--- Log opened Fri Jul 26 00:00:40 2013 | ||
--- Day changed Fri Jul 26 2013 | ||
@sonney2k | gsomix, rule #1 fix issues ASAP | 00:00 |
---|---|---|
gsomix | sonney2k, ASAP? | 00:00 |
gsomix | ok | 00:00 |
gsomix | one | 00:01 |
gsomix | two | 00:01 |
gsomix | sonney2k, https://github.com/shogun-toolbox/shogun/pull/1313 | 00:01 |
gsomix | ta-dam :) | 00:01 |
shogun-notifier- | shogun: Evgeniy Andreev :develop * 31f4073 / src/shogun/lib/CircularBuffer.cpp: https://github.com/shogun-toolbox/shogun/commit/31f407358a4e7714e35b4ae256f0a0ab353cb79a | 00:01 |
shogun-notifier- | shogun: fast (slow) fix of unused variable warning | 00:01 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * ee96957 / src/shogun/lib/CircularBuffer.cpp: https://github.com/shogun-toolbox/shogun/commit/ee9695791b0c221f02571c40ab31d79ea4721da6 | 00:01 |
shogun-notifier- | shogun: Merge pull request #1313 from gsomix/feature/parser | 00:01 |
shogun-notifier- | shogun: | 00:01 |
shogun-notifier- | shogun: fix of unused variable warning | 00:01 |
gsomix | not critical bug, huh | 00:01 |
@iglesiasg | sonney2k: building python, it should be ready pretty soon... | 00:01 |
@sonney2k | iglesiasg, what is your machine btw? i5? | 00:02 |
@iglesiasg | i7 | 00:02 |
@sonney2k | iglesiasg, you said sth like a day of battery life? | 00:02 |
@sonney2k | i7 !! | 00:02 |
@sonney2k | wooaaah! | 00:02 |
@sonney2k | iglesiasg, how much €€€€? | 00:02 |
@iglesiasg | sonney2k: not a whole day, but pretty much all the night while working and keyboard light on | 00:02 |
@iglesiasg | sonney2k: it was in $$$$ actually :P | 00:03 |
@iglesiasg | bought it from amazon US | 00:03 |
@sonney2k | iglesiasg, how much and how many hours? | 00:03 |
@iglesiasg | sonney2k: check around here http://www.amazon.com/Lenovo-Thinkpad-Carbon-Ultrabook-Professional/dp/B00AGPWGF6/ref=pd_sim_sbs_pc_6 | 00:04 |
@iglesiasg | sonney2k: I worked about 6 hours that night I think | 00:04 |
@iglesiasg | sonney2k: the battery was not over after that though, but my bus arrived to destination :) | 00:05 |
@iglesiasg | battery was something around 10% IIRC | 00:05 |
@sonney2k | iglesiasg, I cannot work more than 1 hour on this one now | 00:05 |
@iglesiasg | but that was when I was using arch linux | 00:05 |
@iglesiasg | I think that this linux mint I put to try out is wasting my battery really badly | 00:05 |
gsomix | iglesiasg, ultracool | 00:05 |
@sonney2k | iglesiasg, where did you ship this to then? Imean you have to pay taxes on top right? | 00:06 |
@iglesiasg | that one is better than mine actually, my was older version with 4GB | 00:06 |
@iglesiasg | sonney2k: yes, I got to know that the hard way.... shipped to Sweden | 00:06 |
@sonney2k | iglesiasg, I still have a slow and hot core 2 duo. | 00:08 |
@iglesiasg | c'mon you should definitely update! | 00:09 |
@iglesiasg | sonney2k: aah but you are apple guy, I understand why not updating :P | 00:09 |
@sonney2k | iglesiasg, I am not really an apple guy. I run linux on all macs I ever owned (starting with titanium...) | 00:10 |
@iglesiasg | sonney2k: if you are actually using a laptop as a laptop is to be used (moving it a lot around) -- I strongly suggest you to get an ultra book | 00:10 |
@sonney2k | iglesiasg, it is fast enough still | 00:11 |
@iglesiasg | sonney2k: when I checked it there were three that seemed powerful enough as to do important job with them | 00:11 |
@sonney2k | iglesiasg, well I use it mainly to compile shogun :) | 00:11 |
@iglesiasg | this lenovo, mac book air, and one asus | 00:11 |
@sonney2k | so I like cores and speed | 00:11 |
@iglesiasg | but lot of people complained about the asus | 00:11 |
@sonney2k | don't care so much about weight | 00:11 |
@iglesiasg | aham, maybe an ultrabook is not the best thing then | 00:12 |
@sonney2k | iglesiasg, ugh! | 00:12 |
@sonney2k | iglesiasg, look at this http://www.shogun-toolbox.org/doc/en/latest/ | 00:12 |
@sonney2k | http://www.shogun-toolbox.org/doc/en/current/ | 00:12 |
@sonney2k | examples are no longer there | 00:12 |
@iglesiasg | what happened with examples? | 00:12 |
@sonney2k | IDK | 00:13 |
@sonney2k | iglesiasg, what is the build status for the fix? | 00:13 |
@iglesiasg | sonney2k: we will have to wait long for travis to build it | 00:13 |
@iglesiasg | sonney2k: here it builded fine, I am checking examples at the moment | 00:13 |
@iglesiasg | didn't do unit tests (nothing is related to the change I believe) | 00:14 |
@iglesiasg | I am doing unit tests in any case | 00:14 |
@sonney2k | iglesiasg, just run the two failing examples - if it works now then we are fine | 00:15 |
-!- travis-ci [~travis-ci@ec2-54-224-45-29.compute-1.amazonaws.com] has joined #shogun | 00:16 | |
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/9480819 | 00:16 |
-!- travis-ci [~travis-ci@ec2-54-224-45-29.compute-1.amazonaws.com] has left #shogun [] | 00:16 | |
@iglesiasg | sonney2k: they run fine here before too though | 00:16 |
@sonney2k | iglesiasg, did you valgrind? | 00:16 |
@sonney2k | hard to believe | 00:16 |
gsomix | http://www.youtube.com/watch?v=O05mMuDvnLc | 00:17 |
@iglesiasg | sonney2k: I didn't valgrind them, I tend to avoid to do it with python :) | 00:17 |
@iglesiasg | sonney2k: I am lazy to get this suppressions files, but I probably should | 00:17 |
@iglesiasg | sonney2k: yeah I know but believe me, I was yesterday talking with Sebastien (maybe you recognize his name from the mailing list) | 00:18 |
@sonney2k | iglesiasg, ahh 0e440ca936a517acf62ce8d333bdbfaf76a66bc2 is breaking the doc | 00:18 |
@iglesiasg | sonney2k: he got this crash and I thought it was because of his machine since here everything worked fine. He was using a VM, so he just created a new one, compiled in the same history revision as in the other machine and the example worked fine that time | 00:18 |
@iglesiasg | I thought something was wrong with the configuration of hist 1st VM, so I was surprised when you showed it to me from the buildbot | 00:19 |
@sonney2k | iglesiasg, I don't use supressions anymore with python. I just open the log in an editor and serach for shogun | 00:19 |
@iglesiasg | sonney2k: doing that atm for structure_discrete_hmsvm_bmrm.py | 00:20 |
@iglesiasg | sonney2k: so the commit for the doc examples was wrong? | 00:23 |
@sonney2k | iglesiasg, yeah fixing | 00:28 |
@sonney2k | iglesiasg, man I want one of these https://www.tomtom.com/en_us/products/your-sports/tomtom-multi-sport-gps-watch/dark-grey/index.jsp | 00:28 |
@sonney2k | slightly too late now | 00:28 |
@sonney2k | I didn't know there are watches counting swimming strokes | 00:29 |
@sonney2k | iglesiasg, but I am all the time wondering what time it is when swimming and what distance I've done etc | 00:30 |
-!- wiking_ [~wiking@info2k1.hu] has joined #shogun | 00:31 | |
@iglesiasg | sonney2k: that seems suitable for you then! | 00:32 |
@iglesiasg | didn't know there those existed either | 00:32 |
@iglesiasg | sonney2k: but do you normally swim all year round? | 00:32 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Ping timeout: 248 seconds] | 00:34 | |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Ping timeout: 248 seconds] | 00:34 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 00:35 | |
@sonney2k | iglesiasg, yes but only indoors in winter then of course | 00:37 |
@sonney2k | usually may-september | 00:37 |
@sonney2k | but I don't have that much time and weather has to match | 00:38 |
@sonney2k | (no thunderstorm ;) | 00:38 |
@iglesiasg | hehe | 00:38 |
@iglesiasg | I miss swimming a bit, I used to play water polo at high school | 00:39 |
-!- travis-ci [~travis-ci@ec2-54-225-12-124.compute-1.amazonaws.com] has joined #shogun | 00:39 | |
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/9486615 | 00:39 |
-!- travis-ci [~travis-ci@ec2-54-225-12-124.compute-1.amazonaws.com] has left #shogun [] | 00:39 | |
@sonney2k | iglesiasg, I am not that fast btw but I just enjoy it :) | 00:42 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * ff18c6c / .gitignore,doc/pages/Examples.mainpage: https://github.com/shogun-toolbox/shogun/commit/ff18c6c34d0a7f7994e31fe0b8ecad4ca14c4066 | 00:42 |
shogun-notifier- | shogun: re-add the main Examples page to get things back working | 00:42 |
shogun-buildbot | build #1399 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1399 blamelist: Soeren Sonnenburg <sonne@debian.org> | 00:43 |
@iglesiasg | sonney2k: why is it that sometimes make in tests/unit/ causes recompilation of Shogun? | 00:44 |
@iglesiasg | one needs to sudo make install from src first maybe? | 00:45 |
@sonney2k | iglesiasg, I am pretty much looking forward to this even http://media.tib1848ev.de/upload/pdf-dateien/veranstaltungen/mss/flyer_mss_2013_(klein).pdf | 00:45 |
@sonney2k | iglesiasg, 3.5km across the lake :) | 00:45 |
@sonney2k | iglesiasg, no idea - but IMHO no more need to install things | 00:46 |
@iglesiasg | sonney2k: it looks nice | 00:46 |
@sonney2k | iglesiasg, we now have thoralfs make check-examples-local etc | 00:46 |
@iglesiasg | aah I didn't notice there is one rule for unit tests as well | 00:47 |
@sonney2k | iglesiasg, so whats the status? bug fixed? | 00:47 |
@sonney2k | iglesiasg, I usually do make in the unit tests folder | 00:47 |
@sonney2k | no idea what that does - wiking_'s territory | 00:48 |
@iglesiasg | for me it sometimes forces compilation of everything | 00:48 |
@iglesiasg | but sometimes it doesn't, I believe it depends on whether it was installed but I am not sure | 00:48 |
@sonney2k | iglesiasg, did you git pull inbetween? | 00:49 |
@iglesiasg | sonney2k: mm no, but I might have rebased indeed | 00:50 |
@iglesiasg | mm well not really, I just did git flow feature publish | 00:50 |
@iglesiasg | that does not rebase | 00:50 |
@sonney2k | iglesiasg, no but whenever HEAD changes it will update the version | 00:52 |
@iglesiasg | it can be that then | 00:52 |
@sonney2k | iglesiasg, anyway just push your change now | 00:52 |
@sonney2k | hmmhh I like that watch | 00:53 |
@sonney2k | for this season it comes too late | 00:53 |
@sonney2k | guess I will get one next easter | 00:53 |
@iglesiasg | sonney2k: I already pushed and did PR time ago | 00:53 |
@iglesiasg | https://github.com/shogun-toolbox/shogun/pull/1314 | 00:53 |
@iglesiasg | I was waiting for the check to merge | 00:53 |
shogun-notifier- | shogun: Fernando Iglesias :develop * 7c35471 / src/shogun/ (3 files): https://github.com/shogun-toolbox/shogun/commit/7c35471acf923285397f06a618e8dde26d10bdf9 | 00:54 |
shogun-notifier- | shogun: Minor refactoring in SGMatrixList. | 00:54 |
shogun-notifier- | shogun: Methods don't return refs to matrix elements and a setter is provided | 00:54 |
shogun-notifier- | shogun: instead | 00:54 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 86f58e3 / src/shogun/ (3 files): https://github.com/shogun-toolbox/shogun/commit/86f58e389325ea7ae484f3cf9ec4591f509147e4 | 00:54 |
shogun-notifier- | shogun: Merge pull request #1314 from iglesias/feature/fix_matrixlist | 00:54 |
shogun-notifier- | shogun: | 00:54 |
shogun-notifier- | shogun: Fix crashers occurring with SGMatrixList. | 00:54 |
@sonney2k | iglesiasg, it only fails on the buildbot anyways so lets see how it likes the change | 00:55 |
@iglesiasg | all right | 00:55 |
shogun-notifier- | shogun: Evgeniy Andreev :develop * 656b2a4 / src/shogun/io/Parser.cpp,src/shogun/io/Parser.h,tests/unit/io/Parser_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/656b2a412b5e0773e7bf985e9cb6c60e13d5eb7e | 00:55 |
shogun-notifier- | shogun: added Parser | 00:55 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 325ec68 / src/shogun/io/Parser.cpp,src/shogun/io/Parser.h,tests/unit/io/Parser_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/325ec684459312e6ed92928e14b4f97bfc9b9baa | 00:55 |
shogun-notifier- | shogun: Merge pull request #1315 from gsomix/feature/parser | 00:55 |
shogun-notifier- | shogun: | 00:55 |
shogun-notifier- | shogun: Added Parser class | 00:55 |
shogun-buildbot | build #1081 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1081 blamelist: Soeren Sonnenburg <sonne@debian.org> | 00:55 |
@sonney2k | gsomix, well done! go go ! | 00:56 |
gsomix | sonney2k, z-z-z-z | 00:56 |
gsomix | new CFile interface and preliminary CSVFile will be next | 00:56 |
gsomix | good night guys | 00:57 |
gsomix | 3AM at my clock | 00:57 |
@sonney2k | gsomix, cu tomorrow | 00:59 |
@sonney2k | iglesiasg, nite from my side too! | 00:59 |
shogun-buildbot | build #1569 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1569 blamelist: Evgeniy Andreev <gsomix@gmail.com>, Soeren Sonnenburg <sonne@debian.org>, Thoralf Klein <thoralf.klein@zib.de> | 01:00 |
shogun-buildbot | build #1285 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1285 blamelist: Soeren Sonnenburg <sonne@debian.org> | 01:00 |
@iglesiasg | sonney2k: good night | 01:01 |
@iglesiasg | sonney2k: it is seg faulting here now :( | 01:03 |
-!- travis-ci [~travis-ci@ec2-54-225-12-124.compute-1.amazonaws.com] has joined #shogun | 01:03 | |
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/9486688 | 01:03 |
-!- travis-ci [~travis-ci@ec2-54-225-12-124.compute-1.amazonaws.com] has left #shogun [] | 01:03 | |
shogun-buildbot | build #1571 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1571 blamelist: Soeren Sonnenburg <sonne@debian.org>, Fernando Iglesias <fernando.iglesiasg@gmail.com> | 01:21 |
shogun-buildbot | build #1400 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1400 blamelist: Soeren Sonnenburg <sonne@debian.org> | 01:27 |
shogun-buildbot | build #1572 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1572 blamelist: Evgeniy Andreev <gsomix@gmail.com>, Soeren Sonnenburg <sonne@debian.org> | 01:45 |
-!- travis-ci [~travis-ci@ec2-54-225-12-124.compute-1.amazonaws.com] has joined #shogun | 01:49 | |
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/9492081 | 01:49 |
-!- travis-ci [~travis-ci@ec2-54-225-12-124.compute-1.amazonaws.com] has left #shogun [] | 01:49 | |
shogun-notifier- | shogun: hushell :develop * 5438f02 / / (2 files): https://github.com/shogun-toolbox/shogun/commit/5438f0273b531e2ecb1292ff358989d2feb14c9b | 02:13 |
shogun-notifier- | shogun: FactorGraph_unittest memory leak fixed | 02:13 |
shogun-notifier- | shogun: Fernando Iglesias :develop * 47040a7 / / (2 files): https://github.com/shogun-toolbox/shogun/commit/47040a74ee04e083c47445e16b68f11985079fe3 | 02:13 |
shogun-notifier- | shogun: Merge pull request #1312 from hushell/develop | 02:13 |
shogun-notifier- | shogun: | 02:13 |
shogun-notifier- | shogun: FactorGraph_unittest memory leak fixed | 02:13 |
-!- travis-ci [~travis-ci@ec2-54-224-45-29.compute-1.amazonaws.com] has joined #shogun | 02:13 | |
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/9493452 | 02:13 |
-!- travis-ci [~travis-ci@ec2-54-224-45-29.compute-1.amazonaws.com] has left #shogun [] | 02:13 | |
shogun-buildbot | build #1082 of cyg1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1082 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 02:20 |
shogun-buildbot | build #1286 of bsd1 - libshogun is complete: Failure [failed test test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1286 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 02:33 |
-!- hushell [~hushell@8-92.ptpg.oregonstate.edu] has quit [Ping timeout: 276 seconds] | 02:34 | |
shogun-buildbot | build #1573 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1573 blamelist: hushell <hushell@hushell-U510.(none)> | 02:39 |
-!- travis-ci [~travis-ci@ec2-54-224-45-29.compute-1.amazonaws.com] has joined #shogun | 02:39 | |
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/9493685 | 02:39 |
-!- travis-ci [~travis-ci@ec2-54-224-45-29.compute-1.amazonaws.com] has left #shogun [] | 02:39 | |
-!- pickle27 [~Kevin@d67-193-243-174.home3.cgocable.net] has joined #shogun | 02:51 | |
-!- hushell [~hushell@c-24-21-169-136.hsd1.or.comcast.net] has joined #shogun | 02:51 | |
-!- travis-ci [~travis-ci@ec2-54-224-45-29.compute-1.amazonaws.com] has joined #shogun | 02:58 | |
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/9493735 | 02:58 |
-!- travis-ci [~travis-ci@ec2-54-224-45-29.compute-1.amazonaws.com] has left #shogun [] | 02:58 | |
shogun-buildbot | build #1401 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1401 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 03:09 |
shogun-buildbot | build #404 of nightly_all is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_all/builds/404 | 03:16 |
shogun-buildbot | build #412 of nightly_none is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_none/builds/412 | 03:22 |
-!- nube [~rho@49.244.6.28] has joined #shogun | 03:27 | |
pickle27 | sonney2k: here? | 03:32 |
pickle27 | iglesiasg: ? | 03:35 |
@iglesiasg | pickle27: yep | 03:38 |
pickle27 | had a question about how to use the static interfaces | 03:38 |
pickle27 | I want to write a octave/matlab static example for jade | 03:39 |
pickle27 | its a converter and all the examples are with classifiers | 03:39 |
@iglesiasg | pickle27: you need to write more code in GUI in order to make it available for static interfaces | 03:39 |
@iglesiasg | pickle27: they are not like modular interfaces, where SWIG generate wrappers | 03:40 |
pickle27 | ahh okay I'll poke around in shogun/ui and see if I can understand it | 03:40 |
@iglesiasg | pickle27: good luck! | 03:41 |
@iglesiasg | if you get to do it finally I will be happy to learn from it :) | 03:41 |
@iglesiasg | I never really did anything in static interfaces | 03:41 |
pickle27 | yeah I just want to have everything covered for matlab and R so that if people are doing ICA reserach they use us | 03:41 |
@iglesiasg | pickle27: very nice idea | 03:43 |
@iglesiasg | pickle27: BTW, I came across with this while compiling: http://harmful.cat-v.org/software/c++/I_did_it_for_you_all | 03:46 |
@iglesiasg | I had never read that before, hardcore stuff | 03:46 |
pickle27 | I'll give it a read | 03:46 |
-!- travis-ci [~travis-ci@ec2-54-224-45-29.compute-1.amazonaws.com] has joined #shogun | 03:46 | |
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/9494090 | 03:46 |
-!- travis-ci [~travis-ci@ec2-54-224-45-29.compute-1.amazonaws.com] has left #shogun [] | 03:46 | |
pickle27 | iglesiasg: I have something neat to share too, the open source report card based off your github activity | 03:47 |
pickle27 | check it out | 03:47 |
pickle27 | here is min | 03:47 |
pickle27 | http://osrc.dfm.io/pickle27 | 03:47 |
pickle27 | its kind of neat, you enter any github users name and see what it comes up with | 03:47 |
@iglesiasg | pickle27: hehe yeah, we were playing around with that some time ago | 03:48 |
@iglesiasg | pickle27: oh! we are competing in here, 36% and 37% :P | 03:48 |
pickle27 | iglesiasg: haha | 03:48 |
@iglesiasg | Sergey is da fucking master | 03:49 |
pickle27 | *quickly pushes a large amount of c++ nonsense* | 03:49 |
@iglesiasg | top 3%, he is like famous | 03:49 |
pickle27 | wow | 03:49 |
pickle27 | actually though | 03:49 |
pickle27 | he should be famous | 03:49 |
pickle27 | they should release who codes the most each month or something like for all of github | 03:49 |
@iglesiasg | pickle27: there are some rankings like that | 03:50 |
@iglesiasg | but we have nothing to do in there - javascript people rule those rankings | 03:50 |
@iglesiasg | I saw one top 250 list I think | 03:51 |
@iglesiasg | NONE was C++ | 03:51 |
pickle27 | yeah true | 03:52 |
pickle27 | maybe by language or something | 03:52 |
pickle27 | iglesiasg: interesting read | 04:05 |
@iglesiasg | pickle27: I was sort of O_O | 04:06 |
pickle27 | yeah, im still not sure if he is trolling or what | 04:06 |
pickle27 | I still like c++ lol | 04:06 |
@iglesiasg | hehe | 04:06 |
@iglesiasg | we should do go | 04:07 |
pickle27 | actually yeah | 04:07 |
pickle27 | we should get a wrapper for it | 04:07 |
pickle27 | by get I mean make | 04:07 |
@iglesiasg | I think SWIG already supports go | 04:07 |
pickle27 | although on the topic of c++ it does kind of depend how you write it, some peoples style makes it really awful | 04:08 |
@iglesiasg | http://www.swig.org/Doc2.0/Go.html | 04:08 |
@iglesiasg | pickle27: yeah, I agree | 04:08 |
@iglesiasg | pickle27: and the compile errors | 04:08 |
@iglesiasg | sometimes they are crazy | 04:08 |
pickle27 | yeah whenever templates are involved | 04:08 |
@iglesiasg | let me show you this quote | 04:08 |
@iglesiasg | C++ is probably the only language where the error [message] can be longer than the program — aiju | 04:08 |
pickle27 | I would seriously prefer a simple one line output saying template mismatch line xx | 04:09 |
@iglesiasg | aha and this one | 04:09 |
@iglesiasg | C++ is to C as Lung Cancer is to Lung. | 04:09 |
@iglesiasg | lol | 04:09 |
@iglesiasg | more: http://harmful.cat-v.org/software/c++/ | 04:09 |
pickle27 | iglesiasg: yeah looks like there is a math lib "gomatrix" | 04:11 |
pickle27 | we should totally get on this and be the first big ml toolkit for go | 04:11 |
@iglesiasg | pickle27: nothing out there already? | 04:12 |
pickle27 | iglesiasg: I haven't done a thorough search but nothing the size of shogun | 04:12 |
pickle27 | iglesiasg: I might see if any of the other guys are interested, if someone could point me in the right direction ish I may be interested in starting the wrapper | 04:18 |
@iglesiasg | pickle27: sonney2k and gsomix are the swig experts | 04:19 |
pickle27 | yeah I'll ping them tomorrow | 04:19 |
pickle27 | I also need their help figuring why my r_modular won | 04:19 |
pickle27 | t build | 04:19 |
shogun-buildbot | build #469 of nightly_default is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/469 | 04:20 |
shogun-notifier- | shogun: Fernando Iglesias :develop * fc1c7f1 / src/shogun/ (5 files): https://github.com/shogun-toolbox/shogun/commit/fc1c7f19e8cbda9f3d4ace2e937acb5bbb56915d | 04:25 |
shogun-notifier- | shogun: Fix memory bugs related to the SO framework | 04:25 |
shogun-notifier- | shogun: Fernando Iglesias :develop * 857008d / examples/undocumented/libshogun/ (2 files): https://github.com/shogun-toolbox/shogun/commit/857008d429b6b86f540ba68416a11686f81ae5a3 | 04:25 |
shogun-notifier- | shogun: Add a couple of minimal libshogun examples to study memory bugs | 04:25 |
shogun-notifier- | shogun: Fernando Iglesias :develop * 8d8a3ea / / (7 files): https://github.com/shogun-toolbox/shogun/commit/8d8a3eae3cb3cc7c9488efca6556f9faae0bcc06 | 04:25 |
shogun-notifier- | shogun: Merge pull request #1316 from iglesias/feature/fix_matrixlist | 04:25 |
shogun-notifier- | shogun: | 04:25 |
shogun-notifier- | shogun: Feature/fix matrixlist | 04:25 |
@iglesiasg | all right, I am off to sleep now | 04:25 |
@iglesiasg | finally | 04:25 |
-!- iglesiasg [~Fernando@s83-179-44-135.cust.tele2.se] has quit [Quit: Leaving] | 04:26 | |
shogun-buildbot | build #1083 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1083 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 04:37 |
shogun-buildbot | build #1287 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1287 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 04:43 |
shogun-buildbot | build #1576 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1576 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 04:49 |
-!- travis-ci [~travis-ci@ec2-54-225-12-124.compute-1.amazonaws.com] has joined #shogun | 04:52 | |
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/9495268 | 04:52 |
-!- travis-ci [~travis-ci@ec2-54-225-12-124.compute-1.amazonaws.com] has left #shogun [] | 04:52 | |
-!- hushell [~hushell@c-24-21-169-136.hsd1.or.comcast.net] has quit [Read error: Connection reset by peer] | 05:07 | |
-!- pickle27 [~Kevin@d67-193-243-174.home3.cgocable.net] has quit [Quit: Leaving] | 05:08 | |
-!- hushell [~hushell@c-24-21-169-136.hsd1.or.comcast.net] has joined #shogun | 05:13 | |
-!- hushell [~hushell@c-24-21-169-136.hsd1.or.comcast.net] has quit [Read error: Connection reset by peer] | 05:19 | |
shogun-buildbot | build #1402 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1402 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 05:19 |
-!- travis-ci [~travis-ci@ec2-54-225-12-124.compute-1.amazonaws.com] has joined #shogun | 05:34 | |
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/9495568 | 05:34 |
-!- travis-ci [~travis-ci@ec2-54-225-12-124.compute-1.amazonaws.com] has left #shogun [] | 05:34 | |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun | 05:46 | |
-!- travis-ci [~travis-ci@ec2-54-225-12-124.compute-1.amazonaws.com] has joined #shogun | 05:53 | |
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/9495602 | 05:53 |
-!- travis-ci [~travis-ci@ec2-54-225-12-124.compute-1.amazonaws.com] has left #shogun [] | 05:53 | |
-!- nube [~rho@49.244.6.28] has quit [Quit: Leaving.] | 06:12 | |
-!- nube [~rho@49.244.6.28] has joined #shogun | 06:12 | |
-!- travis-ci [~travis-ci@ec2-54-224-45-29.compute-1.amazonaws.com] has joined #shogun | 06:16 | |
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/9497342 | 06:16 |
-!- travis-ci [~travis-ci@ec2-54-224-45-29.compute-1.amazonaws.com] has left #shogun [] | 06:16 | |
-!- travis-ci [~travis-ci@ec2-54-224-45-29.compute-1.amazonaws.com] has joined #shogun | 06:57 | |
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/9501042 | 06:57 |
-!- travis-ci [~travis-ci@ec2-54-224-45-29.compute-1.amazonaws.com] has left #shogun [] | 06:57 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 07:25 | |
-!- lisitsyn [~lisitsyn@213.87.140.158] has joined #shogun | 07:48 | |
-!- hushell [~hushell@c-24-21-169-136.hsd1.or.comcast.net] has joined #shogun | 08:04 | |
-!- lisitsyn [~lisitsyn@213.87.140.158] has quit [Read error: Connection reset by peer] | 08:16 | |
gsomix | good morning | 08:25 |
-!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has joined #shogun | 09:08 | |
-!- nube [~rho@49.244.6.28] has quit [Ping timeout: 264 seconds] | 09:44 | |
-!- nube [~rho@49.244.84.177] has joined #shogun | 10:00 | |
gsomix | it's rainy today. | 10:25 |
* gsomix *sigh* | 10:25 | |
-!- van51 [~van51@ppp-94-66-56-238.home.otenet.gr] has joined #shogun | 10:26 | |
van51 | yo yo | 10:26 |
gsomix | van51, hi | 10:26 |
gsomix | van51, how are you? | 10:26 |
van51 | gsomix: fine | 10:27 |
van51 | gsomix: just arrived at my vacation point :) | 10:27 |
van51 | gsomix: how about you? | 10:27 |
gsomix | van51, vacation *__* | 10:27 |
gsomix | van51, so, where are you now? | 10:27 |
gsomix | van51, fine. sitting at home - rain, thunderstorm | 10:28 |
van51 | gsomix: in a little island in Greece, called Aegina | 10:28 |
van51 | here it's the exact opposite | 10:28 |
gsomix | van51, cool | 10:28 |
van51 | gsomix: hot actually :p | 10:28 |
* gsomix *sigh* | 10:28 | |
gsomix | van51, nice island. tnx to wikipedia. :) | 10:34 |
van51 | gsomix: hehe | 10:34 |
van51 | gsomix: yeah it's cute :) | 10:34 |
-!- iglesiasg [~Fernando@s83-179-44-135.cust.tele2.se] has joined #shogun | 11:57 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 11:57 | |
@sonney2k | van51, so did your 50m swim already :P | 12:17 |
@sonney2k | iglesiasg, woah! what is segfaulting? | 12:18 |
@iglesiasg | sonney2k: are you saying it because of my comment in the PR_ | 12:19 |
@iglesiasg | or did you see sth in the buildbot? | 12:19 |
@sonney2k | iglesiasg, because of what you said after I went to bed | 12:19 |
@iglesiasg | sonney2k: oh yeah, it was seg faulting but I fixed that | 12:20 |
@iglesiasg | sonney2k: now in my machine valgrind complains about one of the examples, the other one is fine | 12:20 |
@iglesiasg | sonney2k: https://github.com/shogun-toolbox/shogun/pull/1316 | 12:20 |
@sonney2k | iglesiasg, what leaks or has issues? | 12:25 |
@iglesiasg | sonney2k: both | 12:25 |
@sonney2k | iglesiasg, so you added 2 examples for libshogun? | 12:25 |
@sonney2k | sonney2k, and what? | 12:25 |
@iglesiasg | I added these two examples to get more meaningful valgrind messages than with python examples | 12:26 |
@sonney2k | iglesiasg, yeah true much easier to debug | 12:35 |
@sonney2k | iglesiasg, but did you see where which objects did leak then? | 12:36 |
@iglesiasg | sonney2k: I am going to try see if I can fix it | 12:36 |
@iglesiasg | sonney2k: let me show you valgrind's trace, but it is sort of a $%&$% | 12:36 |
@sonney2k | iglesiasg, give it to me | 12:38 |
@sonney2k | van51, if you have liblinear results I would be very interested | 12:38 |
@iglesiasg | sonney2k: http://pastebin.com/Wz1L5yji | 12:39 |
@sonney2k | van51, and I got one more idea how to properly test stuff | 12:39 |
@iglesiasg | sonney2k: after that it seems to halt for a long while | 12:39 |
@sonney2k | iglesiasg, interesting the matrix seems to be too smal | 12:40 |
@sonney2k | index out of bound access | 12:40 |
@iglesiasg | sonney2k: and a infinite loop too seems to be | 12:40 |
@iglesiasg | ==21808== More than 10000000 total errors detected. I'm not reporting any more. | 12:41 |
@iglesiasg | ==21808== Final error counts will be inaccurate. Go fix your program! | 12:41 |
@iglesiasg | hehe | 12:41 |
@sonney2k | iglesiasg, well deserved :D | 12:41 |
@sonney2k | iglesiasg, we don't have tests for the structured output stuff do we? | 12:42 |
@iglesiasg | sonney2k: no | 12:43 |
@iglesiasg | sonney2k: we should do some | 12:43 |
@iglesiasg | sonney2k: I have a question regarding valgrind's output | 12:43 |
@iglesiasg | sonney2k: what does it come ta say Address 0x9ad08c8 is 8 bytes before a block of size 64 alloc'd? | 12:44 |
@sonney2k | sonney2k, that you do access a double array likely with one index below the allowed | 12:46 |
@iglesiasg | sonney2k: I infected you with the virus of talking to oneself :P | 12:49 |
@sonney2k | iglesiasg, I have no idea why that happens :P | 12:50 |
@sonney2k | let me better sleep | 12:50 |
@iglesiasg | early siesta | 12:50 |
-!- HeikoS [~heiko@nat-160-251.internal.eduroam.ucl.ac.uk] has joined #shogun | 12:51 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 12:51 | |
@HeikoS | votjakovr: hi! | 13:28 |
@HeikoS | votjakovr: just to let you know, I will be gone over the weekend (no internet) until Monday evening | 13:28 |
@HeikoS | votjakovr: so how is the state? | 13:28 |
votjakovr | HeikoS: hi! ok, i'm testing logit and meet a problem with integration on (-inf, inf) :( | 13:41 |
@HeikoS | what is the problem? | 13:41 |
@HeikoS | sonney2k, iglesiasg I will be gone over the weekend until Monday night, just to let you know | 13:45 |
@iglesiasg | HeikoS: all right, have fun! :) | 13:45 |
@HeikoS | iglesiasg: thanks! :) | 13:46 |
votjakovr | HeikoS: big error, if function have a big peak in one place and in other places it's close to 0 | 13:46 |
@HeikoS | votjakovr: where does this come from? is this an algorithm problem or does it come from a bug in the implementation? | 13:47 |
votjakovr | HeikoS: it's algorithm problem, but i think we can integrate sigmoid(x)*N(mu, sigma^2) from, say, mu-10*sigma^2 to mu+10*sigma^2 | 13:49 |
@HeikoS | votjakovr: the GPML toolbox uses this too? | 13:50 |
@HeikoS | votjakovr: since I remember they paid some extra care when integrating in extreme parts | 13:50 |
@HeikoS | votjakovr: well this thing is almost Gaussian | 13:50 |
@HeikoS | votjakovr: so that might be a good hack, yes | 13:51 |
@HeikoS | votjakovr: how does GPML does this? | 13:51 |
@HeikoS | votjakovr: if this continues being ugly, we can just use their method | 13:51 |
votjakovr | HeikoS: oh, i've checked GPstuff and they integrate from mu-6*sigma, mu+6*sigma :) | 13:53 |
@HeikoS | votjakovr: haha | 13:53 |
@HeikoS | so, problem solved! | 13:53 |
@HeikoS | votjakovr: just do that, infinities tend to mess up floating point arithmetic anyways | 13:54 |
-!- nube [~rho@49.244.84.177] has quit [Ping timeout: 256 seconds] | 13:54 | |
votjakovr | HeikoS: nope, gpml uses approximation to error function | 13:54 |
@HeikoS | votjakovr: do you think the logit will be done until end of weekend? | 13:54 |
@HeikoS | votjakovr: I didnt get that, what? | 13:55 |
votjakovr | HeikoS: they calculate this integral via a cumulative Gaussian scale mixture approximation | 13:56 |
@HeikoS | votjakovr: ah yeah I remember | 13:56 |
@HeikoS | votjakovr: why not use that? | 13:57 |
@HeikoS | votjakovr: or do you prefer yours | 13:57 |
@HeikoS | votjakovr: question is which one is more accurate and which one is faster | 13:57 |
@HeikoS | votjakovr: in particular on non trivial likelihoods, for sigmoid, things should be similar right? | 13:57 |
votjakovr | HeikoS: yeah, gpml is faster a little bit, but i think less accurate | 13:58 |
votjakovr | HeikoS: in this case | 13:58 |
@HeikoS | votjakovr: so your suggestion? | 13:59 |
-!- FSCV [~FSCV@187.210.54.166] has joined #shogun | 13:59 | |
votjakovr | HeikoS: but we'll use numerical integration for Student's likelihood for example, and i think, that it's better to use it and in logit case | 14:00 |
@HeikoS | votjakovr: ok, agreed, just make sure the results are close then | 14:01 |
@HeikoS | votjakovr: whats next once the integration works? | 14:01 |
votjakovr | HeikoS: i'd like to ask another question first :) | 14:02 |
@HeikoS | votjakovr: shoot, I gotta go soon, so all questions now please :) | 14:03 |
votjakovr | HeikoS: i'd like to have something like vector<pair<float64_t, float64_t> >, that's because i think, that creating new Nx2 SGMatrix in a loop is not a good idea | 14:06 |
@HeikoS | votjakovr: no way to preallocate this matrix ? | 14:06 |
votjakovr | HeikoS: this is for subintervals | 14:07 |
@HeikoS | votjakovr: what about dynamic re-allocation? | 14:07 |
@HeikoS | votjakovr: DynArray implements this | 14:07 |
@HeikoS | and I think one can see this as a 2d array | 14:07 |
votjakovr | HeikoS: yeah, good, i think so :) But the question: what the difference between DynArray and CDynamicArray? | 14:10 |
@HeikoS | DynArray is more basic | 14:11 |
@HeikoS | its a low-level implementation and not a Shogun object | 14:11 |
@HeikoS | so use DynamicArray<float> | 14:11 |
@HeikoS | this can be interpreted as a 2d array and dynamically grows, and uses DynArray internally | 14:11 |
votjakovr | HeikoS: yep i see that CDynamicArray have DynArray inside, but don't understand, what is dim1, dim2, dim3... | 14:14 |
@HeikoS | votjakovr: I think the array can be interpreted in multiple dimensions | 14:14 |
@HeikoS | votjakovr: maybe grep the Shogun code a bit to find out about this | 14:14 |
@HeikoS | or ask sonney2k, he knows these internals very well | 14:14 |
@HeikoS | I never used DynamicArray | 14:14 |
@HeikoS | votjakovr: there are all these setters/getters for elements in different dims | 14:15 |
votjakovr | HeikoS: yep, i see, but i need to use push_back() | 14:16 |
votjakovr | HeikoS: but dim{1,2,3} doesn't change there | 14:17 |
@HeikoS | votjakovr: internally its a 1d array | 14:20 |
@HeikoS | votjakovr: those indices are just mappings | 14:20 |
@HeikoS | votjakovr: so you push back pairs and then ahve to figure out their dim1, dim2 indices | 14:21 |
votjakovr | HeikoS: yep i understand | 14:21 |
@HeikoS | votjakovr: I gotta go now, maybe ask sonney2k if things are unclear | 14:21 |
votjakovr | HeikoS: ok, see you ;) | 14:21 |
@HeikoS | votjakovr: pls try to send not too much time on the integration, need to continue with big picture things | 14:22 |
@HeikoS | votjakovr: but good that we have a good integratior now :) | 14:22 |
@HeikoS | votjakovr: take care! | 14:22 |
-!- HeikoS [~heiko@nat-160-251.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.] | 14:25 | |
-!- iglesiasg [~Fernando@s83-179-44-135.cust.tele2.se] has quit [Quit: Leaving] | 15:11 | |
van51 | sonney2k: hello | 15:29 |
van51 | sonney2k: sorry I was out floating :) | 15:29 |
van51 | sonney2k: here are some results, but with c=10, https://dl.dropboxusercontent.com/u/23851310/liblinear_results.txt | 15:34 |
van51 | sonney2k: because I had left it all night and it hadn't finished the first run with C=1 :/ | 15:34 |
van51 | sonney2k: although from the results now I see that it took a lot time for num_bits = 8 | 15:34 |
van51 | sonney2k: so I'm trying again w/o it | 15:34 |
van51 | sonney2k: also, does svmocas use some parallelization strategy that may influence the running time? | 15:36 |
van51 | sonney2k: also I'm looking forward to your other testing suggestion | 15:38 |
-!- FSCV [~FSCV@187.210.54.166] has quit [Quit: Leaving] | 16:28 | |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 16:55 | |
shogun-notifier- | shogun: Fernando Iglesias :develop * 8813ca6 / / (2 files): https://github.com/shogun-toolbox/shogun/commit/8813ca63114b43b35e87866f9d9936857c34ceeb | 16:55 |
shogun-notifier- | shogun: Fix discrete HM-SVM example | 16:55 |
shogun-notifier- | shogun: Fernando Iglesias :develop * 9361fdd / / (2 files): https://github.com/shogun-toolbox/shogun/commit/9361fdd237329cdc94573b0bf79e54dffdf2f0ea | 16:55 |
shogun-notifier- | shogun: Merge pull request #1317 from iglesias/feature/fix_matrixlist | 16:55 |
shogun-notifier- | shogun: | 16:55 |
shogun-notifier- | shogun: Fix discrete libshogun HM-SVM example | 16:55 |
shogun-buildbot | build #1084 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1084 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 17:06 |
-!- gsomix_ [~gsomix@109.169.151.109] has joined #shogun | 17:06 | |
-!- pickle27 [~Kevin@d67-193-243-174.home3.cgocable.net] has joined #shogun | 17:09 | |
shogun-buildbot | build #1288 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1288 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 17:09 |
pickle27 | sonney2k: hey | 17:10 |
-!- gsomix [~gsomix@95.67.134.30] has quit [Ping timeout: 240 seconds] | 17:10 | |
shogun-buildbot | build #1577 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1577 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 17:20 |
-!- nube [~rho@49.244.84.177] has joined #shogun | 17:21 | |
-!- travis-ci [~travis-ci@ec2-54-224-45-29.compute-1.amazonaws.com] has joined #shogun | 17:32 | |
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/9520676 | 17:32 |
-!- travis-ci [~travis-ci@ec2-54-224-45-29.compute-1.amazonaws.com] has left #shogun [] | 17:32 | |
shogun-buildbot | build #1403 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1403 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 17:38 |
-!- van51 [~van51@ppp-94-66-56-238.home.otenet.gr] has left #shogun ["QUIT :Leaving."] | 18:31 | |
-!- hushell [~hushell@c-24-21-169-136.hsd1.or.comcast.net] has quit [Ping timeout: 264 seconds] | 18:38 | |
-!- van51 [~van51@ppp-94-66-56-238.home.otenet.gr] has joined #shogun | 18:41 | |
-!- nube [~rho@49.244.84.177] has quit [Ping timeout: 276 seconds] | 18:49 | |
-!- hushell [~hushell@8-92.ptpg.oregonstate.edu] has joined #shogun | 19:31 | |
-!- van51 [~van51@ppp-94-66-56-238.home.otenet.gr] has quit [Quit: Leaving.] | 19:43 | |
-!- van51 [~van51@ppp-94-66-56-238.home.otenet.gr] has joined #shogun | 19:52 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 19:55 | |
van51 | sonney2k: I have to go for a while. Send me an email if you'd like about the testing thing in case I miss you later | 20:40 |
van51 | bye guys | 20:40 |
-!- van51 [~van51@ppp-94-66-56-238.home.otenet.gr] has quit [Quit: Leaving.] | 20:40 | |
-!- lisitsyn [~lisitsyn@92-240-133-94.clients.tlt.100megabit.ru] has joined #shogun | 20:57 | |
@sonney2k | pickle27, ho :) | 21:26 |
pickle27 | sonney2k: so about building the r_modular interface | 21:26 |
pickle27 | sonney2k: this is the beginning of the error message http://pastebin.com/1Kjb5Fne | 21:28 |
pickle27 | looks like it isn't linking | 21:28 |
@sonney2k | pickle27, sry but r_modular is a MAJOR undertaking to get to work | 21:32 |
@sonney2k | pickle27, but r_static should | 21:33 |
pickle27 | oh | 21:33 |
pickle27 | what is required other than the r-base-devel pkg | 21:33 |
@sonney2k | pickle27, I need the configure output and the .config | 21:36 |
pickle27 | sonney2k: configure output http://pastebin.com/4AqNjzLc | 21:38 |
pickle27 | sonney2k: .config http://pastebin.com/5PtK4ygr | 21:38 |
pickle27 | lisitsyn: hey | 21:41 |
lisitsyn | hello | 21:41 |
lisitsyn | pickle27: hey | 21:41 |
lisitsyn | I felt that | 21:41 |
pickle27 | lisitsyn: I had some questions about the static interface | 21:41 |
-!- gsomix_ is now known as gsomix | 21:42 | |
pickle27 | I want to create a version of my python graphical jade example for matlab and r both static and modular | 21:42 |
pickle27 | and Im trying to figure out how to do the static one | 21:42 |
lisitsyn | ahh | 21:42 |
lisitsyn | pickle27: ok just give me like 20 minutes | 21:43 |
pickle27 | for sure | 21:43 |
lisitsyn | once I'm back I'll check and say what we should do here | 21:43 |
@sonney2k | pickle27, why do you compile for octave,python & r? | 21:45 |
pickle27 | sonney2k: I want to write and example of using Jade for each | 21:45 |
@sonney2k | lisitsyn, you can one get the real line for linking to debug pickle27's issue? | 21:58 |
lisitsyn | pickle27: what's the issue? | 22:00 |
gsomix | sonney2k, good morning. | 22:01 |
gsomix | sonney2k, how are you? I have small question. | 22:01 |
lisitsyn | gsomix: good what? :D | 22:02 |
gsomix | lisitsyn, awww, where is morning now? =___= | 22:02 |
@sonney2k | gsomix, ask don't ask to ask | 22:05 |
-!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has left #shogun ["Fallen asleep!"] | 22:06 | |
gsomix | sonney2k, were you at concert "The Wall" of Roger Waters at 1990 in Berlin? | 22:06 |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 22:06 | |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * d8d7a85 / tests/unit/Makefile: https://github.com/shogun-toolbox/shogun/commit/d8d7a85cf34c0ba5e50d356d1e17b40e4687a163 | 22:06 |
shogun-notifier- | shogun: use $(PYTHON) for running jinja2 | 22:06 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * d459978 / src/shogun/mathematics/Math.cpp,src/shogun/mathematics/Math.h: https://github.com/shogun-toolbox/shogun/commit/d4599786c90626434de9be7e0d689dcba016f4be | 22:06 |
shogun-notifier- | shogun: use std's isnan/isinf | 22:06 |
lisitsyn | pickle27: okay sorry took a bit more | 22:18 |
lisitsyn | pickle27: let me just check your example | 22:18 |
lisitsyn | pickle27: still here? | 22:31 |
shogun-buildbot | build #1085 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1085 blamelist: Soeren Sonnenburg <sonne@debian.org> | 22:32 |
-!- travis-ci [~travis-ci@ec2-23-22-182-174.compute-1.amazonaws.com] has joined #shogun | 22:32 | |
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/9531049 | 22:32 |
-!- travis-ci [~travis-ci@ec2-23-22-182-174.compute-1.amazonaws.com] has left #shogun [] | 22:32 | |
shogun-buildbot | build #1579 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1579 blamelist: Soeren Sonnenburg <sonne@debian.org> | 22:34 |
shogun-buildbot | build #1289 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1289 blamelist: Soeren Sonnenburg <sonne@debian.org> | 22:35 |
@sonney2k | gsomix, no :) | 22:41 |
pickle27 | lisitsyn: hey back | 22:43 |
lisitsyn | pickle27: okay some things are rather easy | 22:43 |
gsomix | sonney2k, sad. great concert | 22:44 |
pickle27 | yeah | 22:44 |
lisitsyn | pickle27: let me guide you to exact lines | 22:44 |
pickle27 | lisitsyn: okay | 22:45 |
lisitsyn | pickle27: shogun/ui/GUIConverter.h:57 | 22:45 |
lisitsyn | no need to explain here ;) | 22:45 |
pickle27 | yeah I added a method for create_Jade | 22:46 |
lisitsyn | create_jade ;) | 22:46 |
lisitsyn | pickle27: then its implementation to | 22:46 |
lisitsyn | pickle27: shogun/ui/GUIConverter.cpp | 22:46 |
pickle27 | lisitsyn: yup! | 22:46 |
lisitsyn | easy again | 22:46 |
pickle27 | done | 22:46 |
lisitsyn | next | 22:47 |
lisitsyn | a minute | 22:47 |
lisitsyn | pickle27: shogun/ui/SGInterface.cpp:5368 | 22:48 |
lisitsyn | pickle27: I am scared with the number too ;) | 22:48 |
pickle27 | lisitsyn: lol wow | 22:49 |
pickle27 | okay I added a method there | 22:50 |
lisitsyn | pickle27: what method? | 22:50 |
pickle27 | woops meant I added an if | 22:50 |
lisitsyn | pickle27: oh alright | 22:51 |
lisitsyn | pickle27: okay last thing | 22:52 |
lisitsyn | I am thinking | 22:52 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 02d5912 / src/shogun/classifier/vw/VwParser.cpp: https://github.com/shogun-toolbox/shogun/commit/02d5912d83dcd430a2bd1e3b35e7db01d2e09219 | 22:55 |
shogun-notifier- | shogun: use isnan from CMath | 22:55 |
lisitsyn | pickle27: do you need that 'mixing-matrix' method? | 22:55 |
pickle27 | lisitsyn: mmm sort of but the features get transformed so as long as I get them | 22:56 |
lisitsyn | pickle27: I mean in static | 22:56 |
lisitsyn | let me explain | 22:57 |
pickle27 | yeah I see what the problem is sort of | 22:57 |
lisitsyn | pickle27: shogun/ui/SGInterface.h/cpp | 22:57 |
pickle27 | we can probably do without for now | 22:57 |
lisitsyn | what do you need to do | 22:57 |
lisitsyn | is | 22:57 |
lisitsyn | just search for cmd_embed | 22:57 |
lisitsyn | add cmd_apply | 22:57 |
lisitsyn | or rather | 22:57 |
lisitsyn | cmd_apply_converter | 22:57 |
pickle27 | okay | 22:57 |
lisitsyn | and cmd_get_mixing_matrix | 22:57 |
lisitsyn | the implementation is rather simple | 22:58 |
lisitsyn | check cmd_embed again | 22:58 |
pickle27 | okay, I'll leave the mixing matrix part out for now because I may make a new generic converter for ICA | 22:58 |
lisitsyn | you just use usual shogun code to do the things | 22:58 |
pickle27 | then I'll make that | 22:58 |
lisitsyn | and then set_matrix | 22:58 |
lisitsyn | to output the matrix | 22:58 |
lisitsyn | (that's the method to call) | 22:59 |
lisitsyn | pickle27: and the final step | 22:59 |
lisitsyn | pickle27: shogun/ui/GUICommands.h:237 | 22:59 |
lisitsyn | and one more :D | 22:59 |
lisitsyn | pickle27: shogun/ui/SGInterface.cpp:616 | 23:01 |
lisitsyn | just the same thing like embed | 23:02 |
lisitsyn | but apply | 23:02 |
pickle27 | lisitsyn: thanks! | 23:05 |
lisitsyn | pickle27: that's it I hope | 23:05 |
shogun-buildbot | build #1086 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1086 blamelist: Soeren Sonnenburg <sonne@debian.org> | 23:05 |
pickle27 | I'll get this running now then I might change how it later depending on how many ica methods we get | 23:06 |
lisitsyn | pickle27: well that's just apply method | 23:08 |
lisitsyn | it shoudl be here anyway | 23:08 |
shogun-buildbot | build #1290 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1290 blamelist: Soeren Sonnenburg <sonne@debian.org> | 23:09 |
pickle27 | lisitsyn: yeah good point mayble I';; just add if I think ICA needs its own static interface | 23:12 |
pickle27 | lisitsyn: also any idea how to fix my build of r_modular? | 23:12 |
pickle27 | I posted some output for sonney2k about 2 hours ago | 23:12 |
lisitsyn | pickle27: let me check logs then | 23:12 |
@sonney2k | lisitsyn, it woudl help if you give us said how to get the linker output | 23:13 |
lisitsyn | sonney2k: I have no idea what you are talking about :D | 23:13 |
lisitsyn | ahhh that stuff that hides linker | 23:14 |
@sonney2k | lisitsyn, you know when shogun is compiled it first compiles & then links | 23:14 |
lisitsyn | this tihng? | 23:14 |
@sonney2k | we now have the fancy linking | 23:14 |
@sonney2k | line but no longer know what the real thing is | 23:14 |
@sonney2k | maybe we should output that to some file | 23:14 |
lisitsyn | sonney2k: do you want to have exact command? | 23:14 |
lisitsyn | or? | 23:14 |
@sonney2k | sth like compile.log or so | 23:14 |
@sonney2k | yes the full line | 23:15 |
lisitsyn | sonney2k: well just a quick change to do that is | 23:16 |
lisitsyn | sonney2k: Makefile.template:569 | 23:16 |
lisitsyn | remove first @ | 23:16 |
lisitsyn | and next two @ lines too | 23:17 |
@sonney2k | pickle27, please do what lisitsyn said ^ | 23:21 |
lisitsyn | lisitsyn consulting | 23:21 |
pickle27 | and post output / try and understand myself | 23:21 |
@sonney2k | lisitsyn, but having this in some log would be nice | 23:21 |
lisitsyn | sonney2k: yeah thinking how to do that | 23:21 |
@sonney2k | gsomix, how is the .csv stuff going? | 23:24 |
gsomix | sonney2k, working, refactoring. I hope I'll sent PR after 0:00 UTC - want streak at github. :) | 23:25 |
gsomix | sonney2k, but today was rainy, depressing (dispersing?) and unproductive day. all hopes to night. | 23:26 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 5bc9bff / tests/unit/base/clone_unittest.cc.py: https://github.com/shogun-toolbox/shogun/commit/5bc9bff834299e7a90f57d38c2149703b88ece45 | 23:32 |
shogun-notifier- | shogun: create some dummy unit test in absence of jinja2 | 23:32 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 72865ad / src/shogun/io/SGIO.cpp,src/shogun/io/SGIO.h: https://github.com/shogun-toolbox/shogun/commit/72865ad678802c88358f5f0196511f853b501b35 | 23:32 |
shogun-notifier- | shogun: move some inline functions to .cpp | 23:32 |
@sonney2k | gsomix, sounds better than having 30+ C and having to work | 23:32 |
gsomix | sonney2k, hm, I'd love heat and sun. | 23:33 |
@sonney2k | gsomix, and no way to enjoy it beacuse you have to work? | 23:34 |
gsomix | sonney2k, yep (I should say "yep", I should say "yep", not "no"), because I like sitting at home (office, lab). =____= | 23:38 |
-!- travis-ci [~travis-ci@ec2-23-22-182-174.compute-1.amazonaws.com] has joined #shogun | 23:39 | |
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/9532827 | 23:39 |
-!- travis-ci [~travis-ci@ec2-23-22-182-174.compute-1.amazonaws.com] has left #shogun [] | 23:39 | |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * d1be1ca / src/shogun/structure/FactorGraph.cpp: https://github.com/shogun-toolbox/shogun/commit/d1be1cae03e399fe5ee3ca444b156bba2b449355 | 23:41 |
shogun-notifier- | shogun: ifdef refcount debug access to fix build failure when refcounting is not available | 23:41 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 702c6af / src/shogun/distributions/HMM.cpp: https://github.com/shogun-toolbox/shogun/commit/702c6afd961defbd9982b04b5f80bfe7721da7f3 | 23:46 |
shogun-notifier- | shogun: fix compile error w/ HMM parallel structures enabled | 23:46 |
-!- van51 [~van51@ppp-94-66-56-238.home.otenet.gr] has joined #shogun | 23:48 | |
@sonney2k | van51, evening! | 23:49 |
van51 | sonney2k: hey! | 23:49 |
@sonney2k | van51, hmmhh C=10 is not a useful comparison I must admit | 23:50 |
@sonney2k | van51, what you could do is run the original experiment for a real comparison | 23:50 |
van51 | sonney2k: I ran it for c=1 but with a higher number of bits | 23:50 |
van51 | and it was pretty much the same | 23:51 |
van51 | I didn't keep results, but I can run it again tomorrow | 23:51 |
@sonney2k | van51, ../examples/undocumented/python_modular/kernel_comm_ulong_string_modular.py | 23:51 |
@sonney2k | van51, it would also be interesting if you used your converter once to give us an idea how much faster it is with the sparse representation (and results should be the same so good test too) | 23:52 |
shogun-buildbot | build #1405 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1405 blamelist: Soeren Sonnenburg <sonne@debian.org> | 23:53 |
@sonney2k | van51, in this python script above you would use the StrinFeatures<uint64_t> get them from char features of order 8 | 23:53 |
@sonney2k | 0 gap not reversed | 23:53 |
@sonney2k | and then with the CCommUlongStringKernel train a CSVMLight with C=1 | 23:53 |
van51 | sonney2k: and after that use the hasheddocfeatures? | 23:54 |
@sonney2k | van51, sorry what? | 23:55 |
@sonney2k | I am missing context | 23:55 |
@sonney2k | van51, what did that refer to? | 23:55 |
van51 | sonney2k: I though I would use that python example to compare results with the HashedDoc features | 23:55 |
@sonney2k | van51, you can do 2 things | 23:56 |
@sonney2k | van51, a) get your result with hasheddocfeatures on webspam and with the CommUlongStringkernel and SVMLight and compare | 23:56 |
@sonney2k | van51, b) just compare the kernel matrices - with once using the commulongstring kernel and once the polykernel d=1 on your hashed features | 23:57 |
@sonney2k | b) is similar to what you have done already | 23:57 |
@sonney2k | a) is to really see the speed comparison old vs new | 23:58 |
van51 | sonney2k: ok got it | 23:58 |
van51 | sonney2k: I can try them both | 23:58 |
van51 | I 'll do it tomorrow, I'm a bit sleepy now.. | 23:59 |
van51 | sonney2k: btw I'm also downloading documents from wikipedia atm | 23:59 |
--- Log closed Sat Jul 27 00:00:49 2013 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!