IRC logs of #shogun for Monday, 2012-09-03

« Sunday, 2012-09-02 Index Tuesday, 2012-09-04 »
--- Log opened Mon Sep 03 00:00:17 2012
bern4rdsonney2k: parser.py can be removed is a beta version of parserHTML.py. A parser object is created in views.py as a general variable, when creating the object the path newsfolder is set to /news. then the method parseFolder is called each time the home page is loaded00:42
bern4rdin the method parseFolder i check if there is any file in /news that has been modified after the last date i uploaded the news00:43
-!- bern4rd [533886e2@gateway/web/freenode/ip.83.56.134.226] has quit [Ping timeout: 245 seconds]01:12
-!- bern4rd [533886e2@gateway/web/freenode/ip.83.56.134.226] has joined #shogun01:23
-!- in3xes_ [~in3xes@122.174.83.206] has quit [Ping timeout: 246 seconds]02:11
-!- bern4rd [533886e2@gateway/web/freenode/ip.83.56.134.226] has quit [Quit: Page closed]02:33
-!- av3ngr [av3ngr@nat/redhat/x-qlfpklgncbazlikg] has joined #shogun03:38
-!- av3ngr [av3ngr@nat/redhat/x-qlfpklgncbazlikg] has quit [Client Quit]03:38
-!- uricamic [~uricamic@2001:718:2:1634:d0c1:4e83:6062:c909] has joined #shogun08:42
-!- sonne|work [~sonnenbu@194.78.35.195] has joined #shogun09:14
-!- alexlovesdata [~binder@eduroam7.first.fraunhofer.de] has joined #shogun09:53
-!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has joined #shogun10:45
blackburnoh mighty sonne|work is back10:45
blackburnuricamic: hey10:46
blackburnalexlovesdata: hey10:46
sonne|workblackburn: hey yeah!10:46
sonne|workblackburn: look at shogun.7nn.de ...10:46
sonne|worksome updates10:46
blackburnthings go better10:47
blackburnsonne|work: did you make any progress on the foundation thing?10:48
sonne|workno10:50
blackburnI see10:50
blackburnI've got new commercial task to participate in10:51
sonne|work?10:51
uricamicblackburn: hi10:51
blackburnsome guys need a stereo based pedestrian *counter*10:52
blackburnso CRFs and other stuff are waiting for me10:59
blackburnsonne|work: scikits going to release tomorrow10:59
sonne|workblackburn: so we release today :D10:59
sonne|workblackburn: nice application11:00
blackburnsonne|work: yes and happily paid again :D11:00
sonne|workblackburn: btw the road sign detector would also be some nice application for the apps folder11:01
blackburnsonne|work: yes I planned to integrate that11:02
blackburnbut not now, no more time left :)11:02
sonne|workblackburn: but we need the buildbot to be happy11:03
blackburnsonne|work: what's wrong with bb?11:04
blackburnall green11:04
blackburnbattlesystem operational11:04
blackburnehm there are a few issues with integration aka regression tests11:05
blackburnuricamic: could you please check whether libshogun BMRM example still works in a correct way?11:06
uricamicblackburn: sure11:06
blackburnuricamic: I ran it yesterday a few times and it was so random and its convergence path was way too curvy :)11:06
uricamicblackburn: hmm, on what data?11:07
blackburnuricamic: default11:07
sonne|workuricamic: whatever you gave it in the example11:08
uricamici.e. randomly generated11:08
sonne|workbtw it seemed pretty random to me11:08
sonne|workuricamic: maybe better seed the random number generator11:08
sonne|workuricamic: CMath::init_random(17);11:08
uricamicsonne|work: ok, it was there by purpose to get more testing data11:09
sonne|workuricamic: yeah but not good if you want to see if this works reliable / gives same result11:09
uricamicsonne|work: ok, I will fix it11:09
-!- av3ngr [~av3ngr@60-241-222-244.static.tpgi.com.au] has joined #shogun11:11
-!- av3ngr [~av3ngr@60-241-222-244.static.tpgi.com.au] has left #shogun []11:13
sonne|workblackburn: btw did you fix regression_lars.py?11:32
blackburnsonne|work: not yet11:32
sonne|workblackburn: but apps are good now right?11:33
sonne|workor what is missing?11:33
blackburnnot all11:33
blackburn:D11:33
blackburnoh well I have to test splicer etc11:33
sonne|workso what do you still want to do before release?11:33
blackburnsonne|work: msplicer fails for sure11:33
sonne|workyeah well that cannot work sure11:33
blackburnsonne|work:okay we need to11:33
blackburnget applications working again11:34
blackburnfix all graphical examples11:34
blackburnand get tester of static interfaces working again too11:34
sonne|workIMHO that is doable for today11:34
blackburnsure11:35
sonne|workI think I can do some more website polishing and then we should be able to release11:35
blackburnbut I am not available for serious bugfixing next 4-5 hours11:35
sonne|workI guess then we get bug reports and will have to do some 2.0.1 at some point11:35
blackburnsure we will have to fix and fix11:35
sonne|workbut more important we need some nice blog entry describing all the new features11:36
blackburnthat's eternal11:36
sonne|workand progress updates on our twitter :D11:36
sonne|workblackburn: btw click the twitter icon on the homepage :)11:36
-!- hoijui [~hoijui@dslb-092-078-182-230.pools.arcor-ip.net] has joined #shogun11:36
blackburnyeah I am already following our twitter11:36
sonne|workheh11:37
CIA-27shogun: Michal Uricar master * rf874098 / examples/undocumented/libshogun/so_multiclass_BMRM.cpp : so_multiclass_BMRM libshogun example random number generator seed - http://git.io/Sn2KaA11:38
CIA-27shogun: Soeren Sonnenburg master * rdc3534b / examples/undocumented/libshogun/so_multiclass_BMRM.cpp : Merge pull request #768 from uricamic/BM_SOL_EXAMPLE - http://git.io/Cn0Fyw11:38
blackburnhttp://www.codinghorror.com/blog/2012/07/new-programming-jargon.html13:17
-!- n4nd0 [53b32c87@gateway/web/freenode/ip.83.179.44.135] has joined #shogun13:19
n4nd0hi blackburn, sonne|work13:20
sonne|workblackburn: *lol*13:21
sonne|workn4nd0: hey13:21
n4nd0did you get some time to take a look to the SGVector's destructor call from the constructor fired by QDA example?13:21
sonne|workn4nd0: so we release today right?13:21
sonne|workn4nd0: no - where what?13:21
n4nd0sonne|work: do you remember, graphical qda crashes?13:22
sonne|workyes13:22
n4nd0ok13:22
n4nd0let me show you13:23
n4nd0https://github.com/shogun-toolbox/shogun/blob/master/src/shogun/multiclass/QDA.h13:23
n4nd0the trace follows something like13:23
n4nd0get_mean -> SGVector constructor -> ~SGVector13:24
n4nd0no idea why the destructor is called13:24
blackburnunbeliebug13:25
sonne|workn4nd0: I understand it13:25
sonne|workit is just a ptr13:26
sonne|workso in get_mean it needs an additional false13:26
sonne|worka ptr into an sgmatrix13:26
sonne|workI mean13:26
n4nd0oh fuck13:27
n4nd0to mark that the vector passed in the constructor shouldn't be deleted later right?13:27
blackburnyes13:27
blackburnI recognize now too13:28
n4nd0thank you guys!13:28
n4nd0I am going to apply the change and test13:28
blackburnwe've got a lot of smurf conventional naming13:28
blackburn:D13:28
-!- n4nd0 [53b32c87@gateway/web/freenode/ip.83.179.44.135] has quit [Quit: Page closed]13:31
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun13:31
-!- alexlovesdata [~binder@eduroam7.first.fraunhofer.de] has quit [Ping timeout: 268 seconds]14:25
sonne|workblackburn: and ninja comments!14:36
blackburnsonne|work: a lot of ninja comments14:36
sonne|workblackburn: hmmhh maybe we should use http://www.mathjax.org/demos/tex-samples/ in our doxygen14:40
blackburnsonne|work: mathjax is cool, I've been using it on one stie14:50
sonne|workblackburn: doxygen has a flag to enable using it instead of manually generating latex images14:52
CIA-27shogun: Soeren Sonnenburg master * rcb4b2b5 / src/shogun/multiclass/QDA.h : disable reference counts for bare ptrs - http://git.io/09AkmA14:57
CIA-27shogun: Soeren Sonnenburg master * r51aecfe / doc/Doxyfile : assume eigen3 / nlopt are available when generating doxygen doc - http://git.io/G3SuBQ14:57
CIA-27shogun: Soeren Sonnenburg master * rcd23d43 / src/Makefile.template : fix typo -> templatE - http://git.io/cvu3Ig14:57
sonne|workblackburn: nlopt is finally in debian \o/15:00
blackburnsonne|work: really?15:01
sonne|workhttp://ftp.debian.org/debian/pool/main/n/nlopt/nlopt_2.2.4+dfsg-2.dsc15:01
blackburnoh it has recent release15:01
blackburn20 july15:01
n4nd0sonne|work: the QDA problem is not fixed yet though15:01
n4nd0after those changes it still crashes here15:01
n4nd0sonne|work: did you test it?15:02
blackburnnot that dead I thought15:02
sonne|workn4nd0: where does it crash?15:02
n4nd0sonne|work: classifier_qda.cpp15:02
n4nd0in the last line15:02
n4nd0retur 0;15:02
n4nd0something being deleted twice I think15:02
n4nd0trying to find out15:02
-!- hoijui [~hoijui@dslb-092-078-182-230.pools.arcor-ip.net] has quit [Ping timeout: 255 seconds]15:13
n4nd0sonne|work: ok, now it is working fine again15:35
n4nd0there was nothing else to change15:35
n4nd0I don't know why but sometimes I keep on getting a Segmentation fault (core dumped) that gets fixed after make clean and build again15:36
blackburnn4nd0: is it in python or C++ examples?15:38
n4nd0C++15:39
n4nd0in python I got glibc instead15:39
blackburnwell each time you change headers you would need to recompile examples as well15:42
blackburnother way it could be dangerous15:42
blackburnbut it doesn't answer why python fails - it shouldn't fail15:42
sonne|workblackburn: now only your example fails :P16:24
sonne|workblackburn: ahh no not true - still regression_gaussian_process_demo.py fails16:25
blackburnsonne|work: mine? which?16:36
sonne|workblackburn: regression_lars.py16:38
sonne|workjust some double free situation :)16:38
sonne|workvector<float64_t> GA1(corr_sign_a);16:38
sonne|workthis line tells it all16:39
blackburnsonne|work: it is example of chiyuan btw :D16:41
sonne|workahh too bad :)16:42
sonne|workI prefer blaming you16:42
blackburnI don't mind to fix it when I get home still16:47
sonne|workblackburn: I had a brief look - I think the fix would be in cholesky_insert - one should rather return the new R (nR) in this function and all good16:50
@sonney2kblackburn, let met try ...17:10
blackburnhah you get home faster17:18
blackburnI'm leaving17:18
-!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has quit [Quit: Page closed]17:18
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 248 seconds]17:28
-!- uricamic [~uricamic@2001:718:2:1634:d0c1:4e83:6062:c909] has quit [Quit: Leaving.]17:30
-!- CIA-27 [cia@cia.vc] has quit [Ping timeout: 272 seconds]17:35
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun17:37
-!- in3xes [~in3xes@122.174.152.235] has joined #shogun17:43
-!- CIA-27 [cia@cia.vc] has joined #shogun17:50
-!- blackburn [~blackburn@188.168.5.75] has joined #shogun18:02
-!- yoo [2eda6d52@gateway/web/freenode/ip.46.218.109.82] has joined #shogun18:13
-!- yoo [2eda6d52@gateway/web/freenode/ip.46.218.109.82] has quit [Client Quit]18:14
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 240 seconds]19:02
CIA-27shogun: Soeren Sonnenburg master * raedcfae / (2 files): let cholesky_{insert/delete} return newly created R matrix - http://git.io/fFUF1g19:03
blackburnsonney2k: okay so did that help?19:07
blackburnI was about to start.. :D19:07
blackburnI'm running examples now19:07
@sonney2kblackburn, try to fix puffins example...19:10
blackburnyes19:10
blackburnalready trying to19:10
blackburnsonney2k: I've got enrolled to coursera statistics one to get less dumb :D19:11
blackburnsonney2k: gp stuff should be related to forward declaration19:16
blackburnI hope so19:16
blackburnsonney2k: fixed19:31
CIA-27shogun: Sergey Lisitsyn master * r1a78d61 / (2 files in 2 dirs): Forward declarations for GPs and modular interfaces - http://git.io/C1RR1w19:31
CIA-27shogun: Sergey Lisitsyn master * r71f80cb / (3 files in 2 dirs): Dropped sailfish application - http://git.io/nr2Wdw19:32
-!- in3xes [~in3xes@122.174.152.235] has quit [Ping timeout: 264 seconds]20:07
CIA-27shogun: Sergey Lisitsyn master * r9bbe807 / applications/ocr/data/ocr.svm.gz : Updated ocr data - http://git.io/CA2u6w20:09
-!- blackburn1 [~blackburn@188.168.3.7] has joined #shogun20:14
-!- blackburn [~blackburn@188.168.5.75] has quit [Ping timeout: 276 seconds]20:14
CIA-27shogun: Sergey Lisitsyn master * r2d57c62 / (2 files in 2 dirs): Updated easysvm - http://git.io/FLhl5g20:16
CIA-27shogun: Sergey Lisitsyn master * r8eac595 / data : Updated data - http://git.io/v_86Hw20:16
CIA-27shogun: Sergey Lisitsyn master * rfd33d77 / applications/asp/signal_detectors.py : A few fixes for asp - http://git.io/fz6hZA20:21
-!- blackburn [~blackburn@83.234.54.39] has joined #shogun20:28
-!- blackburn1 [~blackburn@188.168.3.7] has quit [Ping timeout: 240 seconds]20:28
blackburnsonney2k: where are you ? :)20:30
CIA-27shogun: Sergey Lisitsyn master * re7c3462 / applications/msplicer/signal_detectors.py : Fixes for msplicer - http://git.io/-oksqA20:30
blackburnsonney2k: I have no idea what to do with msplicer - it requires missing dynprog methods20:44
blackburnothers should be ok20:44
blackburnsonney2k: that's the only thing I wanted to do20:44
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun20:46
n4nd0so release today?20:47
blackburnn4nd0: msplicer is left20:49
blackburnn4nd0: it actually depends whether sonney2k would appear today :)20:49
n4nd0blackburn: what is it with msplicer? leaks, crash?20:50
blackburnn4nd0: it requires some dynprog methods that are missing already20:50
n4nd0blackburn: were they removed?20:51
blackburnno idea20:51
blackburn:d20:51
-!- blackburn [~blackburn@83.234.54.39] has quit [Ping timeout: 240 seconds]20:56
-!- blackburn [d55789c3@gateway/web/freenode/ip.213.87.137.195] has joined #shogun21:09
n4nd0going to sleep now21:40
n4nd0I hope I don't miss the release!21:40
n4nd0good night21:40
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving]21:42
-!- blackburn [d55789c3@gateway/web/freenode/ip.213.87.137.195] has quit [Quit: Page closed]22:01
-!- blackburn [~blackburn@188.168.13.98] has joined #shogun22:09
-!- blackburn [~blackburn@188.168.13.98] has quit [Ping timeout: 248 seconds]22:20
--- Log closed Tue Sep 04 00:00:17 2012
« Sunday, 2012-09-02 Index Tuesday, 2012-09-04 »

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