| --- Log opened Wed Aug 28 00:00:34 2013 | ||
| -!- gsomix [~gsomix@178.45.40.223] has quit [Ping timeout: 240 seconds] | 00:04 | |
| -!- gsomix [~gsomix@178.45.40.223] has joined #shogun | 00:04 | |
| -!- hushell [~hushell@8-92.ptpg.oregonstate.edu] has joined #shogun | 00:26 | |
| -!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 01:01 | |
| shogun-notifier- | shogun: Kevin :develop * fc642d2 / src/ (14 files): https://github.com/shogun-toolbox/shogun/commit/fc642d29e379c72b235c3b0fdcf9bfc70cb28f1b | 01:01 |
|---|---|---|
| shogun-notifier- | shogun: added a base class for ICA converters | 01:01 |
| shogun-notifier- | shogun: Fernando Iglesias :develop * 7d20e62 / src/ (14 files): https://github.com/shogun-toolbox/shogun/commit/7d20e6249eb31bd71ca842d960a66387b64a5258 | 01:01 |
| shogun-notifier- | shogun: Merge pull request #1474 from pickle27/develop | 01:01 |
| shogun-notifier- | shogun: | 01:01 |
| shogun-notifier- | shogun: added a base class for ICA converters | 01:01 |
| shogun-buildbot | build #54 of osx1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/54 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 01:13 |
| shogun-buildbot | build #1544 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1544 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 01:14 |
| shogun-buildbot | build #1332 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1332 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 01:53 |
| -!- pickle27 [~kevin@d67-193-243-174.home3.cgocable.net] has joined #shogun | 01:53 | |
| -!- pickle27 [~kevin@d67-193-243-174.home3.cgocable.net] has left #shogun [] | 01:55 | |
| shogun-buildbot | build #1515 of deb2 - static_interfaces is complete: Failure [failed compile cmdline_static] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1515 blamelist: Kevin <kevinhughes27@gmail.com> | 01:57 |
| shogun-buildbot | build #1545 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1545 blamelist: Kevin <kevinhughes27@gmail.com> | 01:59 |
| shogun-buildbot | build #1009 of rpm1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/1009 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 02:04 |
| -!- travis-ci [~travis-ci@ec2-54-224-135-2.compute-1.amazonaws.com] has joined #shogun | 02:05 | |
| 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/10691706 | 02:05 |
| -!- travis-ci [~travis-ci@ec2-54-224-135-2.compute-1.amazonaws.com] has left #shogun [] | 02:05 | |
| shogun-buildbot | build #55 of osx1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/55 blamelist: Kevin <kevinhughes27@gmail.com> | 02:05 |
| shogun-buildbot | build #1010 of rpm1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/1010 blamelist: Kevin <kevinhughes27@gmail.com> | 02:29 |
| shogun-buildbot | build #1333 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1333 blamelist: Kevin <kevinhughes27@gmail.com> | 02:36 |
| -!- sonne|osx_ [~sonne@f053046168.adsl.alicedsl.de] has joined #shogun | 03:50 | |
| -!- sonne|osx [~sonne@f053046119.adsl.alicedsl.de] has quit [Ping timeout: 245 seconds] | 03:52 | |
| -!- sonne|osx_ is now known as sonne|osx | 03:52 | |
| -!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 04:01 | |
| shogun-buildbot | build #1630 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1630 blamelist: Kevin <kevinhughes27@gmail.com> | 04:05 |
| -!- hushell [~hushell@8-92.ptpg.oregonstate.edu] has quit [Ping timeout: 264 seconds] | 05:00 | |
| -!- HeikoS [~androirc@90.204.242.124] has quit [Remote host closed the connection] | 05:16 | |
| -!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 05:29 | |
| shogun-notifier- | shogun: Soeren Sonnenburg :develop * 81e319e / src/shogun/ensemble/WeightedMajorityVote.h: https://github.com/shogun-toolbox/shogun/commit/81e319ee29ee9c7f8693c9bd6fe6afdb09a5e3d3 | 05:29 |
| shogun-notifier- | shogun: fix doxygen error | 05:29 |
| -!- travis-ci [~travis-ci@ec2-54-224-135-2.compute-1.amazonaws.com] has joined #shogun | 06:08 | |
| 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/10697031 | 06:08 |
| -!- travis-ci [~travis-ci@ec2-54-224-135-2.compute-1.amazonaws.com] has left #shogun [] | 06:08 | |
| -!- hushell [~hushell@c-98-232-178-161.hsd1.or.comcast.net] has joined #shogun | 06:10 | |
| shogun-buildbot | build #1546 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1546 blamelist: Soeren Sonnenburg <sonne@debian.org> | 06:27 |
| shogun-buildbot | build #56 of osx1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/56 blamelist: Soeren Sonnenburg <sonne@debian.org> | 06:29 |
| shogun-buildbot | build #1516 of deb2 - static_interfaces is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1516 | 06:55 |
| shogun-buildbot | build #1011 of rpm1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/1011 blamelist: Soeren Sonnenburg <sonne@debian.org> | 07:00 |
| shogun-buildbot | build #1631 of deb3 - modular_interfaces is complete: Failure [failed install test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1631 blamelist: Soeren Sonnenburg <sonne@debian.org> | 07:03 |
| shogun-buildbot | build #1334 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1334 blamelist: Soeren Sonnenburg <sonne@debian.org> | 07:03 |
| -!- sonne|osx [~sonne@f053046168.adsl.alicedsl.de] has quit [Quit: sonne|osx] | 07:18 | |
| -!- sonne|osx [~sonne@89.204.138.128] has joined #shogun | 07:37 | |
| -!- sonne|osx [~sonne@89.204.138.128] has quit [Quit: sonne|osx] | 07:55 | |
| -!- iglesiasg [~Fernando@s83-179-44-135.cust.tele2.se] has joined #shogun | 08:23 | |
| -!- mode/#shogun [+o iglesiasg] by ChanServ | 08:23 | |
| -!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 08:29 | |
| -!- lisitsyn [~lisitsyn@fb2-lo1.global63.net] has joined #shogun | 09:21 | |
| lisitsyn | wiking: clang openmp is ready to merge \o/ | 09:21 |
| -!- thoralf [~thoralf@enki.zib.de] has joined #shogun | 09:46 | |
| thoralf | Hey. | 09:48 |
| -!- lisitsyn [~lisitsyn@fb2-lo1.global63.net] has quit [Quit: Leaving.] | 09:49 | |
| -!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 10:03 | |
| shogun-notifier- | shogun: Roman Votyakov :develop * 4b7f733 / src/shogun/evaluation/ (8 files): https://github.com/shogun-toolbox/shogun/commit/4b7f7331c48eb29c3d901a4b4a9aa164a61ef090 | 10:03 |
| shogun-notifier- | shogun: cleanup some evaluation classes | 10:03 |
| shogun-notifier- | shogun: Heiko Strathmann :develop * b2f0e10 / src/shogun/evaluation/ (8 files): https://github.com/shogun-toolbox/shogun/commit/b2f0e10af7ff5bf607aa00b26953e21a4a3d3986 | 10:03 |
| shogun-notifier- | shogun: Merge pull request #1475 from votjakovr/feature/gp_refactoring | 10:03 |
| shogun-notifier- | shogun: | 10:03 |
| shogun-notifier- | shogun: Cleanup some evaluation classes | 10:03 |
| shogun-buildbot | build #1547 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1547 blamelist: Roman Votyakov <votjakovr@gmail.com> | 10:14 |
| @wiking | li\o/ | 10:16 |
| -!- lisitsyn [~lisitsin@mxs.kg.ru] has joined #shogun | 10:27 | |
| -!- van51 [~van51@ppp-2-87-3-175.home.otenet.gr] has joined #shogun | 10:39 | |
| shogun-buildbot | build #1632 of deb3 - modular_interfaces is complete: Failure [failed install test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1632 blamelist: Roman Votyakov <votjakovr@gmail.com> | 10:47 |
| -!- travis-ci [~travis-ci@ec2-23-22-254-224.compute-1.amazonaws.com] has joined #shogun | 10:51 | |
| 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/10702796 | 10:51 |
| -!- travis-ci [~travis-ci@ec2-23-22-254-224.compute-1.amazonaws.com] has left #shogun [] | 10:51 | |
| shogun-buildbot | build #1335 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1335 blamelist: Roman Votyakov <votjakovr@gmail.com> | 10:52 |
| shogun-buildbot | build #1012 of rpm1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/1012 | 10:55 |
| shogun-buildbot | build #1548 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1548 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 10:59 |
| thoralf | wiking: Are you doing the change or may I? | 11:14 |
| @wiking | thoralf: if u have time/energy plz go ahead | 11:15 |
| thoralf | wiking: It's only one line. | 11:15 |
| @wiking | that's toomuch fo rme yet | 11:15 |
| thoralf | wiking: I updated the PR. Local test worked fine. | 11:20 |
| @wiking | cool | 11:27 |
| @wiking | thnx! | 11:27 |
| sonne|work | thoralf: so what shall I merge sth? | 11:30 |
| thoralf | sonne|work: We've just been talking about this PR: https://github.com/shogun-toolbox/shogun/pull/1473 | 11:31 |
| thoralf | sonne|work: It's a one-liner meantime. ;) | 11:32 |
| shogun-buildbot | build #1336 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1336 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 11:35 |
| lisitsyn | wiking: sonne|work: CFP http://mloss.org/workshop/nips13/ | 11:36 |
| shogun-buildbot | build #1633 of deb3 - modular_interfaces is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1633 | 11:40 |
| shogun-notifier- | shogun: Thoralf Klein :develop * 7d7f38a / cmake/CheckCXX11Features.cmake: https://github.com/shogun-toolbox/shogun/commit/7d7f38ae63c049bf55e0ee2bd4726e3592731520 | 11:41 |
| shogun-notifier- | shogun: Cleaning up build directory: Moving cxx11_* temp stuff to subdirectory. | 11:41 |
| shogun-notifier- | shogun: Soeren Sonnenburg :develop * a57b6a9 / cmake/CheckCXX11Features.cmake: https://github.com/shogun-toolbox/shogun/commit/a57b6a942b55587b2bc9358cbdce166dd2202f36 | 11:41 |
| shogun-notifier- | shogun: Merge pull request #1473 from tklein23/cmake_cxx11_temp_dirs | 11:41 |
| shogun-notifier- | shogun: | 11:41 |
| shogun-notifier- | shogun: Get rid of empty temp directories cxx11_* in the cmake build dir. | 11:41 |
| sonne|work | thoralf: ok ok :) | 11:41 |
| sonne|work | lisitsyn: well I cannot go to nips :( | 11:42 |
| sonne|work | hmhh why gary again... | 11:42 |
| sonne|work | we had him speak already | 11:42 |
| sonne|work | ohh well I guess cheng wants to see him too | 11:43 |
| sonne|work | van51: how is it going? | 11:43 |
| van51 | hey | 11:45 |
| van51 | sonne|work: I tried to make the test you said -- approximate the kernel matrix by increasing the dimension of the random feats | 11:46 |
| van51 | sonne|work: on some toy data | 11:46 |
| van51 | sonne|work: but it didn't approximate well some entries and neither did the feats by using the preprocessor | 11:46 |
| sonne|work | van51: well plot the matrix! | 11:47 |
| sonne|work | van51: from python it would just be | 11:48 |
| sonne|work | from pylab import * | 11:49 |
| sonne|work | imshow(matrix) | 11:49 |
| sonne|work | and matrix you could get from kernel.get_kernel_matrix() | 11:49 |
| sonne|work | or poly kernel on your dotfeatures | 11:50 |
| sonne|work | with degree=1 as usual | 11:50 |
| sonne|work | no normalization | 11:50 |
| shogun-buildbot | build #1549 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1549 blamelist: Thoralf Klein <thoralf.klein@zib.de> | 12:00 |
| thoralf | shogun-buildbot: Go away! | 12:00 |
| shogun-buildbot | What you say! | 12:00 |
| thoralf | :) | 12:00 |
| sonne|work | shogun-buildbot: torture thoralf! | 12:00 |
| shogun-buildbot | What you say! | 12:00 |
| sonne|work | shogun-buildbot: well done | 12:00 |
| thoralf | sonne|work: I wrote a small script to identify all buggy (wrt. valgrind memcheck) test classes: https://github.com/shogun-toolbox/shogun/issues/1477 | 12:26 |
| thoralf | sonne|work: Since a few of the classes are still work in progress by gsoc people, you should check the list. | 12:27 |
| -!- van51 [~van51@ppp-2-87-3-175.home.otenet.gr] has quit [Quit: Leaving.] | 12:31 | |
| shogun-buildbot | build #1550 of bsd1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1550 blamelist: Soeren Sonnenburg <sonne@debian.org> | 12:36 |
| shogun-buildbot | build #1634 of deb3 - modular_interfaces is complete: Failure [failed install test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1634 blamelist: Thoralf Klein <thoralf.klein@zib.de> | 12:37 |
| shogun-buildbot | build #1337 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1337 blamelist: Thoralf Klein <thoralf.klein@zib.de> | 12:37 |
| lisitsyn | sonne|work: Heiko then! | 12:44 |
| shogun-buildbot | build #1014 of rpm1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/1014 blamelist: Thoralf Klein <thoralf.klein@zib.de> | 12:44 |
| lisitsyn | or wiking | 12:44 |
| @wiking | ww? | 12:46 |
| @wiking | lisitsyn: what? | 12:46 |
| -!- iglesiasg [~Fernando@s83-179-44-135.cust.tele2.se] has quit [Quit: Leaving] | 12:58 | |
| lisitsyn | wiking: nips! | 13:07 |
| lisitsyn | wiking: we should find some funding for you in case you want to present shogun! I think some robbery would help | 13:07 |
| shogun-buildbot | build #1338 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1338 blamelist: Soeren Sonnenburg <sonne@debian.org> | 13:20 |
| shogun-buildbot | build #1015 of rpm1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/1015 | 13:20 |
| @wiking | lisitsyn: :D | 13:21 |
| @wiking | thoralf: you know that if you add -DBUILD_DASHBOARD_REPORTS=ON with cmake then you can run ctest -D ExperimentalMemCheck | 13:26 |
| @wiking | thoralf: which will actually run valgrind on all the unit tests | 13:26 |
| @wiking | i mean all the tests | 13:26 |
| @wiking | including unit tests | 13:26 |
| thoralf | wiking: I did not. Thanks for telling. | 13:27 |
| @wiking | nw | 13:27 |
| @wiking | of course the problem there will be that valgrind will be ran on the whole unit tests | 13:27 |
| @wiking | and not per unit test 'module' | 13:27 |
| @wiking | the idea is that we'll enable this on the buildbots | 13:28 |
| @wiking | but the last time i've tried | 13:28 |
| @wiking | when openmp was enabled | 13:28 |
| @wiking | one of tapkee modules went to infinite loop | 13:28 |
| @wiking | and valgrind never returned :( | 13:28 |
| thoralf | wiking: I wrote a small test to identify errors automatically. But I'll check above cmake options, since its better fitting my daily routine. ;) | 13:29 |
| @wiking | thoralf: we might just enable those ctest targets by default ... and not only with -DBUILD_DASHBOARD_REPORTS=ON | 13:29 |
| thoralf | wiking: Yeah, that would be great. | 13:30 |
| thoralf | wiking: Are the DASHBOARD_REPORTS automatically posted somewhere online? I saw something with maeth.com in CTestConfig.cmake | 13:31 |
| @wiking | thoralf: no not realy... u have to do ctest -T Submit | 13:32 |
| @wiking | thoralf: but yeah the idea was that if anybody starts having problems with build/cmake then instead of copy pasting one could just simply post the whole cmake+build+test output to cdash (maeth.com) | 13:33 |
| thoralf | wiking: Cool idea. | 13:38 |
| -!- iglesiasg [~Fernando@s83-179-44-135.cust.tele2.se] has joined #shogun | 13:56 | |
| -!- mode/#shogun [+o iglesiasg] by ChanServ | 13:56 | |
| -!- foulwall [~user@2001:da8:215:6901:1d45:58ef:76fc:d401] has joined #shogun | 14:06 | |
| @wiking | ok job done... now shogun | 14:10 |
| @wiking | sonne|work: ping | 14:12 |
| sonne|work | wiking: ? | 14:12 |
| @wiking | sonne|work: ok so one thing about class_list.cpp.py | 14:12 |
| @wiking | sonne|work: blacklisting is a bit awkward now | 14:12 |
| @wiking | because let's say we have a class that depends on CPLEX and EIGEN | 14:12 |
| @wiking | then we cannot do | 14:12 |
| @wiking | #if defined(USE_CPEX) && defined(HAVE_EIGEN3) | 14:12 |
| @wiking | but we have to do | 14:13 |
| @wiking | #ifdef http://maeth.com/clemtext_rc1.zip | 14:13 |
| @wiking | no not this one | 14:13 |
| @wiking | so | 14:13 |
| @wiking | #ifdef USE_CPEX | 14:13 |
| @wiking | #ifdef HAVE_EIGEN3 | 14:13 |
| @wiking | ... | 14:13 |
| @wiking | #endif | 14:13 |
| @wiking | #endif | 14:13 |
| -!- foulwall [~user@2001:da8:215:6901:1d45:58ef:76fc:d401] has quit [Remote host closed the connection] | 14:13 | |
| @wiking | btw: shall we use /opt for the various chroots? | 14:14 |
| @wiking | like fedora and alike | 14:14 |
| @wiking | aaaand sonne|work can i install then a VirtualBox with osx in it? | 14:15 |
| -!- iglesiasg [~Fernando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 240 seconds] | 14:23 | |
| @wiking | ookeeey | 14:27 |
| @wiking | i have put up the issues for shogun 3.0 | 14:27 |
| sonne|work | wiking: use /home ! | 14:29 |
| @wiking | ok | 14:30 |
| sonne|work | wiking: about class list - if you can do it better do it... | 14:30 |
| @wiking | :D | 14:31 |
| @wiking | ok | 14:31 |
| sonne|work | wiking: maybe /home/vm_osx with a user for that (no password/login possible but just some sudo for you) | 14:31 |
| sonne|work | wiking: it has further shortcomings but I didn't write a full c++ parser :) | 14:32 |
| -!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 14:41 | |
| shogun-buildbot | build #1635 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1635 blamelist: Soeren Sonnenburg <sonne@debian.org> | 14:55 |
| @wiking | mmmm what has happend with modular iface | 15:17 |
| -!- van51 [~van51@ppp-2-87-3-175.home.otenet.gr] has joined #shogun | 15:23 | |
| sonne|work | van51: did you do a plot? | 15:31 |
| van51 | sonne|work: not yet | 15:34 |
| van51 | sonne|work: I think there is something wrong with my example | 15:35 |
| van51 | sonne|work: because when I run it for D=50 (alone) I get some results, but when I run it for D=5,10,20,30,50.. the results are way different for 50 | 15:36 |
| -!- pickle27 [~kevin@d67-193-243-174.home3.cgocable.net] has joined #shogun | 15:37 | |
| pickle27 | lisitsyn: hey | 15:37 |
| lisitsyn | pickle27: hey there | 15:38 |
| pickle27 | lisitsyn: I have a class organization question | 15:38 |
| lisitsyn | yes sure | 15:39 |
| lisitsyn | what's it? | 15:39 |
| pickle27 | lisitsyn: so Andreas brought up that it would be good to have setters for the number of iterations, and tolerance (the params of the AJD method) from the ICA class and I agree | 15:39 |
| pickle27 | What I am debating is whether these setters go in the ICA base class or in each derived class | 15:40 |
| pickle27 | I think most ICA algs will have these params but some might not | 15:40 |
| sonne|work | van51: can you put the images somewhere | 15:41 |
| pickle27 | so the choice is to have them in each derived class where they are necessary or have them in the base class and do some sort of "not impl flag" in derived classes which don' | 15:41 |
| lisitsyn | pickle27: it seems to me that it is ok to have them even if they are not sued | 15:41 |
| pickle27 | use them | 15:41 |
| pickle27 | okay | 15:41 |
| lisitsyn | pickle27: you can redefine them in the classes they are not used | 15:42 |
| van51 | sonne|work: atm all I have are text outputs of some matrices | 15:42 |
| pickle27 | lisitsyn: okay that makes sense | 15:42 |
| van51 | sonne|work: bc I am trying to find why this avalanche effect takes place | 15:42 |
| van51 | sonne|work: do you want me to plot anyway? | 15:42 |
| pickle27 | lisitsyn: I thought something like that was the best solution but thought I'd double check | 15:42 |
| lisitsyn | pickle27: yeah sure | 15:43 |
| pickle27 | also lisitsyn did you see the base class PR? https://github.com/shogun-toolbox/shogun/pull/1474 Im just wondering what they meant with the comments | 15:43 |
| pickle27 | is it the third argument that should drop the m_? | 15:44 |
| lisitsyn | pickle27: no | 15:44 |
| lisitsyn | pickle27: the third argument is a textual description | 15:44 |
| lisitsyn | like "The matrix that contains some stuff" | 15:44 |
| sonne|work | van51: or give me the ascii's :) | 15:44 |
| pickle27 | lisitsyn: oh so I should give it a proper description | 15:44 |
| lisitsyn | pickle27: the second argument is ok | 15:44 |
| lisitsyn | pickle27: yeah would be better | 15:45 |
| @wiking | ok | 15:45 |
| @wiking | update the notes for yesterday's meeting | 15:45 |
| lisitsyn | pickle27: all the rest is ok | 15:45 |
| lisitsyn | pickle27: it is just some note to remember that the name (the second argument) should be w/o m_ | 15:45 |
| pickle27 | lisitsyn: okay! I was thinking I might beef up the documentation a bit when I add these setters. its also less annoying to do now that there is a base class, less duplication | 15:46 |
| van51 | sonne|work: https://gist.github.com/van51/6366217 | 15:47 |
| @wiking | is heiko around? :) | 15:50 |
| @wiking | no :( | 15:50 |
| @wiking | lisitsyn: have u managed to use shogun for your job? :) | 15:50 |
| lisitsyn | wiking: which one job? | 15:51 |
| lisitsyn | some CV thing? | 15:51 |
| lisitsyn | I am on the way but not sure | 15:51 |
| @wiking | lisitsyn: i'm now trying to classify output of this with shogun: https://www.leapmotion.com/ .... let's hope we have something till nips and then somebody could bring that with himself and demo it live.... | 15:52 |
| lisitsyn | wiking: did you but leap? | 15:52 |
| lisitsyn | buy* | 15:52 |
| @wiking | ye | 15:52 |
| lisitsyn | I should have bought it too! hmm | 15:52 |
| @wiking | lisitsyn: i have a coupon | 15:52 |
| @wiking | lisitsyn: if u want to buy it... i think it's like 20USD discount or soemthing | 15:53 |
| lisitsyn | wiking: what would be the price then? | 15:53 |
| @wiking | mmmm | 15:53 |
| @wiking | w8 i'll tell | 15:53 |
| lisitsyn | oh no they not ship to ?ussia | 15:53 |
| lisitsyn | that wass predictable | 15:54 |
| lisitsyn | pickle27: I have a background subtraction problem | 15:54 |
| @wiking | lisitsyn: 70 usd | 15:54 |
| pickle27 | lisitsyn: sure | 15:54 |
| @wiking | + shipping | 15:54 |
| lisitsyn | pickle27: is that you have worked on? | 15:54 |
| pickle27 | lisitsyn: yeah my Thesis was on BGS | 15:54 |
| lisitsyn | I remember you said something | 15:54 |
| lisitsyn | ahh nice | 15:54 |
| lisitsyn | pickle27: now you're a mentor | 15:54 |
| lisitsyn | we change roles haha | 15:54 |
| pickle27 | haha | 15:55 |
| lisitsyn | wiking: I could buy it in SF if I get there | 15:55 |
| lisitsyn | pickle27: I need real-time BGS | 15:55 |
| @wiking | lisitsyn: do u want the code? | 15:55 |
| @wiking | GSoC2014 Ideas | 15:55 |
| @wiking | boosting / ensemble learning methods - port from matlab | 15:55 |
| @wiking | lol | 15:55 |
| @wiking | this is half done | 15:55 |
| @wiking | ;) | 15:55 |
| lisitsyn | gsoc2014 already? | 15:55 |
| lisitsyn | wiking: github! | 15:56 |
| @wiking | lisitsyn: yeah github | 15:56 |
| lisitsyn | wiking: I mean just share it | 15:56 |
| lisitsyn | ;) | 15:56 |
| @wiking | ah u mean the code? | 15:56 |
| lisitsyn | yes | 15:56 |
| @wiking | that's a one time only shit | 15:56 |
| @wiking | lol https://github.com/shogun-toolbox/shogun/wiki/Future-of-Shogun-Brainstorming | 15:56 |
| @wiking | this is like 80% done | 15:56 |
| pickle27 | lisitsyn: wiking for my future involvement with shogun I quite like doing matlab -> shogun or python -> shogun ports | 15:57 |
| pickle27 | fyi | 15:57 |
| lisitsyn | pickle27: ok so mentor me | 15:57 |
| lisitsyn | I need realtime bgs | 15:57 |
| @wiking | lisitsyn: :D | 15:57 |
| pickle27 | so Im happy to help with such thinga | 15:57 |
| pickle27 | ah right wasn;t sure you were done with the question | 15:57 |
| @wiking | bgs? | 15:57 |
| lisitsyn | wiking: background subtraction | 15:57 |
| pickle27 | lisitsyn: video data then? how big is the data? is it indoor or outdoor? | 15:57 |
| @wiking | ah ok | 15:57 |
| @wiking | :) | 15:57 |
| @wiking | lisitsyn: gpu it | 15:58 |
| @wiking | ;) | 15:58 |
| pickle27 | as in resolution of the images | 15:58 |
| lisitsyn | pickle27: let me send you a file | 15:58 |
| pickle27 | okay | 15:58 |
| lisitsyn | https://twitter.com/CVsIntellect/status/372706530178502657 haha see I am kicking asses | 15:59 |
| @wiking | thoralf: btw that was an interesting one that the valgrind was not finding any leaking with Release :( | 15:59 |
| thoralf | wiking: Yeah, I was wondering, too. | 16:00 |
| thoralf | wiking: But I don't know enough about valgrind&co to tell what caused it. | 16:00 |
| thoralf | wiking: btw. was no leak but an invalid read/write | 16:01 |
| lisitsyn | pickle27: please see your inbox | 16:01 |
| pickle27 | got it | 16:01 |
| @wiking | thoralf: make -j install # install required for "make test" | 16:01 |
| @wiking | thoralf: not true | 16:01 |
| thoralf | wiking: But the (python?) examples will fail without make install | 16:02 |
| van51 | sonne|work: it was because of the preprocessor.. it also changes the feature matrix of the CDotFeatures object it is working on | 16:02 |
| @wiking | thoralf: mmm indeed but we should drop that one | 16:02 |
| @wiking | thoralf: because we are switching to from modshogun import ... | 16:02 |
| @wiking | afaik foor shogun 3.0 we should already remove that | 16:03 |
| @wiking | :) | 16:03 |
| @wiking | but we'll see | 16:03 |
| sonne|work | van51: errm ? | 16:03 |
| thoralf | wiking: It's just what I (painfully ;)) learned so far and wanted to share. ;) | 16:03 |
| pickle27 | lisitsyn: okay so indoor scene fairly controlled decently high res video | 16:03 |
| sonne|work | van51: what preprocessor? | 16:03 |
| thoralf | wiking: You can always change it and/or correct me if I'm wrong. | 16:04 |
| van51 | sonne|work: I was testing both the random fourier dot features and the random fourier preprocessor on the same target dimensions | 16:04 |
| lisitsyn | pickle27: yeah all the videos are rather the same | 16:04 |
| pickle27 | lisitsyn: If you can I would use OpenCV probably the gpu bgs module | 16:04 |
| lisitsyn | pickle27: no gpu unfortunately | 16:04 |
| sonne|work | wiking: any news on the workshop video upload? | 16:04 |
| lisitsyn | pickle27: I tried MOG and MOG2 and GMG | 16:04 |
| pickle27 | lisitsyn: they didn't work well enough or were too slow? | 16:05 |
| lisitsyn | pickle27: the main problem is that the color of the vehicle is quite similar to the color of the floor | 16:05 |
| lisitsyn | so it slowly disappears :D | 16:05 |
| lisitsyn | from the bg/fg mask | 16:05 |
| pickle27 | lisitsyn: ah right, the learning params are weird in OpenCVs implementation | 16:05 |
| lisitsyn | pickle27: so I wonder if you know anything possible to be 25fps but based on some context around pixel | 16:06 |
| lisitsyn | may be texture based or so | 16:06 |
| sonne|work | van51: I still don't get it - so you had the preprocessor attached on the random fourier dot features? | 16:06 |
| pickle27 | lisitsyn: I don't think you need something complicated you just need something you have more control over | 16:06 |
| lisitsyn | pickle27: do you think GM model would be ok anyway? | 16:06 |
| van51 | sonne|work: I think this will explain it better : https://gist.github.com/van51/6366231 :) | 16:07 |
| lisitsyn | pickle27: RPCA is not near to realtime right? | 16:07 |
| pickle27 | lisitsyn: specifically an alg that features an update mask so you don't learn from what you just classified as foreground | 16:07 |
| lisitsyn | pickle27: oh good idea I haven't though about that | 16:07 |
| pickle27 | lisitsyn: RPCA is real time after training, but training a model takes forever | 16:07 |
| pickle27 | I'd also recommend a saliency filter for post processing | 16:07 |
| lisitsyn | pickle27: I don't know about these filters | 16:08 |
| lisitsyn | what is it? | 16:08 |
| pickle27 | lisitsyn: its really basic, you run your bgs alg with a high and low thresh | 16:08 |
| lisitsyn | ahh and how do I combine them? | 16:08 |
| pickle27 | then you do a grass fire type approach where you scan the high thresh map and look for a froegorund pixel and then fan out but then the pixels only need to beet the low thresh | 16:09 |
| pickle27 | so that all forground was connected to a high thresh pixel at some point | 16:09 |
| sonne|work | van51: no :D | 16:09 |
| pickle27 | it can let you have a lower thresh and get the obj outline better | 16:09 |
| sonne|work | van51: which line is the culprit? | 16:09 |
| pickle27 | lisitsyn: Im sending a paper | 16:10 |
| van51 | sonne|work: 63. it also set the dense_feats matrix | 16:10 |
| lisitsyn | pickle27: thanks! | 16:10 |
| van51 | sonne|work: I thought it was just returning it | 16:10 |
| pickle27 | lisitsyn: its a survey paper but he does a really nice job describing the common post processing stuff | 16:11 |
| lisitsyn | pickle27: how would you filter out some small blobs? | 16:11 |
| pickle27 | ahh an area filter, I'd use opencv's connected components | 16:11 |
| pickle27 | lisitsyn: Parks releases a lib of some basic techniques which I updated to modern opencv | 16:12 |
| pickle27 | https://github.com/pickle27/libBGS?source=c | 16:12 |
| sonne|work | van51: but you never touch dense_feats afterwards? | 16:12 |
| lisitsyn | pickle27: cool | 16:12 |
| pickle27 | the median alg might be a good one to try, super simple but should work for your indoor scene | 16:12 |
| van51 | sonne|work: I do in the next loop | 16:12 |
| sonne|work | ahh | 16:12 |
| sonne|work | missed the loop | 16:12 |
| pickle27 | lisitsyn: I also have some have done code for the post processing that I'll send you | 16:13 |
| sonne|work | van51: btw the kernel->set_normalizer should be refactored to be void set_normalizer() | 16:13 |
| lisitsyn | pickle27: oh great :) | 16:13 |
| van51 | sonne|work: if it cannot fail, sure :) | 16:13 |
| sonne|work | van51: well if it fails -> SG_ERROR | 16:14 |
| pickle27 | lisitsyn: the last thing I would recommend | 16:14 |
| lisitsyn | pickle27: the main problem I see now is that sometimes some flares are here | 16:14 |
| lisitsyn | and the floor is made of metal | 16:14 |
| sonne|work | this bool return stuff is rather useless | 16:14 |
| pickle27 | lisitsyn: right | 16:14 |
| van51 | sonne|work: new results : https://gist.github.com/van51/6366217 | 16:14 |
| pickle27 | I think this will help with that | 16:14 |
| pickle27 | so | 16:14 |
| lisitsyn | pickle27: I don't know how to handle that yet | 16:14 |
| lisitsyn | ah okay tell me | 16:14 |
| lisitsyn | ;) | 16:14 |
| pickle27 | lisitsyn: the thing with BGS is that when it comes to your application you probably want to have higher recall and lower precision to make sure you don't miss anything | 16:15 |
| pickle27 | like you are probably sending the foreground object to a tracker or something right? | 16:15 |
| sonne|work | van51: please compute max(abs(ref_matrix-random_ff)) | 16:15 |
| lisitsyn | pickle27: no I need to detect it | 16:15 |
| lisitsyn | pickle27: it is just a first step for me | 16:15 |
| pickle27 | lisitsyn: thats what I mean its a first step | 16:16 |
| sonne|work | van51: and please use other data :) | 16:16 |
| lisitsyn | pickle27: ahh yeah okay | 16:16 |
| sonne|work | van51: this kernel is basically a 0 everywhere and 1 on the diagonal | 16:16 |
| pickle27 | lisitsyn: so you really only need your BGS alg to get a good guess | 16:16 |
| pickle27 | I'd recommend making another system on top using machine learning to find the mistakes from BGS | 16:16 |
| pickle27 | some code is being linked in a sec | 16:17 |
| sonne|work | van51: or make the width of the kernel to be about num_dim | 16:17 |
| -!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has joined #shogun | 16:17 | |
| -!- mode/#shogun [+o iglesiasg] by ChanServ | 16:17 | |
| @iglesiasg | hello hello | 16:17 |
| lisitsyn | pickle27: yeah I need some detection on top of that | 16:17 |
| lisitsyn | no idea what kind of though | 16:17 |
| pickle27 | lisitsyn: I'd use something similar to what they use here: https://github.com/MasteringOpenCV/code/tree/master/Chapter5_NumberPlateRecognition | 16:17 |
| @iglesiasg | wiking, more than 10 mails from Viktor Gal in my mailbox! :D | 16:18 |
| pickle27 | They run a segmentation (for you this is BGS) and then use a SVM to take all the candidates and decide which are actually licence plates | 16:18 |
| pickle27 | I used something like this here as well: | 16:18 |
| van51 | sonne|work: ok, will do | 16:18 |
| lisitsyn | pickle27: yeah that sounds good | 16:18 |
| sonne|work | van51: but it looks OKish - nice work! | 16:18 |
| lisitsyn | pickle27: I think some HOG would detect the vehicle | 16:18 |
| pickle27 | here I wrote on my blog about this | 16:19 |
| pickle27 | http://kevinhughes.ca/2013/02/21/frc-ultimate-ascent-vision-system/ | 16:19 |
| pickle27 | what I did was still pretty simple but I think thats the framework you need | 16:19 |
| sonne|work | gsomix: did you fix the locale bug? | 16:19 |
| pickle27 | lisitsyn: also you might find some good src code for newer algs on this site | 16:20 |
| pickle27 | http://changedetection.net/ | 16:20 |
| pickle27 | BSG wise | 16:20 |
| lisitsyn | pickle27: quite cool! | 16:20 |
| lisitsyn | pickle27: thanks a lot | 16:21 |
| pickle27 | lisitsyn: np! I hope that helps | 16:21 |
| lisitsyn | pickle27: actually the task is to determine whether the vehicle carries something | 16:21 |
| pickle27 | lisitsyn: ahh cool | 16:21 |
| lisitsyn | pickle27: I have no idea whether svm can handle that | 16:22 |
| pickle27 | lisitsyn: yeah me either | 16:23 |
| pickle27 | lisitsyn: are you going to try and use Shogun at all? I've been wanting to do something combining opencv and shogun for a while | 16:24 |
| -!- HeikoS [~heiko@nat-166-161.internal.eduroam.ucl.ac.uk] has joined #shogun | 16:24 | |
| -!- mode/#shogun [+o HeikoS] by ChanServ | 16:24 | |
| lisitsyn | pickle27: don't know yet actually | 16:26 |
| lisitsyn | but once I get to machine learning that's the tool to use | 16:26 |
| lisitsyn | quite bad we have no adaboost and stuff though | 16:26 |
| pickle27 | lisitsyn: opencv has adaboost we could probably prot over | 16:26 |
| lisitsyn | pickle27: yeah sure | 16:27 |
| @iglesiasg | pickle27, I think viktor has done some stuff with Shogun and opencv together | 16:33 |
| pickle27 | iglesiasg: sweet! | 16:34 |
| @iglesiasg | I think he said that for the lung detection program (or sth similar) he showed in the workshop | 16:34 |
| pickle27 | iglesiasg: ah I actually remember looking at that briefly | 16:34 |
| -!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 16:36 | |
| shogun-notifier- | shogun: Roman Votyakov :develop * a905d18 / src/shogun/machine/gp/EPInferenceMethod.cpp: https://github.com/shogun-toolbox/shogun/commit/a905d18f500db6116e884d4ecdc6d1cb10bc94b0 | 16:36 |
| shogun-notifier- | shogun: fix invalid memory read in CEPInferenceMethod class | 16:36 |
| shogun-notifier- | shogun: Heiko Strathmann :develop * 8c65f74 / src/shogun/machine/gp/EPInferenceMethod.cpp: https://github.com/shogun-toolbox/shogun/commit/8c65f743ab7830180bab51ea1b88216ccc38f658 | 16:36 |
| shogun-notifier- | shogun: Merge pull request #1478 from votjakovr/feature/gp_refactoring | 16:36 |
| shogun-notifier- | shogun: | 16:36 |
| shogun-notifier- | shogun: Fix invalid memory read in CEPInferenceMethod class | 16:36 |
| -!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has joined #shogun | 16:42 | |
| @iglesiasg | votjakovr, hey | 16:46 |
| @iglesiasg | votjakovr, just tried compiling with -g and still no error :D | 16:46 |
| votjakovr | iglesiasg: hi! hmm... | 16:47 |
| @iglesiasg | votjakovr, it is just for fun, not something really important though hehe | 16:47 |
| @iglesiasg | just curious about it | 16:48 |
| @iglesiasg | but anyway yes, what you said makes complete sense | 16:48 |
| votjakovr | iglesiasg: yeah, but i still have no idea, why valgrind doesn't report an error in your example | 16:49 |
| @iglesiasg | votjakovr, yep, me neither | 16:50 |
| @iglesiasg | votjakovr, let me try if it happens when using SGVector | 16:50 |
| @iglesiasg | HeikoS, I think I forgot checking in that PR the thing of m_ in SG_ADD | 16:52 |
| @HeikoS | iglesiasg: ? | 16:52 |
| @iglesiasg | HeikoS, yeah sorry, in hushell's pull request | 16:53 |
| @HeikoS | ah yeah | 16:53 |
| @iglesiasg | the one you were commenting a minute ago | 16:53 |
| shogun-buildbot | build #1551 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1551 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 16:53 |
| @HeikoS | iglesiasg: yep got it | 16:54 |
| @iglesiasg | HeikoS, thank you | 16:54 |
| @HeikoS | thanks for reviewing all that stuff btw :) | 16:55 |
| @HeikoS | iglesiasg: what do you think about all students reviewing each others notebooks for gsoc? | 16:55 |
| @iglesiasg | HeikoS, very good idea | 16:56 |
| @HeikoS | you suggested that right? | 16:56 |
| @HeikoS | takes some time though | 16:56 |
| @iglesiasg | I don't know | 16:56 |
| @iglesiasg | maybe I did hehe | 16:56 |
| @iglesiasg | but one reading all the others is quite a bit | 16:56 |
| @iglesiasg | I would suggest just doing it randomly | 16:56 |
| @iglesiasg | I mean, if someone wants to read all of them, that's fine | 16:56 |
| @iglesiasg | but maybe each student could be assigned to another project and be expected to comment a little bit about it | 16:57 |
| @HeikoS | when do you think is a good idea to have the thing ready so that there is some time to review? | 16:58 |
| @HeikoS | also would be good if students read each others class documentations | 16:58 |
| @HeikoS | iglesiasg: should we maybe ask every student for a list of class (documentations) that the others should read and suggest improvements | 16:58 |
| @HeikoS | ok what about everyone does two? | 16:58 |
| @HeikoS | who is writing class based code? | 16:58 |
| @HeikoS | or who is not? | 16:58 |
| @HeikoS | foulwall is not right? | 16:58 |
| @HeikoS | but still his stuff should be documented | 16:58 |
| @HeikoS | maybe if we generate random paris of students to review each others dox? | 16:58 |
| @HeikoS | do you think people will like that? | 16:58 |
| @HeikoS | will probably be useful | 16:58 |
| @HeikoS | iglesiasg: yeah its not about reviewing code | 16:58 |
| @HeikoS | more about reeviewing the docs so that people understand | 16:58 |
| @HeikoS | sometimes one has a very narrow view after working on it for months | 16:58 |
| @iglesiasg | in particular foulwall I think it is pretty important his stuff is well documented | 16:59 |
| @HeikoS | yewp | 16:59 |
| @iglesiasg | so that it is easy in the future for other people extend with more demos | 16:59 |
| @HeikoS | ok so then I will ask everyone to have things done one 1-2 weeks before deadlines | 17:00 |
| @HeikoS | so that that last two weeks can be spent on fine-tuning and reviewing own and other's docs | 17:00 |
| @iglesiasg | HeikoS, that sounds good to me | 17:00 |
| votjakovr | iglesiasg: i've figured, if we create an array in heap, then valgrind reports about invalid memory read | 17:01 |
| @iglesiasg | votjakovr, aham that would make sense! | 17:02 |
| lisitsyn | http://www.secretgeek.net/ex_ms.asp | 17:03 |
| @iglesiasg | lisitsyn, haha man that's hilarious | 17:06 |
| @iglesiasg | lisitsyn, are those dir names for real? | 17:06 |
| lisitsyn | iglesiasg: SysWOW64 yes | 17:07 |
| lisitsyn | Documents and Settings | 17:07 |
| lisitsyn | haven't you used windows? :D | 17:08 |
| @iglesiasg | lisitsyn, eeeh when I was young hehe | 17:08 |
| pickle27 | I had a good laugh too | 17:08 |
| pickle27 | I can't stand windows anymore | 17:09 |
| @iglesiasg | "WOW, nature is beautiful!" killed me :D | 17:09 |
| gsomix | sonne|work, yes. forgot to send PR | 17:09 |
| votjakovr | HeikoS: hi! I'm currently thinking about non negative parameters... How is it better to deal with it? Move to log space or something else? What do you think about that? | 17:12 |
| -!- sonne|osx [~sonne@89.204.138.92] has joined #shogun | 17:12 | |
| @HeikoS | votjakovr: how would these come up? | 17:12 |
| @HeikoS | votjakovr: does the ML2 push one to negative places? | 17:13 |
| @HeikoS | votjakovr: in that case: yes log-space | 17:13 |
| votjakovr | HeikoS: currently yes | 17:13 |
| @HeikoS | votjakovr: but that means that we have to change derivatives dont we? | 17:13 |
| shogun-notifier- | shogun: Soeren Sonnenburg :develop * 95c81d3 / CMakeLists.txt: https://github.com/shogun-toolbox/shogun/commit/95c81d37aa006885b2eb4fd9df9c593c0055d420 | 17:14 |
| shogun-notifier- | shogun: add cmdline static status | 17:14 |
| shogun-notifier- | shogun: Soeren Sonnenburg :develop * 64016bc / NEWS: https://github.com/shogun-toolbox/shogun/commit/64016bcb50c8f6eb79a7e5d2c8f6c95c9fc1ba06 | 17:14 |
| shogun-notifier- | shogun: update news to 3.0 | 17:14 |
| votjakovr | HeikoS: in GPs derivatives are in log space already | 17:14 |
| @HeikoS | votjakovr: I see, so little effort moving to log space right? | 17:14 |
| votjakovr | HeikoS: but something like width in Gaussian kernel is not | 17:14 |
| @HeikoS | votjakovr: we cannot change that though | 17:15 |
| @HeikoS | votjakovr: ideas how to deal with that? | 17:15 |
| shogun-notifier- | shogun: Evgeniy Andreev :develop * fea6d2d / / (5 files): https://github.com/shogun-toolbox/shogun/commit/fea6d2d2f56a10eab683bcec455985167b89a5de | 17:16 |
| shogun-notifier- | shogun: probably fix for issues #1401 #1402 | 17:16 |
| shogun-notifier- | shogun: Soeren Sonnenburg :develop * 4f55402 / / (5 files): https://github.com/shogun-toolbox/shogun/commit/4f5540210a22f0300e27b8666c5032648a9adaaf | 17:16 |
| shogun-notifier- | shogun: Merge pull request #1489 from gsomix/feature/libsvm | 17:16 |
| shogun-notifier- | shogun: | 17:16 |
| shogun-notifier- | shogun: Probably fix for issues #1401 #1402 | 17:16 |
| sonne|osx | van51: I guess I can merge the random fourier features now right? | 17:16 |
| votjakovr | HeikoS: no ideas yet, thinking about it... | 17:16 |
| van51 | sonne|osx: ah no | 17:16 |
| sonne|osx | van51: still adding a test ? | 17:17 |
| van51 | sonne|osx: it did have a bug | 17:17 |
| @HeikoS | votjakovr: we could add a slag to the parameters, whether they are in std space or log-space | 17:17 |
| van51 | sonne|osx: and I added support to specify kernel parameters | 17:17 |
| sonne|osx | van51: that you fixed but not in the PR? | 17:17 |
| van51 | sonne|osx: yes I haven't pushed yet | 17:17 |
| sonne|osx | van51: yes please do that and the kernel test would be good too | 17:17 |
| van51 | sonne|osx: yea I will update it soon and let you know :) | 17:17 |
| @HeikoS | votjakovr: or we just say: kernel parameters are in std-space, all others in GP in log-space (thats correct right) | 17:17 |
| @HeikoS | votjakovr: so then, you could just do the transformation by hand, i.e. translate log-parameters back to normal ones for kernels | 17:18 |
| sonne|osx | alrighty | 17:19 |
| sonne|osx | home sweet home time | 17:19 |
| votjakovr | HeikoS: ok | 17:19 |
| -!- sonne|osx [~sonne@89.204.138.92] has quit [Quit: sonne|osx] | 17:19 | |
| votjakovr | HeikoS: than i move all GPs parameters to log space, right? | 17:20 |
| @HeikoS | votjakovr: yes how much work is that? | 17:20 |
| gsomix | sonne|osx, ok. can we force osx buildbot to build? | 17:20 |
| votjakovr | HeikoS: i'll do this quickly | 17:21 |
| shogun-buildbot | build #1636 of deb3 - modular_interfaces is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1636 | 17:22 |
| votjakovr | HeikoS: is it ok, to name them something like: log_sigma? | 17:23 |
| @HeikoS | votjakovr: do you think they should be renamed? | 17:26 |
| @HeikoS | votjakovr: maybe do it | 17:26 |
| votjakovr | HeikoS: yeah, i think it's hard to understand what this output means: Quantity: [205.046388] Gradient: [width: -0.126531; scale: 1.860658; sigma: 96.633392] | 17:27 |
| shogun-buildbot | build #1552 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1552 blamelist: Soeren Sonnenburg <sonne@debian.org> | 17:27 |
| votjakovr | HeikoS: i mean what is width, scale, sigma, etc.. | 17:28 |
| votjakovr | HeikoS: and i probably parameters should have meaningful names or not? | 17:30 |
| @HeikoS | votjakovr: yes | 17:30 |
| -!- lisitsyn [~lisitsin@mxs.kg.ru] has quit [Quit: Leaving.] | 17:30 | |
| @HeikoS | do it then | 17:30 |
| @HeikoS | votjakovr: good thing | 17:30 |
| @HeikoS | votjakovr: but its just a subtlety | 17:31 |
| shogun-buildbot | build #1339 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1339 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 17:32 |
| -!- travis-ci [~travis-ci@ec2-23-22-254-224.compute-1.amazonaws.com] has joined #shogun | 17:32 | |
| 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/10715248 | 17:32 |
| -!- travis-ci [~travis-ci@ec2-23-22-254-224.compute-1.amazonaws.com] has left #shogun [] | 17:32 | |
| shogun-buildbot | build #1553 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1553 blamelist: Roman Votyakov <votjakovr@gmail.com> | 17:33 |
| @HeikoS | votjakovr: I am curious whether we can get rid of the negative parameters this way | 17:34 |
| @HeikoS | votjakovr: or whether there is a bug thats causing it | 17:34 |
| @HeikoS | votjakovr: we can always compare against gpml ML2 I guess | 17:35 |
| votjakovr | HeikoS: yeah, i think there is a bug, i'll do an investigation | 17:36 |
| @HeikoS | votjakovr: cool | 17:36 |
| @HeikoS | votjakovr: also try to figure out a way to initialise the parameters with values | 17:36 |
| @HeikoS | votjakovr: I think the starting point should always be the "current" parameters | 17:36 |
| @HeikoS | not some specified by a tree | 17:37 |
| votjakovr | HeikoS: absolutely agree! | 17:37 |
| @HeikoS | votjakovr: and this brings us to the biggest problem I currently have with this | 17:37 |
| @HeikoS | votjakovr: the way users have to specify the parameter trees | 17:37 |
| @HeikoS | votjakovr: this is not really necessary. I usually want to optimize all parameters right? | 17:38 |
| @HeikoS | votjakovr: but maybe the possibility to optimize just a few is good | 17:38 |
| @HeikoS | votjakovr: so there should be a nice way to do this .... | 17:38 |
| @HeikoS | maybe you can think of something | 17:38 |
| shogun-buildbot | build #1554 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1554 blamelist: Evgeniy Andreev <gsomix@gmail.com>, Soeren Sonnenburg <sonne@debian.org> | 17:38 |
| @HeikoS | votjakovr: but first priority is to make it work | 17:38 |
| @HeikoS | cosmetics afterwards :) | 17:38 |
| votjakovr | HeikoS: ok :) | 17:39 |
| pickle27 | iglesiasg: did Georg ever send you the gene expression data? | 17:57 |
| @iglesiasg | pickle27, gene expression? | 17:57 |
| @iglesiasg | was that data for your project or for my project? | 17:57 |
| pickle27 | iglesiasg: it was for yours but he mentioned on the mailing list a while back that people have used ICA to pre process | 17:58 |
| pickle27 | so that maybe I could use it too | 17:58 |
| pickle27 | Im just considering making another notebook | 17:58 |
| @iglesiasg | pickle27, aham ok. So he has just sent me a mail a few minutes ago, we said he would have the data ready for Friday | 17:58 |
| @iglesiasg | but I am unsure if it is the one you mean | 17:59 |
| @iglesiasg | maybe you can mail him about it | 17:59 |
| pickle27 | iglesiasg: will he be at the meeting? I could juts talk to him then | 17:59 |
| @iglesiasg | pickle27, so last time I checked the doodle I think he had written he couldn't make it at the time that was finally established. But from this mail I got the idea he will be | 18:00 |
| pickle27 | iglesiasg: okay well either way I'll wait til friday | 18:00 |
| @iglesiasg | pickle27, all right | 18:00 |
| @iglesiasg | pickle27, he told me he is very busy these days so probably it is a good idea to ask him about it in the meeting | 18:01 |
| pickle27 | iglesiasg: okay, its deffs not a big deal if he doesn't have data that can work for me I just thought it might be neat | 18:02 |
| @iglesiasg | pickle27, yeah sure | 18:02 |
| -!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 240 seconds] | 18:11 | |
| shogun-buildbot | build #1638 of deb3 - modular_interfaces is complete: Failure [failed install test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1638 blamelist: Roman Votyakov <votjakovr@gmail.com> | 18:11 |
| shogun-buildbot | build #1340 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1340 blamelist: Evgeniy Andreev <gsomix@gmail.com>, Soeren Sonnenburg <sonne@debian.org>, Roman Votyakov <votjakovr@gmail.com> | 18:16 |
| -!- lambday [67157d4c@gateway/web/freenode/ip.103.21.125.76] has joined #shogun | 18:20 | |
| -!- gsomix [~gsomix@178.45.40.223] has quit [Quit: Leaving] | 18:41 | |
| -!- travis-ci [~travis-ci@ec2-54-224-40-163.compute-1.amazonaws.com] has joined #shogun | 18:43 | |
| 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/10716726 | 18:43 |
| -!- travis-ci [~travis-ci@ec2-54-224-40-163.compute-1.amazonaws.com] has left #shogun [] | 18:43 | |
| -!- van51 [~van51@ppp-2-87-3-175.home.otenet.gr] has quit [Quit: Leaving.] | 18:54 | |
| -!- gsomix [~gsomix@178.45.40.223] has joined #shogun | 19:06 | |
| -!- travis-ci [~travis-ci@ec2-54-224-40-163.compute-1.amazonaws.com] has joined #shogun | 19:12 | |
| 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/10716786 | 19:12 |
| -!- travis-ci [~travis-ci@ec2-54-224-40-163.compute-1.amazonaws.com] has left #shogun [] | 19:12 | |
| -!- pickle27 [~kevin@d67-193-243-174.home3.cgocable.net] has left #shogun [] | 19:38 | |
| @wiking | HeikoS: yo man | 19:39 |
| -!- hoijui [~hoijui@wlan-141-23-94-144.tubit.tu-berlin.de] has joined #shogun | 19:41 | |
| @HeikoS | wiking: yoyoy | 19:42 |
| @HeikoS | whazup? | 19:42 |
| @wiking | mmm you'll ee | 19:42 |
| @wiking | ;) | 19:42 |
| @wiking | in my github pr comment :D | 19:42 |
| @HeikoS | what? | 19:43 |
| @HeikoS | ee? | 19:43 |
| @HeikoS | see | 19:43 |
| @HeikoS | wiking: ah colpack | 19:43 |
| @HeikoS | well, as long as there is a quickfix, I am fine | 19:43 |
| @HeikoS | I in fact cloned lambdays branch to work with it | 19:43 |
| shogun-notifier- | shogun: Viktor Gal :develop * 7711764 / src/shogun/base/class_list.cpp.py: https://github.com/shogun-toolbox/shogun/commit/7711764498e59a07d740c27c90b2ffac9782feea | 19:44 |
| shogun-notifier- | shogun: Add HAVE_COLPACK to config_tests in class_list.cpp.py | 19:44 |
| @wiking | there | 19:44 |
| @wiking | that should fix lambday's problem... but he'll have to revert the last changes i requested :( | 19:44 |
| @HeikoS | wiking: merge his PR when as soon as you dare ;) | 19:44 |
| @HeikoS | wiking: he will get over it | 19:44 |
| @wiking | HeikoS: i'll wait for the reversion ;) | 19:44 |
| @HeikoS | wiking: we can do that thing properly ater on | 19:44 |
| @HeikoS | wiking: now I just need the sampler | 19:44 |
| @HeikoS | wiking: excited, because my PBS swarm works | 19:45 |
| @HeikoS | all I need is this log-determinant | 19:45 |
| @wiking | votjakovr: didn't u wanted NLOPT on travis? | 19:45 |
| @HeikoS | wiking: thats votjakovr | 19:45 |
| @HeikoS | he needs it | 19:45 |
| @HeikoS | so yes | 19:45 |
| @wiking | HeikoS: awesooome where's the python script | 19:45 |
| @wiking | HeikoS: ok let me generate the shiatz | 19:45 |
| -!- sonne|osx [~sonne@f053046168.adsl.alicedsl.de] has joined #shogun | 19:45 | |
| @wiking | votjakovr: nlopt 2.3 is good enough? | 19:46 |
| @HeikoS | wiking: its quite some code | 19:47 |
| @HeikoS | wiking: I will put into a git soon | 19:47 |
| @HeikoS | just very raw prototype | 19:47 |
| @HeikoS | I dont think we should extend it | 19:48 |
| @HeikoS | rather use it as a basis to develop this stuff in shogun | 19:48 |
| lambday | wiking: hi just saw your comment | 19:48 |
| @wiking | lambday: sorry about that | 19:48 |
| lambday | wiking: lol no no not a problem | 19:48 |
| lambday | the only problem is, I never reverted to an older commit before.. | 19:48 |
| lambday | :-/ | 19:48 |
| @wiking | is there any special stuff from nlopt we need? | 19:49 |
| lambday | with git | 19:49 |
| lambday | I am figuring it out | 19:49 |
| @wiking | lambday: just do the changes again | 19:49 |
| @wiking | and commit it | 19:49 |
| @wiking | reversion is a bitch | 19:49 |
| hushell | HeikoS: Hi, may I ask a short question? | 19:49 |
| @HeikoS | hushell: shoot | 19:49 |
| @wiking | and i'm afraid your branch is going to die | 19:49 |
| @wiking | in a way | 19:49 |
| lambday | wiking: that's easier but there was something for git.. | 19:49 |
| @wiking | so rather just redo and recommit | 19:49 |
| hushell | HeikoS: why should we init every member in the init()? | 19:49 |
| lambday | okay making these changes is easier actually - gotta change 5 lines | 19:49 |
| hushell | HeikoS: I mean for different constructor, some of them may be initialized differently | 19:50 |
| @HeikoS | hushell: because everyone (including me and you) keep on forgetting to initialized members in classes and that causes memory errors which I spent many hours fixing recently | 19:50 |
| @HeikoS | hushell: you still can do that, just make sure you call init before | 19:51 |
| @HeikoS | then *everything* is initialised | 19:51 |
| shogun-buildbot | build #57 of osx1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/57 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Soeren Sonnenburg <sonne@debian.org>, Thoralf Klein <thoralf.klein@zib.de>, Evgeniy Andreev <gsomix@gmail.com>, Roman Votyakov | 19:51 |
| shogun-buildbot | <votjakovr@gmail.com> | 19:51 |
| @HeikoS | and then you can override | 19:51 |
| hushell | HeikoS: I see, for safty reason | 19:51 |
| @HeikoS | but doing it when registering is good practice to guaranty that they are registered | 19:51 |
| @HeikoS | hushell: in the near future, forgetting it will break unit-tests but not yet | 19:51 |
| hushell | HeikoS: SGVector is not necessarily being initialized, isn't it? | 19:51 |
| @HeikoS | hushell: its just a good practice thing | 19:51 |
| votjakovr | wiking: yeah, nlopt 2.3 is ok | 19:52 |
| @HeikoS | hushell: SGVector does not need to be initialised since its on the stack and the default constructor does this for you | 19:52 |
| @HeikoS | hushell: only basic types and CSGObject | 19:52 |
| gsomix | wiking, hey. it seems bug #1401 #1402 is fixed, right? http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/57/steps/test/logs/stdio | 19:52 |
| hushell | HeikoS: okay, I got the point! Thanks! | 19:52 |
| @wiking | gsomix: dunno, but i've just restarted the travis osx job as well | 19:53 |
| @wiking | gsomix: but most likely yes | 19:53 |
| @HeikoS | hushell: not a big thing, just helps me in getting rid of all memory errors :) | 19:53 |
| shogun-buildbot | build #1555 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1555 blamelist: Viktor Gal <viktor.gal@maeth.com> | 19:54 |
| shogun-buildbot | build #58 of osx1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/58 blamelist: Viktor Gal <viktor.gal@maeth.com> | 19:56 |
| votjakovr | wiking: and i'd like to have it on travis :) | 19:56 |
| lambday | changed | 19:56 |
| lambday | testing on the new machine | 19:56 |
| @wiking | votjakovr: yeah first let me make it locally | 19:58 |
| @wiking | :) | 19:58 |
| @wiking | votjakovr: which class should use NLopt? | 19:59 |
| votjakovr | wiking: GradientModelSelection | 19:59 |
| @wiking | votjakovr: ok lets see if it compiles | 19:59 |
| @wiking | ;) | 19:59 |
| @sonney2k | whose is this RationalApproximationCGMJob.compute failing on osx? | 20:01 |
| @wiking | sonney2k: lambday | 20:01 |
| lambday | wiking: yes | 20:01 |
| @sonney2k | lambday, http://shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/58/steps/test/logs/stdio | 20:01 |
| lambday | sonney2k: checking | 20:02 |
| lambday | direct_result evaluates to -inf !!!! | 20:02 |
| lambday | direct result is what we get using eigen3 | 20:03 |
| lambday | wait I am checking more | 20:03 |
| lambday | sonney2k: the reference value I am checking it against is the value computed using a direct eigen3 linear solver - which is giving -inf... I am not sure why that is... I am testing | 20:06 |
| lambday | I never got this error on linux :-/ | 20:06 |
| lambday | sonney2k: a temporary solution would be to put the actual value to check it with, instead of using eigen3 to generate that value.... | 20:07 |
| @sonney2k | HeikoS, still around? | 20:10 |
| shogun-notifier- | shogun: Viktor Gal :develop * 4a6422d / .travis.yml,CMakeLists.txt,cmake/external/NLopt.cmake: https://github.com/shogun-toolbox/shogun/commit/4a6422d9305773925c89e4c788b45e88e492ce32 | 20:10 |
| shogun-notifier- | shogun: Add BUNDLE_NLOPT option to build and link NLopt as external project | 20:10 |
| shogun-notifier- | shogun: enable BUNDLE_NLOPT on travis | 20:10 |
| @wiking | votjakovr: ^ | 20:10 |
| @HeikoS | sonney2k: yes but about to go, whats up? | 20:11 |
| @sonney2k | lambday, hmmhh maybe it is a numerically not so stable case? | 20:11 |
| @sonney2k | HeikoS, about booking flights? | 20:11 |
| @HeikoS | yes | 20:11 |
| @sonney2k | HeikoS, if we want to share a car then we should sync | 20:11 |
| @sonney2k | if not then not :) | 20:11 |
| lambday | sonney2k: HeikoS I don't think that is so... this is a simple diagonal matrix :( | 20:11 |
| @HeikoS | sonney2k: we already agreed we want that right? I would also like to share the flight, otherwise its boring | 20:12 |
| @HeikoS | we can do shogun hacking in the plane :D | 20:12 |
| votjakovr | wiking: thank you very much!!! :) | 20:12 |
| @sonney2k | so my plans were to leave on friday and then go back friday or so | 20:12 |
| @HeikoS | lambday: since when do you have the error? | 20:12 |
| @sonney2k | HeikoS, so I should bring 3 notebooks just in case :D | 20:12 |
| lambday | HeikoS: I didn't... sonney2k discovered this error on osx :( | 20:12 |
| @HeikoS | sonney2k: I am pretty much fine with everything | 20:12 |
| @HeikoS | flexible | 20:12 |
| @HeikoS | sonney2k: I will bring mine certainly :) | 20:12 |
| lambday | I never noticed it | 20:12 |
| @sonney2k | lambday, I can run stuff on this osx machine if it helps... | 20:12 |
| @HeikoS | lambday: so its arch dependent? | 20:13 |
| lambday | sonney2k: alright.. I will test and will let you know | 20:13 |
| @HeikoS | then it cannot be a numerical issue right? | 20:13 |
| @sonney2k | HeikoS, do you know when sergey is flying? | 20:13 |
| lambday | HeikoS: its a unit-test.. I am checking a value generated from our algo against a direct eigen3 linear solver for complex.... the problem is... in osx, the eigen3 one is giving -inf as answer | 20:13 |
| lambday | HeikoS: I don't have any idea why is that | 20:14 |
| @sonney2k | HeikoS, so friday-friday is ok for you? | 20:14 |
| @HeikoS | sonney2k: so we two can go together | 20:14 |
| @HeikoS | sonney2k: are you maybe flying over london? | 20:14 |
| @HeikoS | then I can meet you there? | 20:14 |
| @HeikoS | sonney2k: better ask him yourself, I dont think he has planned yet, | 20:14 |
| @HeikoS | sonney2k: which dates are those? | 20:15 |
| @HeikoS | 18-25 | 20:15 |
| @HeikoS | yes | 20:16 |
| @HeikoS | its fine | 20:16 |
| @HeikoS | time wise, I would prefer something human | 20:16 |
| @HeikoS | sonney2k: question is: are you going over london? | 20:16 |
| @HeikoS | or am I going over Berlin | 20:16 |
| @sonney2k | HeikoS, well we have to leave in the morning to arrive around 4-5 pm at the hotel an be in time for dinner | 20:17 |
| @sonney2k | HeikoS, are there direct flights from london? | 20:17 |
| @HeikoS | sonney2k: no idea, I can check tonight though | 20:17 |
| @HeikoS | which city? | 20:17 |
| @HeikoS | are we going to? | 20:17 |
| @sonney2k | HeikoS, SFO | 20:18 |
| shogun-buildbot | build #1639 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1639 blamelist: Evgeniy Andreev <gsomix@gmail.com>, Soeren Sonnenburg <sonne@debian.org> | 20:18 |
| lambday | you guys are going for mentor summit? :) | 20:18 |
| @HeikoS | sonney2k: btw what about leaving one day earlier? to have things a bit more relaxed? | 20:19 |
| @sonney2k | lambday, yeah | 20:19 |
| lambday | :) | 20:19 |
| @HeikoS | sonney2k: there are direct flights from london | 20:19 |
| @sonney2k | HeikoS, when? | 20:19 |
| @HeikoS | sonney2k: what when? | 20:19 |
| @sonney2k | HeikoS, when are the flights? | 20:20 |
| @sonney2k | I mean arrival / departure | 20:20 |
| @HeikoS | sonney2k: well we can choose | 20:20 |
| @HeikoS | http://www.britishairways.com/travel/fx/public/en_gb?eId=111011×tamp=0828061850# | 20:20 |
| @HeikoS | does that link work for you? | 20:20 |
| @HeikoS | 1145 - 1435 | 20:20 |
| @HeikoS | for example | 20:20 |
| @HeikoS | on 18th | 20:20 |
| @HeikoS | that would also work for you to come here | 20:21 |
| @sonney2k | no doesn't | 20:21 |
| @HeikoS | since 1145 is not too early | 20:21 |
| @HeikoS | quite cheap | 20:22 |
| @HeikoS | is the one 14==-1650 | 20:22 |
| @HeikoS | 1400 | 20:22 |
| @HeikoS | 263 punds | 20:22 |
| @HeikoS | the 11-45-1435 is 464 | 20:22 |
| @HeikoS | you can search for yourself | 20:22 |
| @HeikoS | some connection flights are also cheaper | 20:23 |
| @HeikoS | but I prefer a direct one | 20:23 |
| @HeikoS | sonney2k: have a look what would work for you, I have to go home now, should we discuss tomorrow afternoon? | 20:23 |
| @sonney2k | HeikoS, flights are pretty expensive over heathrow :/ | 20:25 |
| @HeikoS | sonney2k: how about from berlin? | 20:25 |
| @HeikoS | sonney2k: this is british airways which is a bit more exp | 20:25 |
| @HeikoS | sonney2k: I will google a bit tonight to find cheaper ones | 20:26 |
| @sonney2k | 907 EUR | 20:26 |
| @sonney2k | HeikoS, how much is yours? | 20:26 |
| @HeikoS | cheapest is about 600 | 20:27 |
| @HeikoS | more expensive is about 900-1000 | 20:27 |
| @HeikoS | depends | 20:27 |
| @sonney2k | HeikoS, 700-800 is what I find otherwise | 20:27 |
| @HeikoS | sonney2k: I will google a bit tonight and let you know tomorrow | 20:27 |
| @HeikoS | you can do the same for berlin | 20:27 |
| @HeikoS | then we can compare | 20:27 |
| @HeikoS | maybe it makes sense to fly seperately | 20:27 |
| @sonney2k | HeikoS, well I selected the one that gets us to SFO at 14:35 | 20:27 |
| @HeikoS | sonney2k: what about arriving one day earlier? | 20:27 |
| @HeikoS | then its easier to sync | 20:28 |
| @HeikoS | sonney2k: gotta run off now, sorry. see you tomorrow | 20:28 |
| lambday | going for dinner - sonney2k: I will check into that issue.. will send you some tests for osx probably :/ | 20:28 |
| @sonney2k | HeikoS, ok tomorrow then cu | 20:29 |
| @sonney2k | lambday, yeah I am not around today much longer anyway | 20:29 |
| lambday | sonney2k: will mail you :) | 20:30 |
| shogun-buildbot | build #59 of osx1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/59 blamelist: Viktor Gal <viktor.gal@maeth.com> | 20:30 |
| shogun-buildbot | build #1556 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1556 blamelist: Viktor Gal <viktor.gal@maeth.com> | 20:30 |
| lambday | sonney2k: on the machine you gave the ssh login yesterday, all tests are working fine | 20:31 |
| lambday | alright I will check | 20:31 |
| lambday | wiking: yay have_colpack test seems to have worked fine... | 20:31 |
| @wiking | lambday: \o | 20:31 |
| @wiking | awesome | 20:31 |
| lambday | wiking: I just checked on the new machine that sonney2k gave us access to... and it doesn't have colpack still compiles all fine and all the tests too | 20:32 |
| lambday | I didn't check what travis said though | 20:32 |
| shogun-buildbot | build #1341 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1341 blamelist: Viktor Gal <viktor.gal@maeth.com> | 20:33 |
| -!- HeikoS [~heiko@nat-166-161.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.] | 20:34 | |
| lambday | brb | 20:35 |
| shogun-buildbot | build #1640 of deb3 - modular_interfaces is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1640 | 20:36 |
| @wiking | lambday: ok that should be fine i'll merge it | 20:43 |
| -!- travis-ci [~travis-ci@ec2-54-224-40-163.compute-1.amazonaws.com] has joined #shogun | 20:44 | |
| 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/10716786 | 20:44 |
| -!- travis-ci [~travis-ci@ec2-54-224-40-163.compute-1.amazonaws.com] has left #shogun [] | 20:44 | |
| shogun-notifier- | shogun: lambday :develop * debe6ac / / (11 files): https://github.com/shogun-toolbox/shogun/commit/debe6acb2a9e23597794e07ac8590a74c18cc29d | 20:47 |
| shogun-notifier- | shogun: probing sampler added (log-det) | 20:47 |
| shogun-notifier- | shogun: Viktor Gal :develop * 6eba0f7 / / (11 files): https://github.com/shogun-toolbox/shogun/commit/6eba0f75a6bfcbe0da7fa555c59764a567ae4d04 | 20:47 |
| shogun-notifier- | shogun: Merge pull request #1470 from lambday/feature/log_determinant | 20:47 |
| shogun-notifier- | shogun: | 20:47 |
| shogun-notifier- | shogun: probing sampler added (log-det) | 20:47 |
| -!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has joined #shogun | 20:49 | |
| -!- lisitsyn [~lisitsyn@fb2-lo1.global63.net] has joined #shogun | 20:56 | |
| lisitsyn | wow | 20:56 |
| iglesiasg | lisitsyn, what is it? :) | 20:56 |
| lisitsyn | iglesiasg: 17 ppl here | 20:57 |
| iglesiasg | hehe yeah! | 20:57 |
| @wiking | lisitsyn: gsoc ;) | 20:57 |
| iglesiasg | lisitsyn, although we have three soerens and three buildbots :P | 20:57 |
| lisitsyn | wiking: yeah lets see january | 20:57 |
| @wiking | lisitsyn: november will be like dead channel :P | 20:57 |
| lisitsyn | I think january is more dead | 20:58 |
| @wiking | heheh yeah january is crazy empty :d | 20:58 |
| lisitsyn | but probably it is my russian feeling | 20:58 |
| lisitsyn | as we don't work like 10 days in january | 20:58 |
| @wiking | yeah i remember 2012 january i came along :) | 20:58 |
| iglesiasg | this year there will be lot of users after Shogun 3.0 | 20:58 |
| iglesiasg | let's be optimistic guys :P | 20:58 |
| @wiking | iglesiasg: u wish :)) | 20:58 |
| lisitsyn | I remember one year we didn't work 17 days in january | 20:58 |
| @wiking | lisitsyn: yeah ortodox ppl... same in serbia | 20:59 |
| shogun-buildbot | build #60 of osx1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/60 | 20:59 |
| lisitsyn | wiking: first you have new year then christmas | 20:59 |
| lisitsyn | etc | 20:59 |
| lisitsyn | :D | 20:59 |
| shogun-buildbot | build #1557 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1557 blamelist: Viktor Gal <vigsterkr@gmail.com> | 20:59 |
| @sonney2k | lisitsyn, when did you book your flights? | 20:59 |
| lisitsyn | sonney2k: flight?? | 20:59 |
| @sonney2k | lisitsyn, I mean when are your flights | 20:59 |
| @sonney2k | lisitsyn, going by ship? | 20:59 |
| lisitsyn | sonney2k: yes | 20:59 |
| lisitsyn | train actually | 21:00 |
| lisitsyn | sonney2k: well I didn't buy tickets yet | 21:00 |
| lisitsyn | should I? | 21:00 |
| @sonney2k | lisitsyn, ahh ok I thought you did | 21:00 |
| lisitsyn | no no they can reject my visa | 21:00 |
| lisitsyn | then I'll lose it | 21:00 |
| @sonney2k | lisitsyn, when will you get your visa? | 21:00 |
| lisitsyn | sonney2k: just right after my interview | 21:00 |
| @sonney2k | lisitsyn, we are planning to go from friday to friday | 21:00 |
| lisitsyn | morning of 6 sept | 21:00 |
| @sonney2k | and if we sync somehow we can share a car | 21:01 |
| lisitsyn | sonney2k: it seems it would be quite impossible to share a flight though | 21:01 |
| lisitsyn | :D | 21:02 |
| lisitsyn | well I mean I thought it is possible | 21:02 |
| lisitsyn | sonney2k: where do you want to stay the next week? | 21:02 |
| @wiking | ok my commit stats will go down in the next 7 days | 21:02 |
| @wiking | cya guys in a week | 21:03 |
| lisitsyn | wiking: wedding? | 21:03 |
| @wiking | yeah i'm leaving tomorrow morning | 21:03 |
| shogun-buildbot | build #61 of osx1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/61 blamelist: lambday <heavensdevil6909@gmail.com> | 21:03 |
| lisitsyn | wiking: it must be some hardcore stuff :D | 21:03 |
| lisitsyn | like russian wedding | 21:03 |
| iglesiasg | wiking, enjoy :) | 21:03 |
| lisitsyn | we will miss ya | 21:03 |
| @wiking | lisitsyn: heheh no... french wedding | 21:03 |
| @wiking | lisitsyn: imho it's going to be boring | 21:03 |
| lisitsyn | wiking: just get drunk and get the party started | 21:04 |
| lisitsyn | table dance etc | 21:04 |
| lisitsyn | wiking: in russia it is not ok to not have a fight in wedding programme | 21:04 |
| lisitsyn | so please transfer the experience | 21:04 |
| shogun-buildbot | build #1558 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1558 blamelist: lambday <heavensdevil6909@gmail.com> | 21:05 |
| lisitsyn | sonney2k: okay I am ok with friday to friday - but it will depend when I'll get appropriate flight | 21:06 |
| @wiking | lisitsyn: heheh yeah | 21:06 |
| @wiking | lisitsyn: but u know... french ppl | 21:06 |
| @wiking | they r too civilized :P | 21:06 |
| lisitsyn | wiking: no I don't know french people at all | 21:06 |
| lisitsyn | :D | 21:06 |
| @wiking | but none the less i'll get drunk | 21:06 |
| lisitsyn | I know german people only | 21:06 |
| lisitsyn | a bit | 21:06 |
| lisitsyn | :) | 21:06 |
| lisitsyn | wiking: iglesiasg: I'd say I know serbian and spanish too but you guys are way too mobile | 21:08 |
| lisitsyn | especially you wiking, you have been to all the countries in the world I guess | 21:08 |
| lisitsyn | ;) | 21:08 |
| -!- gsomix_ [~gsomix@178.45.35.45] has joined #shogun | 21:08 | |
| iglesiasg | lisitsyn, hehe I am maybe 70% typical Spanish :) | 21:08 |
| lisitsyn | iglesiasg: how do ya measure that :D | 21:09 |
| iglesiasg | lisitsyn, LMNN | 21:09 |
| @wiking | lisitsyn: nooo there's a lot of countries still that i haven't visited :) | 21:10 |
| lisitsyn | wiking: zambia? angola? | 21:11 |
| lisitsyn | :) | 21:11 |
| lisitsyn | iglesiasg: good! | 21:11 |
| -!- gsomix [~gsomix@178.45.40.223] has quit [Ping timeout: 245 seconds] | 21:12 | |
| @wiking | lisitsyn: yeah haven't been to africa ever | 21:13 |
| @wiking | lisitsyn: neither south-america | 21:13 |
| @wiking | and just a few places in asia | 21:13 |
| @wiking | sooo there's still a lot to see :) | 21:13 |
| @wiking | anyways i'm off | 21:15 |
| @wiking | cyaz | 21:15 |
| lisitsyn | see you | 21:16 |
| shogun-buildbot | build #1342 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1342 blamelist: Viktor Gal <viktor.gal@maeth.com> | 21:17 |
| -!- lisitsyn [~lisitsyn@fb2-lo1.global63.net] has quit [Ping timeout: 264 seconds] | 21:31 | |
| shogun-buildbot | build #1021 of rpm1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/1021 blamelist: Viktor Gal <vigsterkr@gmail.com> | 21:32 |
| -!- gsomix__ [~gsomix@178.45.41.63] has joined #shogun | 21:32 | |
| -!- gsomix_ [~gsomix@178.45.35.45] has quit [Ping timeout: 268 seconds] | 21:35 | |
| -!- hushell [~hushell@c-98-232-178-161.hsd1.or.comcast.net] has quit [Ping timeout: 256 seconds] | 21:45 | |
| shogun-buildbot | build #1643 of deb3 - modular_interfaces is complete: Failure [failed install test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1643 blamelist: lambday <heavensdevil6909@gmail.com> | 21:45 |
| -!- lambday [67157d4c@gateway/web/freenode/ip.103.21.125.76] has quit [] | 21:52 | |
| shogun-buildbot | build #1343 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1343 blamelist: lambday <heavensdevil6909@gmail.com>, Viktor Gal <vigsterkr@gmail.com> | 22:01 |
| -!- hoijui [~hoijui@wlan-141-23-94-144.tubit.tu-berlin.de] has quit [Ping timeout: 256 seconds] | 22:06 | |
| -!- lisitsyn [~lisitsyn@fb2-lo1.global63.net] has joined #shogun | 22:07 | |
| -!- pickle27 [~kevin@d67-193-243-174.home3.cgocable.net] has joined #shogun | 22:19 | |
| shogun-buildbot | build #1022 of rpm1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/1022 | 22:20 |
| -!- hushell [~hushell@8-92.ptpg.oregonstate.edu] has joined #shogun | 22:21 | |
| -!- pickle27 [~kevin@d67-193-243-174.home3.cgocable.net] has left #shogun [] | 22:26 | |
| shogun-buildbot | build #1641 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1641 blamelist: Viktor Gal <viktor.gal@maeth.com> | 22:38 |
| -!- pickle27 [~kevin@d67-193-243-174.home3.cgocable.net] has joined #shogun | 23:07 | |
| -!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 23:47 | |
| -!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has quit [Quit: WeeChat 0.4.0] | 23:47 | |
| --- Log closed Thu Aug 29 00:00:35 2013 | ||
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!