--- Log opened Thu Oct 18 00:00:05 2012 | ||
--- Day changed Thu Oct 18 2012 | ||
blackburn | how much? | 00:00 |
---|---|---|
wiking | i still want to see what mosek does though | 00:00 |
wiking | one loss function for my case: 34.59 | 00:00 |
wiking | the published one best case is 35.6 | 00:00 |
blackburn | what's the difference between your method and previous? | 00:01 |
wiking | well the underlying so-solver | 00:01 |
blackburn | hmm so just a solver issue? | 00:01 |
wiking | yep | 00:01 |
blackburn | are you using BMRM? | 00:01 |
wiking | p-bmrm | 00:01 |
blackburn | proximal point? | 00:02 |
wiking | p3bmrm for that matter | 00:02 |
wiking | not ppbmrm | 00:02 |
blackburn | aham | 00:02 |
wiking | but the p3bmrm | 00:02 |
wiking | with multiple cps | 00:02 |
blackburn | well you should publish it then :D | 00:02 |
wiking | hehehe | 00:03 |
wiking | i need to get the other benchmarking data set | 00:03 |
wiking | but it not only gives me better results | 00:03 |
wiking | but much less training time is needed as well | 00:03 |
wiking | :P | 00:03 |
wiking | it's like 2/3 | 00:03 |
wiking | so pretty good | 00:03 |
blackburn | here comes a title "Efficient latent structured support vector machines training via multiple cutting plane models" | 00:04 |
blackburn | :D | 00:04 |
wiking | ::DDD | 00:05 |
wiking | lolface | 00:05 |
blackburn | lets write an abstract :D | 00:05 |
blackburn | Nowadays latent models of visual concept recognition are widespread and efficient. However blablabla | 00:06 |
blackburn | :D | 00:06 |
wiking | nyihahahaha | 00:07 |
wiking | btw freak is already included in opencv 2.4.2 | 00:08 |
wiking | afaik we have a timer function in shogun right? ?:) | 00:10 |
blackburn | wiking: CTime right | 00:14 |
blackburn | wiking: yes freak is in for a while already | 00:14 |
blackburn | wiking: I tried it as a dense descriptor | 00:16 |
blackburn | works ok but in the end I started to believe it is just wrong | 00:17 |
-!- blackburn [~blackburn@83.234.54.218] has quit [Quit: Leaving.] | 01:05 | |
shogun-buildbot | build #119 of nightly_none is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_none/builds/119 | 03:02 |
shogun-buildbot | build #136 of nightly_default is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/136 | 04:13 |
-!- in3xes [~in3xes@122.174.96.165] has joined #shogun | 07:36 | |
-!- in3xes [~in3xes@122.174.96.165] has quit [Remote host closed the connection] | 08:34 | |
wiking | sonne|work: online? | 09:15 |
sonne|work | y | 09:15 |
wiking | sonne|work: can u give me an access to a buildslave where there is clang? | 09:16 |
sonne|work | that machine is doing all mailinglist, webhosting, buildslave | 09:17 |
sonne|work | so no | 09:17 |
wiking | i will not vandalize it | 09:17 |
sonne|work | but I should have a bit of time once on the mentor summit to do these changes | 09:17 |
wiking | just want to fix the static analyzer | 09:17 |
wiking | it should be a 2liner :) | 09:18 |
wiking | or is there sense to add another buildslave? | 09:18 |
-!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has joined #shogun | 09:18 | |
wiking | since i have a spare ubuntu server running around | 09:19 |
wiking | so there we could have static analyzer | 09:19 |
blackburn | wiking da servers master | 09:19 |
wiking | sonne|work: maybe that sounds better? and then that would be the static analyzer bot? | 09:19 |
sonne|work | wiking: we don't have an ubuntu buildslave yet | 09:20 |
sonne|work | so it would make sense | 09:20 |
wiking | sonne|work: oh ok | 09:20 |
sonne|work | and yes give me the 2 liner :) | 09:20 |
wiking | sonne|work: i'll do the thingy | 09:20 |
wiking | w8 | 09:20 |
wiking | dont go away! | 09:20 |
sonne|work | wiking: I don't have access to the bot from work but can maybe do it form the airport tomorrow | 09:21 |
wiking | sonne|work: buildslave create-slave shogun-buildbot www.shogun-toolbox.org shogun-ubuntu1 <passwd> | 09:31 |
wiking | that sounds good for you? | 09:31 |
sonne|work | probably good - w/o access to the bots I have no clue (brain == empty) | 09:33 |
wiking | sonne|work: okok i've tried to follow what's in the freebsd bot config | 09:33 |
wiking | i've sent you in /msg the details, or should i send it via email? | 09:34 |
sonne|work | ha | 09:35 |
sonne|work | yes | 09:35 |
wiking | i'll run now a test with static analyzer on that host on shogun | 09:37 |
wiking | blackburn: here? | 09:38 |
wiking | blackburn: what packages should i have on ubuntu for a decent shougn build? | 09:39 |
sonne|work | wiking: apt-get build-dep shogun | 09:40 |
sonne|work | and also the newer ones | 09:40 |
wiking | mmm something is weird with atlas/lapack | 09:42 |
wiking | installed libatlas-dev | 09:42 |
wiking | but doesn't recongise it | 09:42 |
blackburn | wiking: libatlas3gf-base | 09:42 |
sonne|work | wiking, well you need arpack, nlopt, superlu, snappy etc too | 09:45 |
wiking | Checking for Atlas support ... no | 09:45 |
sonne|work | wiking: do you want to just build libshogun or the whole thing with octave, java, lua etc... | 09:45 |
wiking | libshogun at first | 09:46 |
sonne|work | k | 09:47 |
blackburn | wiking: install libatlas* :D | 09:58 |
wiking | ok i've done with it | 09:58 |
blackburn | everything that has atlas word :) | 09:58 |
wiking | now it's doing the static-analyzer | 09:58 |
wiking | which package is NLOPT? | 10:02 |
wiking | i couldn't find anything with apt-cache search nlopt | 10:02 |
blackburn | wiking: bingo, it is nonexistent | 10:03 |
blackburn | :D | 10:03 |
wiking | :> | 10:03 |
wiking | i see | 10:03 |
wiking | ok then | 10:03 |
wiking | u install it by hand? | 10:03 |
wiking | brb | 10:04 |
blackburn | yes | 10:07 |
sonne|work | wiking: there is one in debian unstable - so probably newer ubuntus will have it | 10:18 |
wiking | blackburn: yesterday i've tried to do a SGSparseVector -> Eigen::SparseVector mapping but it seems it's more complicated than i thought... i.e. it would take an awful lot of copy elements... and i think this way we'd loose the performance gain of eigen | 10:19 |
wiking | sonne|work: noup :( i have 12.04.1 | 10:19 |
wiking | i.e. latest stable ubuntu and doesn't have it :( | 10:19 |
blackburn | 12.10 is out :) | 10:19 |
wiking | oh woah | 10:19 |
blackburn | wiking: well yeah it is a problem | 10:19 |
sonne|work | wiking: then grab the .deb from debian | 10:20 |
wiking | blackburn: 12.10 release date today :DDD | 10:21 |
wiking | blackburn: 0day you are | 10:21 |
wiking | :) | 10:21 |
wiking | has n4nd0 been around lately? | 10:21 |
wiking | finally i've managed to progress with his mosek solver as well | 10:21 |
blackburn | wiking: no, he is rather in gmail than in irc | 10:22 |
wiking | blackburn: yeah i see him on g+ | 10:23 |
wiking | i've found the bug in his code | 10:24 |
wiking | but it's rather strange | 10:24 |
wiking | i think there's something around CList | 10:24 |
wiking | :( | 10:24 |
wiking | with referencing... | 10:24 |
wiking | yey it worked: http://categorizer.tmit.bme.hu/~wiking/shogun-sa/ | 10:28 |
blackburn | static thing rocks heh | 10:30 |
wiking | no mem-leak 8X | 10:36 |
wiking | almost cannot believe :D | 10:36 |
wiking | ok here's a little PR for mosek :) | 10:40 |
wiking | merge plz https://github.com/shogun-toolbox/shogun/pull/795 | 10:52 |
sonne|work | wiking: write him an email | 10:57 |
sonne|work | wiking: whats wrong with CList? | 10:57 |
sonne|work | did he use the thread safe functions or the non-safe ones? | 10:57 |
wiking | sonne|work: already discussed with him.... he's studying for an exam he has tomorrow | 10:58 |
wiking | so we'll talk after | 10:58 |
sonne|work | k | 10:58 |
wiking | sonne|work: https://github.com/vigsterkr/shogun/blob/b146efb45b7410c370de277f4d1aa02d66f7707a/src/shogun/structure/PrimalMosekSOSVM.cpp#L130 | 10:59 |
wiking | starting from there | 10:59 |
wiking | till line 177 | 10:59 |
wiking | as i see it | 10:59 |
wiking | i'd expect that result has m_refcount =1 | 10:59 |
wiking | on line 1777 | 11:00 |
wiking | *177 | 11:00 |
wiking | but that's not the case | 11:00 |
sonne|work | the non thread safe ones | 11:00 |
wiking | ? | 11:01 |
wiking | how can i have non-safe one? | 11:01 |
sonne|work | wiking: it doesn't matter here - there are no threads | 11:03 |
wiking | yeah :) | 11:03 |
wiking | but anyhow | 11:03 |
sonne|work | wiking: shouldn' thte refcount be 2? | 11:03 |
wiking | an insert_element should ++ the refcount | 11:03 |
wiking | expecially that the list is delete_element=true | 11:04 |
sonne|work | yes | 11:04 |
sonne|work | but he also uses argmax to get it | 11:04 |
sonne|work | so it might be 1 already there | 11:04 |
sonne|work | wiking: what is the actual count? | 11:05 |
wiking | not that https://github.com/vigsterkr/shogun/blob/b146efb45b7410c370de277f4d1aa02d66f7707a/src/shogun/structure/PrimalMosekSOSVM.cpp#L125 doesn't add a ref | 11:05 |
sonne|work | so it has 0 there? | 11:05 |
wiking | yes | 11:07 |
wiking | ok then i'm stupid :D | 11:07 |
wiking | becuase the insert -> 1 | 11:08 |
wiking | and l177 -> 0 | 11:08 |
shogun-buildbot | build #449 of deb2 - static_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/449 blamelist: Viktor Gal <viktor.gal@maeth.com> | 11:10 |
wiking | wasn'tme | 11:11 |
-!- adoniscik [~emre@c-67-180-103-118.hsd1.ca.comcast.net] has quit [Ping timeout: 240 seconds] | 11:28 | |
sonne|work | blackburn: you? | 11:36 |
sonne|work | or clue? | 11:36 |
blackburn | sonne|work: heiko | 11:36 |
blackburn | or me | 11:36 |
blackburn | :D | 11:36 |
blackburn | rather heiko | 11:36 |
shogun-buildbot | build #558 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/558 blamelist: Viktor Gal <viktor.gal@maeth.com> | 11:37 |
sonne|work | wiking: could you please ask heiko then? otherwise you will be drowning in emails | 11:38 |
blackburn | heh I bet wiking already set up his spam filter :D just as I did | 11:39 |
blackburn | well joking gmail puts all these mails to the special folder I never read | 11:39 |
blackburn | but it is not called 'spam' | 11:39 |
shogun-buildbot | build #559 of deb3 - modular_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/559 blamelist: Soeren Sonnenburg <sonne@debian.org> | 11:43 |
shogun-buildbot | build #450 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/450 | 11:43 |
wiking | :D | 11:51 |
wiking | i think i've solved the mosek thing... but it performs really bad :S | 12:11 |
wiking | so maybe i didn't | 12:11 |
-!- naywhayare [~ryan@spoon.lugatgt.org] has quit [Ping timeout: 246 seconds] | 15:32 | |
-!- naywhayare [~ryan@spoon.lugatgt.org] has joined #shogun | 15:32 | |
-!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has quit [Quit: Page closed] | 15:37 | |
-!- sonne|work [~sonnenbu@194.78.35.195] has left #shogun [] | 16:45 | |
-!- adoniscik [~emre@c-50-131-104-102.hsd1.ca.comcast.net] has joined #shogun | 19:28 | |
-!- blackburn [~blackburn@31.28.57.136] has joined #shogun | 20:22 | |
wiking | blackburn: lol man | 20:27 |
wiking | blackburn: have u heard about this stuff: | 20:27 |
wiking | http://about.travis-ci.org/ | 20:27 |
wiking | and btw: best one of the best PRs https://github.com/matplotlib/matplotlib/pull/1329 | 20:28 |
wiking | :DDD | 20:28 |
wiking | note the bottom of the PR description: Good to merge -- The Travis build passed (Details) | 20:28 |
blackburn | hehe | 20:28 |
wiking | here's a nice blog post about it | 20:29 |
wiking | http://jakevdp.github.com/blog/2012/10/07/xkcd-style-plots-in-matplotlib/ | 20:29 |
wiking | but back to the travis thing | 20:29 |
wiking | i'll try it now on my own repo | 20:30 |
wiking | of shogun to set it up | 20:30 |
blackburn | I do not understand CI :D | 20:30 |
wiking | but this PR testing sounds good | 20:30 |
wiking | well this travis shit is basically another type of buildbot | 20:30 |
wiking | but it's well integrated with github | 20:30 |
wiking | so you have things like | 20:30 |
wiking | once you have a PR | 20:31 |
wiking | it'll automatically try to build it | 20:31 |
wiking | and test it for that matter | 20:31 |
blackburn | ahh | 20:31 |
wiking | looks nice | 20:31 |
blackburn | okay I see | 20:32 |
wiking | but i want to test this PR for mpl :D | 20:34 |
wiking | looks hilarious | 20:34 |
wiking | blackburn: https://travis-ci.org/#!/vigsterkr/shogun/jobs/2843350 | 20:43 |
wiking | hehehe | 20:43 |
blackburn | heh | 20:44 |
wiking | niiice | 20:45 |
wiking | even the dependencies are installing | 20:45 |
wiking | :)))) | 20:45 |
wiking | but i think i'll have to add more packages there as that's for an old shogun | 20:45 |
wiking | woah this one even takes mo | 20:46 |
wiking | mono | 20:46 |
wiking | :d | 20:46 |
wiking | oooh noooo | 20:46 |
wiking | i forgot a cd src | 20:46 |
wiking | :D | 20:46 |
wiking | 2nd try: https://travis-ci.org/#!/vigsterkr/shogun/jobs/2843516 | 20:52 |
blackburn | http://researchinprogress.tumblr.com/ wiking it is totally awesome :D | 20:55 |
wiking | woohooo | 20:55 |
wiking | building! | 20:55 |
wiking | blackburn: ahahahha | 20:55 |
blackburn | omg | 20:57 |
blackburn | :D | 20:57 |
wiking | i want this for SHOGUN! | 20:57 |
wiking | it's basically an extra buildbot | 20:58 |
wiking | integrated in github | 20:58 |
wiking | :) | 20:58 |
wiking | heheh python_modular is building | 20:59 |
wiking | yey | 20:59 |
blackburn | wiking: where is it run and why it is free? :) | 21:01 |
wiking | ok clang droped dead | 21:02 |
wiking | blackburn: dunno :D | 21:02 |
wiking | blackburn: let's be happy that it's still free | 21:02 |
wiking | This test series was run on a worker box sponsored by Engine Yard | 21:02 |
wiking | ok wtf is this: | 21:03 |
wiking | In file included from modshogun_wrap.cxx:9527: | 21:03 |
wiking | 2853/usr/lib/gcc/i686-linux-gnu/4.6/../../../../include/c++/4.6/cstring:101:3: error: | 21:03 |
wiking | 2854 declaration conflicts with target of using declaration already in scope | 21:03 |
wiking | 2855 memchr(void* __s, int __c, size_t __n) | 21:03 |
wiking | 2856 ^ | 21:03 |
wiking | 2857/usr/include/string.h:95:14: note: target of using declaration | 21:03 |
wiking | 2858extern void *memchr (__const void *__s, int __c, size_t __n) | 21:03 |
wiking | 2859 ^https://travis-ci.org/#!/vigsterkr/shogun/jobs/2843516/L2852 | 21:03 |
wiking | ok | 21:03 |
wiking | https://travis-ci.org/#!/vigsterkr/shogun/jobs/2843516/L2852 | 21:03 |
wiking | this | 21:03 |
wiking | :> | 21:03 |
blackburn | uhuh | 21:05 |
wiking | that's with clang | 21:05 |
wiking | gcc 4.6 is still running | 21:06 |
wiking | that seems to be working till now | 21:11 |
wiking | 1943Executing your script (./configure && make && make test) took longer than 25 minutes and was terminated. | 21:30 |
wiking | :) | 21:30 |
blackburn | huh | 21:35 |
blackburn | we are pretty big to compile in 25 mins | 21:35 |
wiking | i'm gonna tweak around a bit (some says that make -j 4 helps to speed up compilation, even though there's only one core in the VM) | 21:36 |
wiking | but the other option is to only use certain interfaces in travis | 21:37 |
wiking | status.github.com :Partial service outage | 22:11 |
wiking | :S | 22:11 |
blackburn | oops | 22:11 |
wiking | battlestation doooown | 22:20 |
wiking | Major service outage | 22:20 |
wiking | somebody brokez z internet | 22:20 |
wiking | We are experiencing issues due to a DDOS attack, working hard to restore service... | 22:35 |
wiking | :DDD | 22:35 |
wiking | lol | 22:35 |
wiking | and doubleROTFL for not using ipython till now | 22:35 |
blackburn | is it cool? | 22:36 |
wiking | yeah i like it | 22:36 |
wiking | at least it has completion when i start hitting the <tab> | 22:37 |
wiking | heheh | 22:39 |
wiking | wtf | 22:41 |
wiking | sf is hyper slow as well for me | 22:41 |
blackburn | youtube is down :) | 22:41 |
wiking | worksforme | 22:42 |
wiking | woah motherfucker | 22:44 |
wiking | i've just realised that stupid macports | 22:44 |
wiking | is keeping the distfiles if i do a port clean | 22:44 |
wiking | my distfiles dir is currently 6 gigs | 22:44 |
wiking | what a waste... | 22:45 |
blackburn | why not overwritten? | 22:45 |
wiking | no i mean it keeps the tar.gz for all the installed ports | 22:46 |
blackburn | I see | 22:46 |
blackburn | okay I'm out of power | 22:47 |
blackburn | see you | 22:47 |
wiking | laterz | 22:48 |
-!- blackburn [~blackburn@31.28.57.136] has quit [Quit: Leaving.] | 22:49 | |
--- Log closed Fri Oct 19 00:00:17 2012 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!