--- Log opened Mon Jun 23 00:00:36 2014
--- Day changed Mon Jun 23 2014
-!- thor|alf [~thoralf@91-65-141-60-dynip.superkabel.de] has quit [Quit: Konversation terminated!]00:00
shogun-buildbot_build #2381 of deb3 - modular_interfaces is complete: Failure [failed test python modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/2381  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Fernando Iglesias <fernando.iglesiasg@gmail.com>, Parijat Mazumdar <mazumdarparijat@gmail.com>00:23
@HeikoSany webmaster around?00:23
@HeikoSwiking: lambday is on fixing them00:23
@HeikoSwiking: could you prepare a video-stream for the main day of the workshop? and then I just connect to it like last year?00:37
-!- vivekrai [uid10305@gateway/web/irccloud.com/x-znnetyqdtpqyulex] has quit [Quit: Connection closed for inactivity]00:40
@HeikoSwiking: see email, voting on topics, high priority01:36
@HeikoSthoralf: I sent a mail, some bits are missing for the texts (only links though), so once fernando has set up a page, you can set up the events01:37
-!- HeikoS [~heiko@] has quit [Quit: Leaving.]01:49
-!- HeikoS [~heiko@] has joined #shogun01:50
-!- mode/#shogun [+o HeikoS] by ChanServ01:50
-!- HeikoS [~heiko@] has quit [Quit: Leaving.]01:55
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]01:55
-!- PirosB3 [~pirosb3@host76-224-dynamic.52-79-r.retail.telecomitalia.it] has joined #shogun02:01
-!- travis-ci [~travis-ci@ec2-54-242-2-55.compute-1.amazonaws.com] has joined #shogun02:11
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/2817924202:11
-!- travis-ci [~travis-ci@ec2-54-242-2-55.compute-1.amazonaws.com] has left #shogun []02:11
-!- khalednasr [~k.nasr92@] has quit [Quit: Leaving]02:23
-!- PirosB3 [~pirosb3@host76-224-dynamic.52-79-r.retail.telecomitalia.it] has quit [Quit: PirosB3]02:44
shogun-buildbot_build #729 of nightly_all is complete: Failure [failed compile test]  Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_all/builds/729  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Wu Lin <yorker.lin@gmail.com>, Fernando Iglesias <fernando.iglesiasg@gmail.com>, Parijat Mazumdar <mazumdarparijat@gmail.com>03:05
-!- pickle27 [~pickle27@192-0-136-118.cpe.teksavvy.com] has joined #shogun04:13
-!- pickle27 [~pickle27@192-0-136-118.cpe.teksavvy.com] has quit [Quit: Leaving]04:59
-!- zxtx_ [~zv@] has quit [Ping timeout: 244 seconds]05:27
-!- vivekrai [uid10305@gateway/web/irccloud.com/x-ttgkfgdtbkonbvrv] has joined #shogun06:24
shogun-buildbot_build #833 of nightly_default is complete: Failure [failed test notebooks]  Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/833  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Wu Lin <yorker.lin@gmail.com>, Fernando Iglesias <fernando.iglesiasg@gmail.com>, Parijat Mazumdar <mazumdarparijat@gmail.com>06:38
-!- lisitsyn1 [~qdrgsm@] has quit [Ping timeout: 240 seconds]09:31
-!- PirosB3 [~pirosb3@host76-224-dynamic.52-79-r.retail.telecomitalia.it] has joined #shogun09:50
-!- lisitsyn [~qdrgsm@] has joined #shogun09:52
@besser82wiking, does it work now?09:53
@besser82wiking, the sg-ca I mean...09:53
-!- thoralf [~thoralf@91-65-216-252-dynip.superkabel.de] has quit [Quit: Konversation terminated!]09:56
@wikingbesser82: lemme check09:57
@wikingbesser82: noup same thing09:58
@besser82wiking, you tried port 8000 / 8080  ?09:58
@besser82wiking, and opened it using iptables?09:58
@wikingi'm trying now 8009:59
@wikingjust a sec09:59
@wiking80 works09:59
@besser82If you are fine with 80, this seems to be resolved?!?10:00
@besser82wiking, ^^10:00
@wiking8080 works as well10:00
@besser82wiking, so problem is resolved  ^^10:00
@wikingyeah it's cool10:01
@besser82wiking, np  ;)10:01
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun10:34
shogun-notifier-shogun-data: Saurabh :master * bd2b5ba / toy/diabetes_scale.svm: https://github.com/shogun-toolbox/shogun-data/commit/bd2b5ba3b816393f0680a2d5c3d1da5522af4d6710:34
shogun-notifier-shogun-data: diabetes dataset added10:34
shogun-notifier-shogun-data: Fernando Iglesias :master * e3ff192 / toy/diabetes_scale.svm: https://github.com/shogun-toolbox/shogun-data/commit/e3ff192d86d0fbed8f54f358b7eaec909188588c10:34
shogun-notifier-shogun-data: Merge pull request #58 from Saurabh7/master10:34
shogun-notifier-shogun-data: diabetes dataset added10:34
-!- HeikoS [~heiko@] has joined #shogun10:38
-!- mode/#shogun [+o HeikoS] by ChanServ10:38
-!- HeikoS [~heiko@] has quit [Ping timeout: 240 seconds]11:06
shogun-notifier-shogun-data: lambday :master * 5ff0db6 / testsuite/ (12 files): https://github.com/shogun-toolbox/shogun-data/commit/5ff0db6f766fd796d6596f75a35c1fe540ad588f11:31
shogun-notifier-shogun-data: Added old data to develop for fixing build temporarily11:31
shogun-notifier-shogun-data: Soumyajit De :master * a10320f / testsuite/ (12 files): https://github.com/shogun-toolbox/shogun-data/commit/a10320f2494706f071d9c3cfe4967092dd59bfef11:31
shogun-notifier-shogun-data: Merge pull request #59 from lambday/master11:31
shogun-notifier-shogun-data: Added old data to develop for fixing build temporarily11:31
-!- HeikoS [~heiko@nat-172-18.internal.eduroam.ucl.ac.uk] has joined #shogun11:44
-!- mode/#shogun [+o HeikoS] by ChanServ11:44
-!- PirosB3 [~pirosb3@host76-224-dynamic.52-79-r.retail.telecomitalia.it] has quit [Quit: PirosB3]12:11
-!- soumyaC [uid15286@gateway/web/irccloud.com/x-amaguduemdekoqys] has joined #shogun12:22
-!- PirosB3 [~pirosb3@host76-224-dynamic.52-79-r.retail.telecomitalia.it] has joined #shogun12:44
-!- HeikoS [~heiko@nat-172-18.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.]13:42
-!- Saurabh7 [~Saurabh7@] has joined #shogun13:42
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]14:31
-!- lisitsyn [~qdrgsm@] has quit [Ping timeout: 244 seconds]14:34
-!- PirosB3 [~pirosb3@host76-224-dynamic.52-79-r.retail.telecomitalia.it] has quit [Quit: PirosB3]14:52
-!- HeikoS [~heiko@nat-173-40.internal.eduroam.ucl.ac.uk] has joined #shogun15:45
-!- mode/#shogun [+o HeikoS] by ChanServ15:46
-!- kislay [~Abhijeet@] has joined #shogun16:09
-!- sonne|work [~sonnenbu@sams-office-nat.tomtomgroup.com] has quit [Quit: Leaving.]16:35
-!- pickle27 [~pickle27@] has joined #shogun17:05
-!- Saurabh7 [~Saurabh7@] has quit [Ping timeout: 244 seconds]17:32
-!- Saurabh7 [~Saurabh7@] has joined #shogun17:44
-!- Saurabh7 [~Saurabh7@] has quit [Ping timeout: 255 seconds]17:52
-!- pickle27 [~pickle27@] has quit [Quit: This computer has gone to sleep]17:53
@HeikoSwiking: hey!17:58
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun18:03
shogun-notifier-shogun-demo: Saurabh :master * bdaa4d4 / / (5 files): https://github.com/shogun-toolbox/shogun-demo/commit/bdaa4d43a3b982295d487e1fdd08e3cd9969728e18:03
shogun-notifier-shogun-demo: add diabetes dataset18:03
shogun-notifier-shogun-demo: Heiko Strathmann :master * 9bc285c / / (5 files): https://github.com/shogun-toolbox/shogun-demo/commit/9bc285c569727cc4f92b48b5db31bb7c8f1fa41918:03
shogun-notifier-shogun-demo: Merge pull request #44 from Saurabh7/dataset18:03
shogun-notifier-shogun-demo: add diabetes dataset18:03
@wikingHeikoS: ho18:03
@HeikoSwiking: hey good to catch you. saw my mail?18:04
@HeikoSwiking: could you help me a bit with this?18:04
@HeikoSwould be great18:04
@wikingHeikoS: with what exactly?18:04
@wikinga poll?18:04
-!- Saurabh7 [~Saurabh7@] has joined #shogun18:04
@HeikoSwiking: see emaoil18:05
@HeikoSwiking: there is a list of topics in the googledoc18:05
@HeikoStalk topics18:05
@HeikoSwe just need a vote on them18:05
@HeikoSwith email and name, institution18:05
@wikingHeikoS: and how can i help in that18:05
@wikingi mean the last time i used some free service18:05
@wikingto do that18:05
@HeikoS(and a checkbox for each "attending main day" "attending hands on"18:05
@HeikoSyeah fine18:05
@HeikoSjust to distribute the work a bit18:05
@HeikoSwiking: yeah see our website for how the thing looked like18:06
@HeikoSwiking: btw there is a new issue on github that me and dino just ran into this morhing, could you also have a look, maybe you know a nice fix (that doesnt involve re-compile python as google suggests)?18:07
@wikingno iea18:07
@HeikoSwiking: our website has the workshop 2013 page18:07
@wikingno idea about that18:07
@HeikoSwiking: the issue?18:08
@HeikoSor the poll?18:08
@HeikoSah I see, some ubuntu bug of the new "stable" version18:08
@wikingyou mean the registration form/18:08
@wikingit wasn't done by me18:08
@wikingdunno who created taht18:09
@wikingbut seems that's like a google spreadshite somehow embedded into the page18:09
@HeikoSwiking: if you could figure that out and do, that would be of great help18:10
@wikingbut i have no idea how that was done18:10
shogun-notifier-shogun: Abhijeet :develop * f7f8a21 / tests/unit/classifier/LDA_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/f7f8a21a938f23d83fcacc6e394af41315fe8ecc18:11
shogun-notifier-shogun: CLDA unittest added18:11
shogun-notifier-shogun: Heiko Strathmann :develop * 804bb9f / tests/unit/classifier/LDA_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/804bb9ffc3a741b61bdc324d4911160de9e0afd318:11
shogun-notifier-shogun: Merge pull request #2331 from kislayabhi/unittest_old_lda18:11
shogun-notifier-shogun: CLDA unittest added18:11
@HeikoSwiking: any other idea for a simple vote?18:11
@HeikoSdoesnt hav eto be super nice18:11
@wikingi reckon it'd be better to find out how this was done and just copy-paste the gdocs18:13
@HeikoSwiking: yeah, exactly18:13
@wikingi reckon it's better to ask about this on the mailinglist :)18:14
@HeikoSwiking: just am so busy with organising the other things about the workshop. Would be cool to get some help with this Could you push this to be done today or tomorrow? we need this asap18:14
@HeikoSwiking: mailing list doesnt reply currently, better figure out self18:14
shogun-buildbot_build #2392 of bsd1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/2392  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Abhijeet <abhijeetkislay@gmail.com>18:15
@wikingHeikoS: do we want somebody to give a talk about collaborative filtering?18:18
shogun-buildbot_build #365 of deb4 - python3 is complete: Failure [failed test python modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb4%20-%20python3/builds/365  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Abhijeet <abhijeetkislay@gmail.com>18:21
@HeikoSwiking: for the talks, I think its necessary that a) somebody knows about stuff (though I am happy to talk about anything) b) it is in shogun18:24
@HeikoSwiking: so does shogun have filtering things?18:24
@HeikoSwe dont even have PCA/SVD with missing values :)18:24
@HeikoS(though I issued that a while ago for parijat ;) )18:25
-!- jiaolong [9e6d09e4@gateway/web/freenode/ip.] has quit []18:25
@wikingHeikoS: isn't somebody working on ALS stuff as gsoc?18:26
@wikingi saw movielens dataset imported18:27
@wikingor that was pre-gsoc18:27
@wikingwhat's wu lin working on?18:28
@wikingah no he is gp+kl -ing18:28
@HeikoSwiking: yeah, one can use this for CF, but we kind of are focussing on other stuff (its in the list of talks)18:29
@HeikoSwiking: ok thanks for pushing the vote, if you see fernando, dont let him go until he did that (and the website workshop update)18:29
-!- parijat [75dc2abb@gateway/web/freenode/ip.] has joined #shogun18:30
parijatHeikoS: Hello... do you have a min?18:33
parijatHeikoS: can you please have a look at this PR : https://github.com/shogun-toolbox/shogun/pull/2316 .18:34
@HeikoSparijat:  sure18:34
parijatHeikoS: the SG_ADD is failing SerializationAscii test and SerializationJSON tests18:34
parijatHeikoS: see travis here : https://travis-ci.org/shogun-toolbox/shogun/jobs/2820435018:35
@HeikoSparijat: whats the type of m_delta?18:35
parijatHeikoS: float64_t18:35
@HeikoSparijat: implement std constructor and make it call init18:36
@HeikoSthat should solve it18:36
@HeikoSits a check for that, makes sure you initialise your variables for all constructors ;)18:36
@HeikoSparijat: in the code, you only initialise with non-std constructor18:37
parijatHeikoS: but this constructor is supposed to be std constructor overloaded..18:37
parijatHeikoS: isn't it?18:38
@HeikoSparijat: I might be wrong, but I dont see18:38
@HeikoSCHuberLoss::CHuberLoss() {init;}18:38
@HeikoSCHuberLoss::CHuberLoss() {init();}18:39
parijatHeikoS: I actually overloaded 2 constructors by using default value : CHuberLoss(float64_t delta=CMath::MAX_REAL_NUMBER)18:40
-!- Saurabh7 [~Saurabh7@] has quit [Remote host closed the connection]18:40
@HeikoSparijat: I see, sorry18:40
@HeikoSparijat: double check whether the delta is initialised when you call emptry18:41
@HeikoSjust to be sure18:41
@HeikoSotherwise you will have to debug to find out18:41
@HeikoSonly thing that I can think of18:41
@HeikoScheck whats the difference in the unit test18:41
parijatHeikoS: ok let me see... The error you say is mostly caused by values not being initialized properly right?18:42
@HeikoSparijat: yes18:44
parijatHeikoS: alright! Thanks for helping out.. :)18:45
shogun-buildbot_build #2382 of deb3 - modular_interfaces is complete: Failure [failed test python modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/2382  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Abhijeet <abhijeetkislay@gmail.com>18:47
-!- yorkerlin [c0008056@gateway/web/freenode/ip.] has joined #shogun18:51
-!- pickle27 [~pickle27@] has joined #shogun18:57
-!- pickle27 [~pickle27@] has quit [Quit: This computer has gone to sleep]19:43
-!- pickle27 [~pickle27@] has joined #shogun19:46
-!- kislay [~Abhijeet@] has quit [Quit: Leaving]19:54
@HeikoSpickle27: do you have access to the website?19:55
pickle27I do not19:56
pickle27I believe only sonney2k does19:56
@HeikoSpickle27: no fernando also did some editing19:56
@HeikoSpickle27: alright19:56
pickle27I have write access on the github repo19:56
@HeikoSpickle27: what does this mean?19:56
pickle27I can merge PRs on github that change the code19:56
pickle27some people have access to the admin and can edit the db entries19:57
pickle27but only sonney2k can redeploy the page afaik19:57
@HeikoSpickle27: I want a new page, content is already written19:57
pickle27I think sonney2k can give you access to the django admin and then you can add it there19:57
-!- iglesiasg [524ae0a7@gateway/web/freenode/ip.] has joined #shogun19:58
@HeikoSiglesiasg: hi!19:58
@HeikoSiglesiasg: good to see you! :)19:58
iglesiasgHeikoS: hey19:59
@HeikoSiglesiasg: were just talking about you and sent you an email19:59
iglesiasgHeikoS: yep, I saw :)19:59
iglesiasgHeikoS: I am checking Parijat's PR19:59
iglesiasgHeikoS: anyhow19:59
@HeikoSiglesiasg: the weorkshop thing should have priority #1 for now19:59
@HeikoSat least until the registration works and program is online19:59
@HeikoSrest can be figured out later19:59
@HeikoSbut we need the website for the registration event and the call for participation email20:00
iglesiasgHeikoS: I was going to mail Soeren and Kevin about that because I cannot alter any content on the website20:00
@HeikoSiglesiasg: just talked to pickle27, he also cannot do that20:00
iglesiasgHeikoS: this is a pending issue for some weeks now20:00
@HeikoSiglesiasg: just emailed s?ren, lets hope he gets back20:00
@HeikoSiglesiasg: Ah,20:00
iglesiasgyeah, kevin told me that. We need Soeren it seems20:00
pickle27its a very complicated setup at the moment20:01
iglesiasgpickle27: but I am wondering -- so it isn't possible to add new users either?20:01
@HeikoSiglesiasg: you were changing things earlier didnt you?20:01
pickle27a lot of that website is specifically configured for that server20:01
pickle27it is possible but I think soeren has to do it, at least I don't know how20:02
@HeikoSpickle27: yeah, but for now that doesnt matter, we just need to add one item20:02
iglesiasgHeikoS: yes, but then pickle27 took over for the website -- which was very good since I suck at web20:02
@HeikoSpickle27, iglesiasg just mailed him twice20:02
pickle27I also don't have access at the moment because my admin account on the site got cleared when we shipped the update this fall20:02
pickle27I have a totally new version that I am working on20:02
@HeikoSiglesiasg, pickle27 thing is we really need to add this site somehow. So lets email s?ren until he gives rights to all of us :)20:02
pickle27it will be easier to change :)20:02
@HeikoSsonney2k: website access !!!20:04
@HeikoSsonney2k: see my emails20:04
pickle27HeikoS, I haven't shown you yet20:04
iglesiasgthe thing is that my username and password still work20:04
pickle27here is the new version of the site I've been playing with20:04
iglesiasgbut I get this message20:04
@HeikoSpickle27: cool, looking forward to see this :)20:04
iglesiasgSite administration  You don't have permission to edit anything.20:04
pickle27iglesiasg, I don't even have an account any more20:05
iglesiasgargh weird stuff20:05
@HeikoSpickle27: ah nice, looking forward to use this one20:05
@HeikoSiglesiasg: could you send sonney2k an email?20:05
@HeikoSpickle27: and you too?20:05
pickle27I replied to that email20:05
@HeikoSpickle27: saw it, lets hope :)20:06
pickle27HeikoS, yeah it is much simpler and has no database so to edit content you just make a PR to github20:06
@HeikoSyeah totally20:06
pickle27less layers will be easier to maintain20:06
@HeikoSiglesiasg: did you do the voting last year?20:07
@HeikoSiglesiasg: independent of yes or no, could you set one up? asked wiking but he did not do this last time20:07
@HeikoSiglesiasg: also very urgent20:07
iglesiasgHeikoS: Sergey did this20:08
iglesiasgHeikoS: he used some google doc I think20:08
iglesiasghe set up a form using google docs20:08
@HeikoSiglesiasg: would you mind doing this this year?20:08
iglesiasgHeikoS: with voting you mean for the topics of the talks?20:08
@HeikoSiglesiasg: yes we have a list in the googledoc20:08
iglesiasgHeikoS: ok20:09
@HeikoSiglesiasg: also feel free to change/add20:09
@HeikoSiglesiasg: cool thanks20:09
@HeikoSiglesiasg: about RF talk20:13
@HeikoSiglesiasg: if we add this to the votes and poeple vote on it, would you be interested in talking about basics? and show sshogun notebook demos?20:13
@HeikoSiglesiasg: its just 30 mins, so you can only cover basics anyways20:13
@HeikoSiglesiasg: no problem if not, since you also did the LMNN20:13
-!- abinash_panda [~abinash@] has joined #shogun20:14
iglesiasgHeikoS: describe basics :)20:16
iglesiasgparijat: hey!20:17
iglesiasgparijat: keep me up to date with what Travis says :)20:17
-!- parijat [75dc2abb@gateway/web/freenode/ip.] has quit [Ping timeout: 246 seconds]20:17
-!- khalednasr [~k.nasr92@] has joined #shogun20:31
@HeikoSyorkerlin: hey!20:35
-!- lisitsyn [~qdrgsm@] has joined #shogun20:36
iglesiasgHeikoS, pickle27 we got an answer20:46
pickle27we did but Im confused ...20:46
yorkerlin@HeikoS I am working on some unit tests of the dual KL inference. Once they are passed, I will send  PRs for the KL method. (including variational Dual Likelihood) Maybe tonight or tomorrow.  If you can see some Matlab code I written if you want to take a quick look.20:50
@HeikoSiglesiasg: another one, could you try20:51
@HeikoSiglesiasg: works?20:52
iglesiasgHeikoS: didn't try to edit anything, but now I see the normal menus -- so yeah, it must work now20:52
@HeikoSiglesiasg: better double check :)20:52
iglesiasgHeikoS: working http://shogun-toolbox.org/page/Events/workshop201320:53
iglesiasgfind the easter egg if you like20:54
@HeikoSiglesiasg: found it :)20:54
@HeikoSiglesiasg: awesome!20:54
@HeikoSiglesiasg: could you add the program?20:54
@HeikoS(see email from yesterday)20:54
iglesiasgthe program in google doc, sure20:55
iglesiasgHeikoS: have a look at the new registration form if you have not done so yet, please20:55
@HeikoSiglesiasg: what do you mean?20:55
@HeikoSiglesiasg: sorry confused, where is that?20:56
iglesiasgHeikoS: just mailed it few minutes ago20:56
iglesiasgHeikoS: Workshop TODO thread20:56
@HeikoSiglesiasg: ah, cool!20:56
iglesiasgpretty much recycled from last year20:57
@HeikoSiglesiasg: I will mail some suggestions20:57
@HeikoSiglesiasg: we want to do registration over event-brite20:58
@HeikoSso this is just for voting20:58
@HeikoSiglesiasg: But I guess its still fine20:58
iglesiasgno idea what event-brite is, but ok20:59
@HeikoSiglesiasg: approved! Could you put this into the google doc and let thoralf know?20:59
@HeikoSiglesiasg: its a platform for registering events in an oragnised way20:59
@HeikoSiglesiasg: sends out reminders etc20:59
@HeikoSiglesiasg: gives directions ...20:59
@HeikoSand people click through the offers21:00
iglesiasgHeikoS: I will share the form with you guys in a second -- working on the webpage right now21:00
@HeikoSiglesiasg: amazing man! :)21:00
iglesiasgHeikoS: but what do you want into the Google doc??21:00
@HeikoSiglesiasg: link to form21:01
@HeikoSiglesiasg: for thoralf to see21:01
@HeikoSiglesiasg: he will set up the eventbrite21:02
shogun-buildbot_build #142 of osx2 - python is complete: Failure [failed test python modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx2%20-%20python/builds/142  blamelist: Wu Lin <yorker.lin@gmail.com>21:02
@HeikoSyorkerlin: what about the other PRs21:03
yorkerlindual variational likelihood21:04
yorkerlinmodified lbfgs21:04
@HeikoSyorkerlin: I see21:05
@HeikoSyorkerlin: this dual method, is it already working for you locally?21:05
yorkerlinyes. But I have to test it and clean-up the code.  I can send you the Matlab code first.21:06
yorkerlinThe Matalb code is tested21:07
yorkerlinThe Matlab is based on Poisson distrbution, which has the close form of varational expection.21:09
@HeikoSI see21:09
@HeikoSyorkerlin: just wondering how this scales up in practice, how many points can you run this with locally?21:09
yorkerlinBasically, it like the FullDiagonal method. It means O(n^3) but seems have fast convergence rate21:11
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]21:11
yorkerlinIn Latent Gaussian model, n is the dim of the latent space. In GP, n is the number of sample point21:12
yorkerlinCholesky KL method has O(n^2) at each L-BFGS update. ApproxDiagonal has O(n) at each L-BFGS update.  Fast Dual KL method seems has fast convergence rate of L-BFGS compared to FullDiagonal KL method.21:15
yorkerlinI only test on small dataset. I will test on some real dataset which has more then 1000 points21:17
yorkerlinDo you have any suggestion?21:18
-!- PirosB3 [~pirosb3@host76-224-dynamic.52-79-r.retail.telecomitalia.it] has joined #shogun21:21
@HeikoSyorkerlin: yes, I would like to see what happens when you use 10000 pooints21:24
@HeikoSyorkerlin: but go ahead with your sutff for now21:24
@HeikoSyorkerlin: we should in fact stress test all your code in such larger regimes and put the scripts into benchmark21:25
yorkerlinOK. I will test them locally first once I finish my sutff.21:26
iglesiasgHeikoS: http://shogun-toolbox.org/page/Events/workshop201421:27
iglesiasgHeikoS: http://shogun-toolbox.org/page/Events/workshop2014_program21:27
@HeikoSiglesiasg: nice21:27
@HeikoSiglesiasg: what about - "Give your data a treat - Vol. 2"21:27
iglesiasgHeikoS: let me know if you want to see anything change21:27
iglesiasgHeikoS: sure21:27
iglesiasgHeikoS: if you don't want to edit just let me know anything - doesn't matter how tiny the detail is ;-)21:28
@HeikoSiglesiasg: TODO link should be replaced with your vote system.21:28
iglesiasgHeikoS: sure21:28
@HeikoSiglesiasg: maybe even remove the list of topics in there, just put the lunk21:28
iglesiasgit makes sense21:28
@HeikoSiglesiasg: all fine apart from that. thanks!21:28
iglesiasgHeikoS: but I understand that in the form I created I should just keep the topic votation21:29
@HeikoSok, now thoralf can set up the events on event brite21:29
iglesiasgHeikoS: shall I remove everyting else then?21:29
@HeikoSiglesiasg: let me check21:29
iglesiasgbtw, for the brite -- we should ask people about veg food21:29
iglesiasglisitsyn: sup man21:29
shogun-buildbot_build #143 of osx2 - python is complete: Failure [failed test python modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx2%20-%20python/builds/143  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Fernando Iglesias <fernando.iglesiasg@gmail.com>, Abhijeet <abhijeetkislay@gmail.com>, Parijat Mazumdar <mazumdarparijat@gmail.com>21:29
lisitsynI sent a few mails back then21:29
lisitsynto all the participants21:29
lisitsyncan do the same this time21:29
lisitsynjust before the workshop21:29
@HeikoSiglesiasg: lets just order/provide a few veg things21:29
@HeikoSiglesiasg: do you have the vote link again?21:30
lisitsyniglesiasg: IIRC we had veg issue21:30
@HeikoSiglesiasg: main page of workshop should contain the text call for participation in the google doc, and the link to the vote site21:31
@HeikoSlisitsyn: ok, we can do thi21:31
iglesiasgHeikoS: docs.google.com/forms/d/1GHve9BAxnJDe4LvRdplt-Bs9ARWCB90uZqPFHojSliE/viewform21:31
@HeikoSiglesiasg: I think its fine. What about: "Vegetarian food?"21:32
iglesiasgHeikoS: but the thing is -- are we going to use this form or this event brite thing for all these questions?21:33
@HeikoSiglesiasg: goo dquestion21:33
@HeikoSiglesiasg: thoralf brought this up for the hands on21:33
iglesiasglast I understood was the current form should only contain topics21:33
@HeikoSiglesiasg: maybe we should only have this for the hands-on and do the main registration over this form21:33
@HeikoSiglesiasg: otherwise three things to fiull in21:33
iglesiasgwell, since it is thoralf who would set up this event brite, let's wait for his opinion21:33
iglesiasgyeah, I agree. Already more than one thing to fill in is too much21:34
@HeikoSiglesiasg: thoralf wanted to have this eventbrite to get a list of people, but we would get this anyways21:34
@HeikoSiglesiasg: and to advertise21:34
iglesiasgwith this we get list of people in an spreadsheet in drive. Pretty handy if you ask me21:34
@HeikoSiglesiasg: ok, what about having a checkbox for each "plan to attend workshop main day" and "plan to attend hands-on session ?21:34
@HeikoSiglesiasg:  I agree, maybe lets ditch event-brite21:34
@HeikoSiglesiasg: yeah then add three things: main day (28 c-base), hands-on (27, researchgate)21:35
@HeikoSand veg21:35
@HeikoSiglesiasg: thoralf can add the eventbrite thing to get more prople involved, but we dont need to advertise I gues21:36
@HeikoSiglesiasg: lets wait for his opinon. could you write an email and ask=?21:36
iglesiasgHeikoS: yes21:36
@HeikoSiglesiasg: in fact, lets just announce using this, we can put the eventbrite thing afterwards21:36
@HeikoSiglesiasg: let me know once things are done, then I will send out some mails21:36
-!- yorkerlin [c0008056@gateway/web/freenode/ip.] has quit [Ping timeout: 246 seconds]21:37
@besser82HeikoS, Is the reg. form already working?21:38
@HeikoSiglesiasg: could you put the call for contributions stuff at the main workshop site?21:38
@HeikoSbesser82: not yet21:38
@HeikoSbesser82: when I send an email then it is21:38
@besser82HeikoS, allrighty ;)21:38
@HeikoSbesser82: planning to come?21:39
@besser82HeikoS, sure  :D21:39
@HeikoSnice! :D21:39
@besser82HeikoS, wanna all you folks in person ^^21:39
@besser82      ^^meet21:39
@besser82HeikoS, is there a schedule,yet?  like begins - ends?21:40
@HeikoSbesser82: coming21:40
@HeikoS10-18 if you wanna help21:40
@HeikoSin fact, then its 9-22 :)21:40
@HeikoSiglesiasg: gotta go now, will do the remainders tomorrow, thanks so much for your help21:40
iglesiasgbesser82: preliminar stuff - http://shogun-toolbox.org/page/Events/workshop2014_program21:40
iglesiasgHeikoS: no problem, we are in this together21:41
iglesiasgHeikoS: thanks to you for pushing us. You are doing the real hard job21:41
@besser82HeikoS, will need some hotel then, I guess  ;)21:41
@HeikoSiglesiasg: one more push: Let's aim to have this done by tomorrow! :)21:41
@HeikoSiglesiasg: cool!21:41
@HeikoSbesser82: yeah, I just booked mine a few days ago airbnb21:42
@besser82HeikoS, thx ^^  Will try my luck  ;)21:42
@HeikoSiglesiasg: could you maybe answer in the workshop thread once you are done to keep everyone posted. also on the not using eventbrite thing?21:43
iglesiasgHeikoS: let's wait a couple of hours for thoralf though since the eventbrite was his thing21:44
iglesiasgHeikoS: putting the link to the form in the website and making it live so people can start voting should take no time (like one minute)21:45
@HeikoSiglesiasg: yeah thats why I suggested the email. Maybe we can have it optionally for the hands-on21:45
iglesiasgHeikoS: but I will do it before tomorrow morning since I cannot do it once I get to job21:45
@HeikoSiglesiasg: yeah, but I want to send out everything together, so lets see what he says21:46
@HeikoSiglesiasg: ok fine21:46
@HeikoSiglesiasg: cu!21:46
-!- HeikoS [~heiko@nat-173-40.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.]21:46
iglesiasgbye bye21:46
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun22:11
shogun-notifier-shogun: Parijat Mazumdar :develop * 81eb291 / src/shogun/loss/ (7 files): https://github.com/shogun-toolbox/shogun/commit/81eb291ef93dc9aac3b3e0991fe7452f99c9248a22:11
shogun-notifier-shogun: loss functions added for Gradient Boosting22:11
shogun-notifier-shogun: Parijat Mazumdar :develop * a78268d / / (5 files): https://github.com/shogun-toolbox/shogun/commit/a78268de6e7707d59cb2ed164f524c6c4c48cbc722:11
shogun-notifier-shogun: Loss functions : minor changes and unittests22:11
shogun-notifier-shogun: Parijat Mazumdar :develop * bb6b834 / src/shogun/loss/HuberLoss.h: https://github.com/shogun-toolbox/shogun/commit/bb6b8348ab6b2f26c720359aca6e33cfe246a19722:11
shogun-notifier-shogun: minor changes in constructor22:11
shogun-notifier-shogun: Fernando Iglesias :develop * 764d192 / / (8 files): https://github.com/shogun-toolbox/shogun/commit/764d19280d2e0d338f40e0ba7b771a0bc6b5bab122:11
shogun-notifier-shogun: Merge pull request #2316 from mazumdarparijat/gradient22:11
shogun-notifier-shogun: loss functions added for Gradient Boosting22:11
lisitsynhah yes22:16
shogun-buildbot_build #2393 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/2393  blamelist: Parijat Mazumdar <mazumdarparijat@gmail.com>22:20
iglesiasgand midterm filled in!22:26
shogun-buildbot_build #2394 of bsd1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/2394  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>, Parijat Mazumdar <mazumdarparijat@gmail.com>22:29
iglesiasglisitsyn: so you maybe know about this C++ x-file22:31
lisitsynwhat x-file?22:31
iglesiasglisitsyn: https://github.com/shogun-toolbox/shogun/pull/231622:31
iglesiasglisitsyn: let me sum it up22:31
iglesiasgbasically there was a constructor with a default value22:31
iglesiasgCHuberLoss(float64_t m_delta=default_value)22:31
iglesiasgthat constructor calls default constructor on base class22:32
iglesiasgcalls init that initialises the attirbute and SG_ADD22:32
iglesiasgwhen SG_ADD is left, then serialization tests fail22:32
iglesiasgwhen it is removed, they work22:32
iglesiasgproblem is solved by adding the default constructor that calls init, in addition to default constructor of base class22:33
iglesiasgso it must be definitely be that the default constructor that is implicitly created was called22:33
iglesiasginstead of using CHuberLoss(float64_t) with its default value22:33
lisitsynsorry ok22:34
lisitsynthis is confusing22:34
lisitsyniglesiasg: that's quite interesting22:36
lisitsyniglesiasg: can we test whether default constructor is called?22:39
lisitsynI mean with breakpoint or debug print in the non-default constructor22:39
iglesiasglisitsyn: it is already fixed, but we could do it locally I guess22:40
iglesiasgbut it is confusing yes22:40
iglesiasgI am reading22:40
iglesiasgIn the case of the default constructor, the compiler will not generate a default constructor if a class is defined with any constructors.22:41
iglesiasgfrom: http://stackoverflow.com/questions/7469468/defaulted-default-constructor-in-n3290-draft22:41
lisitsynit shouldn't generate22:41
iglesiasgit seems though that C++11 uses other rules22:41
shogun-buildbot_build #366 of deb4 - python3 is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb4%20-%20python3/builds/36622:50
-!- PirosB3 [~pirosb3@host76-224-dynamic.52-79-r.retail.telecomitalia.it] has quit [Quit: PirosB3]22:51
-!- PirosB3 [~pirosb3@host76-224-dynamic.52-79-r.retail.telecomitalia.it] has joined #shogun22:55
shogun-notifier-shogun: Parijat Mazumdar :develop * e010948 / / (5 files): https://github.com/shogun-toolbox/shogun/commit/e01094895efa62156d1c1d34e91fd293f58fd7c023:06
shogun-notifier-shogun: stochastic gradient boosting added, feature added for automagically choosing problem type in CART23:06
shogun-notifier-shogun: Fernando Iglesias :develop * 3dfae58 / / (5 files): https://github.com/shogun-toolbox/shogun/commit/3dfae583894921254e716b5a95d368a4589b368a23:06
shogun-notifier-shogun: Merge pull request #2327 from mazumdarparijat/gd23:06
shogun-notifier-shogun: stochastic gradient boosting added, feature added for automagically choosing problem type in CART23:06
shogun-buildbot_build #144 of osx2 - python is complete: Failure [failed test python modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx2%20-%20python/builds/144  blamelist: Parijat Mazumdar <mazumdarparijat@gmail.com>23:12
shogun-notifier-shogun: jiaolong :develop * 8360ae7 / src/shogun/structure/GraphCut.cpp: https://github.com/shogun-toolbox/shogun/commit/8360ae7fdc7f593d687884d17edb4b76c83e8b8a23:30
shogun-notifier-shogun: fixed defects found by Coverity Scan23:30
shogun-notifier-shogun: Fernando Iglesias :develop * b4b96d1 / src/shogun/structure/GraphCut.cpp: https://github.com/shogun-toolbox/shogun/commit/b4b96d1741e9c8dc841cab5930c7dbb451567df823:30
shogun-notifier-shogun: Merge pull request #2328 from Jiaolong/graph_cut23:30
shogun-notifier-shogun: fixed defects found by Coverity Scan23:30
-!- travis-ci [~travis-ci@ec2-54-82-51-114.compute-1.amazonaws.com] has joined #shogun23:34
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/2825624223:34
-!- travis-ci [~travis-ci@ec2-54-82-51-114.compute-1.amazonaws.com] has left #shogun []23:34
shogun-buildbot_build #2383 of deb3 - modular_interfaces is complete: Failure [failed test python modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/2383  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>, Parijat Mazumdar <mazumdarparijat@gmail.com>23:36
shogun-buildbot_build #145 of osx2 - python is complete: Failure [failed test python modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx2%20-%20python/builds/145  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>, Parijat Mazumdar <mazumdarparijat@gmail.com>23:46
-!- iglesiasg [524ae0a7@gateway/web/freenode/ip.] has quit [Quit: Page closed]23:50
-!- hushell [~hushell@c-71-56-149-179.hsd1.wa.comcast.net] has joined #shogun23:53
shogun-buildbot_build #367 of deb4 - python3 is complete: Failure [failed test python modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb4%20-%20python3/builds/367  blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>, Parijat Mazumdar <mazumdarparijat@gmail.com>23:55
-!- pickle27 [~pickle27@] has quit [Quit: This computer has gone to sleep]23:58
--- Log closed Tue Jun 24 00:00:46 2014