Open in new window / Try shogun cloud
--- Log opened Wed Aug 08 00:00:10 2012
--- Day changed Wed Aug 08 2012
@sonney2kso better keep it (besides it is easier...)00:00
n4nd0ok00:00
n4nd0anyway, I fill first fix the pull request you commented yesterday00:00
n4nd0and after that I will do another one for the plifs00:00
@sonney2kblackburn, there it is: [1]: <U+FEFF>Gretton, A., Fukumizu, K., Teo, C., & Song, L. (2008).00:01
blackburnhehe00:01
blackburnwhere does he get these nice chars?00:01
@sonney2kn4nd0, thanks00:01
@sonney2k0 width space!!!00:01
@sonney2kwhy did people invent that?00:01
n4nd0good night00:01
-!- n4nd0 [~nando@119.Red-2-137-47.dynamicIP.rima-tde.net] has quit [Quit: leaving]00:01
blackburnsonney2k: I need to pass indices to evaluation00:02
blackburnany idea?00:02
blackburnadding them to evaluate as optional paramers could be badd00:02
@sonney2kblackburn, I wish we did not had subsets as members in features but as parameter to get_feature_vector etc00:03
@sonney2kthis would save a lot of trouble00:03
blackburnsounds ineffective00:03
shogun-buildbot_build #272 of deb3 - modular_interfaces is complete: Failure [failed compile python_modular]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/272  blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com>00:03
@sonney2k(but requires changine *every* learning algorithm)00:03
@sonney2kblackburn, why?00:03
blackburnsonney2k: how to do subset stack stuff this way?00:04
@sonney2kblackburn, you give the stacked subset (that is precomputed elsewere)00:05
blackburnyeah but as you said this would require changing everytihng00:05
blackburnsonney2k: in multitask indexing became a pain for me00:06
@sonney2kblackburn, yeah it is horrible :/00:06
@sonney2kblackburn, so what do you propose? I mean features and labels support subset stacks00:07
@sonney2kso in evaluation you can directly use this00:07
blackburnI do not propose anything00:07
blackburn:)00:07
@sonney2kso no problem with indices - just use subset stacks00:07
blackburnI think I will add a method that is implemented only in multitask evaluations00:08
blackburnand NOP in others00:08
blackburncomputing the map from test indices to vector of vectors of indices00:08
blackburnsonney2k: I need to compute ROC of different tasks separately00:09
blackburnso with subset [0,4,5,6,14,15] and tasks [0,1,2,3,4]; [5,6,7,8];[14,15,16] I would need to compute [0,4]; [5,6]; [14,15]00:10
CIA-18shogun: Soeren Sonnenburg master * r5025ef1 / src/shogun/statistics/HSIC.h : remove zero width unicode whitespace causing build failures - https://github.com/shogun-toolbox/shogun/commit/5025ef152461e231d8ce3d2136a0783749067fc100:10
@sonney2ksounds ok00:12
shogun-buildbot_build #273 of deb3 - modular_interfaces is complete: Failure [failed compile java_modular]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/273  blamelist: Soeren Sonnenburg <sonne@debian.org>00:32
@sonney2kblackburn, that was easily fixed http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/273/steps/compile%20java_modular/logs/stdio00:34
@sonney2kblackburn, what was that macro again?00:35
blackburnsonney2k: SERIALIZABLE_DUMMY00:35
@sonney2kblackburn, can you do it?00:35
blackburnyeah sure00:35
@sonney2kyhx00:35
@sonney2kthx00:35
blackburnsonney2k: I do not undrestand why it has no serialization00:37
blackburnsonney2k: it is a different issue I think00:38
CIA-18shogun: Sergey Lisitsyn master * r5bc1a7d / src/shogun/multiclass/tree/RelaxedTree.h : Attempt to fix relaxed tree compilation - https://github.com/shogun-toolbox/shogun/commit/5bc1a7d0de026de3721e6f99f6d2eac2f3f73c5000:41
shogun-buildbot_build #266 of deb1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/266  blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com>00:46
CIA-18shogun: Sergey Lisitsyn master * r081ebb3 / (2 files): An another attempt to fix compilation - https://github.com/shogun-toolbox/shogun/commit/081ebb3cd0facbe7fd3e2697a786c0ff7fc620e300:50
shogun-buildbot_build #267 of deb1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/267  blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com>00:55
@sonney2kblackburn, hmmhh might be due to the templates that these functions are somehow hidden00:58
blackburnyeah00:58
blackburnsonney2k:00:59
blackburn../../shogun/multiclass/tree/RelaxedTree.h:32: Warning 401: Nothing known about base class 'CTreeMachine< RelaxedTreeNodeData >'. Ignored.00:59
blackburn../../shogun/multiclass/tree/RelaxedTree.h:32: Warning 401: Maybe you forgot to instantiate 'CTreeMachine< RelaxedTreeNodeData >' using %template.00:59
blackburnI have it here00:59
blackburnoh got it00:59
CIA-18shogun: Sergey Lisitsyn master * ra099c28 / (2 files in 2 dirs): Yet another attempt to fix compilation with relaxed tree - https://github.com/shogun-toolbox/shogun/commit/a099c287a9fda3c40332625ee2b1fe6a32e8d90e01:00
blackburnargh damn01:01
blackburnI do not understand it01:04
shogun-buildbot_build #268 of deb1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/26801:05
blackburnso did I solve that?01:06
blackburncrazy01:06
blackburnprobably I didn't01:07
blackburnit doesn't work here01:08
shogun-buildbot_build #274 of deb3 - modular_interfaces is complete: Failure [failed compile python_modular]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/274  blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com>01:09
blackburndamned puzzle01:15
blackburnI give up01:18
CIA-18shogun: Sergey Lisitsyn master * r23feaca / (2 files): Removed relaxed tree from modular - https://github.com/shogun-toolbox/shogun/commit/23feacaffca10e6a5825a3e2a10143e53399ff3301:24
shogun-buildbot_build #275 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/275  blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com>01:46
-!- av3ngr [~av3ngr@60-241-222-244.static.tpgi.com.au] has joined #shogun02:01
-!- av3ngr [~av3ngr@60-241-222-244.static.tpgi.com.au] has quit [Client Quit]02:01
-!- zxtx [~zv@c-71-227-187-90.hsd1.wa.comcast.net] has quit [Ping timeout: 246 seconds]02:05
-!- blackburn [~blackburn@109.226.80.43] has quit [Quit: Leaving.]02:48
shogun-buildbot_build #47 of nightly_none is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_none/builds/4703:11
shogun-buildbot_build #53 of nightly_default is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/5303:35
shogun-buildbot_build #43 of nightly_all is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_all/builds/4303:59
-!- n4nd0 [~nando@119.Red-2-137-47.dynamicIP.rima-tde.net] has joined #shogun08:36
-!- uricamic [~uricamic@2001:718:2:1634:8c0f:6f6f:d7f8:cf10] has joined #shogun08:44
-!- pluskid [~pluskid@66.85.151.216] has joined #shogun09:40
-!- n4nd0 [~nando@119.Red-2-137-47.dynamicIP.rima-tde.net] has quit [Ping timeout: 255 seconds]09:47
-!- n4nd0 [~nando@119.Red-2-137-47.dynamicIP.rima-tde.net] has joined #shogun09:49
-!- pluskid [~pluskid@66.85.151.216] has quit [Ping timeout: 246 seconds]10:05
-!- n4nd0 [~nando@119.Red-2-137-47.dynamicIP.rima-tde.net] has quit [Quit: leaving]10:12
-!- pluskid [~pluskid@111.120.34.161] has joined #shogun10:25
-!- pluskid [~pluskid@111.120.34.161] has quit [Ping timeout: 248 seconds]10:43
-!- pluskid [~pluskid@66.85.151.216] has joined #shogun10:43
-!- alexlovesdata [~binder@goldenezahl.ml.tu-berlin.de] has joined #shogun10:48
-!- pluskid [~pluskid@66.85.151.216] has quit [Ping timeout: 246 seconds]10:51
-!- heiko [~heiko@host86-178-85-194.range86-178.btcentralplus.com] has joined #shogun10:57
-!- heiko [~heiko@host86-178-85-194.range86-178.btcentralplus.com] has left #shogun []11:01
-!- pluskid [~pluskid@111.120.34.161] has joined #shogun11:07
CIA-18shogun: Soeren Sonnenburg master * r6dbd12d / src/shogun/statistics/KernelTwoSampleTestStatistic.h : add missing $ to fix doc - https://github.com/shogun-toolbox/shogun/commit/6dbd12df36d2b2fc8396fc7ffdf49c5efacd129311:15
CIA-18shogun: Soeren Sonnenburg master * r5ab7f26 / src/shogun/evaluation/CrossValidation.h : remove get_mean function (mean is public anyways) - https://github.com/shogun-toolbox/shogun/commit/5ab7f26f4f08c4ac9b5a2a18da7a15c6416b79a011:15
CIA-18shogun: Soeren Sonnenburg master * rcb99343 / src/shogun/features/AttributeFeatures.h : fix const mismatch in attribute features get_feature type/class size - https://github.com/shogun-toolbox/shogun/commit/cb99343a78bfff58faf0e8fed0edd5f151525b4211:15
-!- blackburn [~blackburn@109.226.80.43] has joined #shogun11:20
@sonney2kblackburn, about the c# compile failure11:20
blackburnsonney2k: yeah I see11:21
@sonney2kblackburn, in csharp / swig any public variable will get a get_variable() / set_variable() function11:21
@sonney2kso there is a clash11:21
blackburnyes I see11:21
@sonney2kwe cannot have a public get_mean() and a public mean varialbe11:21
blackburngood thing to have in python11:22
@sonney2kwhat?11:22
blackburnI feel tired adding getters/setters sometimes11:22
@sonney2kI think having getters setters instead of *any* public fields is preferable11:22
CIA-18shogun: Soeren Sonnenburg master * raa0bab7 / (29 files in 7 dirs): Merge pull request #692 from pluskid/fix-typo (+5 more commits...) - https://github.com/shogun-toolbox/shogun/commit/aa0bab77b4df81f3ef7d60db64835ecc360188f711:23
blackburnomg11:23
blackburnso euclidean now?11:24
@sonney2kblackburn, ?11:26
@sonney2kwant manhatten?11:26
blackburnsonney2k: no, I used to euclidian name :D11:27
blackburnbut actually it is kind of wrong11:27
-!- n4nd0 [~nando@119.Red-2-137-47.dynamicIP.rima-tde.net] has joined #shogun11:30
@sonney2kblackburn, heh - we had lots of other typos for years11:34
@sonney2klike mathmatics or os11:35
@sonney2k*so11:35
blackburnmathermatics11:35
blackburn:D11:35
@sonney2kand one really gets used to mattermattics :D11:35
@sonney2kpluskid, could you please fix these warnings in your code  http://shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/173/steps/compile/logs/warnings%20%2815%29 ?11:37
shogun-buildbot_build #276 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/276  blamelist: Soeren Sonnenburg <sonne@debian.org>11:37
pluskidsonney2k: ok11:37
@sonney2kthansk11:38
@sonney2kthansk11:38
@sonney2kthanks11:38
pluskidyuo aer welcmoe!11:39
blackburnwelcme!11:39
@sonney2khaha11:39
@sonney2kpluskid, actually there are tons of undocumented fields/classes/functions in your code too that the buildbot complains about herehttp://shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/276/steps/compile%20python_modular/logs/warnings%20%28381%2911:40
blackburnpluskid: theer weer isuse wtih reldaxed tere, hve u sene?11:40
@sonney2kpluskid, not all yours of course - blackburn also has his share :D11:40
blackburnsonney2k: I hazz11:41
blackburnich habe a lot of der issues11:41
pluskidhmm, I'll try handle those11:41
@sonney2kblackburn, I can see you BIG cat eyes wide open11:41
pluskidbut super frist11:41
@sonney2kblackburn, konetschno11:41
@sonney2kpluskid, super fooding!11:41
pluskid:D11:41
blackburnsonney2k: bezuslovno!11:42
@sonney2kmaybe we manage to create some new chinrusdeutschenglish11:42
@sonney2kI guess n4nd0 could even make things more interesting with swedspanchinrusdeutschenglish11:42
n4nd0lol11:43
blackburnsie pretty interesnyi SPRACHE11:43
@sonney2kn4nd0, you have to contribute some spanenglish word!11:44
@sonney2kn4nd0, and swenglish too!11:44
gsomix_sonney2k, Lables and SGVector work,  ta-dam...11:46
gsomix_*works11:46
blackburnwork11:46
@sonney2kgsomix_, great job!11:46
@sonney2kgsomix_, what was it?11:46
gsomix_sonney2k, strange python behavior11:47
-!- gsomix_ is now known as gsomix11:47
@sonney2kgsomix, could you explain more?11:47
blackburngsomix consider renaming to m_gsomix according to our code style11:47
gsomixsonney2k, yep. look, http://docs.python.org/dev/c-api/buffer.html#Py_buffer this is structure, that we fill, when want to use internal buffer.11:49
gsomixand only view->internal is guaranteed to remain constant11:50
@sonney2kok11:50
@sonney2kand now?11:50
gsomixanother pointers may be corrupted11:51
gsomixlike shape and strides11:51
gsomixthat I try to delete in releasebuffer11:51
gsomixboo11:51
gsomixm11:51
alexlovesdataprinting the draft of my thesis :)11:51
@sonney2kgsomix, ok but how do you know when to not delete these fields then?11:52
alexlovesdataok now I have time for you all again O:-)11:52
@sonney2kalexlovesdata, congrats11:52
n4nd0alexlovesdata: yeah, congrats! that's cool :)11:52
alexlovesdatalets see what my prof will say about it :D11:52
gsomixsonney2k, now I store this pointers in view->internal.11:53
gsomixugly but works11:54
blackburngsomix: how do you handle unreference?11:57
gsomixblackburn, nohow11:58
blackburnshouldbe11:58
blackburnsomehow11:58
gsomixmaybe it's not needed? ref and unref, I mean11:58
gsomixbecause we have pythno ref/unref11:59
gsomix*python11:59
blackburnI am unsure we can avoid leaks without ref/unref here11:59
blackburncan you put sgvector to the internal?11:59
blackburnor at least pointer to the sgvector12:00
gsomixyep, I can12:00
blackburnhmm then there could be a way12:01
blackburnleaving now12:01
-!- blackburn [~blackburn@109.226.80.43] has quit [Quit: Leaving.]12:01
@sonney2kgsomix, if you put sgvector in internal then all you have to do to free it is to overwrite it with an empty SGVector12:08
@sonney2kuricamic, please also have a look at http://shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/276/steps/compile%20python_modular/logs/warnings%20%28381%29 - there are a couple of fields / functions not documented in your code12:13
@sonney2k^ n4nd0 you too btw12:13
n4nd0sonney2k: ok12:13
shogun-buildbot_build #277 of deb3 - modular_interfaces is complete: Failure [failed test python_modular]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/277  blamelist: Soeren Sonnenburg <sonne@debian.org>12:17
CIA-18shogun: Soeren Sonnenburg master * r3bf7da4 / (doc/tutorial src/shogun/lib/slep/slep_solver.cpp): remove cleanup exit point - https://github.com/shogun-toolbox/shogun/commit/3bf7da4a5413ef2f263aab6f37de558631c1280c12:19
shogun-buildbot_build #278 of deb3 - modular_interfaces is complete: Failure [failed compile libshogun]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/278  blamelist: Chiyuan Zhang <pluskid@gmail.com>12:20
shogun-buildbot_build #174 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/174  blamelist: Soeren Sonnenburg <sonne@debian.org>, Chiyuan Zhang <pluskid@gmail.com>12:24
@sonney2kpluskid, ^ seen these buildbot failures? Seems you missed to convert a few includes / examples.12:25
pluskidsonney2k: oh, I forget about examples12:26
pluskidack -i euclidian | wc -l12:32
pluskid28412:32
pluskidI want to die...12:32
pluskidmaybe I should use sed12:32
uricamicsonney2k: ok12:49
n4nd0pluskid: I have just updated the PR with SGMatrixList refactoring12:51
n4nd0pluskid: ups sorry12:51
n4nd0that was for sonney2k ^12:52
n4nd0:D12:52
pluskidlet's name today the TYPO-day12:52
n4nd0hehe12:53
shogun-buildbot_build #279 of deb3 - modular_interfaces is complete: Failure [failed test libshogun]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/279  blamelist: Soeren Sonnenburg <sonne@debian.org>, Chiyuan Zhang <pluskid@gmail.com>12:54
-!- pluskid is now known as shogun-buildbot12:57
shogun-buildbot_build #230 of deb2 - static_interfaces is complete: Failure [failed test libshogun]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/230  blamelist: Soeren Sonnenburg <sonne@debian.org>, Chiyuan Zhang <pluskid@gmail.com>12:58
shogun-buildbotbuild #299 of deb3 - modular_interfaces is complete: Failure [failed test libshogun]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/279  blamelist: Shogun Buildbot <buildbot@shogun.com>12:59
n4nd0shogun-buildbot: your revenge?13:00
n4nd0shogun-buildbot_: dance baby13:00
shogun-buildbot_<(^.^<)13:00
shogun-buildbot_<(^.^)>13:00
shogun-buildbot_(>^.^)>13:00
shogun-buildbot_(7^.^)713:00
shogun-buildbot_(>^.^<)13:00
shogun-buildbotomg, I didn't know you can dance...13:00
n4nd0haha13:00
-!- shogun-buildbot is now known as pluskid13:01
shogun-buildbot_build #175 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/175  blamelist: Soeren Sonnenburg <sonne@debian.org>13:02
-!- n4nd0 [~nando@119.Red-2-137-47.dynamicIP.rima-tde.net] has quit [Quit: leaving]13:10
-!- gsomix is now known as m_gsomic13:15
-!- m_gsomic is now known as m_gsomix13:15
shogun-buildbot_build #280 of deb3 - modular_interfaces is complete: Failure [failed test libshogun]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/280  blamelist: Soeren Sonnenburg <sonne@debian.org>13:22
shogun-buildbot_build #231 of deb2 - static_interfaces is complete: Failure [failed test libshogun]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/231  blamelist: Soeren Sonnenburg <sonne@debian.org>13:25
CIA-18shogun: Chiyuan Zhang master * r5495ec0 / (2 files in 2 dirs): fix some warnings in relaxed tree. - https://github.com/shogun-toolbox/shogun/commit/5495ec0bd82ab51864b1bac4dc473c50becbded214:48
CIA-18shogun: Chiyuan Zhang master * rd80c8e9 / (91 files in 13 dirs): euclidian => euclidean in examples. - https://github.com/shogun-toolbox/shogun/commit/d80c8e9d4fb5a2c26860c1b79c6f66e6e2fb4b6114:48
CIA-18shogun: Soeren Sonnenburg master * rbe0b772 / (93 files in 15 dirs): Merge pull request #693 from pluskid/multiclass - https://github.com/shogun-toolbox/shogun/commit/be0b7726c8ee987a6487d87eb0335bad0cd9882d14:48
shogun-buildbot_build #176 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/176  blamelist: Chiyuan Zhang <pluskid@gmail.com>14:57
shogun-buildbot_build #177 of bsd1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/17715:03
shogun-buildbot_build #281 of deb3 - modular_interfaces is complete: Failure [failed test libshogun]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/281  blamelist: Chiyuan Zhang <pluskid@gmail.com>15:15
pluskidshogun-buildbot_: dance baby15:16
shogun-buildbot_<(^.^<)15:16
shogun-buildbot_<(^.^)>15:16
shogun-buildbot_(>^.^)>15:16
shogun-buildbot_(7^.^)715:16
shogun-buildbot_(>^.^<)15:16
pluskidgood15:16
shogun-buildbot_build #232 of deb2 - static_interfaces is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/23215:21
shogun-buildbot_build #282 of deb3 - modular_interfaces is complete: Failure [failed test python_modular]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/282  blamelist: Soeren Sonnenburg <sonne@debian.org>, Chiyuan Zhang <pluskid@gmail.com>15:45
-!- pluskid [~pluskid@111.120.34.161] has quit [Quit: Leaving]15:54
-!- n4nd0 [~nando@119.Red-2-137-47.dynamicIP.rima-tde.net] has joined #shogun16:04
-!- n4nd0 [~nando@119.Red-2-137-47.dynamicIP.rima-tde.net] has quit [Quit: leaving]16:32
-!- ckwidmer [8ca3fe9d@gateway/web/freenode/ip.140.163.254.157] has quit [Quit: Page closed]16:42
-!- uricamic [~uricamic@2001:718:2:1634:8c0f:6f6f:d7f8:cf10] has quit [Quit: Leaving.]17:25
-!- blackburn [~blackburn@109.226.80.43] has joined #shogun17:52
-!- blackburn is now known as cblackburn17:57
-!- cblackburn is now known as CBlackBurn17:59
-!- yoo [575b08cb@gateway/web/freenode/ip.87.91.8.203] has joined #shogun18:03
yoohi all18:03
-!- heiko [~heiko@host86-178-85-194.range86-178.btcentralplus.com] has joined #shogun18:04
-!- yoo [575b08cb@gateway/web/freenode/ip.87.91.8.203] has quit [Client Quit]18:06
CIA-18shogun: Heiko Strathmann master * rc9ba8b6 / (3 files): added debug messages - http://git.io/zlsq5w18:26
CIA-18shogun: Heiko Strathmann master * r9296faf / examples/undocumented/libshogun/statistics_hsic.cpp : added test for bootstrapping - http://git.io/Ag3N_w18:26
CIA-18shogun: Heiko Strathmann master * r5bb8506 / (4 files in 2 dirs): Merge pull request #694 from karlnapf/master - http://git.io/EG3hlQ18:26
-!- Netsplit *.net <-> *.split quits: shogun-buildbot_, octopine18:33
-!- Netsplit over, joins: shogun-buildbot_, octopine18:34
-!- Netsplit *.net <-> *.split quits: shogun-buildbot_, octopine18:45
-!- Netsplit over, joins: shogun-buildbot_, octopine18:45
-!- alexlovesdata [~binder@goldenezahl.ml.tu-berlin.de] has left #shogun []18:47
shogun-buildbot_build #283 of deb3 - modular_interfaces is complete: Failure [failed test python_modular]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/283  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>18:56
-!- n4nd0 [~nando@119.Red-2-137-47.dynamicIP.rima-tde.net] has joined #shogun19:12
n4nd0I am willing to see how many times shogun-buildbot_ is going to blame me when the PR is merged :D19:18
@sonney2kn4nd0, well one usually should not do any merges until the bb is good19:21
@sonney2kbtw, heiko please don't insert zero width unicode spaces in your doc :D19:21
heikosonney2k, did I do this again? :)19:22
heikoits my evil second identity that does these things :D19:22
n4nd0sonney2k: what is the bb?19:23
@sonney2kbuildbot19:24
n4nd0all right19:25
@sonney2kheiko, btw a few warnings are also on you http://shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/283/steps/compile%20python_modular/logs/warnings%20%28381%2919:25
@sonney2kdoes anyone know where puffin is?19:25
n4nd0no idea19:25
@sonney2khe is the #1 warning producer currently19:25
heikosonny2k, ok thanks for noticing, I will start on warnings tomorrow19:26
shogun-buildbot_build #284 of deb3 - modular_interfaces is complete: Failure [failed test python_modular]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/284  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>19:26
n4nd0the internet connection I have here gets me on my nerves19:27
CBlackBurnn4nd0: imagine you are a martian19:31
-!- heiko [~heiko@host86-178-85-194.range86-178.btcentralplus.com] has quit [Quit: Leaving.]19:33
n4nd0CBlackBurn: haha I think this interplanetary internet connection NASA is about use is better than mine's19:33
-!- shogun-t1olbox [~shogun@7nn.de] has quit [Ping timeout: 245 seconds]19:33
--- Log closed Wed Aug 08 19:33:56 2012
--- Log opened Wed Aug 08 19:34:09 2012
-!- shogun-toolbox [~shogun@7nn.de] has joined #shogun19:34
-!- Irssi: #shogun: Total of 10 nicks [1 ops, 0 halfops, 0 voices, 9 normal]19:34
!moorcock.freenode.net [freenode-info] channel flooding and no channel staff around to help? Please check with freenode support: http://freenode.net/faq.shtml#gettinghelp19:34
-!- Irssi: Join to #shogun was synced in 7 secs19:34
-!- ypp [575b08cb@gateway/web/freenode/ip.87.91.8.203] has joined #shogun19:40
CIA-18shogun: Soeren Sonnenburg master * r47a593d / (2 files): add tree business to modular interfaces - http://git.io/ppcQpw19:43
@sonney2k%rename(BlackBurn) CBlackBurn;19:43
@sonney2km_gsomix, how about going global?19:44
@sonney2kg_gsomix :D19:44
@sonney2km_gsomix, I guess you will send a PR tonight now that all problems are resolved?19:44
-!- CBlackBurn is now known as BlackBurn19:45
-!- BlackBurn is now known as blackburn19:45
-!- n4nd0 [~nando@119.Red-2-137-47.dynamicIP.rima-tde.net] has quit [Ping timeout: 252 seconds]19:48
-!- puffin444 [62e3926e@gateway/web/freenode/ip.98.227.146.110] has joined #shogun19:53
-!- Netsplit *.net <-> *.split quits: m_gsomix19:55
-!- Netsplit over, joins: m_gsomix19:58
shogun-buildbot_build #285 of deb3 - modular_interfaces is complete: Failure [failed compile java_modular]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/285  blamelist: Soeren Sonnenburg <sonne@debian.org>20:03
blackburnsonney2k: why bigger C is harder to solve but bigger L1 regularization is easier?20:59
-!- gsomix_ [~gsomix@178.45.75.43] has joined #shogun21:16
-!- m_gsomix [~gsomix@178.45.40.178] has quit [Ping timeout: 252 seconds]21:19
-!- gsomix__ [~gsomix@109.169.158.254] has joined #shogun21:28
-!- puffin444 [62e3926e@gateway/web/freenode/ip.98.227.146.110] has quit [Quit: Page closed]21:29
-!- gsomix_ [~gsomix@178.45.75.43] has quit [Ping timeout: 268 seconds]21:32
-!- ypp [575b08cb@gateway/web/freenode/ip.87.91.8.203] has quit [Quit: Page closed]21:40
@sonney2kblackburn, what is bigger L1?21:41
blackburnsonney2k: well lambda with L1 or anything21:42
@sonney2kblackburn, lambda == 1/C ?21:44
blackburnno, just C analogue21:45
blackburnsonney2k: in other words in logistic regression when I regularize more it goes faster21:47
blackburnsonney2k: I am curious why SVMs are slower with stronger regularization21:48
blackburnin L1 logit regression21:48
@sonney2kno idea why L1 gets faster - maybe solution is more sparse?21:49
blackburnyeah it is21:49
@sonney2kblackburn, with larger C I meant21:51
blackburnsonney2k: I am trying to find an interpretation - is larger C a tighter solution space?21:52
blackburnsearch space I mean21:53
CIA-18shogun: Soeren Sonnenburg master * rd1d2d12 / (3 files in 2 dirs): mention BaseMulticlassMachine in Relaxed tree to get python_modular - http://git.io/q_rIag22:01
@sonney2kblackburn, for SVM alphas can range from 0...C (so larger space)22:03
blackburnsonney2k: ahh I see22:04
@sonney2klets thing loudly22:06
@sonney2kthe higher C the less training errors are tolerated22:07
blackburnlets think?22:07
@sonney2kthink!22:07
@sonney2kso it absolutely tries to have 0 training error22:07
blackburnyeah that's clear22:07
blackburnstop I am idiot22:07
blackburnyour lambda=1/C was true :D22:07
blackburnI kept wrong thing in mind22:08
blackburnin logistic regression I worked with I increase regularization so no surprise it works faster22:08
blackburnbecause it becomes pretty tolerant to errors :D22:08
blackburntotally tolerant at some point22:08
@sonney2kbah :/22:09
blackburnsonney2k: probably larger C = more time holds for every loss+regularizer algo22:10
-!- gsomix__ [~gsomix@109.169.158.254] has quit [Quit: Ex-Chat]22:23
@sonney2kblackburn, I wouldn't say it is always like this - sometimes larger C creates a much simpler / easier solution (sparser...) and things can go quicker but in general yes22:28
@sonney2kwhat you say22:28
shogun-buildbot_build #286 of deb3 - modular_interfaces is complete: Failure [failed compile java_modular]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/286  blamelist: Soeren Sonnenburg <sonne@debian.org>22:29
@sonney2kblackburn, do you need this?22:29
blackburnI see22:29
@sonney2khttp://shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/181/steps/compile/logs/warnings%20%281%2922:29
blackburnno crap22:29
blackburnI will remove22:29
@sonney2kblackburn, can you again help with this too http://shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/286/steps/compile%20java_modular/logs/stdio22:30
@sonney2kneeds your macro again22:30
@sonney2kpluskids code is causing quite some headaches this time :/22:30
blackburnyeah22:31
@sonney2kwiking, so?22:31
blackburnsonney2k: so bad in java we can't come with a solution without  copying22:32
@sonney2kblackburn, yeah there is no way22:32
@sonney2kblackburn, even jblas copies...22:32
@sonney2kblackburn, I would already wish we had some way of getting serialization to work for java w/o the exceptions ...22:35
blackburnsonney2k: what do you mean w/o exceptions?22:35
blackburnsonney2k: it fails because of non sgobjects exposed to interface22:40
blackburnactually it is a *test*22:40
blackburn:D22:40
CIA-18shogun: Sergey Lisitsyn master * r05719ba / src/interfaces/modular/modshogun_ignores.i : Added a few non-sgobject ignores to modular - http://git.io/4Hr5LA22:43
-!- n4nd0 [~nando@31.Red-2-137-62.dynamicIP.rima-tde.net] has joined #shogun22:47
wikingwellsonfixed22:52
wikingsonney2k: fixed22:53
CIA-18shogun: Viktor Gal master * r5e2d783 / src/shogun/io/SGIO.h : Mountain Lion changed API of scandir - http://git.io/xScUbA22:55
CIA-18shogun: Viktor Gal master * r2cebf0a / src/shogun/io/SGIO.h : check for osx version a little bit smarter - http://git.io/1HCzSg22:55
CIA-18shogun: Soeren Sonnenburg master * rf342b7d / src/shogun/io/SGIO.h : Merge pull request #695 from vigsterkr/master - http://git.io/zGrJOA22:55
@sonney2kwiking, thx - hope you are otherwise progressing well with your so latent stuff?22:55
wikingsonney2k: yeah crazy times... wanna get everything PR-ed by this friday22:56
wikingso that we can have it a) for pencil down b) next relase22:56
wikingrls22:56
@sonney2kwiking, I need to add - you too have some warnings to fix... http://shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/287/steps/compile%20python_modular/logs/stdio22:56
wikingbut just as u seen22:56
@sonney2k(like everyone here...)22:56
wikingi had now a dist-upgrade22:56
@sonney2kwiking, well good luck - I hope you make it22:57
wikingthat broke a bit my system for a while22:57
wikingbut now it's back22:57
@sonney2k388 warnings - and we need to get it down to 022:59
shogun-buildbot_build #287 of deb3 - modular_interfaces is complete: Failure [failed compile java_modular]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/287  blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com>23:03
@sonney2kblackburn, ^ bah - I think you need more ignore / other magic23:04
blackburnsonney2k: I think it requires the magic you added23:04
n4nd0sonney2k: do you have any reference or do you know something about risk functions for hm-svm?23:08
n4nd0my computer is going to die in Spain during these holidays :D23:10
n4nd0blackburn: I just issued a dumb PR with a few warnings fixed23:14
blackburnnice23:14
shogun-buildbot_build #288 of deb3 - modular_interfaces is complete: Failure [failed compile java_modular]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/288  blamelist: Viktor Gal <viktor.gal@maeth.com>23:29
CIA-18shogun: iglesias master * r8c1eeea / (7 files in 4 dirs): * minor fixes for doc warnings - http://git.io/5edJYA23:35
CIA-18shogun: Sergey Lisitsyn master * rac8c5ba / (7 files in 4 dirs): Merge pull request #696 from iglesias/master - http://git.io/5Elhpg23:35
-!- n4nd0 [~nando@31.Red-2-137-62.dynamicIP.rima-tde.net] has quit [Quit: leaving]23:40
shogun-buildbot_build #289 of deb3 - modular_interfaces is complete: Failure [failed compile java_modular]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/289  blamelist: Soeren Sonnenburg <sonne@debian.org>, Viktor Gal <viktor.gal@maeth.com>23:46
--- Log closed Thu Aug 09 00:00:17 2012