Open in new window / Try shogun cloud
--- Log opened Tue Nov 05 00:00:11 2013
-!- pickle27 [~kevin@24-212-221-132.cable.teksavvy.com] has joined #shogun00:06
-!- zxtx [~zv@129-79-241-148.dhcp-bl.indiana.edu] has quit [Ping timeout: 268 seconds]00:11
-!- pickle27 [~kevin@24-212-221-132.cable.teksavvy.com] has left #shogun []00:13
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun00:30
shogun-notifier-shogun: Fernando Iglesias :develop * ae23d41 / doc/ipython-notebooks/statistics/mmd_two_sample_testing.ipynb: https://github.com/shogun-toolbox/shogun/commit/ae23d415f4286b8abc9b9cf161f35539003500d800:30
shogun-notifier-shogun: Shorter abstract in hypothesis testing notebook00:30
shogun-notifier-shogun: so that it fits better into the website00:30
shogun-notifier-shogun: Fernando Iglesias :develop * efc1c8c / doc/ipython-notebooks/statistics/mmd_two_sample_testing.ipynb: https://github.com/shogun-toolbox/shogun/commit/efc1c8c351160ff7975937f454c36044db433f1c00:30
shogun-notifier-shogun: Merge pull request #1749 from iglesias/develop00:30
shogun-notifier-shogun:00:30
shogun-notifier-shogun: Shorter abstract in hypothesis testing notebook00:30
shogun-buildbotbuild #1931 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1931  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>00:41
shogun-buildbotbuild #1932 of bsd1 - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/193200:46
-!- zxtx [~zv@c-98-223-198-47.hsd1.in.comcast.net] has joined #shogun00:51
-!- travis-ci [~travis-ci@ec2-184-72-203-210.compute-1.amazonaws.com] has joined #shogun01:18
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/1349892101:18
-!- travis-ci [~travis-ci@ec2-184-72-203-210.compute-1.amazonaws.com] has left #shogun []01:18
shogun-buildbotbuild #103 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/103  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>01:33
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has quit [Quit: Ex-Chat]01:35
shogun-buildbotbuild #102 of clang34 - thread analysis is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/102  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>01:49
-!- Netsplit *.net <-> *.split quits: lisitsyn01:58
-!- Netsplit over, joins: lisitsyn01:59
shogun-buildbotbuild #121 of clang34 - static analysis is complete: Failure [failed analyse]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20static%20analysis/builds/121  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>02:31
shogun-buildbotbuild #104 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/104  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>02:42
shogun-buildbotbuild #103 of clang34 - thread analysis is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/103  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>02:57
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]03:30
shogun-buildbotbuild #122 of clang34 - static analysis is complete: Failure [failed analyse]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20static%20analysis/builds/122  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>03:41
-!- benibadman [~benibadma@port-92-206-52-177.dynamic.qsc.de] has joined #shogun07:05
-!- benibadman [~benibadma@port-92-206-52-177.dynamic.qsc.de] has quit [Remote host closed the connection]07:05
-!- benibadman [~benibadma@port-92-206-52-177.dynamic.qsc.de] has joined #shogun07:05
-!- benibadman [~benibadma@port-92-206-52-177.dynamic.qsc.de] has quit [Ping timeout: 245 seconds]07:10
-!- benibadman [~benibadma@port-92-206-52-177.dynamic.qsc.de] has joined #shogun07:31
-!- benibadman [~benibadma@port-92-206-52-177.dynamic.qsc.de] has quit [Remote host closed the connection]07:54
-!- benibadman [~benibadma@port-92-206-52-177.dynamic.qsc.de] has joined #shogun07:54
-!- benibadman [~benibadma@port-92-206-52-177.dynamic.qsc.de] has quit [Ping timeout: 240 seconds]07:59
-!- sonne|osx [~sonne@89.204.139.126] has joined #shogun08:19
sonne|osxbesser82: whats up?08:20
-!- sonne|osx [~sonne@89.204.139.126] has quit [Quit: sonne|osx]08:35
-!- lisitsyn1 [~lisitsin@mxs.kg.ru] has joined #shogun08:55
-!- benibadman [~benibadma@94.135.236.129] has joined #shogun09:01
-!- new_lido [~walid@41.218.173.235] has joined #shogun09:02
-!- benibadman [~benibadma@94.135.236.129] has quit [Remote host closed the connection]09:02
-!- benibadman [~benibadma@94.135.236.129] has joined #shogun09:03
-!- sonne|work [~sonnenbu@24-134-74-216-dynip.superkabel.de] has quit [Quit: Leaving.]09:29
-!- sonne|work [~sonnenbu@24-134-74-216-dynip.superkabel.de] has joined #shogun09:48
-!- new_lido [~walid@41.218.173.235] has quit [Ping timeout: 264 seconds]09:53
-!- new_lido [~walid@41.218.181.144] has joined #shogun10:07
-!- benibadm_ [~benibadma@94.135.236.129] has joined #shogun10:11
-!- benibadm_ [~benibadma@94.135.236.129] has quit [Remote host closed the connection]10:12
-!- benibadman [~benibadma@94.135.236.129] has quit [Ping timeout: 246 seconds]10:15
-!- iglesiasg [~iglesias@2001:6b0:1:1da0:2089:9fad:841f:6fd9] has joined #shogun10:44
-!- mode/#shogun [+o iglesiasg] by ChanServ10:44
-!- lisitsyn1 [~lisitsin@mxs.kg.ru] has quit [Quit: Leaving.]12:32
-!- Saurabh7 [~Saurabh7@115.248.130.148] has joined #shogun13:21
-!- new_lido [~walid@41.218.181.144] has quit [Quit: Leaving]13:52
-!- taylan [d5f4a885@gateway/web/freenode/ip.213.244.168.133] has joined #shogun14:06
taylanHi All,14:07
taylanDoes any body know the role of directors?14:07
taylanCan't we extend regular classes, ie. ContingencyTableEvaluation instrad of DirectorContingencyTableEvaluation14:08
taylanA second question, I want to use an external evaluation, depending on the input dataset. Does set_indices function do what I want? i.e if I pass some cost annotation to the dataset, can I use the indices set by set_indices to access those global annotations?14:11
@iglesiasghey taylan14:25
@iglesiasgtaylan, regarding the director classes, the use case is more complex than that. They are useful to extend functionality of the libshogun interface using a modular interface14:26
@iglesiasgtaylan, so that, for example, code that you have implemented in a modular interface like Python can be called from anywhere in the libshogun interface14:27
taylanah, ok..so I don't need the directors for extending the class in python..14:28
taylanThank you for the answer..you got any clues for me for the second question?14:29
taylan Sorry, the question should be like this: If I pass some cost annotation to the evaluation class, can I use the indices set by set_indices to access those global annotations?14:30
@iglesiasgMm, I am sorry but I don't think I get the point14:31
taylanok, so instead of evaluating my classifier performance with the existing evaluators, I want to extend the Evaluation class, passing some costs for each label. This means, some samples are more costly to wrongly classify, and some are less..14:34
@iglesiasgall right14:34
@iglesiasgso I have never done this myself using Shogun directly but I would bet there must be something for sure to do this14:35
@iglesiasgsince it is something quite common14:35
taylanso I was wondering whether the set_indices method was passing me the real indices of the labels. This is most probably the case, I just wanted to make sure..14:36
taylanok, I will anyway give it a try and write back here.14:36
@iglesiasgsure, try it better14:37
@iglesiasgin case you don't get an answer about it right, I suggest you to send a mail to the mailing list14:37
@iglesiasgHeiko should be able to explain this without a trouble, he implemented this part I think14:37
taylanok, thanks a lot!14:48
@iglesiasgtaylan, no problem, you are welcome!14:51
-!- thoralf [~thoralf@enki.zib.de] has joined #shogun15:16
thoralfHeyho.15:17
-!- benibadman [~benibadma@94.135.236.129] has joined #shogun15:19
@wikingyo15:20
sonne|workwiking: another question15:30
sonne|workwiking: I want to display the images from the slide thing on the cloud website15:30
sonne|workwiking: problem is that we need the notebook directory somewhere15:31
sonne|workon the cloud server15:31
sonne|workI mean the pre-rendered ones15:31
sonne|workso I can automagically extract the images15:31
@wikingsonne|work: images?15:32
sonne|workwiking: the thumbnails we have in the slide thing at the bottom15:32
@wikingsonne|work: so you want to actually add the prerendered images?15:32
sonne|workwiking: so you can just click on them15:33
@wikingclick on them where?15:33
sonne|workand that would open the notebook15:33
sonne|workwiking: cloud.*15:33
@wikingah before login?15:33
@wikingat the login page?15:33
sonne|workinstead of this login window you would see as main page the notebook15:33
sonne|workbasically just the images in a grid15:33
@wikingand then a login on the button?15:34
sonne|workyeah15:34
@wiking*buttom15:34
sonne|workbottom15:34
sonne|workyes15:34
@wikingmmm15:34
@wikingok well modify the template15:34
sonne|work?15:34
@wikingindex.html is the template for the login page15:34
sonne|workI know15:34
sonne|workI started doing this already15:34
@wikingso just modify it and that's all15:34
sonne|workno15:34
sonne|workit is not all15:35
@wikingwhy not?15:35
sonne|workI need the prerendered notebooks15:35
@wikingit's on the shogun website15:35
@wikingstatic/..15:35
sonne|workthat doesn't help15:35
@wikingwhy not?15:35
sonne|workwe need them locally on the machine running cloud.*15:35
@wikingwhy?15:35
sonne|workbecause I parse the notebooks and extract the images from them15:36
@wikingi mean you can add an <img src="whatever" />15:36
sonne|workno static image15:36
@wikingwhy would it be a non-static15:36
sonne|workwiking: well it is like this on www.* etc too15:37
@wikingso?15:37
sonne|workI need the notebooks15:37
sonne|workthat is all15:37
@wikingjquery15:37
@wiking?15:38
sonne|work??15:38
@wikingfetch the notebook and parse it with jquery15:38
@wikingthe preprendered ipynb + html is available15:38
@wikingon www.15:38
@wikingjust fetch it with jquery15:38
sonne|workyou want some dependency on www.* on cloud.* ?15:38
sonne|workis that clever?15:38
@wikingwhat's the problem?15:38
sonne|workwiking: the dependency that is all15:49
sonne|workwiking: you need www.* to work - if it is down you don't have cloud.* working15:50
-!- lisitsyn [~lisitsyn@80.252.20.67] has quit [Ping timeout: 245 seconds]16:07
-!- lisitsyn [~lisitsyn@80.252.20.67] has joined #shogun16:11
thoralfWhen using the python_modular interface, is there anything I need to consider to avoid memory leaks?16:26
thoralfIt seems like me programm is heavily leaking memory (using liblinear in a loop)16:26
sonne|workthoralf: errm it shouldn't what are youdoing?16:30
sonne|workin any case w/ python16:30
sonne|workdel object16:30
sonne|workas often as possible...16:30
sonne|workwell early16:30
thoralfsonne|work: I'm loading a big data file (csc matrix) and creating sparse features + labels.  Running a loop hundreds of times on the same data - and already "del predicted_labels" and "del svm" after every run.16:32
thoralfresource.struct_rusage().ru_maxrss grows by 400-500 MB in every iteration.16:32
sonne|workthoralf: well only chance is to do some valgrind thing on a small data set to check if there is no leak16:33
thoralfsonne|work: Okay, I thought I possibly did something stupid and could avoid valgrind...16:34
thoralfThanks ;)16:34
sonne|workthoralf: I am not aware of a leak in liblinear / sparse etc16:35
sonne|workso no idea16:35
-!- taylan [d5f4a885@gateway/web/freenode/ip.213.244.168.133] has quit [Ping timeout: 250 seconds]16:39
-!- Saurabh7 [~Saurabh7@115.248.130.148] has quit [Ping timeout: 268 seconds]16:43
-!- zxtx [~zv@c-98-223-198-47.hsd1.in.comcast.net] has quit [Ping timeout: 246 seconds]17:12
-!- taylan [d5f4a885@gateway/web/freenode/ip.213.244.168.133] has joined #shogun17:22
-!- benibadman [~benibadma@94.135.236.129] has quit [Remote host closed the connection]17:27
-!- iglesiasg [~iglesias@2001:6b0:1:1da0:2089:9fad:841f:6fd9] has quit [Quit: Ex-Chat]17:31
-!- sonne|osx [~sonne@89.204.130.113] has joined #shogun17:56
-!- zxtx [~zv@129-79-241-148.dhcp-bl.indiana.edu] has joined #shogun17:58
-!- sonne|osx [~sonne@89.204.130.113] has quit [Quit: sonne|osx]18:07
@wikingthoralf: u in tuberlin?19:08
thoralfwiking: No, Zuse-Institute-Berlin19:08
-!- sonne|osx [~sonne@f053044146.adsl.alicedsl.de] has joined #shogun19:26
thoralfsonne|work: Maybe you can tell me whats wrong: https://github.com/shogun-toolbox/shogun/issues/175019:26
thoralfHey sonne|osx19:27
thoralfsonne|osx: I can reproduce the leak in a minimal example.19:28
@wikingthoralf: oh man i thought that that leak was fixed already19:34
thoralfwiking: So it's already known?  Do you maybe know a hotfix for this?19:37
@wikingthoralf: i've fixed this19:38
@wikingbut now i've realised19:38
@wikingthat actually it wasn't this19:38
@wiking:(19:38
@wikingit was in multimachine19:38
@wikingletmecheck19:38
@wikingmaybe the same thing happened ther19:38
@wikingoh man this looks weird19:40
@wikingthoralf: i guess u haven't tried with c++19:40
thoralfwiking: I'm about to...19:41
thoralfwiking: Hmm.  I don't need to.  I've been using LibLinear all the time, I'm memchecking all my programs, so either new or it should be working in c++19:44
-!- sonne|osx [~sonne@f053044146.adsl.alicedsl.de] has quit [Quit: sonne|osx]19:49
-!- sonne|osx [~sonne@f053044146.adsl.alicedsl.de] has joined #shogun20:01
thoralfwiking: All heap blocks were freed -- no leaks are possible20:18
-!- sonne|osx [~sonne@f053044146.adsl.alicedsl.de] has quit [Quit: sonne|osx]20:24
-!- FSCV [~FSCV@50.7.50.60] has joined #shogun23:12
--- Log closed Wed Nov 06 00:00:12 2013