Open in new window / Try shogun cloud
--- Log opened Wed Dec 25 00:00:21 2013
shogun-buildbotbuild #659 of nightly_default is complete: Failure [failed notebooks]  Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/65904:19
@sonney2kbesser82, merry x-mas to you too!07:42
-!- Saurabh7 [~Saurabh7@117.200.189.126] has joined #shogun07:53
Saurabh7hi , anybody around?07:59
@sonney2kSaurabh7, moin08:00
Saurabh7sonney2k, hello ,so i was tinkering with some xval stuff and i found08:03
Saurabh7shouldn't https://github.com/shogun-toolbox/shogun/blob/develop/examples/undocumented/libshogun/splitting_standard_crossvalidation.cpp#L7708:03
Saurabh7be   get_label(subset.vector[j]) ??08:04
@sonney2kSaurabh7, IMHO yes - but it is Heiko who wrote the code - so better ask him08:08
Saurabh7 ok sure, thanks!08:16
-!- Saurabh7 [~Saurabh7@117.200.189.126] has quit [Ping timeout: 264 seconds]09:21
-!- iglesiasg [~iglesiasg@211.Red-83-40-129.dynamicIP.rima-tde.net] has joined #shogun11:44
-!- mode/#shogun [+o iglesiasg] by ChanServ11:44
@sonney2kiglesiasg, one question - when you call apply_structured - will it *always* generate a new label?12:11
@iglesiasgsonney2k, tell me more12:17
@sonney2kiglesiasg, well I thought you are the expert...13:03
@sonney2kI just need to know if apply_structured will create a new object (like all the other apply* methods) or not13:03
@iglesiasgsonney2k, yes, it will13:04
-!- bb [458fc8de@gateway/web/freenode/ip.69.143.200.222] has joined #shogun13:13
@sonney2kiglesiasg, hmmh then I don't know13:18
@sonney2konly that the fgm notebook never worked reliably13:19
-!- iglesiasg [~iglesiasg@211.Red-83-40-129.dynamicIP.rima-tde.net] has quit [Quit: Leaving]13:48
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun14:09
shogun-notifier-shogun: Soeren Sonnenburg :develop * 5c498fa / doc/ipython-notebooks/multiclass/KNN.ipynb: https://github.com/shogun-toolbox/shogun/commit/5c498fadbdfd0abecb77ad2b5d09f9753e1affca14:09
shogun-notifier-shogun: improve KNN notebook14:09
-!- travis-ci [~travis-ci@ec2-54-224-203-27.compute-1.amazonaws.com] has joined #shogun14:57
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/1595840014:57
-!- travis-ci [~travis-ci@ec2-54-224-203-27.compute-1.amazonaws.com] has left #shogun []14:57
shogun-buildbotbuild #144 of clang34 - undefined behaviour analysis is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20undefined%20behaviour%20analysis/builds/144  blamelist: Soeren Sonnenburg <sonne@debian.org>15:09
shogun-buildbotbuild #142 of clang34 - thread analysis is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/142  blamelist: Soeren Sonnenburg <sonne@debian.org>15:25
shogun-buildbotbuild #145 of debian wheezy - memcheck is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/debian%20wheezy%20-%20memcheck/builds/14515:53
shogun-buildbotbuild #160 of clang34 - static analysis is complete: Failure [failed analyse]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20static%20analysis/builds/160  blamelist: Soeren Sonnenburg <sonne@debian.org>16:08
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]17:09
-!- siddharth950 [~siddharth@triband-mum-59.182.146.35.mtnl.net.in] has joined #shogun17:31
siddharth950anyone there?17:32
siddharth950Hi, I'm Siddharth, I have some experience of coding in C++ and have some knowledge of mahine learning17:58
siddharth950*machine17:58
siddharth950I'm interested in contributing to Shogun's development17:59
siddharth950can anyone guide me to do this best?17:59
-!- siddharth950 [~siddharth@triband-mum-59.182.146.35.mtnl.net.in] has quit [Quit: Leaving]18:38
-!- dsockwell [~twgs@199.167.199.97] has quit [Ping timeout: 272 seconds]19:11
-!- dsockwell [~twgs@199.167.199.97] has joined #shogun19:12
@sonney2kwiking, lisitsyn, besser82 do you know https://coveralls.io/ ?20:19
@sonney2kand besser82 how is my x-mas present doing ;)20:20
-!- bb [458fc8de@gateway/web/freenode/ip.69.143.200.222] has quit [Ping timeout: 272 seconds]20:24
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun21:54
shogun-notifier-shogun: Soeren Sonnenburg :develop * 870ade1 / doc/ipython-notebooks/multiclass/KNN.ipynb: https://github.com/shogun-toolbox/shogun/commit/870ade11e7ab6db28b27c12e5bb01943f367b1e321:54
shogun-notifier-shogun: add small comparison section to MC svms21:54
@wikingsonney2k: i've setup the thing for shogun (coveralls...)21:56
@wikingsonney2k: but doesn't really work i think21:56
shogun-notifier-shogun: Soeren Sonnenburg :develop * 3d4688a / doc/ipython-notebooks/multiclass/naive_bayes.ipynb: https://github.com/shogun-toolbox/shogun/commit/3d4688a8fe45e5589402cd8d6bc6a28eb420f16022:05
shogun-notifier-shogun: improve naive bayes tutorial22:05
@sonney2kwiking, maybe it needs https://github.com/eddyxu/cpp-coveralls22:10
@sonney2kI recall lisitsyn wanting to implement that for some time but ...22:11
* sonney2k enjoys doing notebooks22:11
@wikingsonney2k: did that22:11
@sonney2kwiking, you did? did we enable gcov when compiling them?22:11
@wikingsonney2k: i mean i really tried eveyrthing to setup coveralls.io for shogun like 6 months ago22:12
@wikingsonney2k: yes it was part of my cmake 'patch'22:12
@wikingsonney2k: plz give me some credit when i said i tried it ;P22:12
@wikingsonney2k: btW: merry xmas22:12
@sonney2kwiking, same to you22:12
@sonney2kwiking, well I have no idea what you do/did so apologies22:13
@wikingnw22:13
@sonney2kI think we need some kernels tutorial22:13
@wikingi'm just saying that i've tried but the stats22:13
@sonney2kand string kernels22:13
@sonney2kand mkl22:13
@wikingwere not really working..22:13
@sonney2kand performance measures22:13
@sonney2kdoing this makes me feel like being back in the 60's22:14
@sonney2k:D22:14
@wikingnotebook?22:14
@sonney2kwiking, btw about the whole license discussion22:15
@sonney2kwiking, yes22:15
@wikingsonney2k: mmm idont even know what's the current status22:15
@sonney2kwiking, I think we either stick with GPL or BSD22:15
@sonney2kall the rest is a waste22:15
@sonney2kmy interest in switching to anything is rather limited22:15
@wiking4 votes = 3 for lgpl 1 for bsd22:15
@sonney2kyeah but doesn't make sense22:16
@wikingso not many people voteD: P22:16
@sonney2keither we want to team up with scikits people22:16
@sonney2kand so need BSD22:16
@sonney2kor stay as we are22:16
@wikingwell imho we'll stay like this for at least 2 more releases22:16
@sonney2kI mean the only reason to switch licenses would be to somehow be the C++ backend of scikits22:17
@wikingas it'll take a long time regardless of license we will go with (if we change)22:17
@sonney2kI mean seriously22:17
@sonney2keveryone is using scikits for quick stuff22:17
@wikingmmm dunno if scikits ppl would be interested in that22:17
@sonney2kwhen using python22:17
@sonney2kif not then why change?22:17
@wikingas that's the good part of scikits22:17
@sonney2kit is not getting us users22:17
@wikingfast implementation22:17
@sonney2kand who cares about M$ forking shogun?22:18
@wikingmmm it's more about22:18
@sonney2kI mean it usually is the training algorithm that is complex22:18
@wikingsmaller entities using shogun22:18
@sonney2kall the rest is soo trivial22:18
@sonney2kso no need for more 'liberal' licenses22:18
@wikingsonney2k: as lisitsyn pointed out we could at least give a BSD-like inference interface22:18
@sonney2kyeah that makes the most sense I think22:19
@wikingyeah that requires horizontal slicing of the library22:19
@sonney2kit is tough still since it includes all kernels etc etc22:19
@sonney2kyeah and that22:19
@wikingbut we would need a vertical one as well22:19
@wikingmake it more modular22:19
@sonney2kbut no one knows how22:20
@sonney2kIMHO we should go d-ptr's22:20
@wikingnot as one big bundle22:20
@wikingsonney2k: yeah that's like step 122:20
@sonney2kand then maybe we see how a split could work22:20
@wikingwe need to define how to do dptrs22:20
@sonney2kwiking, btw *I* cannot fix proto for 3.1 for osx22:20
@wikingsonney2k: no macbook air around?22:21
@sonney2kwiking, and btw I have a new idea for meetings22:21
@sonney2kwiking, just no time22:21
@wikingsonney2k: ah ok22:21
@sonney2kwiking, IMHO we should do the meetings like this:22:21
@wikingsonney2k: for dptrs we should write a small howto22:21
@sonney2kwiking, every monday evening from 20-24hrs we have our shogun evening22:22
@wikingsonney2k: and then core developers can start implementing that22:22
@sonney2kwiking, so everyone who has time can attend22:22
@sonney2kand we chat and code22:22
@wikingsonney2k: and maybe even starte a feature branch...22:22
@sonney2keveryone tries to be there22:22
@wikingsonney2k: sounds reasonable22:22
@sonney2kand if not well not22:22
@sonney2kit is kind of what we call 'Stammtisch' :)22:22
@sonney2kin German22:22
@wikingwe could implement that22:22
@wikingbut not before 1st of jan22:23
@wikingor of course we can22:23
@wikingbut i think the attendance might be low on the 30th22:23
@wikingso 6th would be the first one ... that's reasonable22:24
@sonney2kI mean I know people are overloaded / busy with other stuff but hey better a long stretched fun meeting than none22:24
@sonney2kwiking, and if we want to be in GSoC 2014 we have to start serious planning already now...22:25
@sonney2kso I will release 3.1 before end of vacations and then we can do other stuff22:25
@sonney2kwiking, btw what happened to your libarchive endeavour?22:26
@wikingsonney2k: i had quite a problem with libarchive supporting raw file22:28
@wikingsonney2k: but i do have the feature branch localhost22:29
@wikingsonney2k: btw for dptrs we would just need so set some very basic rules: file and class naming convention22:29
@wikingsonney2k: and create some macros supporting dptrs22:29
@wikinglike the qt stuff: Q_DECLARE_PRIVATE and Q_DECLARE_PUBLIC22:31
@wikingand i think we could write about this a small page on github22:32
shogun-buildbotbuild #2106 of deb3 - modular_interfaces is complete: Failure [failed test libshogun]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/2106  blamelist: Soeren Sonnenburg <sonne@debian.org>22:34
@sonney2kwiking, makes sense22:35
@sonney2kwiking, if you have time draft that thing - we should start that right after 3.1 ... and btw I don't expect new major releases after end of gsoc'14 :/22:37
@wikinghttps://github.com/shogun-toolbox/shogun/wiki/Dpointers-in-Shogun22:37
@wikingthere22:37
@wikingwe should put there the main rules22:37
@sonney2kI mean only if we have just a few new features d-ptr's are heavily intrusive already bringing stuff into a non-releasable state22:38
@wikingsonney2k: mmm ok i'll try to scetch up something based on qt's ideas22:38
@sonney2kwiking, maybe you can try that on some small class to see if it works22:38
@wikingsonney2k: yeah will do on classes that i've implemented22:38
@sonney2kwiking, I personally would want to do notebooks for a while atm - I am really enjoying that part22:39
@wikingsonney2k: go ahead and have fun22:39
@wikingsonney2k: i'm trying to digest xmas food atm22:39
@wiking:P22:39
@wikingsonney2k: check those awesome social mining notebooks i've sent the other day22:42
@wikingsonney2k: if you could hack in there some shogun that would be amazing22:42
lisitsynho ho22:43
@sonney2kwiking, seen them but we far away from that still - we first need lots of pieces of the puzzle before we get there22:45
-!- travis-ci [~travis-ci@ec2-54-227-179-156.compute-1.amazonaws.com] has joined #shogun22:45
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/1596879622:45
-!- travis-ci [~travis-ci@ec2-54-227-179-156.compute-1.amazonaws.com] has left #shogun []22:45
@sonney2kand mr ho ho lisitsyn needs to contribute his stuff as notebooks too :D22:45
lisitsynsounds like a great time to do that haha22:45
lisitsynoh gosh will this stuff ever end22:46
@sonney2klisitsyn, you just wrote too much shogun code22:47
@sonney2klisitsyn, now you have to document it :P22:47
lisitsynsonney2k: no I mean my current status22:47
@sonney2klisitsyn, all your fault ;)22:47
lisitsynsonney2k: totally22:47
@sonney2klisitsyn, become the paid shogun e.V. secretary ;)22:48
lisitsynsonney2k: 100 EUR/year? ;)22:48
@sonney2klisitsyn, exactly :P22:49
lisitsynsonney2k: I totally like your stammwhatever idea22:55
lisitsynsonney2k: ah yeah and I like idea of visiting berlin once again though I'd like to visit other eu cities too :D23:01
lisitsynsonney2k: do you want to visit that py conference?23:01
@sonney2klisitsyn, don't count on me for anything - I am no better time wise23:03
lisitsynsonney2k: probably even worse23:03
@sonney2klisitsyn, I was hoping Heiko would give a talk at europython23:03
lisitsynI see23:03
@sonney2kand that we could have some workshop before or after23:03
lisitsynsonney2k: what about gsoc?23:03
lisitsynsonney2k: I remember you said you will have much less time23:04
@sonney2kmustn't be of the standard we set last year but still some talks or so and discussions23:04
@sonney2kyeah Heiko has the notes23:04
@sonney2kwe had like 100 ideas ...23:04
@sonney2kbut we need some mentors / organizers23:04
@sonney2kso the thing runs even with me disappearing23:05
lisitsynsonney2k: I see - I'll try to reserve as much time as I can23:05
@sonney2kIMHO heiko should be admin too23:05
lisitsynsonney2k: why too?23:05
@sonney2kbecause he almost was one last year23:06
@sonney2kfrom what he did I mean23:06
lisitsynsonney2k: too like who?23:06
@sonney2kwhat?23:06
@sonney2kyou me heiko?23:07
lisitsynsonney2k: I mean he is kinda the only admin23:07
lisitsyn:D23:07
lisitsynah23:07
lisitsynapart from you I am totally useless23:07
@sonney2kwhy that?23:07
lisitsynso its he who could handle that if he has time23:07
@sonney2kyeah - I would wish to just be a mentor23:08
@sonney2kshogun and teaching is what I would like to pursue...23:08
lisitsynsonney2k: you should have been a professor ;)23:09
shogun-buildbotbuild #2107 of deb3 - modular_interfaces is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/210723:09
@sonney2kbed time!23:09
lisitsynno bed for me today :D23:10
lisitsynsonney2k: good night23:10
-!- travis-ci [~travis-ci@ec2-54-224-203-27.compute-1.amazonaws.com] has joined #shogun23:17
travis-ci[travis-ci] it's Soeren Sonnenburg's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/1596923923:17
-!- travis-ci [~travis-ci@ec2-54-224-203-27.compute-1.amazonaws.com] has left #shogun []23:17
shogun-buildbotbuild #145 of clang34 - undefined behaviour analysis is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20undefined%20behaviour%20analysis/builds/145  blamelist: Soeren Sonnenburg <sonne@debian.org>23:23
shogun-buildbotbuild #143 of clang34 - thread analysis is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/143  blamelist: Soeren Sonnenburg <sonne@debian.org>23:40
--- Log closed Thu Dec 26 00:00:23 2013