--- Log opened Thu Aug 22 00:00:25 2013 | ||
-!- iglesiasg [~iglesias@89-77-4-38.dynamic.chello.pl] has quit [Quit: Ex-Chat] | 00:02 | |
shogun-buildbot | build #1460 of deb2 - static_interfaces is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1460 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 00:04 |
---|---|---|
shogun-buildbot | build #1461 of deb2 - static_interfaces is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1461 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com>, Kevin <kevinhughes27@gmail.com> | 00:04 |
shogun-buildbot | build #958 of rpm1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/958 blamelist: hushell <hushell@hushell-U510.(none)> | 00:06 |
shogun-buildbot | build #26 of osx1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/26 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com>, Kevin <kevinhughes27@gmail.com> | 00:08 |
shogun-buildbot | build #1492 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1492 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com>, Kevin <kevinhughes27@gmail.com> | 00:08 |
shogun-buildbot | build #1576 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/1576 | 00:15 |
shogun-buildbot | build #1282 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1282 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 00:17 |
-!- travis-ci [~travis-ci@ec2-54-234-140-21.compute-1.amazonaws.com] has joined #shogun | 00:20 | |
travis-ci | [travis-ci] it's Sergey Lisitsyn'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/10469697 | 00:20 |
-!- travis-ci [~travis-ci@ec2-54-234-140-21.compute-1.amazonaws.com] has left #shogun [] | 00:20 | |
shogun-notifier- | shogun: Roman Votyakov :develop * d711ce4 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/d711ce4e423e2f12628bc98ee302c202918d376b | 00:33 |
shogun-notifier- | shogun: add get_diagonal_vector() method to SGMatrix class | 00:33 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 4918010 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/49180104d1ec213ebbe1663478330cbb1abcc5f1 | 00:33 |
shogun-notifier- | shogun: Merge pull request #1451 from votjakovr/feature/get_diagonal_vector | 00:33 |
shogun-notifier- | shogun: | 00:33 |
shogun-notifier- | shogun: Add get_diagonal_vector() method to SGMatrix class | 00:33 |
shogun-buildbot | build #1462 of deb2 - static_interfaces is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1462 blamelist: Roman Votyakov <votjakovr@gmail.com> | 00:50 |
shogun-buildbot | build #959 of rpm1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/959 | 00:52 |
shogun-buildbot | build #1493 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1493 blamelist: Roman Votyakov <votjakovr@gmail.com> | 00:53 |
shogun-buildbot | build #27 of osx1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/27 blamelist: Roman Votyakov <votjakovr@gmail.com> | 00:58 |
shogun-buildbot | build #1494 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1494 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 00:58 |
shogun-buildbot | build #1283 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1283 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com>, Kevin <kevinhughes27@gmail.com> | 00:59 |
shogun-buildbot | build #28 of osx1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/28 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 01:03 |
-!- travis-ci [~travis-ci@ec2-174-129-166-170.compute-1.amazonaws.com] has joined #shogun | 01:08 | |
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/10469839 | 01:08 |
-!- travis-ci [~travis-ci@ec2-174-129-166-170.compute-1.amazonaws.com] has left #shogun [] | 01:08 | |
shogun-buildbot | build #1284 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1284 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Roman Votyakov <votjakovr@gmail.com> | 01:27 |
shogun-buildbot | build #1463 of deb2 - static_interfaces is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1463 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 01:53 |
shogun-buildbot | build #1577 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/1577 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com>, Kevin <kevinhughes27@gmail.com> | 02:19 |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Read error: Connection reset by peer] | 02:37 | |
-!- wiking_ [~wiking@info2k1.hu] has joined #shogun | 02:37 | |
-!- pickle27 [~kevin@d67-193-243-174.home3.cgocable.net] has joined #shogun | 02:52 | |
-!- hushell [~hushell@8-92.ptpg.oregonstate.edu] has quit [Ping timeout: 276 seconds] | 03:07 | |
-!- travis-ci [~travis-ci@ec2-174-129-166-170.compute-1.amazonaws.com] has joined #shogun | 03:24 | |
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/10472615 | 03:24 |
-!- travis-ci [~travis-ci@ec2-174-129-166-170.compute-1.amazonaws.com] has left #shogun [] | 03:24 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 03:33 | |
-!- pickle27 [~kevin@d67-193-243-174.home3.cgocable.net] has left #shogun [] | 03:34 | |
-!- hushell [~hushell@c-98-232-178-161.hsd1.or.comcast.net] has joined #shogun | 03:49 | |
-!- zxtx [~zv@rrcs-76-79-81-162.west.biz.rr.com] has quit [Ping timeout: 276 seconds] | 05:25 | |
-!- hushell [~hushell@c-98-232-178-161.hsd1.or.comcast.net] has quit [Quit: WeeChat 0.3.7] | 07:48 | |
-!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has joined #shogun | 08:25 | |
-!- wiking_ is now known as wiking | 08:27 | |
-!- wiking [~wiking@info2k1.hu] has quit [Changing host] | 08:27 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 08:27 | |
-!- mode/#shogun [+o wiking] by ChanServ | 08:27 | |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 08:33 | |
shogun-notifier- | shogun: Evgeniy Andreev :develop * 83e4b44 / / (4 files): https://github.com/shogun-toolbox/shogun/commit/83e4b444bd5972fce5a36620f463b9e036abeb1b | 08:33 |
shogun-notifier- | shogun: add LibSVMFile | 08:33 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * a004341 / / (4 files): https://github.com/shogun-toolbox/shogun/commit/a004341409aae8a396e70018769a2941eab70d0c | 08:33 |
shogun-notifier- | shogun: Merge pull request #1450 from gsomix/feature/libsvm | 08:33 |
shogun-notifier- | shogun: | 08:33 |
shogun-notifier- | shogun: Add LibSVMFile | 08:33 |
shogun-notifier- | shogun: van51 :develop * 0c2023b / / (6 files): https://github.com/shogun-toolbox/shogun/commit/0c2023b9170c4cd5b08b6c7667b671cae7b1ca93 | 08:33 |
shogun-notifier- | shogun: Added quadratic support in HashedDocDotFeatures | 08:33 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 30caf36 / / (6 files): https://github.com/shogun-toolbox/shogun/commit/30caf361c095e206d356dd983f0e6d6b8a5f84f6 | 08:33 |
shogun-notifier- | shogun: Merge pull request #1342 from van51/feature/quadratic | 08:33 |
shogun-notifier- | shogun: | 08:33 |
shogun-notifier- | shogun: K-skip ngrams for quadratic support in HashedDoc classes | 08:33 |
shogun-buildbot | build #1464 of deb2 - static_interfaces is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1464 blamelist: van51 <vangelis_51@hotmail.com> | 08:44 |
gsomix | sonney2k, around? | 08:48 |
shogun-buildbot | build #1495 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1495 blamelist: van51 <vangelis_51@hotmail.com> | 08:51 |
shogun-buildbot | build #29 of osx1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/29 blamelist: van51 <vangelis_51@hotmail.com> | 08:51 |
-!- lisitsyn [~lisitsin@mxs.kg.ru] has joined #shogun | 09:22 | |
-!- travis-ci [~travis-ci@ec2-54-234-140-21.compute-1.amazonaws.com] has joined #shogun | 09:23 | |
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/10482544 | 09:23 |
-!- travis-ci [~travis-ci@ec2-54-234-140-21.compute-1.amazonaws.com] has left #shogun [] | 09:23 | |
shogun-buildbot | build #1285 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1285 blamelist: van51 <vangelis_51@hotmail.com> | 09:25 |
shogun-buildbot | build #1465 of deb2 - static_interfaces is complete: Failure [failed git] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/1465 blamelist: Evgeniy Andreev <gsomix@gmail.com>, Soeren Sonnenburg <sonne@debian.org> | 09:27 |
shogun-buildbot | build #1580 of deb3 - modular_interfaces is complete: Failure [failed git] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1580 blamelist: Evgeniy Andreev <gsomix@gmail.com>, Soeren Sonnenburg <sonne@debian.org> | 09:27 |
shogun-buildbot | build #1496 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1496 blamelist: Evgeniy Andreev <gsomix@gmail.com>, Soeren Sonnenburg <sonne@debian.org> | 09:29 |
shogun-buildbot | build #30 of osx1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/osx1%20-%20libshogun/builds/30 blamelist: Evgeniy Andreev <gsomix@gmail.com>, Soeren Sonnenburg <sonne@debian.org> | 09:29 |
shogun-buildbot | build #961 of rpm1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/961 blamelist: van51 <vangelis_51@hotmail.com> | 09:30 |
-!- travis-ci [~travis-ci@ec2-174-129-166-170.compute-1.amazonaws.com] has joined #shogun | 09:31 | |
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/10482586 | 09:31 |
-!- travis-ci [~travis-ci@ec2-174-129-166-170.compute-1.amazonaws.com] has left #shogun [] | 09:31 | |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun | 09:48 | |
shogun-buildbot | build #1286 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1286 blamelist: Evgeniy Andreev <gsomix@gmail.com>, Soeren Sonnenburg <sonne@debian.org> | 09:54 |
thoralf | sonney2k: regarding https://github.com/shogun-toolbox/shogun/pull/1424: you wrote "keep the static one" - what do you mean? | 09:55 |
-!- gsomix_ [~gsomix@80.234.19.54] has joined #shogun | 09:56 | |
-!- gsomix [~gsomix@109.169.154.231] has quit [Ping timeout: 264 seconds] | 10:00 | |
-!- foulwall [~user@2001:da8:215:6901:ec1a:e681:d5fa:d54e] has joined #shogun | 10:07 | |
shogun-buildbot | build #962 of rpm1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/962 blamelist: Evgeniy Andreev <gsomix@gmail.com>, Soeren Sonnenburg <sonne@debian.org> | 10:11 |
-!- foulwall [~user@2001:da8:215:6901:ec1a:e681:d5fa:d54e] has quit [Ping timeout: 245 seconds] | 11:27 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 11:33 | |
@wiking | why do we assume in case of serialisation that a template class will always get a builtintype as template arg? | 11:43 |
lisitsyn | wiking: all template classes are specialized in C++ IIRC | 11:44 |
lisitsyn | err | 11:44 |
lisitsyn | .cpp | 11:44 |
lisitsyn | :D | 11:44 |
lisitsyn | not C++ | 11:44 |
@wiking | lisitsyn: i'm talking about class_list.cpp | 11:46 |
@wiking | lisitsyn: so it's generated in a way that if there's a TEMPLATE | 11:46 |
@wiking | then it'll generate like TemplateClass<PT_*> | 11:46 |
@wiking | e.g. TemplateClass<bool> etc. | 11:47 |
lisitsyn | yes | 11:47 |
lisitsyn | what's wrong? | 11:47 |
@wiking | but what if a template is not like that?:) | 11:47 |
lisitsyn | MyClass<OtherClass>? | 11:47 |
@wiking | so that it's template argument is actually not a PT_* | 11:47 |
@wiking | yeah something like that | 11:47 |
lisitsyn | throw new FuckYouException(); | 11:48 |
lisitsyn | jk :D | 11:48 |
@wiking | :D | 11:48 |
lisitsyn | wiking: well it is not supported as Soeren used templates only to provide various feature types | 11:48 |
-!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has left #shogun ["WeeChat 0.4.0"] | 13:01 | |
@sonney2k | thoralf, the one from CMath:: | 13:49 |
@sonney2k | thoralf, it is very convenient to just write CMath::random / shuffle w/o having to care about anything | 13:50 |
@sonney2k | gsomix_, pon | 13:50 |
@sonney2k | g | 13:50 |
thoralf | sonney2k: Okay, thanks. | 13:50 |
@sonney2k | wiking, and pong to you too | 13:53 |
@sonney2k | wiking, any news on the workshop videos uploaded? | 13:54 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Ping timeout: 248 seconds] | 13:55 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 13:55 | |
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has joined #shogun | 15:06 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 15:06 | |
lisitsyn | iglesiasg: welcome to .se again | 15:12 |
@iglesiasg | lisitsyn, thank you! | 15:12 |
-!- mode/#shogun [+o lisitsyn] by iglesiasg | 15:12 | |
-!- mode/#shogun [-o lisitsyn] by ChanServ | 15:12 | |
-!- mode/#shogun [+o lisitsyn] by iglesiasg | 15:12 | |
-!- mode/#shogun [-o lisitsyn] by ChanServ | 15:12 | |
@iglesiasg | lisitsyn, ChanServ doesn't let you be admin, it is a bad guy | 15:13 |
lisitsyn | iglesiasg: I am too lazy to identify | 15:14 |
-!- pickle27 [~kevin@d67-193-243-174.home3.cgocable.net] has joined #shogun | 16:43 | |
lisitsyn | pickle27: yo howdy | 16:46 |
pickle27 | lisitsyn: hey! | 16:46 |
pickle27 | lisitsyn: I actually have a question for you | 16:47 |
pickle27 | I was looking into making a getter for the time separated covariance matrices as per the emails with Andreas | 16:48 |
pickle27 | but they are SGNDArrays | 16:48 |
pickle27 | so I made the getter but it is of no use from python_modular | 16:48 |
pickle27 | so wondering what the options are | 16:49 |
lisitsyn | pickle27: what do you mean by is of no use? | 16:49 |
pickle27 | well it returns an SGNDArray swig obj | 16:49 |
pickle27 | but there is no way to get the data from it | 16:49 |
lisitsyn | ahh | 16:49 |
pickle27 | the get_matrix returns a pointer | 16:49 |
pickle27 | so one option is to make a new method for NDArray that returns SGMatrix | 16:50 |
pickle27 | or make the getter take a number and return that MAtrix from the Array and not change the Array class | 16:50 |
lisitsyn | pickle27: something must be wrong with typemaps | 16:51 |
lisitsyn | that's essentially it | 16:51 |
lisitsyn | pickle27: could you show me the method? | 16:52 |
pickle27 | I don't think anything is wrong with the typemaps this situation just isn't accounted for | 16:52 |
pickle27 | here | 16:52 |
pickle27 | SGNDArray<float64_t> CFFSep::get_covs() const | 16:52 |
pickle27 | { | 16:52 |
pickle27 | return m_covs; | 16:52 |
pickle27 | } | 16:52 |
pickle27 | if I used it in python and print the return val I get: | 16:54 |
pickle27 | <Swig Object of type 'shogun::SGNDArray< double > *' at 0x245cdc0> | 16:54 |
pickle27 | and if I try to access the data using .array or using get_matrix(0) I get something like this: | 16:55 |
pickle27 | <Swig Object of type 'shogun::T_ALPHA_BETA_TABLE *' at 0x251adf8> | 16:55 |
lisitsyn | pickle27: looks quite strange indeed | 17:15 |
lisitsyn | alpha beta table is something from the outer space | 17:17 |
pickle27 | haha | 17:17 |
lisitsyn | how can that be here :) | 17:17 |
lisitsyn | I didn't even know we have such a class | 17:17 |
pickle27 | lisitsyn: I think its because its a pointer | 17:17 |
-!- foulwall [~user@2001:da8:215:c252:f035:404c:df34:eab8] has joined #shogun | 17:18 | |
lisitsyn | pickle27: as a workaround I can suggest to add get_cov(int) | 17:18 |
pickle27 | lisitsyn: thats what I was thinking | 17:18 |
lisitsyn | that returns slices of the n-d array | 17:18 |
pickle27 | lisitsyn: the other option would be to create a new method for SGNDArray that does the same | 17:19 |
pickle27 | but how to name it to avoid confusion with the C++ method that returns a pointer | 17:19 |
lisitsyn | pickle27: actually as we have typemaps we just have to fix it | 17:22 |
lisitsyn | iglesiasg: have you used nd array typemaps? | 17:22 |
pickle27 | lisitsyn: oh we actually do have a typemap for NDArray | 17:23 |
pickle27 | I didn't think we did | 17:23 |
pickle27 | lisitsyn: looks like we might only have a "from" and we need a "to" | 17:24 |
lisitsyn | pickle27: yeah I thought about it | 17:26 |
lisitsyn | pickle27: could you please check if it is really true | 17:26 |
lisitsyn | just check what is missed for sgndarray compared to sgmatrix | 17:26 |
pickle27 | lisitsyn: yeah I think I see it | 17:27 |
pickle27 | lisitsyn: this will have to be done for each modular interface yes? | 17:27 |
lisitsyn | pickle27: yeah unfortunately | 17:28 |
pickle27 | lisitsyn: okay I'll try and do for python today and maybe octave | 17:28 |
pickle27 | I may just submit an issue on github for the remaining for now | 17:29 |
lisitsyn | pickle27: yeah please submit it | 17:31 |
-!- lisitsyn [~lisitsin@mxs.kg.ru] has quit [Quit: Leaving.] | 17:39 | |
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 245 seconds] | 18:38 | |
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has joined #shogun | 18:47 | |
iglesiasg | lisitsyn, no, I don't think I have, why? | 18:48 |
pickle27 | iglesiasg: we were trying to see if anyone had tested them | 18:56 |
pickle27 | iglesiasg: I think I found the issue though - there was only an in typemap and not an out | 18:56 |
iglesiasg | pickle27, I just read the logs and got a bit updated | 18:56 |
iglesiasg | pickle27, how is it going with it? | 18:56 |
pickle27 | iglesiasg: actually just ran it for python and it looks good | 18:56 |
pickle27 | wiking: is there a way to force the interfaces to rebuild? I seem to have to keep rebuilding from scratch in order for them to update | 19:03 |
-!- lambday [67157e37@gateway/web/freenode/ip.103.21.126.55] has joined #shogun | 19:12 | |
-!- foulwall [~user@2001:da8:215:c252:f035:404c:df34:eab8] has quit [Ping timeout: 264 seconds] | 19:15 | |
-!- HeikoS [~heiko@nat-176-108.internal.eduroam.ucl.ac.uk] has joined #shogun | 19:46 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 19:46 | |
@HeikoS | hi | 19:47 |
@HeikoS | sonney2k, wiking do you know whether and how to do modular access to sparse matrices? | 19:47 |
@HeikoS | is that even supported yet? | 19:47 |
@HeikoS | gsomix_: are you around? | 19:51 |
gsomix_ | HeikoS, hey | 19:55 |
@HeikoS | gsomix_: hi | 19:55 |
gsomix_ | HeikoS, I should add our libsvm reader to modular interfaces. | 19:55 |
@HeikoS | just wondering what to do to make swig accept sparse matrices? | 19:55 |
@HeikoS | gsomix_: yep that would be nice, but currently even using python to create them | 19:55 |
@HeikoS | csc_matrix? | 19:56 |
@HeikoS | ah yes that did the job for me | 19:56 |
@HeikoS | gsomix_: I am asking you since you did some typemaps for sparse matrices | 19:57 |
gsomix_ | HeikoS, ahh, sorry. I thought you were about reading. =___= | 19:57 |
@HeikoS | gsomix_: will be interesting very soon for me! :) | 19:57 |
@HeikoS | but for now, just how to map | 19:57 |
gsomix_ | HeikoS, there is a little doc in swig_typemaps.i for python modular interface. | 19:59 |
gsomix_ | HeikoS, spmatrix_from_numpy func | 19:59 |
@HeikoS | yep there I found the csc | 20:00 |
@HeikoS | cool thanks | 20:00 |
gsomix_ | ok | 20:00 |
-!- lambday [67157e37@gateway/web/freenode/ip.103.21.126.55] has quit [Ping timeout: 250 seconds] | 20:04 | |
@sonney2k | HeikoS, yes well just use a sparse structure | 20:11 |
@HeikoS | sonney2k: doesnt work, have to convert to csc | 20:11 |
@sonney2k | HeikoS, any news about your bank account? | 20:11 |
@HeikoS | but that works | 20:11 |
@HeikoS | sonney2k: parents return end of august | 20:12 |
@sonney2k | HeikoS, so one more week | 20:12 |
@HeikoS | yep | 20:12 |
@HeikoS | the it works immediately | 20:12 |
@sonney2k | HeikoS, which language python? | 20:12 |
@HeikoS | yes | 20:12 |
@sonney2k | HeikoS, then look at the python examples that use sparse | 20:12 |
@HeikoS | thats what I did finally | 20:13 |
@sonney2k | HeikoS, but I guess I don't really understand the question | 20:13 |
@sonney2k | well you should do that as a *first* step always right? | 20:13 |
@sonney2k | so it works now right? | 20:13 |
pickle27 | hey guys, I was working on the typemaps earlier, I added an out typemap for SGNDArray in python | 20:13 |
@HeikoS | the mapping SGSparseMatrix to scipy.sparsematrix doesnt work | 20:13 |
pickle27 | then I started looking at octave | 20:14 |
@HeikoS | only the csc(scipy.sparse) | 20:14 |
@HeikoS | anyway | 20:14 |
pickle27 | and I got a bit confused | 20:14 |
pickle27 | sonney2k: got a second to look at it with me? | 20:14 |
@sonney2k | HeikoS, yes sure only csc sparse matrix works | 20:14 |
@HeikoS | sonney2k: I did not know that :) | 20:15 |
@sonney2k | HeikoS, there are 3 formats and csc is the one we support | 20:15 |
@HeikoS | ok, going home now, see you! | 20:15 |
@sonney2k | HeikoS, wait | 20:15 |
@HeikoS | yes? | 20:15 |
@sonney2k | HeikoS, did you plan your trip already? | 20:15 |
@HeikoS | nope | 20:15 |
@HeikoS | nothing | 20:15 |
@sonney2k | I am asking because renting a car would make sense if we all use it | 20:16 |
@HeikoS | sonney2k: I am totally up for that! | 20:16 |
@HeikoS | are you around tomorrow? I am a bit late here and should go soon | 20:16 |
@HeikoS | ok gotta go, bye! I will be around tomorrow all day | 20:18 |
-!- HeikoS [~heiko@nat-176-108.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.] | 20:19 | |
pickle27 | sonney2k: can I get your help with the octave typemaps for a sec | 20:19 |
pickle27 | Im trying to add support for SGNDArray | 20:19 |
@sonney2k | pickle27, what is the issue? | 22:11 |
@sonney2k | gsomix_, and what was your question? | 22:16 |
@sonney2k | wiking, ? | 22:16 |
@sonney2k | wiking, what do you use to chat with under osx? | 22:17 |
@wiking | sonney2k: mmm i used to use | 22:17 |
@wiking | colloquy | 22:17 |
@wiking | but i'm irssi now | 22:17 |
@sonney2k | wiking, woah hardcore :) | 22:18 |
@sonney2k | wiking, did you manage to uploade some videos? | 22:19 |
gsomix_ | sonney2k, about SGVector. | 22:22 |
gsomix_ | sonney2k, is there nice method for work with SGVector+offsets? | 22:23 |
@sonney2k | gsomix_, errm?? | 22:23 |
@sonney2k | gsomix_, as in vec[i] ? | 22:23 |
@sonney2k | or vec[i+offs] ? | 22:23 |
gsomix_ | nope. look. I need substrings from SGVector. | 22:25 |
gsomix_ | without copying | 22:25 |
gsomix_ | on same memory | 22:25 |
gsomix_ | but with offset. | 22:25 |
gsomix_ | something like SGVector<char> substring=string.get_vector_with_offset(offset, length), where string is SGVector<char> | 22:26 |
@sonney2k | gsomix_, you could do dangerous stuff like offsetted= SGVector(vec.vector+offs, vec.vlen, false) | 22:26 |
gsomix_ | sonney2k, yep, without refcount. | 22:26 |
@sonney2k | yeah | 22:26 |
@sonney2k | with vec.vlen-offs | 22:26 |
gsomix_ | I'm just curious is there refcounted solution. | 22:27 |
@sonney2k | gsomix_, impossible! | 22:27 |
@sonney2k | well hmmhh | 22:27 |
gsomix_ | hm, I can imagine extended SGVector, that contains vector, physical length, offset and logical length. but in this case SGVector is not lightweight | 22:28 |
gsomix_ | but it would be very useful for parsing | 22:29 |
gsomix_ | now there are many, many copying and mallocs. | 22:30 |
gsomix_ | but, of course, I can use unsafe solution | 22:31 |
gsomix_ | sonney2k, ok, and I have another question. is there better solution for parsing numbers from strings than atoi/strtol? | 22:33 |
gsomix_ | I mean atoi/strtol requires zero-terminated strings | 22:33 |
gsomix_ | I cannot use it on pure memory block of data that read from file | 22:34 |
gsomix_ | sonney2k, maybe do you know solution for non-zero-terminated strings? for strings with defined length | 22:35 |
@sonney2k | gsomix_, no sorry - I would just put a '\0' there and change it back afterwards :/ | 22:42 |
@sonney2k | gsomix_, well a safe solution would be to keep one SGVector around | 22:43 |
@sonney2k | and then work with an offsetted SGVector | 22:43 |
@sonney2k | well or you derive some SGVectorView class from SGVector where you in addition store such an offset | 22:45 |
@sonney2k | SGVector is not *that* lightweight since it has a vtable already... | 22:45 |
gsomix_ | sonney2k, ok, thanks! | 22:46 |
@sonney2k | actually not derive but make SGVector a member of SGVectorView | 22:46 |
gsomix_ | sonney2k, it would be good solution for parsing. | 22:47 |
gsomix_ | I mean then I can store SGVectorView in dynarray's or so | 22:48 |
gsomix_ | sonney2k, one question about get_sparse_vector. | 22:49 |
gsomix_ | I should just read one line from libsvm file and parse it, right? | 22:49 |
gsomix_ | there is no reason to read whole file, hm | 22:50 |
@sonney2k | gsomix_, when you want to store that in dyn* you need to add some stuff to lib/memory.{cpp,h} | 22:53 |
@sonney2k | gsomix_, yeah for sparse it doesn't help because we store vectors separately anyway | 22:53 |
gsomix_ | sonney2k, do you mean I should enable SG_SPECIALIZED_MALLOC for new class? because it's not SGObject, right? | 22:56 |
@sonney2k | gsomix_, well except if you parse it into a dense SGMatrix | 22:56 |
@sonney2k | gsomix_, it is because it needs inplace new operator :/ | 22:59 |
gsomix_ | sonney2k, ok, I'll try to get into it. | 23:02 |
gsomix_ | now need finish sparse reading for libsvm and csv | 23:02 |
@wiking | sonney2k: yeah i have the python script + client api code from googs | 23:03 |
@wiking | so i'm on it | 23:03 |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 23:22 | |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 7a45b84 / NEWS: https://github.com/shogun-toolbox/shogun/commit/7a45b84e395f4de3c8a5ee6e583b40433febc807 | 23:22 |
shogun-notifier- | shogun: update news | 23:22 |
@wiking | sonney2k what ya think about breaking up the swig modular interface into smaller _wrapper.cxx-es | 23:28 |
@wiking | at least then the compilation could be done parallel (with make -jX) | 23:29 |
gsomix_ | wiking, I'm not Soeren, but would be great. | 23:30 |
gsomix_ | wiking, but how? | 23:30 |
@wiking | sonney2k: and i was thiking that we'd need need a nightly of osx | 23:30 |
@wiking | gsomix_: http://www.swig.org/Doc2.0/Modules.html#Modules_introduction | 23:30 |
--- Log closed Fri Aug 23 00:00:27 2013 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!