IRC logs of #shogun for Sunday, 2013-10-06

--- Log opened Sun Oct 06 00:00:29 2013
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 240 seconds]01:22
-!- HeikoS [~heiko@05453f70.skybroadband.com] has left #shogun []03:15
-!- sonne|osx_ [~sonne@f053032016.adsl.alicedsl.de] has joined #shogun03:43
-!- sonne|osx [~sonne@f053034139.adsl.alicedsl.de] has quit [Ping timeout: 245 seconds]03:45
-!- sonne|osx_ is now known as sonne|osx03:45
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun08:45
shogun-notifier-shogun: Fernando Iglesias :develop * 7159bef / doc/ipython-notebooks/template.ipynb: https://github.com/shogun-toolbox/shogun/commit/7159befbd092a83a1589cc14591330783ef501b608:45
shogun-notifier-shogun: Add ipython notebook template08:45
shogun-notifier-shogun: Soeren Sonnenburg :develop * d4d23ee / doc/ipython-notebooks/template.ipynb: https://github.com/shogun-toolbox/shogun/commit/d4d23ee1bf5327107f0f01964dc9a5510db77d5508:45
shogun-notifier-shogun: Merge pull request #1709 from iglesias/feature/notebook_template08:45
shogun-notifier-shogun:08:45
shogun-notifier-shogun: Add ipython notebook template08:45
shogun-buildbotbuild #69 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/69  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>09:39
shogun-buildbotbuild #71 of clang34 - thread analysis is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/71  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>09:54
shogun-buildbotbuild #89 of clang34 - static analysis is complete: Failure [failed analyse]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20static%20analysis/builds/89  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>10:36
shogun-buildbotbuild #70 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/70  blamelist: Soeren Sonnenburg <sonne@debian.org>10:47
shogun-buildbotbuild #72 of clang34 - thread analysis is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/72  blamelist: Soeren Sonnenburg <sonne@debian.org>11:02
-!- besser82 [~besser82@fedora/besser82] has joined #shogun11:41
shogun-buildbotbuild #90 of clang34 - static analysis is complete: Failure [failed analyse]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20static%20analysis/builds/90  blamelist: Soeren Sonnenburg <sonne@debian.org>11:44
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]11:45
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has joined #shogun14:03
-!- mode/#shogun [+o iglesiasg] by ChanServ14:03
@iglesiasghey guys14:06
lisitsynhey iglesiasg14:15
-!- HeikoS [~heiko@nat-181-182.internal.eduroam.ucl.ac.uk] has joined #shogun15:19
-!- mode/#shogun [+o HeikoS] by ChanServ15:19
@HeikoShello!15:21
lisitsynHeikoS: he15:37
lisitsyny15:37
@HeikoShi!15:37
@HeikoShow are things?15:37
lisitsynHeikoS: fine, what about you? working on submission?15:39
lisitsynI am reading it right now btw15:39
@HeikoSyeah doing that15:39
@HeikoSpimping notebooks a bit15:39
@HeikoSwould be good if we had the other things also, might also be good for release15:40
lisitsynHeikoS: I have no concerns with text15:40
lisitsynmay be mention LibSVM and Liblinear15:40
lisitsynand LibOCAS may e15:40
@HeikoSisnt that in there?15:42
@HeikoSyeah in the methods15:42
lisitsynHeikoS: only libsvm15:42
@HeikoSthats in the external software15:42
lisitsynmay be liblinear makes sense too15:42
@HeikoSbut I did not want to make them overlapping15:42
lisitsynah I see15:42
@HeikoSlibsvm=liblinear for most people isnt it?15:42
lisitsynI don't know how people see it to be honest :)15:43
@HeikoSok will add15:44
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving]16:09
-!- lisitsyn [~lisitsyn@213.178.35.165] has quit [Quit: Leaving.]16:10
-!- lisitsyn [~lisitsyn@fb2-lo1.global63.net] has joined #shogun19:29
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun19:43
shogun-notifier-shogun: Heiko Strathmann :develop * b688979 / doc/ipython-notebooks/statistics/mmd_two_sample_testing.ipynb: https://github.com/shogun-toolbox/shogun/commit/b6889794d12f91e74422fac822dc81ef4b0d293819:43
shogun-notifier-shogun: massive update to the MMD part19:43
-!- lambday [67157d37@gateway/web/freenode/ip.103.21.125.55] has joined #shogun19:53
@HeikoSlambday: hey!19:53
lambdayHeikoS: hi how are you :)19:54
lambdayam I still there :-/19:55
@HeikoSlambday: pretty good and you?19:55
@HeikoSyes19:55
lambdayHeikoS: pretty bad :( just hoping that things will get better :-s19:56
@HeikoSwhats wrong?19:56
lambdayHeikoS: when is the  release date?19:56
@HeikoSwe wanted to do it over the weekend, but didnt ;) soon!19:56
@HeikoSwe need to fix many things19:56
lambdayHeikoS: institute related shits :-/19:57
@HeikoSbut we also need the release for the nips application19:57
@HeikoSlambday: ah, do you have trouble with your supervisors?19:57
@HeikoSwiking: around?19:57
@HeikoSwiking: I need a link for the cloud shogun for the application19:57
@HeikoSbesser82: ^19:57
lambdayHeikoS: yeah! when I told the idea about working regarding the idea the exact words he told was "don't you try to shove down my throat some work that you worked for Google"!!19:58
lisitsynhaven't you ever thought of changing supervisor ;)19:59
@HeikoSlambday: you did not even work for google, but for Shogun!19:59
@HeikoSanyway, lambday try to calm things down, that always helps19:59
@HeikoSlisitsyn: whats the state on the opencv example?20:00
lisitsynbuy a machine gun20:00
@HeikoSI would like to reference that20:00
lisitsynehm I don't know20:00
lisitsynHeikoS: what do you want to have?20:01
@HeikoScurrently I can only reference the gist of pickle20:01
lisitsynI can do something using that20:01
lisitsynbut I don't know what20:01
@HeikoSjust the thing itself20:02
@HeikoSwould be great to have in the repository20:02
-!- lambday_ [67157e37@gateway/web/freenode/ip.103.21.126.55] has joined #shogun20:02
@HeikoSlambday_:  hi again!20:02
lambday_HeikoS: sorry I got dc20:02
lisitsynwhat exactly? converting matrices?20:02
lambday_argh my net :(20:02
@HeikoSlisitsyn:  https://gist.github.com/pickle27/675172220:02
-!- lambday [67157d37@gateway/web/freenode/ip.103.21.125.55] has quit [Ping timeout: 250 seconds]20:02
lambday_HeikoS: do you think we should get rid of arprec?20:03
@HeikoSlambday_: not sure, if we can get rid of the memory bug, then no, if thats not possible then yes20:03
@HeikoSlambday_: but thats no focus20:03
@HeikoSfocus is release for now, only things that are very important20:03
lambday_HeikoS: anything that I can help with20:03
lambday_HeikoS: I really reduce the runtime of the notebok20:04
lambday_book*20:04
@HeikoSfixing github issues  would be a great help20:04
lambday_should*20:04
@HeikoSlambday_: ah I think the notebook currently fails20:04
@HeikoSthat is important to work20:04
lambday_ah!!20:04
lambday_:'(20:04
lambday_HeikoS: alright.. so I am thinking to reducing the samples from 10 to 120:04
@HeikoScant you also reduce the number of colours?20:05
lambday_HeikoS: and dropping the part that computes exact log-det cause that takes a lot of memory20:05
@HeikoSlambday_: yeah just comment out for now or so20:05
lambday_HeikoS: that won't eb good20:05
lambday_HeikoS: we gotta use all the colors20:05
lambday_HeikoS: we can reduce the number of shifts20:05
@HeikoSlambday_: ok shifts then is good20:06
@HeikoSand the exact stuff20:06
lambday_yeah20:06
@HeikoSlambday_: just focus on making it work with the stuff that is there20:06
lambday_alright20:06
lambday_alright'20:06
lambday_sonne|osx said 7 mins20:06
lambday_that would be a bit tough! but let's see20:06
@HeikoSyeah, small steps towards there ;)20:07
lambday_HeikoS: preconditioned CG is required in many many applications man! we gotta make outr implementation top notch!20:08
@HeikoSlambday_: yeah I know, it is essential20:08
@HeikoSlambday_: so would be great to have, also its pretty interesting to integrate that into our framework20:08
lambday_HeikoS: with chain of operators...20:09
@HeikoSyeah, but lambday_ we really have to make the notebook work asap20:09
@HeikoSits blocking the others on the page20:09
lambday_HeikoS: alright.. yeah that's first priority20:09
lambday_HeikoS: I will surely try to make this work in coming 2-3 daysw20:10
@HeikoSlambday_: well have to exclude it from the release if it doesnt work and that would be a shame20:10
lambday_yeah! :(20:10
lambday_it looks pretty awesome! :(20:11
lambday_(thanks to you)20:11
@HeikoSit does, maybe just comment out some parts for now?20:11
lambday_HeikoS: yeah!20:11
@HeikoSand then commit this, and then fix in the background?20:11
@HeikoScould you do that?20:11
lambday_HeikoS: checking now20:11
-!- lambday_ [67157e37@gateway/web/freenode/ip.103.21.126.55] has quit [Ping timeout: 250 seconds]20:15
-!- lambday [67157d37@gateway/web/freenode/ip.103.21.125.55] has joined #shogun20:19
lambdayguys! I can't build shogun :(20:26
lambdaycc1plus: warning: unrecognized command line option "-Wno-c++11-narrowing" [enabled by default]20:26
lambdayand then it stops!20:26
lambdayjust did rebase and this is what I get20:26
@HeikoSlambday: manually uninstall, remove build dir, and do git clean -dfx that fixes it20:27
lambdayHeikoS: alright20:27
@HeikoSlambday: gotta go home now, see you around! :) thanks for helping on the fix20:27
lambdayHeikoS: no problem man!20:28
lambdayHeikoS: really want this notebook to work!20:28
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has joined #shogun20:30
-!- lambday [67157d37@gateway/web/freenode/ip.103.21.125.55] has quit [Ping timeout: 250 seconds]20:32
-!- HeikoS [~heiko@nat-181-182.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.]20:33
shogun-buildbotbuild #71 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/71  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>20:33
shogun-notifier-shogun: Soeren Sonnenburg :develop * 3665b00 / doc/ipython-notebooks/statistics/mmd_two_sample_testing.ipynb: https://github.com/shogun-toolbox/shogun/commit/3665b001e2b836456f7d7c5ef83334eeef445eab20:36
shogun-notifier-shogun: fix notebook to follow agreed style20:36
shogun-buildbotbuild #73 of clang34 - thread analysis is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/73  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>20:48
-!- devl82 [5e45348a@gateway/web/freenode/ip.94.69.52.138] has joined #shogun21:03
devl82hello21:03
iglesiasghi devl8221:04
devl82i have an installation problem using ubuntu13.04+shogun3.0+matlab13a21:04
devl82specifically ccmake tells me "Could not find Matlab, which is required for compiling matlab_static    interface.  Try setting MATLAB_ROOT enviroment variable to the right path"21:05
hushelliglesiasg: hey21:05
devl82however using sexport MATLAB_ROOT='/usr/local/MATLAB'21:06
devl82i meant export21:06
devl82and testing it with echo "unix('env');quit" | matlab -nojvm -nodisplay | grep MATLAB21:07
devl82i should have been ok ..any suggestions?21:07
lisitsyndevl82: hey21:18
lisitsyndevl82: did you try to run cmake again after your export?21:18
devl82hey lisitsyn ..yeap exactly the same message21:19
lisitsyndevl82: what about removing build folder and running cmake?21:19
lisitsyndid you try that too?21:19
devl82nope i will try that.. thank you21:20
devl82very much:)21:20
lisitsyndevl82: let me know whether it works21:21
devl82ok..21:21
iglesiasghey hushell, how are you doing?21:25
shogun-buildbotbuild #91 of clang34 - static analysis is complete: Failure [failed analyse]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20static%20analysis/builds/91  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>21:31
-!- devl82 [5e45348a@gateway/web/freenode/ip.94.69.52.138] has quit [Ping timeout: 250 seconds]21:32
-!- devl82 [5e426536@gateway/web/freenode/ip.94.66.101.54] has joined #shogun21:49
devl82hi again.. @lisitsyn: no luck(s)21:59
devl82this is my pastebin:: http://pastebin.com/wx5SQRjv for evidence21:59
lisitsyndevl82: still cant find it?22:00
devl82nope22:00
lisitsyndevl82: ok let me see22:01
devl82also tried MATLAB_ROOT=/usr/local/MATLAB/R2013a and MATLAB_ROOT=/usr/local/MATLAB/R2013a/bin -after deleting build folder and running again ccmake22:02
shogun-buildbotbuild #72 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/72  blamelist: Soeren Sonnenburg <sonne@debian.org>22:03
lisitsyndevl82: do you have extern folder22:07
lisitsynin /usr/local/MATLAB/R2013a?22:07
devl82yes i have22:09
lisitsyndevl82: okay then what if you do22:09
lisitsyncmake -DMATLAB_ROOT=/usr/local/MATLAB/R2013a ..22:10
devl82ok one sec to check22:10
devl82omg it worked & -D parameter was already mentioned in readme22:14
lisitsynah I forgot about it22:14
lisitsyn;)22:14
lisitsynokay I am glad it worked22:14
devl82i should have guessed it..sorry for crunching your time22:14
devl82and thank you for all your suggestions22:14
shogun-buildbotbuild #74 of clang34 - thread analysis is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/74  blamelist: Soeren Sonnenburg <sonne@debian.org>22:17
lisitsynyou are welcome22:18
shogun-notifier-shogun-web: Soeren Sonnenburg :master * ef416df / static/css/style.css,static/js/overlay.js,templates/page.html: https://github.com/shogun-toolbox/shogun-web/commit/ef416df4d00d4ceaa758bff0ceda8324dbd6c27b22:24
shogun-notifier-shogun-web: don't use ul but separate div's for overlay22:24
@sonney2kshogun-buildbot, force build --branch=develop nightly_default22:24
shogun-buildbotThe build has been queued, I'll give a shout when it starts22:24
hushelliglesiasg: I am good :)22:26
hushelliglesiasg: how is the PR going? Is it good to merge?22:26
iglesiasghushell: let me see22:27
iglesiasghushell: so you removed the normalization you had added in the bundle methods at the end?22:29
hushelliglesiasg: yes, I feel in this way might be better, not touching the original code22:33
hushelliglesiasg: but when using verbose mode, the code will be 2x slower, I guess not too bad22:33
iglesiasghushell: yes, that is ok22:33
shogun-notifier-shogun: hushell :develop * e07efbe / / (14 files): https://github.com/shogun-toolbox/shogun/commit/e07efbe0cccabe25c961a452588ca31c5d075ca622:36
shogun-notifier-shogun: add helper functions to StructuredOutputMachine22:36
shogun-notifier-shogun: Fernando Iglesias :develop * 6060258 / / (14 files): https://github.com/shogun-toolbox/shogun/commit/6060258276085bde9e474d301663158c43066a4122:36
shogun-notifier-shogun: Merge pull request #1699 from hushell/develop22:36
shogun-notifier-shogun:22:36
shogun-notifier-shogun: add helper functions to StructuredOutputMachine22:36
hushelliglesiasg: Thanks! :) finally terminate it22:37
-!- mode/#shogun [+o lisitsyn] by ChanServ22:41
-!- ChanServ changed the topic of #shogun to: Shogun Machine Learning Toolbox | Devs are usually around between 10am-10pm UTC | When asking questions please be patient or use the mailing list. We always answer but it might take a while. | Channel logs http://bit.ly/eQDk8Q22:41
iglesiasghushell: :)22:42
@lisitsyngsoc is over right? ;)22:42
iglesiasgyes22:45
@sonney2klisitsyn, can you do some fixes?22:45
@lisitsynsonney2k: yes22:45
@lisitsynwhat fixes?22:45
shogun-buildbotbuild forced [ETA 29m09s]22:46
shogun-buildbotI'll give a shout when the build finishes22:46
@sonney2klisitsyn, hmmhh I am wondering if it is possible to pass float32 feature types to the Chi2Kernel's init()22:47
@sonney2klisitsyn, forget it22:47
@sonney2klisitsyn, everything OK22:47
@sonney2kwe nicely catch this22:48
@lisitsynsonney2k: that's design gap22:48
@sonney2kno all good22:48
@lisitsynsonney2k: no that's a gap ;)22:49
@lisitsynonce you get float32 features you are very restricted22:49
@sonney2klisitsyn, ?22:57
@sonney2knot very no22:57
@lisitsynsonney2k: how to use chi2 with float32?22:57
@sonney2knot possible22:58
@lisitsynsonney2k: and you say it is not gap? ;)22:58
@sonney2kyes22:58
@sonney2kgaussian, poly, linear, etc etc kernel work with any feature type22:58
@lisitsynbecause they are using dotfeatures, right?22:58
@sonney2kand I would rather not implement further ops in dotfeatures22:59
@sonney2kyes22:59
@lisitsynbut once we are using densefeatures shit happens22:59
@lisitsynthat's a major flaw22:59
@sonney2kWhy?22:59
@sonney2kit is like saying a string kernel should work with double23:00
@lisitsynno, but chi2 should work with float3223:00
shogun-buildbotbuild #92 of clang34 - static analysis is complete: Failure [failed analyse]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20static%20analysis/builds/92  blamelist: Soeren Sonnenburg <sonne@debian.org>23:01
-!- devl82 [5e426536@gateway/web/freenode/ip.94.66.101.54] has quit [Ping timeout: 250 seconds]23:02
@sonney2klisitsyn, where is the design flaw then?23:04
@sonney2klisitsyn, kernels are meant to be optimized for a single data type23:04
@sonney2kif we want high speed stuff there is no way around this23:04
@lisitsynsonney2k: no I am not that sure it is impossible23:05
@lisitsynsonney2k: at least it could work for both types23:06
@sonney2khow?23:06
@lisitsyndispatching the type once is not that difficult23:07
@lisitsynand it won't disable high speed23:07
@sonney2kugly!23:07
@sonney2kand won't work currently23:07
@lisitsynsonney2k: isn't she ugly right now?23:08
@sonney2ksince we require certain feature types in kernels23:08
@sonney2knot when drunk!23:08
iglesiasglol23:08
@lisitsynit must be stevie wonder's song about chi223:08
@lisitsynI still don't have any solution but it is flawed23:09
@lisitsynwell I know how to solve it23:10
@lisitsynbut you won't like it23:10
@lisitsyn:D23:10
@sonney2k?23:11
@lisitsynexpression templates etc23:11
@lisitsyncompile time will raise more23:11
@lisitsynalthough it may fail to resolve typing problem23:12
shogun-buildbotHey! build nightly_default #568 is complete: Success [build successful]23:16
shogun-buildbotBuild details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/56823:16
shogun-notifier-shogun-web: Soeren Sonnenburg :master * 05a0de4 / pages/views.py: https://github.com/shogun-toolbox/shogun-web/commit/05a0de41b8b998b8bed35ebc3fbb27d2c89a26e423:34
shogun-notifier-shogun-web: don't display notebooks w/o images23:34
--- Log closed Mon Oct 07 00:00:30 2013

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