl and svms is pretty tough thing!00:31
PhilTilletAgreed :) I wanted to do SVM rather than NN at first, but SVM is just on the Math point of view more complicated00:31
@sonney2kproblem with opencl (or any graphics card based stuff) is that you need to fit data into GPUs00:31
@sonney2kand I mean 100%00:31
@sonney2kotherwise you waste all your time in i/o00:32
@sonney2kand then you need to have a linear memory access pattern to be fast00:32
PhilTilletWell, I thought about that indeed, and my NN stopped working for very big datasets00:32
PhilTilletbecause all the data did not fit in memory00:32
PhilTilletbut transfering from GPU to CPU is very fast00:33
blackburnsonney2k: I am going to extract multitask kernel normalizers to shogun/multitask00:33
@sonney2kyeah - so it can speed up very expensive algorithms00:33
@sonney2kblackburn, not before the release I hope00:33
blackburnsonney2k: in my fork!00:33
blackburn;)00:33
@sonney2kdo wahtever you want in your fork :)00:34
blackburnyou should be happy - my multitask will be developed under pull request approach ;)00:34
@sonney2kPhilTillet, YMMV - for example when one trains an algorithm like perceptron or SVMSGD - it usually just does one pass over the data to do the whole learning00:35
blackburnsonney2k: can I integrate linux kernel in my fork? ;)00:36
@sonney2kso in this case memory access is everything00:36
@sonney2kblackburn, my fork has it already :D00:36
blackburnsonney2k: well it would be nice to have kernel in shogun/kernel00:36
blackburndon't you think so?00:36
blackburnI believe it is a kernel so it should be in kernel00:36
blackburn;)00:36
@sonney2kPhilTillet, how big is the memory on GPUs nowadays? 4G?00:37
PhilTilletsonney2k: Well, on some high end graphic cards yes00:37
@sonney2kPhilTillet, blackburn I have an idea though: mini-batches for liblinear00:38
@sonney2ke.g. one loads say '1GB' into GPU memory00:38
@sonney2kthen does multiple liblinear passes00:38
@sonney2kand then continues with the next chunk00:38
@sonney2kuntil convergence00:38
blackburnwould it converge? ;)00:39
@sonney2ksure00:39
PhilTilletyes, seems like a good idea indeed00:39
blackburnmulticlass may not00:39
@sonney2kbut I suspect kernel based SVMs would benefit most00:39
blackburnsonney2k: my idea was to use PhilTillet experience with opencl but not as main project :)00:41
@sonney2kblackburn, well look they even have a related paper about this http://www.csie.ntu.edu.tw/~cjlin/papers/kdd_disk_decomposition.pdf00:41
@sonney2kso it can surely work00:41
blackburnok ;)00:42
PhilTilletKernel Based SVMs on GPU would sure be more complicated than NN :p00:43
@sonney2kPhilTillet, no there has also been a paper maybe 5 years back about libsvm on GPU00:46
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun00:47
@sonney2kisn't opencl C-like?00:47
@sonney2kthen it shouldn't be too difficult00:47
PhilTillettrue that00:47
PhilTilletjust the way to think algorithm is different00:47
PhilTilletbut for basic kernels it shouldn't be that difficult indeed00:48
PhilTilletand the benchmarks on CUDA are indeed great : http://mklab.iti.gr/project/GPU-LIBSVM00:48
@sonney2kPhilTillet, the most expensive operation that is done for SVMs is usually this one http://shogun-toolbox.org/doc/en/current/classshogun_1_1CKernelMachine.html00:49
blackburnwrong link ;)00:50
@sonney2kPhilTillet, so if you manage to pre-load these coefficients alpha_i and examples (aka support vectors) x_i into memory00:50
PhilTilletwell, this operation is naturally parallel :p00:50
@sonney2kand have a fast implementation of kernel in opencl then it should be fast00:50
@sonney2kPhilTillet, I guess that would be a good starting point - try to get this as fast as possible with gaussian kernel http://shogun-toolbox.org/doc/en/current/classshogun_1_1CGaussianKernel.html00:51
PhilTilletcan it replace the "small patch" ? :p00:52
@sonney2kwith big patch you mean :-)00:52
PhilTilletWell, can I use viennacl::norm2 ? :D00:53
PhilTillet(as I said in my mail, I'm a contributor to an OpenSource project which implements already every linear algebra operations)00:55
PhilTillet(so if I can reuse this code, the patch will be really small :) )00:56
@sonney2kPhilTillet, well just use it - I guess later one can extract just the needed parts or make viennacl an optional dependency00:57
PhilTilletthis one is more challenging : http://shogun-toolbox.org/doc/en/current/classshogun_1_1CGaussianShiftKernel.html00:58
PhilTillet:p00:58
@sonney2kthere are other more important kernels00:59
@sonney2klike linear and poly kernel (easy again)00:59
PhilTilletOkay01:00
PhilTilletso my homework is to benchmark these against an OpenCL implementation?01:00
@sonney2kPhilTillet, benchmark and then it would of course be nice to have this as a patch to shogun01:01
PhilTilletokay01:01
PhilTilletcould I just have some pointer concerning the involved files/folders ?01:02
@sonney2kthe above doxygen doc's point to the file names01:02
@sonney2kthe apply() method is the one in kernel machine that does this a_i y_i K(x_i, x)01:03
@sonney2ksumming01:03
PhilTilletoh yes, I see01:03
@sonney2kbut no problem - just start an isolated dry-run first01:03
@sonney2kwhen a kernel machine is applied you have to assume x is setting in CPU memory (and x_i on GPU)01:04
@sonney2kblackburn, any idea where we could squeeze in PhilTillet ?01:05
blackburnsonney2k: squeeze?01:05
blackburnI do not understand..01:06
n4nd0sonney2k: hi! I have a couple of doubts related to the SO project, I asked Nico by mail but probably he didn't get time to answer yet, maybe you can help me though01:08
@sonney2kn4nd0, when did you email him?01:09
n4nd0sonney2k: last Thursday, maybe I have not been patient enough :S01:10
n4nd0sonney2k: we already exchanged a first mail01:11
@sonney2kI susspect he will answer on monday...01:15
PhilTilletsonney2k: just a quick question that can save me a lot of time : in the CFeatures* argument to apply(), are the features stored in lines or column?01:17
n4nd0all right, then there is no reason to worry about :)01:17
blackburnPhilTillet: in simplefeatures features are stored column-wise01:17
n4nd0I will continue improving an example for qda01:17
blackburnvector by vector01:18
PhilTilletOkay :) And, internally it is stored in a ublas::matrix ?01:18
blackburnhuh strange guess01:18
blackburn;)01:18
PhilTillet:D01:19
blackburnjust float64_t*01:19
PhilTilletokay :D01:19
PhilTilletplain old matrix :p01:19
PhilTilletokay, homework seems tough enough :)01:21
@sonney2kPhilTillet, well do the benmark first with some simulated data...01:22
PhilTilletyes :)01:24
PhilTilletGaussian Kernel is rather straightforward, but more complicated is the apply() formula :D01:25
PhilTilletbut I like challenge !01:26
blackburnwhy is it more complicated?01:26
PhilTillethmmm01:28
PhilTilletyou're right, it's not01:29
PhilTilletdepends on how efficient i want the code to be01:30
PhilTilletas always01:30
PhilTilletcan I have an order of magnitude of the size of a feature vectors, and the size of a dataset?01:35
blackburnyou mean in apply?01:36
PhilTilletyes01:37
PhilTilletthe N of the formula, and the size of x_i01:37
blackburnKernelMachine?01:37
PhilTilletyes01:37
blackburnline 24801:38
PhilTilletI meant, in practice01:39
PhilTillet:p01:39
blackburnhmm?01:39
PhilTillethmmmm, concretely, does the feature matrix have more row or more columns?01:40
blackburndepends on data01:40
PhilTilletok, :p01:41
PhilTillet(yes my question was weird)01:41
blackburnno general answer here01:41
PhilTilletokay :)01:42
blackburnit may be 4 2000d vectors01:42
blackburnor 2000 4d01:42
blackburnok I'm falling asleep01:46
blackburnPhilTillet: in fact you may rewrite compute_batch thing in kernel01:47
PhilTilletwell i'll firstly do it on randomly generated data :p01:47
PhilTilleti'm falling asleep too01:47
PhilTillet:D01:47
blackburnI mean no need to add anything to kernelmachine01:48
PhilTilletokay :)01:49
PhilTilletIt was an interesting conversation, but I need to sleep if I want to survive ! I'll probably come back tomorrow ! good night :)01:58
blackburnsure01:58
-!- PhilTillet [52e82691@gateway/web/freenode/ip.82.232.38.145] has quit [Quit: Page closed]01:58
blackburnme too01:58
-!- blackburn [~qdrgsm@188.122.253.192] has quit [Ping timeout: 246 seconds]02:21
-!- GenX__ is now known as GenX02:26
-!- vikram360 [~vikram360@117.192.166.196] has joined #shogun02:36
-!- vikram360 [~vikram360@117.192.166.196] has quit [Ping timeout: 260 seconds]02:44
-!- vikram360 [~vikram360@117.192.166.195] has joined #shogun02:45
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking]02:47
-!- vikram360 [~vikram360@117.192.166.195] has quit [Ping timeout: 264 seconds]03:52
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 240 seconds]03:55
-!- vikram360 [~vikram360@117.192.161.63] has joined #shogun05:07
-!- vikram360 [~vikram360@117.192.161.63] has quit [Ping timeout: 260 seconds]05:15
-!- vikram360 [~vikram360@117.192.161.63] has joined #shogun05:16
-!- harshit_ [~harshit@182.68.139.59] has joined #shogun05:33
-!- harshit_ [~harshit@182.68.139.59] has quit [Remote host closed the connection]05:43
-!- adneus [~anuj@117.195.38.237] has joined #shogun05:45
-!- vikram360 [~vikram360@117.192.161.63] has quit [Ping timeout: 264 seconds]06:08
-!- vikram360 [~vikram360@117.192.185.187] has joined #shogun07:12
-!- adneus [~anuj@117.195.38.237] has quit [Quit: Leaving]07:29
-!- vikram360 [~vikram360@117.192.185.187] has quit [Ping timeout: 252 seconds]07:39
-!- vikram360 [~vikram360@117.192.190.55] has joined #shogun07:45
-!- vikram360 [~vikram360@117.192.190.55] has quit [Ping timeout: 244 seconds]08:03
-!- vikram360 [~vikram360@117.192.183.103] has joined #shogun08:10
-!- vikram360 [~vikram360@117.192.183.103] has quit [Ping timeout: 250 seconds]08:19
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun09:02
-!- vikram360 [~vikram360@117.192.166.74] has joined #shogun09:15
-!- blackburn [~qdrgsm@188.122.253.192] has joined #shogun09:41
n4nd0hi people!09:41
n4nd0I just pulled and found a compilation error09:41
n4nd0swig error : Unrecognized option -builtin09:41
n4nd0Use 'swig -help' for available options.09:41
n4nd0make[1]: *** [modshogun_wrap.cxx] Error 109:41
n4nd0any clue about it?09:42
blackburnn4nd0: you've got old swig09:42
blackburn<2.0.409:42
n4nd0oh, yes it looks like that09:43
n4nd0whereis swig09:43
n4nd0swig: /usr/bin/swig /usr/share/swig1.309:43
blackburnubuntu 11.10?09:43
n4nd0yes09:43
blackburnuse package swig2.009:43
n4nd0the one in the repositories is ok?09:43
n4nd0ok, thanks!09:44
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun10:08
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 244 seconds]10:26
-!- surtani [0e8b5206@gateway/web/freenode/ip.14.139.82.6] has joined #shogun10:38
-!- harshit_ [~harshit@182.68.171.83] has joined #shogun10:46
-!- harshit_ [~harshit@182.68.171.83] has quit [Quit: Leaving]10:53
-!- harshit_ [~harshit@182.68.171.83] has joined #shogun10:53
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun11:09
-!- vikram360 [~vikram360@117.192.166.74] has quit [Ping timeout: 276 seconds]11:50
-!- vikram360 [~vikram360@117.192.167.167] has joined #shogun11:52
-!- puneetgoyal [~puneet@115.242.75.148] has joined #shogun12:02
-!- harshit_ [~harshit@182.68.171.83] has quit [Remote host closed the connection]12:06
-!- puneetgoyal [~puneet@115.242.75.148] has quit [Ping timeout: 260 seconds]12:17
-!- puneetgoyal [~puneet@101.63.34.173] has joined #shogun12:34
-!- in3xes [in3xes@49.249.17.196] has joined #shogun13:11
-!- puneetgoyal [~puneet@101.63.34.173] has quit [Quit: Leaving]13:13
-!- in3xes_ [~in3xes@180.149.49.230] has joined #shogun13:18
-!- in3xes [in3xes@49.249.17.196] has quit [Ping timeout: 250 seconds]13:21
-!- in3xes_ [~in3xes@180.149.49.230] has quit [Ping timeout: 252 seconds]13:31
-!- GenX [~Sambhav@14.139.82.6] has quit [Read error: Connection reset by peer]13:56
-!- GenX [~Sambhav@14.139.82.6] has joined #shogun13:56
-!- xiangwang [~xiangwang@123.116.114.145] has joined #shogun14:09
-!- in3xes [in3xes@49.249.17.138] has joined #shogun14:30
-!- xiangwang [~xiangwang@123.116.114.145] has quit [Quit: Leaving]14:32
-!- vikram360 [~vikram360@117.192.167.167] has quit [Ping timeout: 252 seconds]14:41
-!- surtani [0e8b5206@gateway/web/freenode/ip.14.139.82.6] has quit [Ping timeout: 245 seconds]14:52
-!- adneus [~anuj@117.195.34.137] has joined #shogun15:10
-!- gsomix [~gsomix@188.168.14.44] has joined #shogun15:20
gsomixhi15:20
-!- Varagrawal [7bc955a2@gateway/web/freenode/ip.123.201.85.162] has joined #shogun15:54
VaragrawalHey there!15:55
VaragrawalI am going through the tutorial and the first code minimal.cpp is giving me undefined reference error. :(15:57
VaragrawalI just can't get what's wrong15:57
VaragrawalI ran sudo apt-get install libshogun-dev and then opened up vi and wrote that code15:58
Varagrawalfinally when I compile, errors15:58
blackburnyou should use latest git15:59
blackburnshogun is outdated in the repo15:59
Varagrawalah16:00
VaragrawalI have the git16:00
Varagrawalcompile by source it is then, thanks :D16:00
Varagrawalcompiled by source, stii getting the same error :(16:14
gsomixVaragrawal, maybe the code will say more? http://pastebin.com/16:21
n4nd0blackburn: I prepared a colorful example with multiclass qda :P16:58
blackburnnice16:58
n4nd0blackburn: take a look, what do you think http://dl.dropbox.com/u/11020840/multiclass_qda.png16:58
blackburncool!16:58
n4nd0:)16:59
n4nd0blackburn: matplotlib has lot of cool options, are you used to work with it?17:02
blackburnsure17:02
n4nd0I just worked with matlab/octave plotting stuff before17:03
blackburnI have never worked with it :)17:03
n4nd0so it is quite new for me, but I like it :)17:03
blackburnit is *very* flexible17:04
n4nd0yes17:05
-!- in3xes [in3xes@49.249.17.138] has quit [Read error: Connection reset by peer]17:06
n4nd0I'll try to learn a bit more of it17:06
n4nd0I think it is always nice to have some kind of nice visualization of the stuff17:06
blackburnn4nd0: I bet this picture is not possible with matlab/octave: http://dl.dropbox.com/u/10139213/shogun/lle.png17:07
n4nd0blackburn: haha I don't really know!17:11
n4nd0I've never seen it before in matlab at least17:12
n4nd0it's reallly cool17:12
n4nd0with one of your dimensionality reduction algos?17:12
blackburnyes, LLE17:13
-!- vikram360 [~vikram360@117.192.188.74] has joined #shogun17:25
-!- Varagrawal [7bc955a2@gateway/web/freenode/ip.123.201.85.162] has quit [Ping timeout: 245 seconds]17:58
-!- varagrawal [0e8b7952@gateway/web/freenode/ip.14.139.121.82] has joined #shogun18:28
varagrawalHey118:28
varagrawalI forked the repo and installed from source18:29
varagrawalbut when I compile the minimal.cpp program, I get an undefined reference error18:29
varagrawalany help?18:29
-!- karlnapf [~heiko@host86-181-153-176.range86-181.btcentralplus.com] has joined #shogun18:36
karlnapfblackburn, around=?18:36
n4nd0varagrawal: hey! can you copy paster the error somewhere?18:36
blackburnyeah18:36
n4nd0varagrawal: try here http://snipt.org/18:37
karlnapfhi18:37
karlnapfso you were interested in the migration stuff ?18:38
blackburnkarlnapf: hi. we probably need to update testsuite18:38
karlnapfok18:38
blackburnhow can we do that?18:38
karlnapfok, I will explain what I did there :)18:38
varagrawalyes18:38
karlnapfI introduced this new flag PARAM_VERSION18:39
varagrawal"/tmp/ccND136M.o: In function `main'"18:39
karlnapfwhich is written into each serialised object18:39
varagrawal"minimal.cpp:(.text+0x24): undefined reference to `shogun::init_shogun(void (*)(_IO_FILE*, char const*), void (*)(_IO_FILE*, char const*), void (*)(_IO_FILE*, char const*), void (*)(bool&, bool&))' minimal.cpp:(.text+0x29): undefined reference to `shogun::exit_shogun()' collect2: ld returned 1 exit status"18:39
-!- in3xes [~in3xes@59.163.196.5] has joined #shogun18:40
karlnapfblackburch, btw do you have time or should we talk later?18:40
blackburnkarlnapf: so you introduced it after testsuite was serialized?18:40
blackburnsure, I have a lot18:40
karlnapfWell so far nothing changed18:40
karlnapfThe goal was this:18:40
karlnapfYou have an isntance of an object which you put a lot of effort into, and you save it to a file18:40
karlnapfthen a new shogun version comes out which has this great feature you want to use18:41
karlnapfhowever the nasty developers added a new variable in the class of that particular instance you saved18:41
karlnapfso if you want to load with the new version18:41
n4nd0varagrawal: ok, so you said you managed to fork the repo and get the code locally right?18:41
karlnapfit does not work18:41
n4nd0varagrawal: did you also manage to compile the project?18:41
varagrawal@n4nd018:41
varagrawalthe install works perfectly18:42
karlnapf(alternative: a type change, matrix from float64 to float32)18:42
blackburnkarlnapf: what if name changed?18:42
varagrawalCompiling the minimal.cpp gives the issue18:42
varagrawalhttp://snipt.org/ugajf918:42
karlnapfsame thing, wouldnt work18:42
blackburnno way to make it work?18:42
karlnapfBut we want to allow people loading old version files18:42
karlnapfyes, that what I did :)18:42
blackburnok more concretely18:42
blackburnI renamed labels to m_labels18:42
karlnapfEvery shogun object now has a parameter map variable18:43
blackburnwhat should I do to make it work18:43
karlnapfThere you can secify changes of parameters in between version18:43
karlnapfs18:43
karlnapfwhen you just change the c++ variable name, nothing will happen, only if you change the registered parameter name by the SG_ADD18:43
karlnapflets have a look at one example18:43
n4nd0varagrawal: did you try with some of the examples that come with the project?18:43
varagrawalI literally copy-pasted the code in the tut18:43
blackburnaha so I need to restore names better18:44
blackburn"labels" I mean18:44
blackburninstead of "m_labels"18:44
varagrawal1 sec, I'll do that18:44
karlnapfyou can change them,18:44
blackburnbtw karlnapf, you have added some new parameter about locking18:44
karlnapfyes18:44
blackburnso we can't read testsuite right now :)18:44
karlnapffrom now on, everytime we ad d the structure of objects, we should put this change into the parameter map18:45
blackburnahha18:45
karlnapfah ok.18:45
blackburnwhere?18:45
karlnapfhave a look at examples/libshogun/base_migration_dropping_and_new.cpp18:45
karlnapfThere you got two classes which are the same except for some parameters that are new/dropped18:46
karlnapfYou need to do two things:18:46
varagrawalI am trying to run the OCR example, but the python script throws an error saying "No Module named modshogun"18:46
karlnapffirst: put the change in the parameter map (same place where parameters are registered)18:46
karlnapfsecond: override the CSGObject::migrate methods18:46
karlnapfmethods18:46
karlnapfmethod18:46
karlnapf(only one)18:47
karlnapfFor simple changes like a name change/new parameters/dropped parameters theres a helper function which does most of the wprlk18:47
karlnapfwork18:47
karlnapfFor new parameters its particulary easy18:47
karlnapfjust need to register the new parameter, then its ignored while loading old files18:47
blackburndo you have time right now to make it work with row_subset18:48
karlnapfdropped parameters work out of the box, just delete them, they are also then ignored while loading18:48
blackburnI guess it is now a new parameter18:48
n4nd0varagrawal: did you build with python support?18:48
karlnapfyes, could do that18:48
karlnapfI have some changes in my branch, let me check them18:48
karlnapfI will get back to you later with a pull-request :)18:49
varagrawalI ran ./configure18:49
blackburnthanks!18:49
varagrawaldidn't specify explicit parameters18:49
blackburnkarlnapf: I guess it is better to get m_labels -> labels as before18:49
blackburnI mean name string18:49
varagrawalthe installation guide said that it would be detected on its own18:49
karlnapfyou can change that, just add to map18:49
karlnapfhave a look into18:50
n4nd0varagrawal: I think you should configure with --interfaces=python_modular so you can use shogun from python18:50
n4nd0varagrawal: take a look here http://www.shogun-toolbox.org/doc/en/current/installation.html18:50
blackburnI understand but I think it is not really needed18:50
varagrawalWill do that18:50
blackburnadding "m_" does not affect anything but codestyle18:50
karlnapfm_number_to_keep variable in the example I mentioned18:50
blackburnyeah I checked it already18:51
karlnapfthis m_ shouldnt be in registered names anyway in my opinion18:51
blackburnha!18:51
blackburnkarlnapf: then could you please add row_subset map?18:52
blackburnI hope that would make things work18:52
karlnapfyes will check that18:52
karlnapfshould :)18:52
karlnapfsorry for breaking it18:52
karlnapfwe should add this to shogun code style: whenever you add variable to existing class, add parameter map and miration method18:52
varagrawal"However, just running  ./configure  will autodetect and configure for the available interfaces."18:53
blackburnpain in the ass thingy18:53
karlnapfor least the mapping, so that loading still works, and initialise with reasonable values in constructor, thats missing sometimes18:53
varagrawalThat's what I did18:53
karlnapfYes, true, however, we are using c++ :(18:53
blackburnvaragrawal: ./configure --interfaces=python_modular18:53
varagrawalYeah, did that18:54
varagrawalDo I need to install swig?18:54
blackburnexactly18:54
CIA-64shogun: Sergey Lisitsyn master * rc087f70 / src/shogun/machine/Machine.cpp : Restored machine parameters names - http://git.io/JZN4fA18:55
varagrawalswig2.0 ?18:55
n4nd0varagrawal: yes18:55
karlnapfbackburn, I really would like to integrate this john platt probability sigmoid fitting, have you thought more about that?18:55
karlnapfcurrently doing by hand for a project18:56
karlnapfbut this sucks18:56
karlnapfand also we should do simila stuff for the one-against-all svms18:56
blackburnkarlnapf: not really.. well we have LR regularized liblinear18:57
blackburnfor 2 class things18:57
karlnapfmmh18:57
karlnapfits not a lot of work18:57
blackburnI mean it is prob output already18:57
karlnapfI remember there has been a probabilities variable in labels18:58
karlnapfone year ago or so18:58
karlnapfWhat about just adding it again18:58
blackburnah18:58
blackburnwait18:58
blackburnthe idea is18:58
blackburnLabels (base class)18:58
varagrawalyowza18:58
varagrawalThink it'll work now :D18:58
blackburnBinaryLabels, ProbabilisticBinaryLabels18:58
blackburnMulticlassLabels, ProbabilisticMulticlassLabels18:58
blackburnMultipleLabels18:59
blackburnor so18:59
blackburnbut for now it is ok to add probabilities in labels18:59
blackburn(as I think)18:59
karlnapfthe other stuff is a lot of work though, many changes18:59
blackburnkarlnapf: I would need some help with other stuff btw18:59
karlnapfother stuff?19:00
blackburncheck machine/MulticlassMachine.cpp19:00
blackburn:)19:00
blackburnOvO training would require some subsetting stuff19:00
blackburncan you implement it?19:00
blackburnit is ~20 lines of code probably though :D19:01
blackburnbut I was lazy to get myself into your subsetting thing19:01
karlnapfI see :)19:02
karlnapfsubsetting of the features you mean19:02
karlnapfor for the locking?19:02
blackburnbut it is not of any priority19:02
blackburnyeah19:02
blackburnno19:02
karlnapfmake it possible to use it for x-val you mean this subsetting? :)19:02
karlnapfok19:02
karlnapfyes I can do this19:02
blackburnI mean there should be subset only for 2 classes19:02
blackburnand machine trained on top of that19:02
karlnapfthough I am ultra busy this week19:03
blackburnah19:03
blackburnnevermind then!19:03
karlnapfIf its non-priority lets talk about it next week or so19:03
blackburnor month19:03
blackburncause I'm getting busier day by day as well19:03
blackburn:D19:03
karlnapfWhat are you currently doing btw?19:03
blackburnwell working and studying19:04
karlnapfanythign special when you say coure getting busier?19:04
blackburnoh yeah I have done nothing yet with some studies19:04
blackburn:D19:04
-!- varagrawal [0e8b7952@gateway/web/freenode/ip.14.139.121.82] has quit [Ping timeout: 245 seconds]19:05
blackburnprobably that is the time to start19:05
karlnapf do you also get this combined kernel example error in python_modular?19:06
blackburnkarlnapf: in fact it is all full of ERROR19:06
blackburnno idea WTF19:07
blackburnkarlnapf: with make run-testsuite you may see errors with params19:08
karlnapfkk19:08
karlnapfWhat about all these new guys around here? :)19:09
blackburnkarlnapf: well #shogun is getting active againg19:10
karlnapfpretty good I guess19:10
karlnapfoh year now I see all the parameter errors...19:10
-!- varagrawal [0e8b7952@gateway/web/freenode/ip.14.139.121.82] has joined #shogun19:11
varagrawalHow do I assign a kernel to the OCR application?19:12
blackburnvaragrawal: it is hardcoded I am afraid19:13
varagrawalI keep getting the error19:14
blackburnwhich?19:14
varagrawal"Please specify a kernel"19:14
karlnapfblackburn, to make this mapping work, we have to increase the parameter version in the current git19:14
blackburnI don't mind increasing anything19:14
blackburn:D19:14
blackburnI don't even mind shogun 1254.019:15
blackburnvaragrawal: aha got it19:15
karlnapfok then change it to 1 :)19:15
varagrawal?19:15
blackburnvaragrawal: I guess we need to retrain it19:16
blackburnand upload new state19:16
-!- vikram360 [~vikram360@117.192.188.74] has quit [Ping timeout: 240 seconds]19:20
karlnapfblackburn, should I update the NEWS file a bit or do you want to do that yourself?19:22
blackburnfeel free to do that ;)19:22
-!- adneus [~anuj@117.195.34.137] has left #shogun ["Leaving"]19:30
karlnapfblackburn, seems like there is a bug around, I got a crash now for the migration19:36
blackburnha!19:36
karlnapfI cannot figure out which code causes it19:36
karlnapfcomes after this:19:36
karlnapfcd ../..//../testsuite/python_modular && \19:36
karlnapf( LD_LIBRARY_PATH=//usr/local/lib \19:36
karlnapfPYTHONPATH="//usr/local/lib/python2.7/dist-packages" python test19:36
blackburnhuh19:37
karlnapfI never touched the test-suite so far, how does it work?19:38
blackburnjust runs .py from examples19:38
blackburnand checks equality19:38
karlnapfehm19:38
blackburncomparing with ones saved before19:39
karlnapfI got no error in make tests19:39
karlnapfbut a memleak in make run-testsuite19:39
karlnapfthat makes no sense then?19:39
blackburnmemleak? which one fails?19:39
blackburnI've got no memleaks19:39
karlnapfaveraged percepton, but I just changed code to make it use the migrations19:39
blackburnhmm19:40
karlnapfweiiird19:40
-!- varagrawal [0e8b7952@gateway/web/freenode/ip.14.139.121.82] has quit [Quit: Page closed]19:45
karlnapfbut it loads some serialised files, doesnt it?19:45
blackburnyeah19:45
blackburnit loads and compares19:45
karlnapfwhere is the code for that?19:45
blackburntestsuite/python_modular19:47
karlnapffound it, messy :)19:47
blackburntester.py19:47
karlnapfaaah19:48
karlnapfthink I got the error19:48
-!- puneetgoyal [~puneet@101.63.95.73] has joined #shogun19:48
-!- varagrawal [0e8b7937@gateway/web/freenode/ip.14.139.121.55] has joined #shogun19:52
varagrawalhey19:52
varagrawalOCR is working19:52
varagrawalBut now when I try to run msplicer19:53
varagrawalI get this19:53
varagrawalERROR IMPORTING MODULES, MAKE SURE YOU HAVE SHOGUN INSTALLED19:53
varagrawal:-/19:53
karlnapfblackburn, doesnt work now, I will get back to you on this. Hopefully next week :)19:54
karlnapfwill go now, bye all19:54
n4nd0karlnapf: bye!19:59
blackburnkarlnapf: oh okay20:00
blackburnsee you20:00
-!- karlnapf [~heiko@host86-181-153-176.range86-181.btcentralplus.com] has left #shogun []20:03
-!- in3xes [~in3xes@59.163.196.5] has quit [Ping timeout: 260 seconds]20:07
-!- in3xes [~in3xes@210.212.58.168] has joined #shogun20:15
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 264 seconds]20:24
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking]20:46
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun21:00
varagrawal:n4nd021:09
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun21:09
varagrawalFor msplicer I am getting the error, ERROR IMPORTING MODULES, MAKE SURE YOU HAVE SHOGUN INSTALLED21:09
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 245 seconds]21:10
-!- varagrawal [0e8b7937@gateway/web/freenode/ip.14.139.121.55] has quit [Quit: Page closed]21:14
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun21:21
@sonney2kblackburn, hey...21:23
blackburnsonney2k: hey21:25
blackburnwhat's up?21:26
@sonney2kis the website change / announcement  / blog  etc I did OK?21:26
@sonney2kor do you have any suggestions?21:26
blackburnor it is ok for sure21:26
blackburnoh*21:26
blackburnwhy do you care so much about that announcement?21:26
@sonney2kthat is how we get students looking at the project?21:27
@sonney2k!21:27
blackburnheh take a look how many students we have already ;)21:27
blackburnI mean they would come anyway21:28
@sonney2kI don't think so - in particular good ones might not. anyway I send announcements to the popular machine learning mailinglists and google+ / debian planet21:29
blackburnaahh21:29
blackburni see21:29
@sonney2kscipy is missing still but then done21:29
blackburnso if I had suggestions21:29
blackburnhow can you edit it21:29
blackburn:D21:29
@sonney2kit is a website21:30
@sonney2kand a blog I can change at any time21:30
blackburnI see21:30
@sonney2kwhy would I be asking otherwise?21:30
@sonney2kanyway, any news regarding release progress?21:30
blackburnsonney2k: yeah we fucked up21:31
blackburn:D21:31
@sonney2kwhich means what?21:31
blackburnsonney2k: ok I have no idea how to fix all the things in testsuite right now21:32
blackburnsonney2k: apart from testsuite it is ok21:32
@sonney2kwithout the testsuite working we cannot release though21:33
@sonney2kregressions are much more annoying than broken new features21:33
blackburnI am pretty sure we've got no new regressions21:34
-!- in3xes [~in3xes@210.212.58.168] has quit [Ping timeout: 240 seconds]21:35
blackburnsonney2k: in fact I receive ERROR for each test in python modular21:35
-!- harshit_ [~harshit@182.68.171.83] has joined #shogun21:37
blackburnsonney2k: do you?21:37
harshit_hi everyone, just one quick question is there any built in function in shogun to multiply two matrices.I found one for vectors in CMath but not for matrices21:39
blackburnuse cblas_dgemm21:39
harshit_is that a library or a function ?21:42
@sonney2kblackburn, compiling now...21:42
blackburnharshit_: BLAS function21:42
blackburnsonney2k: ok, please let me know cause I really see no OK there21:42
-!- puneetgoyal [~puneet@101.63.95.73] has quit [Ping timeout: 244 seconds]21:44
-!- puneetgoyal [~puneet@101.63.95.73] has joined #shogun21:45
harshit_can you please provide me its prototype. i am not able to find it21:45
blackburnhttp://www.netlib.org/blas/dgemm.f21:46
harshit_thanks21:46
-!- l0nr4n [~l0nr4n@unaffiliated/l0nr4n] has quit [Ping timeout: 265 seconds]21:48
-!- l0nr4n [~l0nr4n@g225108058.adsl.alicedsl.de] has joined #shogun21:49
-!- l0nr4n [~l0nr4n@g225108058.adsl.alicedsl.de] has quit [Changing host]21:49
-!- l0nr4n [~l0nr4n@unaffiliated/l0nr4n] has joined #shogun21:49
n4nd0sonney2k: hi21:54
n4nd0sonney2k: have you taken a look to the graphical multiclass qda example?21:54
n4nd0just wanted to know if you think it's ok or have any suggestion21:55
n4nd0you can check how it looks like here http://dl.dropbox.com/u/11020840/multiclass_qda.png21:56
@sonney2kn4nd0, yeah looks great :)22:00
n4nd0sonney2k: :)22:00
blackburnsonney2k: any news?22:11
-!- harshit_ [~harshit@182.68.171.83] has quit [Ping timeout: 250 seconds]22:22
-!- puneetgoyal [~puneet@101.63.95.73] has quit [Ping timeout: 264 seconds]22:27
-!- harshit_ [~harshit@182.68.171.83] has joined #shogun22:42
-!- harshit_ [~harshit@182.68.171.83] has quit [Client Quit]22:43
n4nd0good night people!22:43
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving]22:45
-!- gsomix [~gsomix@188.168.14.44] has quit [Ping timeout: 244 seconds]22:47
-!- novice [b4953264@gateway/web/freenode/ip.180.149.50.100] has joined #shogun23:01
-!- GenX [~Sambhav@14.139.82.6] has quit [Read error: Connection reset by peer]23:10
-!- GenX [~Sambhav@14.139.82.6] has joined #shogun23:10
-!- novice [b4953264@gateway/web/freenode/ip.180.149.50.100] has left #shogun []23:14
-!- blackburn [~qdrgsm@188.122.253.192] has quit [Quit: Leaving.]23:29
-!- in3xes [in3xes@49.248.244.12] has joined #shogun23:30
--- Log closed Mon Mar 19 00:00:19 2012