IRC logs of #shogun for Wednesday, 2013-02-06

--- Log opened Wed Feb 06 00:00:40 2013
-!- wiking [~wiking@info2k1.hu] has joined #shogun00:03
-!- wiking_ [~wiking@info2k1.hu] has quit [Read error: Connection reset by peer]00:03
-!- wiking [~wiking@info2k1.hu] has quit [Read error: Connection reset by peer]00:06
-!- wiking [~wiking@info2k1.hu] has joined #shogun00:08
-!- n4nd0 [53b32c87@gateway/web/freenode/ip.83.179.44.135] has quit [Quit: Page closed]00:42
-!- FSCV_ [~FSCV@187.210.54.165] has quit [Quit: Leaving]03:05
-!- wiking [~wiking@info2k1.hu] has quit [Read error: No route to host]03:37
-!- wiking [~wiking@info2k1.hu] has joined #shogun03:37
-!- blackburn [~blackburn@37.61.179.145] has quit [Quit: Leaving.]07:50
-!- blackburn [~lisitsin@mxs.kg.ru] has joined #shogun08:17
-!- ptizoom [~christian@85.210.88.20] has quit [Ping timeout: 276 seconds]09:26
-!- ptizoom [~christian@79-71-81-137.dynamic.dsl.as9105.com] has joined #shogun09:40
-!- wiking [~wiking@info2k1.hu] has quit [Changing host]10:12
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun10:12
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun10:27
-!- heiko [~heiko@pat-191-250.internal.eduroam.ucl.ac.uk] has joined #shogun12:20
-!- heiko [~heiko@pat-191-250.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.]13:23
-!- sumit [ca4eaca2@gateway/web/freenode/ip.202.78.172.162] has joined #shogun13:49
-!- heiko [~heiko@pat-191-250.internal.eduroam.ucl.ac.uk] has joined #shogun13:59
-!- heiko [~heiko@pat-191-250.internal.eduroam.ucl.ac.uk] has quit [Client Quit]14:01
-shogungit:#shogun- [shogun] sonney2k pushed 2 new commits to master: https://github.com/shogun-toolbox/shogun/compare/8ebb42b4a6c4...5c8e2b102d5314:07
-shogungit:#shogun- shogun/master 792f895 Soeren Sonnenburg: Improve speed of SG_*IO debug functions...14:07
-shogungit:#shogun- shogun/master 5c8e2b1 Soeren Sonnenburg: remove legacy whitespace in all SG_*IO functions...14:07
-!- sumit [ca4eaca2@gateway/web/freenode/ip.202.78.172.162] has quit [Ping timeout: 245 seconds]14:09
n4nd0sonne|work: wow that was a large commit :)14:11
n4nd0did you make an script for that or from vim directly?14:11
n4nd0see you later14:18
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving]14:18
-!- FSCV [~FSCV@187.210.54.165] has joined #shogun14:18
-!- FSCV_ [~FSCV@204.45.132.2] has joined #shogun14:22
-!- FSCV [~FSCV@187.210.54.165] has quit [Ping timeout: 276 seconds]14:24
-!- n4nd0 [82e5860a@gateway/web/freenode/ip.130.229.134.10] has joined #shogun15:02
-!- FSCV [~FSCV@204.45.132.2] has joined #shogun15:13
-!- FSCV [~FSCV@204.45.132.2] has quit [Remote host closed the connection]15:13
-!- heiko [~heiko@pat-191-250.internal.eduroam.ucl.ac.uk] has joined #shogun15:14
-!- FSCV_ [~FSCV@204.45.132.2] has quit [Ping timeout: 264 seconds]15:15
-!- FSCV [~FSCV@187.210.54.165] has joined #shogun15:16
-!- FSCV_ [~FSCV@204.45.132.2] has joined #shogun15:18
-!- FSCV_ [~FSCV@204.45.132.2] has quit [Remote host closed the connection]15:18
-!- heiko [~heiko@pat-191-250.internal.eduroam.ucl.ac.uk] has quit [Client Quit]15:18
-!- FSCV_ [~FSCV@204.45.132.2] has joined #shogun15:18
-!- heiko [~heiko@pat-191-250.internal.eduroam.ucl.ac.uk] has joined #shogun15:19
-!- FSCV [~FSCV@187.210.54.165] has quit [Ping timeout: 260 seconds]15:21
-!- blackburn [~lisitsin@mxs.kg.ru] has quit [Quit: Leaving.]15:22
-!- heiko [~heiko@pat-191-250.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.]15:32
-!- ptizoom [~christian@79-71-81-137.dynamic.dsl.as9105.com] has quit [Read error: Connection reset by peer]15:36
-!- blackburn [~lisitsin@mxs.kg.ru] has joined #shogun15:46
-!- blackburn [~lisitsin@mxs.kg.ru] has quit [Quit: Leaving.]16:11
-!- sumit [ca4eaca2@gateway/web/freenode/ip.202.78.172.162] has joined #shogun16:14
-!- n4nd0 [82e5860a@gateway/web/freenode/ip.130.229.134.10] has quit [Quit: Page closed]17:31
-!- sumit [ca4eaca2@gateway/web/freenode/ip.202.78.172.162] has quit [Ping timeout: 245 seconds]17:44
-!- blackburn [~blackburn@185.6.239.122] has joined #shogun18:01
-shogungit:#shogun- [shogun] behrj opened pull request #857: Fix static interface to dynamic program (master...master) https://github.com/shogun-toolbox/shogun/pull/85718:19
-!- heiko [~heiko@pat-191-250.internal.eduroam.ucl.ac.uk] has joined #shogun18:22
heikosonney2k, around?18:26
heikohad phd interviews over the last week - the reason why I was invisible.18:27
heikoI was just wondering about the call for proposals for the workshop, when should we send this out, what should be in there?18:28
heikolet me know what you think on this18:28
@sonney2kheiko, hey18:34
heikohi18:34
@sonney2kheiko, gunnar is hiding from me so before I get hold of him I cannot do more18:35
@sonney2kheiko, btw I am just investigating evaluation_cross_validation_multiclass_mkl.cpp18:35
blackburnsonney2k: I am afraid that pull request is rather for you :(18:35
@sonney2kit fails with assertation < 0.9 accuracy18:35
@sonney2kblackburn, ?18:35
heikosonney2k, yes I saw that18:35
blackburnsonney2k: https://github.com/shogun-toolbox/shogun/pull/85718:36
heikosonney2k, I am a bit confused since this used to work18:36
@sonney2kheiko, who did add this? does it make sense?18:36
heikosonney2k, well just as a check that no nonsense is returned, I evaluated the accuracy, and then simply added this check which was a lower bound on the accuracy18:36
heikoso if it fails this means that something has changed18:36
heikoand this way, a few problems with the labels were detected18:37
heikoso I consider it quite useful, however, remove it if it doesnt make sense or something18:37
@sonney2kheiko, do you know what accuracy is possible w/ e.g. a 3 class mc svm?18:38
@sonney2kheiko, this example is rather advanced... mkl & multiclass and ... so hmmh18:39
heikobut the data is fixed right?18:39
@sonney2kheiko, it trains on the full data set and then does some stratified x-val again on this dataset giving out accuracy18:43
heikoI agree that through x-.validation, we cannot be sure what the accuacy is18:43
heikohowever, in practice, it makes still sense to check for reasonable accuracy18:43
heikoBut agreed that 0.9 is too large18:44
heikoI will change it to 0.7 or 0.618:44
heikoor even remove it18:44
heikosonney2k, blackburn, on thing, in lapack.h, can I change line 31 to #include <atlas/cblas.h> ? Would that work on your systems? On mine, it does not compile otherwise18:47
blackburnheiko: it seems to be little wrong18:47
blackburnheiko: do you have -I/usr/include/atlas in your .config?18:48
heikono18:49
blackburnbut you have -DHAVE_ATLAS, right?18:49
heikoyes18:49
blackburninteresting18:50
heikoI will just create a link on my system then18:50
blackburnheiko: so that's ubuntu?18:51
heikoyep, any I haven't changed anything, it just stopped working a while ago18:51
blackburnheiko: did you re-configure recently?18:52
heikoyes18:52
heikooften18:52
heikoin fact just a minute ago18:52
blackburnsad18:53
@sonney2kheiko, no that is not what I meant - what accuracy is possible?18:53
heikosonney2k, how do you  mean that?18:53
@sonney2kheiko, if 0.9 is very well possible then this assertion is valid18:54
@sonney2kheiko, about your atlas issue... I have -I/usr/include/atlas in my .config. could you paste your configure.log?18:55
heikosonney2k, I dont know, but I certainly would not have added this if I did not get these values. Cannot remember however, so I just changed it to 0.65 instead18:55
heikosure18:55
@sonney2kheiko, well no then don't change it18:55
heikosonney2k, there were problems with labels before, causing wrong x-validation results18:56
@sonney2kheiko, which test test for this?18:57
@sonney2kor how can we be sure that we don't have this issue again?18:57
heikowell, those accuracy assertions are a straight forward way to at least ensure that no nonsense comes out (assert that performance is better than random or similar)18:58
heikobut in general, we can only be sure if we finally start writing tests for our code ;)  for which I was a bit overoptimistic regarding time to do this18:58
@sonney2kvery true18:59
heikoany updates on the gtest framework btw?18:59
blackburnit is not the matter of framework I believe19:00
@sonney2kheiko, that is wiking's domain19:00
heikoI know, just wondering whether there was progress19:00
@sonney2kheiko, wiking was/is busy with real life stuff19:00
heikosonney2k, here is my configure.log19:01
heikohttp://pastebin.com/PKtumAuR19:01
heikosonney2k, who isn't :)19:01
heikoI finally managed to have two jobs at once :D19:01
@sonney2kheiko, stupid you19:01
@sonney2kget rid of one19:01
@sonney2kthere is no need to get a burn out before becoming 3019:02
blackburnooh two jobs nice19:02
blackburnhah heiko is near IIRC19:02
heikosonney2k, in two months, and its two 50% ones, but you know how that is - at least I will get my name on some nice papers :)19:02
heikosonney2k, no risk of burn out here, I am quite relaxed19:02
heikotake my time to rest etcetc19:03
blackburnI am dying with two areas I am trying to succeed19:03
blackburn:D19:03
@sonney2kheiko, your c++ compiler is weird19:06
heikowhat going on?19:07
@sonney2kit shouldn't compile this test19:07
heikono idea19:09
heikowhat about the atlas issue ?19:10
@sonney2kthat is your atlas issue19:10
heikoI see19:10
heikommh19:10
heikoso?19:10
heikore-install ubuntu? :D19:10
heikothis is usually how I fix those thing19:11
heikos19:11
@sonney2klets try -Werror-implicit-function-declaration19:13
heikowhere to add this best?19:14
@sonney2kheiko, sec19:18
heikook19:18
@sonney2kheiko, please pull and test19:23
heikosonney2k, ok19:23
heikogive me a minute, just committing some other thinfs19:24
@sonney2kheiko, check if -I/usr/include/atlas appears in your .config19:24
heikosonney2k,19:28
heikocat .config  | grep atlas19:28
heikoPOSTLINKFLAGS=  -lm -pthread -lhdf5 -ljson   -lxml2   -llapack -lcblas -llapack_atlas -larpack -lglpk -lm -llzo2 -lz -lbz219:28
heikofrom shogun/src19:28
heikobuilding from latest pull now19:28
heikosonney2k, compiled just fine, you fixed it :)19:30
-!- travis-ci [~travis-ci@ec2-204-236-250-130.compute-1.amazonaws.com] has joined #shogun19:30
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/462259219:30
-!- travis-ci [~travis-ci@ec2-204-236-250-130.compute-1.amazonaws.com] has left #shogun []19:30
@sonney2kwiking, could you fix travis?19:31
@sonney2kit is giving weird erros19:31
@sonney2kand is useless currently19:31
blackburnI can19:32
blackburnuhm19:33
blackburnillegal instruction19:33
blackburnsonney2k: I'd expect we should turn off optimizations19:33
@sonney2kblackburn, so it is a bug in clang?19:37
@sonney2kin any case we should turn of optimzations for travis - speeds up compilation...19:38
blackburnsonney2k: it happen to me with -O4 once19:38
blackburnyes agree19:38
@sonney2kblackburn, I guess it is enough to edit .travis.yml?19:38
blackburnsonney2k: yes already done19:38
@sonney2koops19:39
@sonney2kme too19:40
blackburnsonney2k: irc notify disappeared19:41
blackburnsonney2k: btw we could pull travis status to readme like in tapkee: https://github.com/lisitsyn/tapkee (line before Interface header)19:42
@sonney2kI don't understand why thogh19:42
shogun-buildbotbuild #178 of rpm1 - libshogun is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/178  blamelist: Soeren Sonnenburg <sonne@debian.org>19:44
blackburnsonney2k: what would you say if I said I want to CMake shogun totally?19:45
@sonney2kblackburn, I would say you have too much spare time19:46
blackburntrue19:46
blackburnbut it would be nice still19:47
blackburnsonney2k: ah yes, we need gsoc meeting19:47
@sonney2kquestion is do we want to participate this year?19:49
@sonney2kthere hasn't been a lot of activity here ...19:49
blackburnsonney2k: whoa?19:50
blackburnactivity like?19:51
@sonney2kblackburn, well if we don't have the time to get the build green how can we mentor students *and* keep the build green?19:51
blackburnsonney2k: I can get time to manage students19:52
blackburnwinter is tough (and it was tough twice IIRC)19:53
blackburnit is not a question for me - gsoc-boosted we can keep doing some changes19:54
@sonney2kblackburn, would you do the gsoc admin this year then?20:00
blackburnsonney2k: sure20:01
blackburnif you don't have time I'll back you up20:01
@sonney2kblackburn, I mean we should not rely on me this year20:01
shogun-buildbotbuild #179 of rpm1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/17920:01
-!- heiko [~heiko@pat-191-250.internal.eduroam.ucl.ac.uk] has left #shogun []20:12
@sonney2kblackburn, no idea - irc notifications just don't want to work20:18
blackburnsonney2k: github.20:18
blackburn:D20:18
@sonney2kblackburn, if you want gsoc - start creating a list now and invite everyone to a meeting20:18
@sonney2kannounce it on the mailinglist etc20:18
@sonney2kyou know the drill :D20:18
blackburnsonney2k: yes once gsoc is announced20:18
blackburnsonney2k: but I'd need contacts of all possible mentors20:19
@sonney2kblackburn, no *before* gsoc is announced20:20
blackburnwhy before?20:20
@sonney2kvery important to do everything before20:20
blackburnhow can I announce we are participating in something that is not announced yet?20:20
@sonney2kblackburn, don't announce that we participate but that we want to and are looking for mentors/subjects20:21
blackburnah that's ok20:21
blackburnI'll write an mail tomorrow's morning20:21
blackburnsonney2k: you settled some mentors at summit, right?20:22
blackburnwell, possibilities20:22
@sonney2kyes20:29
blackburnsonney2k: I'd need contacts or mail to them from you20:30
@sonney2kblackburn, no20:30
blackburnno what?20:30
@sonney2kfirst discuss on the mailinglist20:30
@sonney2kthis has time20:31
blackburnyes - but then someone should ask they if deal is here still20:31
@sonney2kyes later20:31
blackburnsonney2k: do you know anything on compressive sensing?20:32
@sonney2kno20:32
blackburnjust about - I read some blog on that by igor carron20:33
blackburnI'll ask him in person if he can guide in compressive sensing20:33
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun21:10
shogun-notifier-shogun: Soeren Sonnenburg :master * 4791238 / src/shogun/ (5 files): https://github.com/shogun-toolbox/shogun/commit/4791238c1617776a7add5339a6b00f5c6b07607221:10
shogun-notifier-shogun: Fix a couple of clang compiler warnings.21:10
shogun-notifier-shogun: Soeren Sonnenburg :master * ad53e41 / src/configure: https://github.com/shogun-toolbox/shogun/commit/ad53e41509b3b2641ba25c5c4eb3d23c842ab4d221:10
shogun-notifier-shogun: enable more strict compiler settings21:10
shogun-notifier-shogun: Sergey Lisitsyn :master * 61197e6 / .travis.yml: https://github.com/shogun-toolbox/shogun/commit/61197e6018fe278bfae1b103c8494903f850b80b21:10
shogun-notifier-shogun: Disabled optimizations in travis21:10
shogun-notifier-shogun: Soeren Sonnenburg :master * 397f027 / .travis.yml: https://github.com/shogun-toolbox/shogun/commit/397f02747e5665e68a00be15858f60312953b3fb21:16
shogun-notifier-shogun: really disable optimizations in travis21:16
@sonney2kblackburn, have you seen this /usr/bin/../lib/gcc/x86_64-linux-gnu/4.7/../../../../include/c++/4.7/cstring:101:3: error: declaration conflicts with target of using declaration already in scope21:19
@sonney2k ?21:19
blackburnsonney2k: no, where?21:20
blackburnsonney2k: ah wait I have seen that21:21
@sonney2kwhen compiling with clang21:21
blackburnsonney2k: we had user reporting about that21:23
shogun-notifier-shogun: Soeren Sonnenburg :master * 37d4736 / src/interfaces/python_modular/protocols_helper.i: https://github.com/shogun-toolbox/shogun/commit/37d473621e388d64e0abd8ae0064cd49283b64e621:24
shogun-notifier-shogun: remove cstring include to hopefully fix clang compile error21:24
blackburnyes it fixes the error21:24
@sonney2kblackburn, why is it not committed then?21:24
@sonney2kuntil now I mean...21:24
blackburnsonney2k: I didn't manage to reproduce that21:24
@sonney2kjust compile with clang21:25
-!- travis-ci [~travis-ci@ec2-204-236-250-130.compute-1.amazonaws.com] has joined #shogun21:25
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/462545721:25
-!- travis-ci [~travis-ci@ec2-204-236-250-130.compute-1.amazonaws.com] has left #shogun []21:25
blackburnit compiles with clang on my machine21:25
shogun-notifier-shogun: Soeren Sonnenburg :master * b2e2b30 / src/shogun/multiclass/LaRank.cpp,src/shogun/multiclass/LaRank.h: https://github.com/shogun-toolbox/shogun/commit/b2e2b301eebfaed45cdcc7fa591f5b9599bfc4f121:26
shogun-notifier-shogun: fix some clang compile errors21:26
blackburnworks on my machine certification programme21:26
blackburnsonney2k: we've got to reduce codebase21:31
blackburnsonney2k: and we should decide whether we are general ml or bioinformatics21:32
blackburnI actually would vote for bioinformatics - there are a lot of tools for other ML21:33
-!- travis-ci [~travis-ci@ec2-204-236-250-130.compute-1.amazonaws.com] has joined #shogun21:36
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/462565021:36
-!- travis-ci [~travis-ci@ec2-204-236-250-130.compute-1.amazonaws.com] has left #shogun []21:36
-!- travis-ci [~travis-ci@ec2-204-236-250-130.compute-1.amazonaws.com] has joined #shogun21:37
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/462564021:37
-!- travis-ci [~travis-ci@ec2-204-236-250-130.compute-1.amazonaws.com] has left #shogun []21:37
blackburnehm21:38
shogun-notifier-shogun: Sergey Lisitsyn :master * ff6d336 / src/shogun/lib/CoverTree.h: https://github.com/shogun-toolbox/shogun/commit/ff6d33624986decf4e085ed02418bf7fd8fbf7b721:46
shogun-notifier-shogun: Removed unused covertree21:46
-!- travis-ci [~travis-ci@ec2-204-236-250-130.compute-1.amazonaws.com] has joined #shogun21:50
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/462621021:50
-!- travis-ci [~travis-ci@ec2-204-236-250-130.compute-1.amazonaws.com] has left #shogun []21:50
-!- n4nd0 [53b32c87@gateway/web/freenode/ip.83.179.44.135] has joined #shogun21:51
shogun-buildbotbuild #782 of deb1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/782  blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com>22:09
shogun-notifier-shogun: Sergey Lisitsyn :master * 48e3489 / src/shogun/lib/FibonacciHeap.cpp,src/shogun/lib/FibonacciHeap.h: https://github.com/shogun-toolbox/shogun/commit/48e3489656524eaa7689af5dc34492922167dfab22:16
shogun-notifier-shogun: Removed duplicate of fibonacci heap22:16
-!- travis-ci [~travis-ci@ec2-204-236-250-130.compute-1.amazonaws.com] has joined #shogun22:21
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/462687622:21
-!- travis-ci [~travis-ci@ec2-204-236-250-130.compute-1.amazonaws.com] has left #shogun []22:21
shogun-buildbotbuild #783 of deb1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/783  blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com>22:21
blackburnwhoops22:22
blackburnkeep calm sonney22:22
blackburnsunny! yesterday my life was filled with rain22:22
shogun-notifier-shogun: Sergey Lisitsyn :master * 18666b1 / src/shogun/ (4 files): https://github.com/shogun-toolbox/shogun/commit/18666b1527465efdd4292b93f0e21a9ccd5e26cc22:26
shogun-notifier-shogun: Removed unused arpack wrapper and removed unused covertree import22:26
shogun-buildbotbuild #784 of deb1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/78422:31
blackburnshogun-buildbot: good job soldier22:33
-!- travis-ci [~travis-ci@ec2-204-236-250-130.compute-1.amazonaws.com] has joined #shogun22:41
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/462717622:41
-!- travis-ci [~travis-ci@ec2-204-236-250-130.compute-1.amazonaws.com] has left #shogun []22:41
--- Log closed Thu Feb 07 00:00:41 2013

Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!