Open in new window / Try shogun cloud
--- Log opened Thu Aug 01 00:00:56 2013
-!- travis-ci [~travis-ci@ec2-23-20-234-231.compute-1.amazonaws.com] has joined #shogun00:02
travis-ci[travis-ci] it's Sergey Lisitsyn's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/970776000:02
-!- travis-ci [~travis-ci@ec2-23-20-234-231.compute-1.amazonaws.com] has left #shogun []00:02
-!- travis-ci [~travis-ci@ec2-50-19-131-239.compute-1.amazonaws.com] has joined #shogun00:30
travis-ci[travis-ci] it's Soeren Sonnenburg's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/970872500:30
-!- travis-ci [~travis-ci@ec2-50-19-131-239.compute-1.amazonaws.com] has left #shogun []00:30
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]01:04
shogun-buildbotbuild #1443 of deb3 - modular_interfaces is complete: Failure [failed test octave_modular]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1443  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>01:22
-!- iglesiasg [~Fernando@s83-179-44-135.cust.tele2.se] has quit [Quit: Leaving]02:13
-!- foulwall` [~user@110.17.3.72] has joined #shogun02:30
-!- foulwall [~user@110.17.4.161] has quit [Ping timeout: 246 seconds]02:32
-!- zxtx [~zv@rrcs-76-79-81-162.west.biz.rr.com] has quit [Ping timeout: 240 seconds]05:12
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun05:43
shogun-notifier-shogun: Viktor Gal :feature/CMake * 1ee5198 / / (12 files): https://github.com/shogun-toolbox/shogun/commit/1ee519836e9d0171df7ec77f9cdc8404660e569205:43
shogun-notifier-shogun: Refactor compile flags setting05:43
shogun-notifier-shogun: add libshogun and python_modular tests via ctest05:43
shogun-notifier-shogun: Travis: change script and add memory check target with ctest05:43
-!- travis-ci [~travis-ci@ec2-50-19-131-239.compute-1.amazonaws.com] has joined #shogun05:47
travis-ci[travis-ci] it's Viktor Gal's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/972061405:47
-!- travis-ci [~travis-ci@ec2-50-19-131-239.compute-1.amazonaws.com] has left #shogun []05:47
shogun-notifier-shogun: Viktor Gal :feature/CMake * 99390eb / / (4 files): https://github.com/shogun-toolbox/shogun/commit/99390eba71e8d02af7176cbb30e8db6488db15ad05:51
shogun-notifier-shogun: Add CMakeLists.txt for examples05:51
shogun-notifier-shogun: Viktor Gal :feature/CMake * 0aae6ea / CMakeLists.txt/ (3 files): https://github.com/shogun-toolbox/shogun/commit/0aae6ea0b0ab925871fabb5a3ce63648944e77ee05:56
shogun-notifier-shogun: Minor fixes in cmake files05:56
-!- travis-ci [~travis-ci@ec2-23-20-234-231.compute-1.amazonaws.com] has joined #shogun05:57
travis-ci[travis-ci] it's Viktor Gal's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/972072905:57
-!- travis-ci [~travis-ci@ec2-23-20-234-231.compute-1.amazonaws.com] has left #shogun []05:57
-!- travis-ci [~travis-ci@ec2-23-22-12-85.compute-1.amazonaws.com] has joined #shogun06:03
travis-ci[travis-ci] it's Viktor Gal's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/972080206:03
-!- travis-ci [~travis-ci@ec2-23-22-12-85.compute-1.amazonaws.com] has left #shogun []06:03
shogun-notifier-shogun: Viktor Gal :feature/CMake * 79e442f / / (3 files): https://github.com/shogun-toolbox/shogun/commit/79e442fcf286aa1be14cb3af5913830a6945643006:30
shogun-notifier-shogun: Fix version date macros in versionstring.h06:30
shogun-notifier-shogun: fix ld_library_path env setting in python_modular example06:30
-!- travis-ci [~travis-ci@ec2-50-19-131-239.compute-1.amazonaws.com] has joined #shogun06:46
travis-ci[travis-ci] it's Viktor Gal's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/972126906:46
-!- travis-ci [~travis-ci@ec2-50-19-131-239.compute-1.amazonaws.com] has left #shogun []06:46
-!- lisitsyn [~lisitsin@mxs.kg.ru] has joined #shogun07:16
shogun-notifier-shogun: Viktor Gal :feature/CMake * c2a6966 / examples/undocumented/libshogun/CMakeLists.txt: https://github.com/shogun-toolbox/shogun/commit/c2a6966c2da3020311e5b3b23bf7f34d32ece31807:16
shogun-notifier-shogun: fix libshogun examples' working directory07:16
-!- travis-ci [~travis-ci@ec2-23-20-234-231.compute-1.amazonaws.com] has joined #shogun07:31
travis-ci[travis-ci] it's Viktor Gal's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/972205507:31
-!- travis-ci [~travis-ci@ec2-23-20-234-231.compute-1.amazonaws.com] has left #shogun []07:31
shogun-notifier-shogun: Viktor Gal :feature/CMake * 2660db7 / tests/unit/CMakeLists.txt: https://github.com/shogun-toolbox/shogun/commit/2660db7faf8f9736f871b581253d6ab8570c0f7c07:43
shogun-notifier-shogun: Build unit tests with 'make' if unit testing is enabled07:43
-!- travis-ci [~travis-ci@ec2-23-20-234-231.compute-1.amazonaws.com] has joined #shogun07:59
travis-ci[travis-ci] it's Viktor Gal's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/972275007:59
-!- travis-ci [~travis-ci@ec2-23-20-234-231.compute-1.amazonaws.com] has left #shogun []07:59
shogun-notifier-shogun: Viktor Gal :feature/CMake * 5ab4f9c / .travis.yml,CMakeLists.txt: https://github.com/shogun-toolbox/shogun/commit/5ab4f9c6ddd1aaa6335e702dba8de717ab7edd0a08:02
shogun-notifier-shogun: Add ENABLE_TESTING option to cmake08:02
shogun-notifier-shogun: ctest should be enabled with ENABLE_TESTING regardless08:02
shogun-notifier-shogun: whether GMOCK or GTEST framework has been found08:02
-!- travis-ci [~travis-ci@ec2-23-22-12-85.compute-1.amazonaws.com] has joined #shogun08:19
travis-ci[travis-ci] it's Viktor Gal's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/972316608:19
-!- travis-ci [~travis-ci@ec2-23-22-12-85.compute-1.amazonaws.com] has left #shogun []08:19
-!- foulwall` [~user@110.17.3.72] has quit [Remote host closed the connection]08:57
shogun-notifier-shogun: Viktor Gal :feature/CMake * 9da75b5 / .travis.yml: https://github.com/shogun-toolbox/shogun/commit/9da75b5a21494d6ea8fae7a30081c2c4ef63352609:57
shogun-notifier-shogun: Use ctest instead of 'make test'09:57
shogun-notifier-shogun: This way we can provide error output of failed test with the --output-on-failure command line arg09:57
@wikingyey09:57
@wikinggetting there09:57
hushellgsomix: hi! Would you mind to have a function T& CMap::operator [](key)?10:06
gsomixhushell, yeah. it might be useful.10:12
hushellgsomix: I'll just copy the code from CMap::get_element()10:13
gsomixhushell, ok, I'll do it either today or tomorrow.10:13
gsomixhushell, ah, ok. :)10:13
hushellgsomix: I tested it, works well :)10:13
-!- travis-ci [~travis-ci@ec2-50-19-131-239.compute-1.amazonaws.com] has joined #shogun10:15
travis-ci[travis-ci] it's Viktor Gal's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/972538810:15
-!- travis-ci [~travis-ci@ec2-50-19-131-239.compute-1.amazonaws.com] has left #shogun []10:15
gsomixhushell, cool! I really don't like my CMap - I think it might contains some not effective memory usage. I'm about access by index and array for nodes.10:17
gsomixit works, but if you found that sth wrong with it - let me know10:18
-!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has joined #shogun10:25
-!- hoijui [~hoijui@adsl-84-227-37-53.adslplus.ch] has joined #shogun10:32
hushellgsomix: sure. I really found Map is useful when we need maintaining a lookup table, but I've been told STL is not allowed for class members10:33
hushellgsomix: we could improve CMap from practice :)10:33
-!- foulwall` [~user@110.17.3.72] has joined #shogun10:41
-!- van51 [~van51@ppp-94-66-76-131.home.otenet.gr] has joined #shogun10:44
-!- thoralf [~thoralf@enki.zib.de] has joined #shogun10:51
thoralfHello everybody.10:54
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun11:28
-!- HeikoS [~heiko@nat-181-93.internal.eduroam.ucl.ac.uk] has joined #shogun11:47
-!- mode/#shogun [+o HeikoS] by ChanServ11:47
-!- gsomix_ [~gsomix@95.67.164.224] has joined #shogun11:50
-!- gsomix [~gsomix@109.169.241.214] has quit [Read error: Connection reset by peer]11:50
-!- iglesiasg [~iglesias@2001:6b0:1:1da0:1464:9f67:69f9:4aba] has joined #shogun11:52
-!- mode/#shogun [+o iglesiasg] by ChanServ11:53
@iglesiasghello hello people11:54
@HeikoSiglesiasg: hi!11:55
@HeikoSvotjakovr: thanks for your comments! :)11:55
votjakovrhi everyone!11:55
votjakovrHeikoS: you are welcome!11:55
votjakovrHeikoS: btw check travis-ci build for your PR11:56
@HeikoSvotjakovr: yes I saw it11:56
votjakovrHeikoS: there is something wrong11:56
@HeikoSvotjakovr: its weird, some eigen issues11:56
@HeikoSvotjakovr: but first let me squash the commits ....11:57
votjakovrHeikoS: anyway this is very useful features, thank you :)11:57
@HeikoSvotjakovr: thank YOU, since I used your code :)11:58
@HeikoSand I literally need this feature *this* week ;)11:58
@HeikoSwe might go a bit in this direction after GSoC if you are interested, there is lots of cool stuff waiting to be discovered :)11:59
votjakovrHeikoS: yeah, i see12:00
@HeikoSanyone here know how to use git? :)12:06
@HeikoSbecause I dont12:06
votjakovrHeikoS: i'm not sure, but what is the problem?12:08
@HeikoSvotjakovr: dont worry I just found what I wanted, I always fight with git because I am so unpatient :)12:09
thoralfHeikoS: Maybe? ;)12:18
-!- van51 [~van51@ppp-94-66-76-131.home.otenet.gr] has left #shogun ["PING 1375353433"]12:37
-!- hoijui [~hoijui@adsl-84-227-37-53.adslplus.ch] has quit [Quit: Leaving]12:37
@wiking\o/12:49
shogun-notifier-shogun: Viktor Gal :feature/CMake * 6ab1015 / / (6 files): https://github.com/shogun-toolbox/shogun/commit/6ab1015c928b57c50f9f047cca2c0f9ee21e49f512:49
shogun-notifier-shogun: Add integration test for python_modular12:49
@wikingthis one is about to be finished12:49
@iglesiasgwiking, cool!12:49
hushellis it possible to disable a default constructor for a class derived from SGObject?12:50
@wikingHeikoS: this is your friend: ctest -D MemoryCheck12:50
@HeikoSwiking: what does it do? :)12:50
@wikingHeikoS: it runs valgrind on all the tests... out of box. i do not fucking have to do anything just define the tests...12:51
@iglesiasghushell, what do you mean with disabling it?12:51
@HeikoSwiking: nice! :)12:51
@HeikoSwiking: add to the makefile and make it default!12:51
@wikingHeikoS: nono... this is cmake12:51
@HeikoSah12:51
@wikingHeikoS: i'm not adding this to makefile ever12:51
@iglesiasghushell, I saw that in FactorGraph you did at least one default constructor that basically raised an error, would that be enough?12:51
@wikingHeikoS: but today cmake env is going to be finished12:51
@HeikoSwiking: would be cool to have unit tests valgrinded by default12:51
@HeikoSwiking: nice!12:51
@iglesiasgin the FactorGraph PR, I mean12:51
@wikingsooo we could merge ;)12:51
hushelliglesiasg: I don't want to have a default constructor, but make it private is not allowed12:52
@HeikoSwiking: I have some work to do, but will look into the ipython stuff later12:52
@wikingHeikoS: well yeah unit tests are part of tests ;)12:52
@HeikoSwiking: next thing that is urgent12:52
@wikingHeikoS: okok12:52
@iglesiasghushell, what if you just don't do the default constructor? If there exists another constructor, if the default is not explicitly implemented then it does not exist (IIRC)12:53
hushelliglesiasg: emm, no this will cause some unit tests fail12:53
@wikingHeikoS: and of course this would be great to have: http://cmake.org/Wiki/CTest/Coverage12:53
hushelliglesiasg: but if someone call it by Class(), will be problematic12:54
@HeikoSwiking: wow that would be awesome to have12:54
@wikingHeikoS: it works with gcov12:54
@HeikoSthen we can start dropping code since we know what is tested and what isnt :)12:54
@wikinganyhow again this can be done by simple 2 liner in cmake ;)12:54
@wikingso it's an urgent fucking thing to switch to cmake12:55
@wikingah and one more thing12:55
@wikingthe way i'm going to test cmake12:55
@wikingis to enable remote reporting ;P12:55
@iglesiasghushell, well that is what you get if you want to disable the default constructor :D12:55
@wikingso once it's ready i'm going to announce to test it on the mailing list12:55
@HeikoSwiking: I totally agree12:55
@wikingand whoever runs the cmake build12:55
@wikingautomatically will submit the buildlog to a site12:55
@wiking:P12:55
@wikinghence i can analyse if something is went wrong12:56
@wikingand nobody has to copypaste shit12:56
hushelliglesiasg: okay, I need mention in the doc that never use default constructor :/12:56
@wikingi'll register here shogun: http://open.cdash.org/index.php?project=PublicDashboard#12:56
@wikingso that we dont have to run our own cdash server12:57
@iglesiasghushell, mmm but can't we make some sort of workaround?12:57
@wikinglisitsyn: ping12:57
@iglesiasghushell, I mean, why should it be not possible to use the default constructor? What is the scenario?12:57
lisitsynwiking: ja12:57
lisitsynwhat's up?12:57
@wikinglisitsyn: did u find out about the octave external lib path?12:57
lisitsynwiking: oh sorry I didn't manage to find anything on that12:58
@wikingwhen u have time could u look at it plz?12:58
lisitsynwiking: yes this evening12:59
@wikinglisitsyn: cool thnx12:59
@wikinglisitsyn: see ctest in action: https://travis-ci.org/shogun-toolbox/shogun/jobs/9730469 (scroll to the bottom)13:00
@wikinglisitsyn: the add_test(name ....) worked nicely13:00
hushelliglesiasg: for example, I have a CDisjointSet, which need to specify the num of elements at very beginning13:00
@iglesiasgwiking, the remote reporting thing sounds awesome13:00
@HeikoSwiking: nice!13:01
@iglesiasghushell, so that number of elements has to be specified because when calling other method of CDisjointSet it must be something, right?13:01
hushelliglesiasg: yes, if the number is nothing, everything goes wrong I guess13:02
@iglesiasghushell, what if the default constructor sets it to 0, -1 or a sentinel value; you provide a setter for the number of elements; and in the methods that need it to be something you ensure that this number is something13:02
@iglesiasghushell, so you cool have two use cases:13:03
hushelliglesiasg: this is I am doing set 1 for default case13:03
hushelliglesiasg: but doesn't make sense indeed13:03
@iglesiasghushell, well then just make sure that if this value has not been set in the constructor or by the setter, the stuff that you guess could go wrong is not executed13:04
@iglesiasgwith a REQUIRE or if clause with SG_ERROR13:04
@iglesiasghushell, what do you think?13:06
hushelliglesiasg: sounds good, set -1 for default constructor and check if > 0 when calling other functions13:06
@iglesiasghushell, yep13:06
hushellThanks for the REQUIRE! :)13:07
shogun-notifier-shogun: Roman Votyakov :develop * 8f65b03 / examples/undocumented/libshogun/ (2 files): https://github.com/shogun-toolbox/shogun/commit/8f65b0370c6378f74e28dcee068023ee75322f6913:10
shogun-notifier-shogun: add very basic libshogun example of GP binary classification13:10
shogun-notifier-shogun: Heiko Strathmann :develop * 5caa7db / examples/undocumented/libshogun/ (2 files): https://github.com/shogun-toolbox/shogun/commit/5caa7db6ccec7f40f6064e58c3a6af4ea5c294b013:10
shogun-notifier-shogun: Merge pull request #1344 from votjakovr/feature/gp_binary_classification13:10
shogun-notifier-shogun:13:10
shogun-notifier-shogun: Add very basic libshogun example of GP binary classification13:10
@iglesiasgHeikoS, wow you are in super-hacker mode these days!13:10
@HeikoSiglesiasg: need this stuff here :)13:10
@HeikoSpython implementation was too slow ;)13:11
@iglesiasgHeikoS, for job?13:11
@HeikoSits now 20 times faster13:11
@HeikoSyes13:11
@iglesiasgHeikoS, that is pretty awesome, that you can use these things for your job13:11
@HeikoSyeah indeed13:11
@HeikoSshogun is starting to get useful13:11
@HeikoSsince I can rely on things now13:11
@iglesiasg:)13:11
@HeikoSdoesnt always have to make sure they work correct13:11
@HeikoSiglesiasg: where can I see the travis for the PR cannot find it13:12
@HeikoSalready re-sent13:12
@iglesiasgHeikoS, it seems travis didn't trigger yet13:13
@HeikoSok lets wait then, there were weird eigen3 issues13:13
@iglesiasgHeikoS, https://travis-ci.org/shogun-toolbox/shogun/builds13:13
@iglesiasgmaybe when those finish13:14
-!- travis-ci [~travis-ci@ec2-23-22-12-85.compute-1.amazonaws.com] has joined #shogun13:14
travis-ci[travis-ci] it's Viktor Gal's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/973046613:14
-!- travis-ci [~travis-ci@ec2-23-22-12-85.compute-1.amazonaws.com] has left #shogun []13:14
shogun-buildbotbuild #1136 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1136  blamelist: Roman Votyakov <votjakovr@gmail.com>13:15
shogun-buildbotbuild #1137 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1137  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>14:02
-!- travis-ci [~travis-ci@ec2-54-235-60-216.compute-1.amazonaws.com] has joined #shogun14:36
travis-ci[travis-ci] it's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/973107214:36
-!- travis-ci [~travis-ci@ec2-54-235-60-216.compute-1.amazonaws.com] has left #shogun []14:36
@HeikoSvotjakovr: any idea why the travis breaks?14:58
@HeikoShttps://travis-ci.org/shogun-toolbox/shogun/jobs/973103814:58
@HeikoSdid you have a similar issue before?14:58
@HeikoSwiking: which eigen version is used on travis, how can I find out?15:00
@wikingHeikoS: it's there15:00
@wikingin the logs15:00
@HeikoSwhere do I have to look?15:01
@wikingSetting up libeigen3-dev (3.0.5-1) ...15:01
@HeikoSah15:01
@wikingwhy we need newer?15:01
@wiking:)15:01
@HeikoSthanks15:02
@HeikoSwiking: yes :)15:02
@HeikoSplease15:02
@HeikoSalthough its bad to need 3.115:02
@wikingmmm well travis will have only that15:02
@HeikoSwhat do you think? Can we make that a dependency?15:02
@HeikoS3.1?15:02
@wikingwell how old is 3.1?15:02
@wikingi mean travis is a simple ubuntu15:02
@wikingnot the latest i think it's LTS15:02
@HeikoS3.2 is out already15:02
@HeikoSstable15:02
@wikingbut still ubuntu is quite frequently updated15:02
@wikingimagine somebody comes with a centos15:03
@wiking100% that he will not have eigen 3.1 within 2 years :P15:03
@HeikoS3.1 was release june 201215:03
@wikingi'm just saying...15:03
@HeikoSmmmh15:03
@wikingwell we can still make it dependent15:03
@HeikoSmmmmmh15:03
@HeikoSwiking: sorry gotta go15:03
@wikingand in worst case cmake can download it for shogun :P15:04
@wikingnw15:04
thoralfwiking: eigen3 3.1.0~beta1-1ubuntu1 was already contained in Queztal (2012-06-13)15:07
-!- van51 [~van51@ppp-94-66-76-131.home.otenet.gr] has joined #shogun15:12
@iglesiasghey thoralf!15:23
@iglesiasgthoralf, I've got a question for you :)15:23
thoralfiglesiasg: Aloah.15:23
@iglesiasgthoralf, you were using eclipse with shogun, right?15:23
thoralfiglesiasg: Yes and no - since last friday it's crashing all the time and I stopped using it. ;)15:24
thoralfiglesiasg: But maybe I can help. ;)15:24
@iglesiasgthoralf, oh I see15:24
@iglesiasgthoralf, I was in the mood of trying java_modular so I thought it might be a good idea to have it set up with eclipse15:25
thoralfiglesiasg: I have never tried to setup a project based on the modular interfaces, sorry.15:25
@iglesiasgthoralf, all right. Thanks anyway!15:27
-!- foulwall` [~user@110.17.3.72] has quit [Remote host closed the connection]15:27
shogun-notifier-shogun: Viktor Gal :feature/CMake * e55bff1 / / (5 files): https://github.com/shogun-toolbox/shogun/commit/e55bff179bc4f52c1df092da7321751837bea5db15:41
shogun-notifier-shogun: Add csharp_modular and java_modular examples support15:41
-!- iglesiasg [~iglesias@2001:6b0:1:1da0:1464:9f67:69f9:4aba] has quit [Quit: Ex-Chat]15:50
-!- travis-ci [~travis-ci@ec2-54-235-60-216.compute-1.amazonaws.com] has joined #shogun16:04
travis-ci[travis-ci] it's Viktor Gal's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/973586616:04
-!- travis-ci [~travis-ci@ec2-54-235-60-216.compute-1.amazonaws.com] has left #shogun []16:04
-!- foulwall [~user@110.17.3.72] has joined #shogun16:09
-!- lisitsyn [~lisitsin@mxs.kg.ru] has quit [Quit: Leaving.]16:14
shogun-notifier-shogun: Viktor Gal :feature/CMake * 7782a14 / examples/undocumented/java_modular/CMakeLists.txt: https://github.com/shogun-toolbox/shogun/commit/7782a14f31aa9b0c993758f40434f8db33a88e9817:04
shogun-notifier-shogun: Fix java_modular examples17:04
shogun-notifier-shogun: Force to build Load.java before any other examples17:04
-!- pickle27 [~Kevin@d67-193-243-174.home3.cgocable.net] has joined #shogun17:14
shogun-notifier-shogun: Roman Votyakov :develop * f6b5c23 / src/shogun/machine/gp/ (4 files): https://github.com/shogun-toolbox/shogun/commit/f6b5c23b0ea6e506e93ef3b96528fa75c82cf01a17:39
shogun-notifier-shogun: move common methods from drived classes of inference methods to base class17:39
shogun-notifier-shogun: Heiko Strathmann :develop * ca5c973 / src/shogun/machine/gp/ (4 files): https://github.com/shogun-toolbox/shogun/commit/ca5c9739f483956555068fa4ac11c2dd76417a4217:39
shogun-notifier-shogun: Merge pull request #1346 from votjakovr/feature/gp_refactoring17:39
shogun-notifier-shogun:17:39
shogun-notifier-shogun: Move common methods from drived classes of inference methods to base class17:39
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has joined #shogun17:42
-!- mode/#shogun [+o iglesiasg] by ChanServ17:42
@iglesiasghushell, how is it going with the inference?17:43
@iglesiasghi pickle27!17:45
@iglesiasgpickle27, I took a quick look at your PR yesterday, I am going to write a couple of comments now about something I didn't quite understand17:45
@iglesiasgpickle27, if you have a moment, help me with the please :)17:45
pickle27iglesiasg: okay17:46
pickle27iglesiasg: I am actually defending my thesis in an hour so I probably won't get to it until tomorrow17:46
@iglesiasgpickle27, oh sure, never mind then. I will write then later, no distraction now!17:46
@iglesiasgpickle27, good luck with it!17:46
pickle27iglesiasg: thanks!17:47
-!- travis-ci [~travis-ci@ec2-54-235-60-216.compute-1.amazonaws.com] has joined #shogun17:52
travis-ci[travis-ci] it's Viktor Gal's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/973930817:52
-!- travis-ci [~travis-ci@ec2-54-235-60-216.compute-1.amazonaws.com] has left #shogun []17:52
shogun-buildbotbuild #1138 of cyg1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1138  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>17:54
pickle27iglesiasg: basically those new lines in SGInterface are telling the interface that there is a rhs return arg17:56
@iglesiasgpickle27, rhs return arg as an argument by reference?17:56
pickle27right hand side17:57
@iglesiasgyes, but what did you mean with return arg?17:57
@iglesiasgrhs agree, but return? :)17:57
pickle27oh so if the shogun call returns something you need to tell the interface17:57
pickle27then it checks for it and returns whatever it is properly17:58
pickle27so its return by value sort of17:58
@iglesiasgpickle27, aham, ok so the return values in static interface are also passed as right hand side arguments?17:58
@iglesiasgsorry if that is a stupid question :)17:58
pickle27iglesiasg: kind of, to be honest I don't know fully how the interface works17:59
@iglesiasgpickle27, hehe ok17:59
@iglesiasgpickle27, anyway what is the logic beneath the new checks?17:59
pickle27I looked at how the 'classify' function worked because it returns data and the saw what was missing from my new function that needed to return data17:59
pickle27iglesiasg: its copied from the classify function18:00
pickle27its telling the interface that its going to have to return something18:00
@iglesiasgpickle27, oh ok, that makes one feel safer indeed hehe18:00
pickle27yeah without that line it doesn't return anything18:00
@iglesiasgoh really? I would never have guessed that18:01
pickle27yeah me either18:01
pickle27it depends on the interface, because the return arg is still set even if you don't tell the interface that one exists18:02
pickle27it actually worked in octave static without it but not in R and thats when I found the fix18:02
@iglesiasgpickle27, it still works in octave though, right?18:02
pickle27yeah18:03
pickle27iglesiasg: you can take a look at cmd_classify line 4584 in SGInterface.cpp18:03
pickle27thats how I figured out what to do18:03
pickle27anyways I've got to go!18:04
pickle27see you guys tomorrow!18:04
-!- pickle27 [~Kevin@d67-193-243-174.home3.cgocable.net] has quit [Quit: Leaving]18:04
@iglesiasgL4584, crazy18:04
@iglesiasggood luck!18:04
-!- Beenish [6c1c8365@gateway/web/freenode/ip.108.28.131.101] has joined #shogun18:36
shogun-buildbotbuild #1139 of cyg1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1139  blamelist: Roman Votyakov <votjakovr@gmail.com>18:40
shogun-notifier-shogun: Kevin :develop * 5c3603f / / (2 files): https://github.com/shogun-toolbox/shogun/commit/5c3603f9e62442cdbadbda0add11fa858c10b04b18:51
shogun-notifier-shogun: fixed converters and the static interface and added an r static example for Jade18:51
shogun-notifier-shogun: Fernando Iglesias :develop * 4ab4174 / / (2 files): https://github.com/shogun-toolbox/shogun/commit/4ab4174e31634ea57a2b98ff2fe7270cdf1ba20e18:51
shogun-notifier-shogun: Merge pull request #1343 from pickle27/develop18:51
shogun-notifier-shogun:18:51
shogun-notifier-shogun: fixed converters and the static interface and added an r static example ...18:51
shogun-notifier-shogun: Viktor Gal :feature/CMake * 9529b5d / / (11 files): https://github.com/shogun-toolbox/shogun/commit/9529b5d347b0672f4f07bf75619b361ea5fb77b318:53
shogun-notifier-shogun: Add ruby, lua, octave example targets18:53
shogun-notifier-shogun: Viktor Gal :feature/CMake * 50b2a64 / .travis.yml: https://github.com/shogun-toolbox/shogun/commit/50b2a640528ab89ab9e68fb9cc1360f55c3267e018:54
shogun-notifier-shogun: Travis: remove rvm variable18:54
-!- travis-ci [~travis-ci@ec2-54-235-60-216.compute-1.amazonaws.com] has joined #shogun18:57
travis-ci[travis-ci] it's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/974068818:57
-!- travis-ci [~travis-ci@ec2-54-235-60-216.compute-1.amazonaws.com] has left #shogun []18:57
shogun-notifier-shogun: Soeren Sonnenburg :develop * 129254b / src/shogun/ (3 files): https://github.com/shogun-toolbox/shogun/commit/129254b6b556335831569c2c864cd1c8bf782e5518:59
shogun-notifier-shogun: convert functions in PLif to use sgvector18:59
shogun-notifier-shogun-data: foulwall :master * e0a2e04 / tapkee/faces_transparent.zip,tapkee/mnist.zip,tapkee/oks.zip: https://github.com/shogun-toolbox/shogun-data/commit/e0a2e04385ecbb5aa7947fe091519f1691b2638d19:04
shogun-notifier-shogun-data: data for tapkee19:04
shogun-notifier-shogun-data: Soeren Sonnenburg :master * d80dec7 / tapkee/faces_transparent.zip,tapkee/mnist.zip,tapkee/oks.zip: https://github.com/shogun-toolbox/shogun-data/commit/d80dec70fd3e7c348f7750f77b72f2beb4c670ef19:04
shogun-notifier-shogun-data: Merge pull request #10 from ZhengyangL/master19:04
shogun-notifier-shogun-data:19:04
shogun-notifier-shogun-data: data for tapkee19:04
shogun-notifier-shogun-demo: foulwall :master * 32c797c / / (16 files): https://github.com/shogun-toolbox/shogun-demo/commit/32c797cad3d9e375e1e8dd25b364b11923fd288519:07
shogun-notifier-shogun-demo: 1. add README 2. merge demos from tapkee website with .zip dataset19:07
shogun-notifier-shogun-demo: Soeren Sonnenburg :master * c155c7f / / (16 files): https://github.com/shogun-toolbox/shogun-demo/commit/c155c7fd2a626ad4d114044d584cba50466a232f19:07
shogun-notifier-shogun-demo: Merge pull request #18 from ZhengyangL/ocr19:07
shogun-notifier-shogun-demo:19:07
shogun-notifier-shogun-demo: 1. add README 2. merge demos from tapkee website with .zip dataset19:07
@sonney2kfoulwall, nice thanks!19:07
@wikingsonney2k: how do u imagine the ./configure -> cmake transition?19:08
@sonney2kwiking, rm configure and related scripts and done19:10
@sonney2kwiking, but I would want your cmake stuff merged first and enable it on the buildbots19:10
@wikingsonney2k: ok so as i thought merge it19:11
@wikingand then use it19:11
@wikingand then if everything is ok remove the old scripts19:11
shogun-buildbotbuild #1140 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1140  blamelist: Soeren Sonnenburg <sonne@debian.org>, Fernando Iglesias <fernando.iglesiasg@gmail.com>, Kevin <kevinhughes27@gmail.com>19:21
@sonney2kiglesiasg, I am a bit hmmhh about http://nn.7nn.de:8000/tapkee/entrance the tapkee dimred examples19:23
@sonney2kiglesiasg, this is totally static right?19:23
@sonney2kiglesiasg, I mean would it be possible to dynamically generate such embedding in *short* time?19:23
@iglesiasgsonney2k, IIRC it takes a bit of time for a demo to generate am embedding like that19:25
@sonney2kiglesiasg, how much?19:27
@sonney2kiglesiasg, I would rather prefer smaller sample sizes etc but a *real* demo19:27
@iglesiasgsonney2k, I am not sure. I don't remember exactly19:27
@iglesiasgsonney2k, I can run it here in a few minutes if you like19:27
-!- FSCV [~FSCV@206.225.135.24] has joined #shogun19:31
@sonney2kiglesiasg, would be nice19:36
@sonney2kfoulwall, we have to improve the dimension reduction / tapkee examples a bit to make them interactive19:38
@sonney2kfoulwall, currently it is just static (boring) stuff19:38
van51sonney2k: hi19:41
@sonney2kfoulwall, another question - could we have some tree?19:41
@sonney2kfoulwall, sth like this https://www.biomedcentral.com/content/figures/1471-2105-5-126-1-l.jpg19:43
van51sonney2k: first of all, for the normalization in the converter, I can't compare it directly to the dotfeatures19:43
@sonney2kvan51, hey!19:43
@sonney2kvan51, why not?19:43
@wikingHeikoS: yo19:43
van51sonney2k: because whenever the hashed vector is needed there we compute it through the converter19:43
@wikingHeikoS: actually what's the status of the foundation?19:43
van51sonney2k: and it only performs its own hashing in the dense_dot and add_to_dense methods19:43
@HeikoSwiking: stalled19:43
@wikingHeikoS: should not19:44
@HeikoSwiking: there are some other more important things unfortunately19:44
@iglesiasgsonney2k, the one with Oksana's faces takes almost no time19:44
@wikingit is important as well!19:44
@HeikoSmy work (1 week intense since NIPS reviews), google PO (I have problems with my bank) and other things19:44
@HeikoSwiking: I know19:44
@HeikoSbut not as important as the other ones19:44
van51sonney2k: so what I have done in the unit test is compare the result of the dense_dot method19:44
@sonney2kiglesiasg, nice then :)19:44
van51sonney2k: but as I said the result must be different due to numerical rounding errors and stuff19:45
@sonney2kvan51, I don't understand why it should not be possible19:45
@iglesiasgsonney2k, about 0.015 seconds in total19:45
@sonney2kiglesiasg, too fast ;)19:45
@sonney2kiglesiasg, no svm is as fast as this19:45
van51sonney2k: well, for instance in the dot method of the dotfeatures, where the full hashed vector is needed19:46
van51sonney2k: it is computed from the converter19:46
@iglesiasgsonney2k, there are no many images though, but I think more than enough for a demo19:46
@sonney2kvan51, you could do dot_features->get_computed_dot_feature_vector() and it should be directly comparable to your sparse vector19:46
@iglesiasgsonney2k, maybe a demo inspired by these images could be cool. Like the user takes a few screenshots with a webcam19:46
@sonney2kiglesiasg, yeah but later I am currently rather only wanting to have static input all the rest is a security nightmare19:47
van51sonney2k: how is that calculated? through which method?19:47
van51sonney2k: i'll check it actually19:47
@iglesiasgsonney2k, I understand19:47
@iglesiasgdinner time now, see you later19:47
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has quit [Quit: Ex-Chat]19:48
@sonney2kvan51, it just does add_to_dense_vec19:48
van51sonney2k: ok after this compilation is finished I'll check it19:48
@sonney2kgsomix_, how was the lecture and any progress?19:53
gsomix_sonney2k, hello. lecture was interesting - about past and future of science in Russia, huh.19:54
gsomix_sonney2k, working now.19:54
-!- travis-ci [~travis-ci@ec2-54-235-60-216.compute-1.amazonaws.com] has joined #shogun19:55
travis-ci[travis-ci] it's Fernando Iglesias's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/974385819:55
-!- travis-ci [~travis-ci@ec2-54-235-60-216.compute-1.amazonaws.com] has left #shogun []19:55
* sonney2k is off for water20:09
-!- travis-ci [~travis-ci@ec2-23-22-38-199.compute-1.amazonaws.com] has joined #shogun20:25
travis-ci[travis-ci] it's Viktor Gal's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/974392620:25
-!- travis-ci [~travis-ci@ec2-23-22-38-199.compute-1.amazonaws.com] has left #shogun []20:25
-!- lisitsyn [~lisitsyn@213.87.133.80] has joined #shogun20:31
@sonney2kvan51, and?20:37
van51sonney2k:  I did a git checkout and had to recompile :/20:37
@sonney2kvan51, but you only do libshogun right?20:37
@sonney2kso it should take like 1-2mins?20:37
van51sonney2k: I've added python_modular as well20:37
van51sonney2k: so I could prototype :p20:38
@sonney2kfor what?20:38
@sonney2kyeah well if your script is python then yes.20:38
van51sonney2k: ok so it outputs the same result :)20:39
@sonney2kvan51, alright then so you should get the same result then for whatever kernel (if you dot is correct - this is not covered in this test)20:41
@sonney2kor the webspam benchmark20:42
van51sonney2k: I have to go now20:42
van51sonney2k: but I'll check it again tonight20:42
van51sonney2k: btw I've also updated the PR on the quadratic for the dotfeatures20:43
van51sonney2k: have a look if you have some time20:43
van51bye cu in a few hours20:46
-!- van51 [~van51@ppp-94-66-76-131.home.otenet.gr] has quit [Quit: Leaving.]20:46
-!- travis-ci [~travis-ci@ec2-54-235-60-216.compute-1.amazonaws.com] has joined #shogun20:50
travis-ci[travis-ci] it's Viktor Gal's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/974399720:50
-!- travis-ci [~travis-ci@ec2-54-235-60-216.compute-1.amazonaws.com] has left #shogun []20:50
votjakovrHeikoS: hi there?21:01
@HeikoSvotjakovr: hi!21:01
@HeikoSvotjakovr: sampling the approximate posterior is quite hard21:01
votjakovrHeikoS: yeah, btw when i changed implementation your example works well :)21:03
@HeikoSvotjakovr: what did you change?21:03
votjakovrHeikoS: solve L*L^T * B = I21:03
@HeikoSvotjakovr: as you had it before?21:04
votjakovrHeikoS: yep21:04
@HeikoSvotjakovr: cool, could you send a patch, then I will try21:04
votjakovrHeikoS: anyway matrix K looks like symetric21:05
@HeikoSvotjakovr: I know its also easy to fix that via taking 0.5*(K+K^T)21:05
@HeikoSbut the problem really are the eigenvalues that sometimes get massively negative, if you try larger kernel sizes you will see21:05
votjakovrHeikoS: so is L*L^t * B = I more appropriate in our case?21:07
@HeikoSvotjakovr: yeah whatever makes this thing work :)21:07
@HeikoSsend a patch, I will have a look and try it tomorrow then21:07
@HeikoSvotjakovr: still a bit weird that this makes such a difference21:08
votjakovrHeikoS: yeah, it's weird, i agree21:11
@HeikoSvotjakovr: in fact, one doesnt need the covariance matrix itself, but rather its cholesky decomposition21:17
@HeikoSvotjakovr: if there was a stable method for that, that would be awesome21:18
@sonney2kwiking, btw why don't you merge your CMakefile.txt?21:19
votjakovrHeikoS: yeah, btw i've sent a PR21:21
shogun-notifier-shogun: Roman Votyakov :develop * 2ddb4a6 / src/shogun/machine/gp/LaplacianInferenceMethod.cpp: https://github.com/shogun-toolbox/shogun/commit/2ddb4a6f53aef14880c277a4709752202e30e7dc21:21
shogun-notifier-shogun: change implementation of posterior aproximation covariance matrix21:21
shogun-notifier-shogun: Heiko Strathmann :develop * c19a677 / src/shogun/machine/gp/LaplacianInferenceMethod.cpp: https://github.com/shogun-toolbox/shogun/commit/c19a67756281747b69a29877b4272d5653c22e0e21:21
shogun-notifier-shogun: Merge pull request #1347 from votjakovr/feature/gp_binary_classification21:21
shogun-notifier-shogun:21:21
shogun-notifier-shogun: change implementation of posterior aproximation covariance matrix21:21
@HeikoSvotjakovr: thanks!21:21
@HeikoSwill try out21:21
-!- travis-ci [~travis-ci@ec2-54-235-60-216.compute-1.amazonaws.com] has joined #shogun21:22
travis-ci[travis-ci] it's Soeren Sonnenburg's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/974417621:22
-!- travis-ci [~travis-ci@ec2-54-235-60-216.compute-1.amazonaws.com] has left #shogun []21:22
votjakovrHeikoS: welcome :)21:23
@HeikoSvotjakovr: will try that out tomorrow, btw how is EP looking?21:23
votjakovrHeikoS: in process...21:24
@HeikoSvotjakovr: ok cool21:25
@HeikoSvotjakovr: going home now, see you tomorrow! have a good evening21:25
votjakovrHeikoS: bye, you too ;)21:25
lisitsynHeikoS: have you seen wiking's effort?21:28
lisitsyn:)21:28
lisitsynexactly what you wanted21:29
@HeikoSlisitsyn: regarding?21:29
@HeikoSsorry so busy today21:29
lisitsynHeikoS: cmake!21:29
@HeikoSlisitsyn: ok yeah21:29
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 248 seconds]21:29
@HeikoSthat is totally cool21:29
lisitsynHeikoS: I guess that solves most of your concerns21:29
@HeikoSlisitsyn: which ones do you mean?21:30
@HeikoS;)21:30
lisitsynHeikoS: installation etc21:30
lisitsyndistribution21:30
@HeikoSyeah21:30
@HeikoShopefully :D21:30
@HeikoSbut would be amazing if these things would be solved now21:30
lisitsynI think they are pretty near21:30
lisitsynto be solved21:30
@HeikoScmake is also just more modern21:31
lisitsynHeikoS: bless wiking :D21:31
@HeikoSso good to have that21:31
lisitsynHeikoS: well cmake is old too21:31
@HeikoSyeah Ill have a beer on him later :)21:31
lisitsynhas some old fashioned stuff21:31
@HeikoSlisitsyn: we could also start thinking about compiling shogun in parts, takes toooo long on my machine21:31
lisitsynHeikoS: yes indeed21:32
lisitsynHeikoS: not only swig btw21:32
lisitsynthe C++ code is a bit slow to compile too now21:32
@HeikoSlisitsyn: yeah talking about that21:32
@HeikoSc++ now takes as long as swig21:33
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has joined #shogun21:33
-!- mode/#shogun [+o iglesiasg] by ChanServ21:33
lisitsynHeikoS: massive eigen etc21:33
lisitsyncompiling eigen code is massive work21:33
@HeikoSlisitsyn: maybe we can do modules or so, but I really gotta go now :) see you21:34
lisitsynHeikoS: yeah ok21:34
lisitsynwiking: okay I'll get to octave in a bit21:35
shogun-buildbotbuild #1141 of cyg1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1141  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>21:40
@sonney2kHeikoS, but at least you can compile with make -j 821:42
@sonney2kmodular is single cpu and won't scale21:42
@HeikoSsonney2k: ah really? thats annoying21:42
@HeikoSsonney2k:  so I need more CPUs21:42
@HeikoSvotjakovr: it works perfectly now :) :)21:42
@HeikoSok, last work from me, going home.....bye21:43
-!- HeikoS [~heiko@nat-181-93.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.]21:43
-!- Beenish [6c1c8365@gateway/web/freenode/ip.108.28.131.101] has quit [Quit: Page closed]21:54
-!- vgorbati [~vgorbati@91.216.173.29] has joined #shogun22:13
-!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has left #shogun ["sld"]22:17
shogun-buildbotbuild #1142 of cyg1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1142  blamelist: Roman Votyakov <votjakovr@gmail.com>22:26
-!- travis-ci [~travis-ci@ec2-23-22-38-199.compute-1.amazonaws.com] has joined #shogun22:49
travis-ci[travis-ci] it's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/974888522:49
-!- travis-ci [~travis-ci@ec2-23-22-38-199.compute-1.amazonaws.com] has left #shogun []22:49
-!- zxtx [~zv@rrcs-76-79-81-162.west.biz.rr.com] has joined #shogun23:02
-!- vgorbati [~vgorbati@91.216.173.29] has quit [Quit: vgorbati]23:19
shogun-notifier-shogun: Soeren Sonnenburg :develop * 07ee542 / src/interfaces/modular/IO.i,src/interfaces/modular/IO_includes.i: https://github.com/shogun-toolbox/shogun/commit/07ee5429dc69c439223e9e77dd5e3948a61fa66723:43
shogun-notifier-shogun: add CSVFile to modular interfaces23:43
--- Log closed Fri Aug 02 00:00:57 2013