Open in new window / Try shogun cloud
--- Log opened Tue Jan 14 00:00:51 2014
-!- FSCV [~FSCV@] has quit [Quit: Leaving]00:47
-!- sonne|osx_ [] has joined #shogun03:40
-!- sonne|osx [] has quit [Ping timeout: 246 seconds]03:41
-!- sonne|osx_ is now known as sonne|osx03:41
shogun-buildbotbuild #680 of nightly_default is complete: Failure [failed test]  Build details are at
-!- zxtx_ [] has quit [Ping timeout: 252 seconds]06:14
-!- sonne|osx [] has quit [Quit: sonne|osx]06:58
-!- zxtx_ [] has joined #shogun06:59
-!- sonne|osx [~sonne@] has joined #shogun08:13
sonne|osxwiking: no it is not that. if I compile without libstdc++ libshogun compiles just fine08:15
sonne|osxwiking: but swig then doesn't08:15
sonne|osxwiking: so we need libc++ for protbuf08:15
sonne|osxand libstdc++ for swig at this moment08:15
sonne|osxshogun-buildbot: force build --branch=develop 'osx2 - python'08:16
shogun-buildbotbuild #6 forced08:16
shogun-buildbotI'll give a shout when the build finishes08:16
shogun-buildbotbuild #6 of osx2 - python is complete: Failure [failed test python modular]  Build details are at
-!- sonne|osx [~sonne@] has quit [Quit: sonne|osx]08:39
-!- lambday [67157c36@gateway/web/freenode/ip.] has joined #shogun09:11
-!- lambday [67157c36@gateway/web/freenode/ip.] has quit [Ping timeout: 272 seconds]10:22
-!- sonne|work [] has quit [Quit: Leaving.]11:27
-!- sonne|work [] has joined #shogun11:46
-!- besser82 [quassel@fedora/besser82] has quit [Remote host closed the connection]14:54
-!- besser82 [quassel@fedora/besser82] has joined #shogun15:12
-!- zxtx_ [] has quit [Ping timeout: 272 seconds]16:27
-!- zxtx_ [] has joined #shogun17:13
-!- zxtx_ [] has quit [Ping timeout: 248 seconds]17:20
@wikinganybody knows a good way to handle a pool or processes where each process has it's own thread pool in python? :)17:24
-!- thoralf [] has joined #shogun18:15
-!- sonne|osx [~sonne@] has joined #shogun18:25
naywhayarehello folks, I have two things today18:26
naywhayarea) I'm going to be running the shogun benchmarks for the latest shogun release in a week or two (I have to wait until the mlpack ones are done, and they take two weeks ...);18:27
-!- sonne|osx [~sonne@] has quit [Client Quit]18:27
naywhayareI'd like to know if anyone would be interested in looking at our python scripts to see if they can be significantly improved or basically just to make sure we're not doing anything stupid18:27
naywhayareb) I know you guys wrap john langford's cover tree implementation for nearest neighbor search, so I thought I would let you know that there's an unfixed bug in it (an invalid prune) so until he fixes it, that code is wrong18:28
-!- thoralf [] has quit [Quit: Konversation terminated!]18:45
-!- zxtx_ [] has joined #shogun18:48
lisitsynnaywhayare: hey!18:48
lisitsynnaywhayare: about a) I think people around would take a look18:49
lisitsynand b) well it is pure bs, shoulda be dropped :D18:49
naywhayarelisitsyn: here are the scripts we have:
naywhayareand that cover tree code... it took me a month to figure out whether or not the bug was actually a bug and develop the minimal test case18:52
naywhayarejohn langford should send me some money for my work... heh...18:52
lisitsynnaywhayare: code looks quite legit from the first sight18:53
lisitsynnaywhayare: do you have your own impl of cover tree?18:54
naywhayarelisitsyn: yeah, but it's not as fast as his19:11
naywhayareI'm working on speeding it up, but it gets difficult.  what I wanted was an implementation of cover trees for any tree-based algorithm, not just nearest neighbor19:12
naywhayareso I have to implement it kind of abstractly, but also at the same time make it as fast as his, and I haven't managed to do that yet...19:12
lisitsynnaywhayare: we could think if we can implement some module we can both share19:28
lisitsyne.g. gsoc task etc ;)19:28
-!- iglesiasg [] has joined #shogun19:45
-!- mode/#shogun [+o iglesiasg] by ChanServ19:45
-!- sonne|osx [~sonne@] has joined #shogun19:53
sonne|osxiglesiasg, wiking, lisitsyn - I am stiing in the c-base right now :)19:53
lisitsynsonne|osx: heh19:54
@iglesiasgsonne|osx, wow, cool19:54
@iglesiasghave a beer for all of us :)19:54
sonne|osxiglesiasg: I will19:54
sonne|osxintend to donate some $$$ to them - long promised but never managed19:55
@iglesiasgoh good19:55
sonne|osxiglesiasg: I looked at the failing tests yesterday - some indeed are just numerical issues easy to fix19:56
@iglesiasgsonne|osx, just re-generating  the data files I guess?19:56
sonne|osxsome yes19:57
sonne|osxnot all19:57
@iglesiasgsonne|osx, what about the others? which ones looked fishy?19:57
sonne|osxiglesiasg: don't remember was busy with osx buildbot19:57
sonne|osxbut osx is so f#%($*%d up...19:59
sonne|osxa) they switched their compiler to clang19:59
sonne|osxb) they are dropping usage of libstdc++19:59
sonne|osxintroducing libc++19:59
sonne|osxwhich is not compatible19:59
sonne|osxso stuff linked against libc++ won't link against libstdc++20:00
-!- zxtx_ [] has quit [Ping timeout: 276 seconds]20:00
sonne|osxwhich is what shogun uses due to some error with swig20:00
naywhayarelisitsyn: perhaps :) the benchmarking system that one of our GSoC students made this past summer could be used by you guys too (HeikoS was interested, but I don't know if he has or will have the time to look into it anytime soon...)20:07
sonne|osxnaywhayare: yes it would be something nice to have20:09
-!- besser82 [quassel@fedora/besser82] has quit [Quit: freedom, friends, features, first --->]20:10
sonne|osxnaywhayare: but who has time to do stuff ...20:10
sonne|osxwiking: what is generating shogun-unit-test_test.cmake ?20:11
sonne|osxcmake on the osx builbot complains about tons of syntax errors leading to a log file of size >30MB...20:12
naywhayaresonne|osx: too true, I haven't even had time to go over the interface for the benchmarks and make it more intuitive, and it's been like five months now ...20:15
sonne|osxnaywhayare: I really wonder why everyone is having the same issue :)20:19
naywhayareI need either clones of myself or a way to dump responsibilities on someone else, but I don't have time to create either !20:20
naywhayareI think farmers used to just have a bunch of kids to solve this problem, but the lead time on that is like 18-20 years20:20
sonne|osxnaywhayare: hehe20:23
-!- besser82 [quassel@fedora/besser82] has joined #shogun20:23
sonne|osxnaywhayare: clone early clone often :)20:25
-!- sonne|osx [~sonne@] has quit [Quit: sonne|osx]20:31
-!- FSCV [~FSCV@] has joined #shogun20:38
-!- iglesiasg [] has quit [Ping timeout: 245 seconds]21:40
-!- sonne|osx [~sonne@] has joined #shogun21:59
-!- zxtx [] has joined #shogun22:03
-!- sonne|osx [~sonne@] has quit [Quit: sonne|osx]22:23
-!- gsomix_ [~gsomix@] has joined #shogun22:44
-!- gsomix [~gsomix@] has quit [Ping timeout: 265 seconds]22:44
-!- zxtx [] has quit [Ping timeout: 252 seconds]23:14
-!- sonne|osx [~sonne@] has joined #shogun23:37
sonne|osxwiking: any idea?23:40
sonne|osxshogun-unit-test_test.cmake:1:119 Argument not separated from preceding token by whitespace.23:40
-!- shogun-notifier- [] has joined #shogun23:45
shogun-notifier-shogun: Soeren Sonnenburg :develop * 9e91e48 / tests/unit/discover_gtest_tests.cpp:
shogun-notifier-shogun: add whitespace to fix cmake warning23:45
sonne|osxshogun-buildbot: force build --branch=develop 'osx2 - python'23:45
shogun-buildbotbuild #7 forced23:45
shogun-buildbotI'll give a shout when the build finishes23:45
-!- sonne|osx [~sonne@] has quit [Quit: sonne|osx]23:59
--- Log closed Wed Jan 15 00:00:52 2014