--- Log opened Thu Sep 05 00:00:45 2013 | ||
-!- travis-ci [~travis-ci@ec2-107-22-29-9.compute-1.amazonaws.com] has joined #shogun | 00:06 | |
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/10990829 | 00:06 |
---|---|---|
-!- travis-ci [~travis-ci@ec2-107-22-29-9.compute-1.amazonaws.com] has left #shogun [] | 00:06 | |
-!- zxtx [~zv@rrcs-74-62-200-195.west.biz.rr.com] has joined #shogun | 00:12 | |
-!- travis-ci [~travis-ci@ec2-54-234-56-9.compute-1.amazonaws.com] has joined #shogun | 00:20 | |
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/10990467 | 00:20 |
-!- travis-ci [~travis-ci@ec2-54-234-56-9.compute-1.amazonaws.com] has left #shogun [] | 00:20 | |
-!- zxtx [~zv@rrcs-74-62-200-195.west.biz.rr.com] has quit [Ping timeout: 245 seconds] | 00:20 | |
-!- lisitsyn [~lisitsyn@fb2-lo1.global63.net] has quit [Quit: Leaving.] | 00:21 | |
-!- hushell [~hushell@8-92.ptpg.oregonstate.edu] has joined #shogun | 00:23 | |
-!- travis-ci [~travis-ci@ec2-107-22-29-9.compute-1.amazonaws.com] has joined #shogun | 00:31 | |
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/10991344 | 00:31 |
-!- travis-ci [~travis-ci@ec2-107-22-29-9.compute-1.amazonaws.com] has left #shogun [] | 00:31 | |
-!- travis-ci [~travis-ci@ec2-107-22-29-9.compute-1.amazonaws.com] has joined #shogun | 00:34 | |
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/10992662 | 00:34 |
-!- travis-ci [~travis-ci@ec2-107-22-29-9.compute-1.amazonaws.com] has left #shogun [] | 00:34 | |
-!- zxtx [~zv@rrcs-76-79-81-162.west.biz.rr.com] has joined #shogun | 00:35 | |
shogun-buildbot | build #1385 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1385 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 00:36 |
-!- travis-ci [~travis-ci@ec2-107-22-29-9.compute-1.amazonaws.com] has joined #shogun | 00:41 | |
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/10993138 | 00:41 |
-!- travis-ci [~travis-ci@ec2-107-22-29-9.compute-1.amazonaws.com] has left #shogun [] | 00:41 | |
-!- hushell [~hushell@8-92.ptpg.oregonstate.edu] has quit [Ping timeout: 241 seconds] | 01:20 | |
-!- hushell [~hushell@8-92.ptpg.oregonstate.edu] has joined #shogun | 01:22 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 02:28 | |
-!- hushell [~hushell@8-92.ptpg.oregonstate.edu] has quit [Ping timeout: 264 seconds] | 02:29 | |
shogun-buildbot | build #514 of nightly_default is complete: Failure [failed configure] Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/514 | 03:28 |
-!- hushell [~hushell@c-98-232-178-161.hsd1.or.comcast.net] has joined #shogun | 03:28 | |
-!- sonne|osx_ [~sonne@f053037017.adsl.alicedsl.de] has joined #shogun | 03:33 | |
-!- sonne|osx [~sonne@f053040086.adsl.alicedsl.de] has quit [Ping timeout: 264 seconds] | 03:35 | |
-!- sonne|osx_ is now known as sonne|osx | 03:35 | |
-!- foulwall [~zhengyang@114.255.40.12] has joined #shogun | 03:57 | |
-!- foulwall [~zhengyang@114.255.40.12] has quit [Client Quit] | 03:57 | |
-!- foulwall [~zhengyang@114.255.40.12] has joined #shogun | 04:05 | |
-!- zxtx [~zv@rrcs-76-79-81-162.west.biz.rr.com] has quit [Ping timeout: 260 seconds] | 04:52 | |
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has quit [Quit: Ex-Chat] | 05:04 | |
-!- foulwall [~zhengyang@114.255.40.12] has quit [Ping timeout: 268 seconds] | 05:25 | |
-!- foulwall [~zhengyang@114.255.40.12] has joined #shogun | 05:25 | |
-!- shogunuser2 [180315fa@gateway/web/freenode/ip.24.3.21.250] has joined #shogun | 05:43 | |
shogunuser2 | Hello | 05:43 |
-!- shogunuser2 [180315fa@gateway/web/freenode/ip.24.3.21.250] has quit [Client Quit] | 05:45 | |
-!- besser82 [~besser82@fedora/besser82] has joined #shogun | 07:13 | |
-!- sonne|osx [~sonne@f053037017.adsl.alicedsl.de] has quit [Quit: sonne|osx] | 07:30 | |
-!- sonne|osx [~sonne@89.204.139.74] has joined #shogun | 08:11 | |
-!- lisitsyn [~lisitsyn@fb2-lo1.global63.net] has joined #shogun | 08:12 | |
lisitsyn | okay I am heading to moscow see you saturday :) | 08:12 |
lisitsyn | sonne|work: I'll mail you whether they approve my visa or not | 08:12 |
-!- lisitsyn [~lisitsyn@fb2-lo1.global63.net] has quit [Quit: Leaving.] | 08:19 | |
-!- sonne|osx [~sonne@89.204.139.74] has quit [Ping timeout: 264 seconds] | 08:28 | |
-!- sonne|osx [~sonne@89.204.139.213] has joined #shogun | 08:31 | |
-!- sonne|osx [~sonne@89.204.139.213] has quit [Client Quit] | 08:35 | |
-!- sonne|osx [~sonne@91-64-72-127-dynip.superkabel.de] has joined #shogun | 09:31 | |
-!- vgorbati [~vgorbati@gw.comsys.kpi.ua] has joined #shogun | 09:32 | |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 09:41 | |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * f53a70b / examples/undocumented/python_modular/ (7 files): https://github.com/shogun-toolbox/shogun/commit/f53a70b870a2274c6a64aa65c1f8f3cc71e6f51c | 09:41 |
shogun-notifier- | shogun: convert more examples to CSVFile / shogun infrastructure | 09:41 |
shogun-buildbot | build #1994 of deb1 - libshogun is complete: Failure [failed configure] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/1994 blamelist: Soeren Sonnenburg <sonne@debian.org> | 09:43 |
-!- sonne|osx [~sonne@91-64-72-127-dynip.superkabel.de] has quit [Ping timeout: 240 seconds] | 09:44 | |
sonne|work | lisitsyn good luck! | 09:47 |
-!- travis-ci [~travis-ci@ec2-23-20-152-199.compute-1.amazonaws.com] has joined #shogun | 09:49 | |
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/11005787 | 09:49 |
-!- travis-ci [~travis-ci@ec2-23-20-152-199.compute-1.amazonaws.com] has left #shogun [] | 09:49 | |
-!- vgorbati [~vgorbati@gw.comsys.kpi.ua] has quit [Read error: Connection reset by peer] | 09:51 | |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun | 09:51 | |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 38347f2 / CMakeLists.txt: https://github.com/shogun-toolbox/shogun/commit/38347f23ff42ea309daadd03c7b8c909b1062474 | 09:58 |
shogun-notifier- | shogun: drop duplicate JSON detection | 09:58 |
-!- vgorbati [~vgorbati@gw.comsys.kpi.ua] has joined #shogun | 10:01 | |
shogun-buildbot | build #1995 of deb1 - libshogun is complete: Success [build successful] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/1995 | 10:10 |
-!- vgorbati [~vgorbati@gw.comsys.kpi.ua] has quit [Quit: vgorbati] | 10:14 | |
-!- vgorbati [~vgorbati@gw.comsys.kpi.ua] has joined #shogun | 10:16 | |
thoralf | Hey. | 10:17 |
sonne|work | Ho HOoa3o4j! | 10:18 |
-!- vgorbati [~vgorbati@gw.comsys.kpi.ua] has quit [Read error: Connection reset by peer] | 10:18 | |
thoralf | sonne|work: You're okay? ;) | 10:25 |
sonne|work | thoralf: mushroom season you know %-) | 10:25 |
thoralf | sonne|work: Yesterday we didn't finish talking about my PR. | 10:25 |
thoralf | sonne|work: lol :) | 10:25 |
thoralf | sonne|work: We can drop it, no problem. But it would be cool to learn whats wrong. | 10:26 |
thoralf | sonne|work: https://github.com/shogun-toolbox/shogun/pull/1525 | 10:27 |
shogun-buildbot | build #1606 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1606 blamelist: Soeren Sonnenburg <sonne@debian.org> | 10:35 |
-!- iglesiasg [~iglesias@2001:6b0:1:1da0:ed30:b56b:ea5b:1557] has joined #shogun | 10:42 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 10:42 | |
-!- travis-ci [~travis-ci@ec2-23-20-152-199.compute-1.amazonaws.com] has joined #shogun | 10:51 | |
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/11006281 | 10:51 |
-!- travis-ci [~travis-ci@ec2-23-20-152-199.compute-1.amazonaws.com] has left #shogun [] | 10:51 | |
@iglesiasg | hello comrades | 10:55 |
shogun-buildbot | build #1689 of deb3 - modular_interfaces is complete: Failure [failed install test python_modular] Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1689 blamelist: Soeren Sonnenburg <sonne@debian.org> | 11:10 |
shogun-buildbot | build #1386 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1386 blamelist: Soeren Sonnenburg <sonne@debian.org> | 11:12 |
-!- van51 [~van51@athedsl-410351.home.otenet.gr] has joined #shogun | 11:19 | |
-!- jey1401 [8268e41e@gateway/web/freenode/ip.130.104.228.30] has joined #shogun | 11:27 | |
jey1401 | Hi everybody | 11:28 |
@iglesiasg | Hi | 11:29 |
jey1401 | I had a hard time installing the r_static interface of shogun 2.1.0 on R3.0.1. I've foud 2 bugs one in the ./configure file and one in the ./src/interfaces/r_static/RInterface.cpp file | 11:30 |
@iglesiasg | ahm I see | 11:30 |
@iglesiasg | We are not supporting the configure script any longer since we switched to CMake | 11:30 |
jey1401 | I already submitted a pull request for the RInterface.cpp file. I'd like to do so for the ./configure file but I can't find it on github | 11:31 |
jey1401 | Mhh ok | 11:31 |
jey1401 | so let's hope it'll be completely fixed in the next release :-) | 11:32 |
jey1401 | thanks ! | 11:32 |
@iglesiasg | Unfortunately we do not have the static interfaces part in CMake yet | 11:32 |
@iglesiasg | But we are looking forward to doing it | 11:32 |
@iglesiasg | Would you be interested in doing that? | 11:33 |
@iglesiasg | None is working on that right now, as far as I know | 11:33 |
thoralf | wiking: Found a bug in our cmake setup: class_list.cpp is generated outside the build directory (and not updated when header files are removed) | 11:34 |
jey1401 | I'm afraid I don't have time and skill to do that. I'm not fluent at all in c/c++... | 11:34 |
@iglesiasg | jey1401, aham I see. I don't think actually that requires much C++, but CMake skills | 11:35 |
@iglesiasg | anyway, I hope we will get static back some time soon | 11:35 |
thoralf | wiking: What triggers the update of class_list.cpp? | 11:37 |
jey1401 | I hope so, shogun is realy a nice tool. Anyway, thanks for your fast answers. | 11:39 |
jey1401 | See you | 11:39 |
@iglesiasg | jey1401, Thanks! Hope you can use Shogun for your work. Good luck! | 11:39 |
-!- jey1401 [8268e41e@gateway/web/freenode/ip.130.104.228.30] has quit [Quit: Page closed] | 11:40 | |
-!- HeikoS [~heiko@nat-174-124.internal.eduroam.ucl.ac.uk] has joined #shogun | 11:46 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 11:46 | |
thoralf | Why are most of the classes prefixed with "C"? Does it have any special meaning? | 11:46 |
thoralf | Hey HeikoS | 11:46 |
@iglesiasg | C stands for Class I believe | 11:46 |
thoralf | iglesiasg: But there is some magic in the compile process, right? | 11:47 |
sonne|work | thoralf: historically but yes - nowadays we serialize classes prefixed with C | 11:47 |
@HeikoS | thoralf: hi! | 11:47 |
thoralf | iglesiasg: It's different if I add a C or remove it. | 11:47 |
@iglesiasg | thoralf, mm not AFAIK | 11:47 |
@HeikoS | nah NIPS rejected :( | 11:47 |
@iglesiasg | HeikoS, oh, I am sorry man | 11:47 |
@iglesiasg | the paper you sent? | 11:48 |
@HeikoS | yes | 11:48 |
@iglesiasg | they are the ones who are missing one your talks now! | 11:48 |
@HeikoS | haha, yeah :) I guess thats life | 11:48 |
@HeikoS | now we pimp it and send it to ICML, which is in Beijing :) | 11:48 |
@iglesiasg | what reason did they give you? | 11:48 |
shogun-notifier- | shogun: Thoralf Klein :develop * b7b0ba7 / src/shogun/lib/SGSparseVector.cpp: https://github.com/shogun-toolbox/shogun/commit/b7b0ba73b8698207d9256ac90b1d1292299c315b | 11:49 |
shogun-notifier- | shogun: Removing code duplication by proper templating&specialization. | 11:49 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * b78662a / src/shogun/lib/SGSparseVector.cpp: https://github.com/shogun-toolbox/shogun/commit/b78662acc2cdcae662cc30bb8bdab701be2a82db | 11:49 |
shogun-notifier- | shogun: Merge pull request #1525 from tklein23/fix_sparse_templates | 11:49 |
shogun-notifier- | shogun: | 11:49 |
shogun-notifier- | shogun: Removing code duplication by proper templating&specialization. | 11:49 |
@HeikoS | iglesiasg: not well presented why its useful | 11:49 |
@HeikoS | iglesiasg: they got a point though, the most important experiments were in the rebuttal | 11:49 |
sonne|work | HeikoS: will you go to nips nevertheless? | 11:49 |
thoralf | sonne|work: With "C" I get strange compiler errors regading to templates - and the disappear when removing the C. | 11:50 |
@HeikoS | sonne|work: yes I think, I will try to give a workshop talk and also this MLOSS thing | 11:50 |
thoralf | sonne|work: Something within class_list.cpp | 11:50 |
@HeikoS | when is wiking back=? | 11:50 |
@iglesiasg | one week wedding + two weeks hangover :P | 11:50 |
@iglesiasg | just kidding | 11:50 |
sonne|work | thoralf: yeah it assumes CMyClass is derived form CSGObject | 11:51 |
sonne|work | HeikoS: so 2 workshop talks :) | 11:51 |
@HeikoS | well dont know whether this works, just trying | 11:51 |
besser82 | HeikoS: Hi! Is there any special benefit when compiling shugon with NLopt? | 11:53 |
thoralf | sonne|work: It actually is. But it's generating code without the needed <template> when calling the constructor. | 11:53 |
sonne|work | thoralf: so these C prefixed classes will be serialized | 11:53 |
thoralf | (I'm not quite sure, since the class_list.cpp is not always rebuilt) | 11:54 |
shogun-buildbot | build #1607 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1607 blamelist: Soeren Sonnenburg <sonne@debian.org> | 12:02 |
@HeikoS | besser82: yes, we use that for GP model-selection | 12:03 |
@HeikoS | which is crucial for GPs | 12:03 |
besser82 | HeikoS: thanks! What about ColPack and LaPack95? Anything essential or important about them? I'm asking because those would be need to packaged for Fedora, too.... | 12:05 |
@HeikoS | besser82: ColPack is for graph colourings and is used by the GSoC project of lambday which involves sampling log-determinants of large, sparse matrices. It is essential for this part of the framework, but the framework itself is not as mainstream as GPs for example | 12:06 |
@HeikoS | lapack is used in many places, so that should be in | 12:06 |
@HeikoS | colpack is bundled by cmake since its also not a debian package | 12:06 |
besser82 | HeikoS: lapack is in avail. ;) But I'm asking specially about LaPack95 | 12:07 |
besser82 | HeikoS: bundling / donwloads during build is not allowed for Fedora, so I need to prepare them seperately. | 12:08 |
@HeikoS | besser82: I see, ColPack would be good to have since the log-det project might be popular once done | 12:08 |
@HeikoS | I have no idea about LaPack95 though | 12:08 |
besser82 | HeikoS: any idea who might have an idea? | 12:09 |
shogun-buildbot | build #1608 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1608 blamelist: Thoralf Klein <thoralf.klein@zib.de> | 12:10 |
@HeikoS | sonne|work: LaPack95, what is that used for | 12:10 |
@HeikoS | besser82: LAPACK95 -- Fortran95 interface to LAPACK | 12:11 |
@HeikoS | I dont think there is any fortran code in shogun | 12:11 |
@HeikoS | currently searching for "fortran95" in all source files, lets see | 12:12 |
besser82 | HeikoS: I know about that ;) But cmake checks for this.... | 12:12 |
@HeikoS | besser82: the only file where this is in is here | 12:13 |
@HeikoS | # This module was modified by CGAL team: | 12:13 |
@HeikoS | # - find libraries for a C++ compiler, instead of Fortran | 12:13 |
@HeikoS | # - added LAPACK_INCLUDE_DIR, LAPACK_DEFINITIONS and LAPACK_LIBRARIES_DIR | 12:13 |
@HeikoS | # - removed LAPACK95_LIBRARIES | 12:13 |
thoralf | HeikoS: Is it possible to exclude a class from your clone_unittest? (Or at least to limit the template classes it tries?) | 12:13 |
@HeikoS | its in the FindLapack.cmake file | 12:13 |
@HeikoS | thoralf: yes there is a blacklist option | 12:13 |
@HeikoS | in the python script that generates them, should be three classes in there | 12:14 |
@HeikoS | thoralf: why do you want to exclude? | 12:14 |
thoralf | HeikoS: I'm using a sparse label class - labels are array subscripts. | 12:14 |
thoralf | HeikoS: Does not make sense to use float. | 12:14 |
thoralf | for example. | 12:14 |
@HeikoS | thoralf: uh | 12:15 |
@HeikoS | I dont know whether thats possible | 12:15 |
@HeikoS | but you can try :) | 12:15 |
thoralf | HeikoS: Just exclude would be enough :) | 12:15 |
besser82 | HeikoS: thanks for the infos. Seems the "LAPACK95_FOUND=FALSE" is a bogus message from cmake then ;) | 12:26 |
@HeikoS | besser82: yep should be fine just ignoring it | 12:26 |
besser82 | HeikoS: since it finds clapack from ATLAS ;) | 12:27 |
@HeikoS | ok cool | 12:27 |
shogun-buildbot | build #1690 of deb3 - modular_interfaces is complete: Success [build successful] Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1690 | 12:32 |
shogun-buildbot | build #1387 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1387 blamelist: Soeren Sonnenburg <sonne@debian.org> | 12:41 |
shogun-buildbot | build #1691 of deb3 - modular_interfaces is complete: Failure [failed install test python_modular] Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1691 blamelist: Thoralf Klein <thoralf.klein@zib.de> | 12:50 |
sonne|work | HeikoS: well lapack is fortran so shogun might use it | 12:54 |
besser82 | sonne|work: I explicitly asked for lapack95 ;) Because cmake tells me it's not found (non-fatal), but detects clapack from ATLAS.... | 13:12 |
besser82 | sonne|work: so i'm wondering if lapack95 there is needed by sth. explicitly.... | 13:12 |
shogun-buildbot | build #1388 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1388 blamelist: Thoralf Klein <thoralf.klein@zib.de> | 13:25 |
-!- vgorbati [c3ee5cb1@gateway/web/freenode/ip.195.238.92.177] has joined #shogun | 13:34 | |
sonne|work | besser82: no atlas includes lapack and this and eigen3 is what you want | 13:43 |
sonne|work | van51: hey any news? | 13:43 |
van51 | sonne|work: hey | 13:44 |
van51 | sonne|work: yeah I have updated the PR and was waiting for travis | 13:44 |
besser82 | sonne|work: so atlas && eigen3 is sufficient?!? | 13:44 |
sonne|work | van51: I meant did you re-run the benchmark? | 13:44 |
van51 | sonne|work: yeah for 100k examples | 13:45 |
van51 | sonne|work: and the dot-variant was still faster | 13:45 |
van51 | sonne|work: I have those results under the benchmark code | 13:45 |
sonne|work | besser82: for numerical libs yes. I mean of course there are some parts that need more. E.g. glpk for 1-norm MKL and IIRC nlopt for some parts of GPs (right HeikoS?) | 13:45 |
sonne|work | van51: ahh ok nice - what did you do? | 13:45 |
@HeikoS | nlopt has to be in for gps, yes | 13:46 |
van51 | sonne|work: https://github.com/van51/shogun/blob/c5d585c57f5789fc06596dbf0fab41fc69d5c905/benchmarks/rf_feats_benchmark.cpp | 13:46 |
besser82 | sonne|work: the glpk-thing is clear; my question was more about atlas && eigen3 is sufficient to "substitute" this obscure lapack95 | 13:47 |
sonne|work | van51: how many CPUs do you have - looks like it paralleized very well (speedup > 8x) | 13:47 |
sonne|work | besser82: yes 110% | 13:47 |
van51 | sonne|work: I 'm using a server from uni with 8 cpus :) | 13:48 |
besser82 | sonne|work: ^^, thanks. | 13:48 |
sonne|work | van51: ahh so speedup is 7.9 only - my bad :) | 13:48 |
sonne|work | but still very good to know | 13:48 |
van51 | I'm going to have lunch now | 13:51 |
van51 | sonne|work: let me know if something is wrong :) | 13:51 |
sonne|work | van51: heh tell me what you did to make it work with arbitrary DF's | 13:51 |
sonne|work | van51: enjoy your meal | 13:51 |
van51 | sonne|work: well the place where it mattered what the underlying features were, was when we wanted to compute the dot product of w' * x | 14:19 |
van51 | sonne|work: so instead of getting the vector x through a get_feature_vector(idx) method and then computing the dot product | 14:19 |
van51 | sonne|work: I just did feats->dense_dot(i , w.vector, w.vlen) | 14:20 |
@iglesiasg | sonne|work, ping | 14:20 |
@iglesiasg | sonne|work, can I get kcachegrind in the fatbot pleease? :) | 14:21 |
@iglesiasg | for some reason valgrind breaks when I compile Shogun in Release and I profiling LMNN with the fatbot | 14:21 |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 14:49 | |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 15:10 | |
shogun-notifier- | shogun: jey1401 :develop * 359c1c4 / src/interfaces/r_static/RInterface.cpp: https://github.com/shogun-toolbox/shogun/commit/359c1c4ab32d6ac3a8d1c38617abb221cc570004 | 15:10 |
shogun-notifier- | shogun: Update RInterface : correction of 'number of arguments' bug | 15:10 |
shogun-notifier- | shogun: | 15:10 |
shogun-notifier- | shogun: Since R3.0 the number of arguments of External calls is checked for each call. sg has an unfixed number of argument. Fixing the number of arguments to -1 in the registration allows to disable that check. | 15:10 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 804d6c9 / src/interfaces/r_static/RInterface.cpp: https://github.com/shogun-toolbox/shogun/commit/804d6c92c3be45aa4ba78a26f57f313cc15cdef8 | 15:10 |
shogun-notifier- | shogun: Merge pull request #1534 from jey1401/patch-1 | 15:10 |
shogun-notifier- | shogun: | 15:10 |
shogun-notifier- | shogun: Update RInterface : correction of 'number of arguments' bug | 15:10 |
-!- van51 [~van51@athedsl-410351.home.otenet.gr] has quit [Quit: Leaving.] | 15:16 | |
shogun-buildbot | build #1609 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1609 blamelist: Soeren Sonnenburg <sonne@debian.org> | 15:28 |
thoralf | sonne|work: Something went wrong when complex64_t was introduced. | 15:38 |
thoralf | sonne|work: When calling CMath::max() I'm getting errors related to complex64_t | 15:38 |
sonne|work | thoralf: well what is max for a complex nr? | 15:39 |
thoralf | shogun-buildbot: I need to specify CMath::max<type> | 15:39 |
thoralf | sonne|work: CMath::max(int32_t, int32_t); | 15:40 |
thoralf | cannot convert 'complex64_t {aka std::complex<double>}' to 'shogun::int32_t {aka short int}' in assignment | 15:41 |
thoralf | I'm providing integers, but it chooses the wrong template. | 15:41 |
sonne|work | thoralf: weird | 15:42 |
thoralf | When removing this max-complex macro, then inference works fine (for the other types). | 15:43 |
thoralf | I think the problem is, that you're overloading max instead of specializing it. | 15:43 |
shogun-notifier- | shogun: Fernando Iglesias :develop * 75aff69 / src/shogun/structure/DualLibQPBMSOSVM.cpp: https://github.com/shogun-toolbox/shogun/commit/75aff69ab83fcb9947a62e11d85d1153a14d3527 | 15:43 |
shogun-notifier- | shogun: Minor warning fix in the structured learning framework. | 15:43 |
shogun-notifier- | shogun: Comparison between signed and unsigned integers. | 15:43 |
shogun-notifier- | shogun: Fernando Iglesias :develop * f40a66a / src/shogun/structure/DualLibQPBMSOSVM.cpp: https://github.com/shogun-toolbox/shogun/commit/f40a66aeff25ca3607b2d1a6bfbfbdf5ae9de9a6 | 15:43 |
shogun-notifier- | shogun: Merge pull request #1535 from iglesias/feature/fix_warning | 15:43 |
shogun-notifier- | shogun: | 15:43 |
shogun-notifier- | shogun: Minor warning fix in the structured learning framework. | 15:43 |
sonne|work | thoralf: so why not overload it all the time then? | 15:45 |
sonne|work | thoralf: hmmh or template the complex thing | 15:47 |
-!- sonne|work [~sonnenbu@91-64-72-127-dynip.superkabel.de] has quit [Quit: Leaving.] | 15:52 | |
-!- sonne|work [~sonnenbu@91-64-72-127-dynip.superkabel.de] has joined #shogun | 15:52 | |
shogun-buildbot | build #1610 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1610 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 15:55 |
shogun-buildbot | build #1611 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1611 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 16:00 |
shogun-buildbot | build #1612 of bsd1 - libshogun is complete: Failure [failed compile] Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1612 blamelist: jey1401 <jey1401@gmail.com> | 16:02 |
shogun-buildbot | build #1692 of deb3 - modular_interfaces is complete: Failure [failed install test python_modular] Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1692 blamelist: Soeren Sonnenburg <sonne@debian.org> | 16:04 |
shogun-buildbot | build #1389 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1389 blamelist: Soeren Sonnenburg <sonne@debian.org> | 16:04 |
thoralf | sonne|work: I tried to fix it without templating the whole class, but it didn't work out. | 16:18 |
thoralf | sonne|work: Some versions didn't compile, others broke everything. ;) | 16:18 |
thoralf | sonne|work: Here is what I tried - maybe you see whats wrong: https://github.com/tklein23/shogun/compare/cmath_max_complex64_t | 16:21 |
thoralf | I think I'll stop trying, since it seems to be more complex then just fix some declarations. | 16:22 |
shogun-buildbot | build #1693 of deb3 - modular_interfaces is complete: Failure [failed install test libshogun test python_modular test java_modular test csharp_modular] Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1693 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 16:26 |
shogun-buildbot | build #1390 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1390 blamelist: jey1401 <jey1401@gmail.com>, Fernando Iglesias <fernando.iglesiasg@gmail.com> | 16:48 |
shogun-buildbot | build #1694 of deb3 - modular_interfaces is complete: Failure [failed install test python_modular] Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1694 blamelist: jey1401 <jey1401@gmail.com> | 16:52 |
shogun-notifier- | shogun: van51 :develop * 2fae496 / / (6 files): https://github.com/shogun-toolbox/shogun/commit/2fae496edc1df5c9d00423f44057e85dd0ca5a03 | 17:11 |
shogun-notifier- | shogun: Random Fourier Features | 17:11 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * a8358cc / / (6 files): https://github.com/shogun-toolbox/shogun/commit/a8358cc7db1a9f879c9adacf05b3de2cf31edf79 | 17:11 |
shogun-notifier- | shogun: Merge pull request #1466 from van51/feature/randomfeat | 17:11 |
shogun-notifier- | shogun: | 17:11 |
shogun-notifier- | shogun: Random Fourier Features | 17:11 |
shogun-notifier- | shogun: Kevin :develop * 7c532d1 / doc/ipython-notebooks/ica/ecg_sep.ipynb: https://github.com/shogun-toolbox/shogun/commit/7c532d1484bea275226358378fd825a791abc751 | 17:12 |
shogun-notifier- | shogun: updated ICA ecg notebook | 17:12 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 2ac823d / doc/ipython-notebooks/ica/ecg_sep.ipynb: https://github.com/shogun-toolbox/shogun/commit/2ac823dcc3a481063545da16264d8d33d49eb0d9 | 17:12 |
shogun-notifier- | shogun: Merge pull request #1532 from pickle27/develop | 17:12 |
shogun-notifier- | shogun: | 17:12 |
shogun-notifier- | shogun: updated ICA ecg notebook | 17:12 |
shogun-notifier- | shogun: Kevin :develop * f713eec / / (29 files): https://github.com/shogun-toolbox/shogun/commit/f713eec5465a000e1d452d1dea5dfed2a940e78b | 17:14 |
shogun-notifier- | shogun: updated ruby modular to use nmatrix, initial commit | 17:14 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 94c1d75 / / (29 files): https://github.com/shogun-toolbox/shogun/commit/94c1d75fb75820e45565afe236deb80c6acbdf7b | 17:14 |
shogun-notifier- | shogun: Merge pull request #1516 from pickle27/nmatrix | 17:14 |
shogun-notifier- | shogun: | 17:14 |
shogun-notifier- | shogun: updated ruby modular to use nmatrix | 17:14 |
shogun-buildbot | build #1695 of deb3 - modular_interfaces is complete: Failure [failed compile libshogun] Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1695 blamelist: Soeren Sonnenburg <sonne@debian.org>, van51 <vangelis_51@hotmail.com> | 17:20 |
shogun-buildbot | build #1613 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1613 blamelist: Soeren Sonnenburg <sonne@debian.org>, van51 <vangelis_51@hotmail.com> | 17:22 |
shogun-buildbot | build #1696 of deb3 - modular_interfaces is complete: Failure [failed configure] Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1696 blamelist: Soeren Sonnenburg <sonne@debian.org>, Kevin <kevinhughes27@gmail.com> | 17:28 |
shogun-buildbot | build #1614 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1614 blamelist: Soeren Sonnenburg <sonne@debian.org>, Kevin <kevinhughes27@gmail.com> | 17:33 |
-!- vgorbati [c3ee5cb1@gateway/web/freenode/ip.195.238.92.177] has quit [Quit: Page closed] | 17:33 | |
@HeikoS | iglesiasg: do you know how to specify my own swig version to shogun? | 17:49 |
@HeikoS | sonne|work, sonney2k ^ | 17:49 |
@iglesiasg | HeikoS, -D SWIG_EXECUTABLE=/usr/bin/swig2.0, for instance | 17:50 |
@iglesiasg | HeikoS, ccmake helps to find these options ;) | 17:50 |
@iglesiasg | I just did ccmake .. from the build directory | 17:51 |
@HeikoS | iglesiasg: thanks! | 17:51 |
@HeikoS | iglesiasg: not installed on this cluster I am trying :) | 17:51 |
@iglesiasg | HeikoS, welcome! :) | 17:51 |
@iglesiasg | HeikoS, cool, are you going to do something large scale in a cluster? | 17:51 |
@HeikoS | iglesiasg: yes | 17:51 |
@HeikoS | at least I am trying | 17:51 |
@HeikoS | going to allocate a few thousand nodes | 17:51 |
@HeikoS | and test my PBS computation engine | 17:52 |
@HeikoS | which implements lambdays interface using batch clusters | 17:52 |
@HeikoS | and I need this for my ozone experiements | 17:52 |
@HeikoS | which I have to present next week ;) | 17:52 |
-!- van51 [~van51@athedsl-410351.home.otenet.gr] has joined #shogun | 17:54 | |
-!- hushell [~hushell@c-98-232-178-161.hsd1.or.comcast.net] has quit [Ping timeout: 240 seconds] | 17:58 | |
shogun-buildbot | build #1391 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1391 blamelist: Soeren Sonnenburg <sonne@debian.org>, van51 <vangelis_51@hotmail.com> | 18:00 |
shogun-buildbot | build #1615 of bsd1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/1615 blamelist: Kevin <kevinhughes27@gmail.com> | 18:08 |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 260 seconds] | 18:35 | |
shogun-buildbot | build #1697 of deb3 - modular_interfaces is complete: Failure [failed install test python_modular] Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1697 blamelist: Kevin <kevinhughes27@gmail.com> | 18:43 |
shogun-buildbot | build #1392 of cyg1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1392 blamelist: Soeren Sonnenburg <sonne@debian.org>, Kevin <kevinhughes27@gmail.com> | 18:43 |
-!- hushell [~hushell@c-98-232-178-161.hsd1.or.comcast.net] has joined #shogun | 18:54 | |
-!- travis-ci [~travis-ci@ec2-50-17-163-98.compute-1.amazonaws.com] has joined #shogun | 19:09 | |
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/11020799 | 19:09 |
-!- travis-ci [~travis-ci@ec2-50-17-163-98.compute-1.amazonaws.com] has left #shogun [] | 19:09 | |
-!- sonne|osx [~sonne@89.204.139.247] has joined #shogun | 20:01 | |
sonne|osx | van51: for some reason github didn't mark your PR updated | 20:03 |
sonne|osx | took me a while to figure out that you actually added a test | 20:03 |
van51 | sonne|osx: hmm was that making a difference? | 20:04 |
sonne|osx | van51: I was waiting for you to update the PR but only realized that you did today afternoon | 20:05 |
van51 | sonne|osx: ah now I understand.. yea I don't think it marks it if you just update | 20:07 |
van51 | sonne|osx: I should have commented afterwards like iglesiasg has suggested at another time | 20:07 |
@iglesiasg | aham! So it is not a faulty GitHub configuration of mine | 20:08 |
van51 | iglesiasg: I'd say not, but I'm not that familiar :p | 20:08 |
-!- sonne|osx [~sonne@89.204.139.247] has quit [Ping timeout: 276 seconds] | 20:10 | |
-!- sonne|osx [~sonne@89.204.135.19] has joined #shogun | 20:11 | |
sonne|osx | grpmf | 20:12 |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 20:14 | |
sonne|osx | van51: please fix the warnings | 20:18 |
sonne|osx | http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1697/steps/compile%20python_modular/logs/warnings%20%2886%29 | 20:18 |
sonne|osx | ^ van51 these ones | 20:18 |
sonne|osx | not all are yours of course but you will recognize the classes that are :) | 20:19 |
van51 | sonne|osx: ah yeah! sure, I'm on it ;) | 20:19 |
* sonne|osx off | 20:19 | |
-!- sonne|osx [~sonne@89.204.135.19] has quit [Quit: sonne|osx] | 20:19 | |
-!- vgorbati [~vgorbati@91.216.173.29] has joined #shogun | 20:28 | |
-!- HeikoS [~heiko@nat-174-124.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.] | 20:35 | |
-!- iglesiasg [~iglesias@2001:6b0:1:1da0:ed30:b56b:ea5b:1557] has quit [Quit: Ex-Chat] | 20:36 | |
-!- sonne|osx [~sonne@f053037017.adsl.alicedsl.de] has joined #shogun | 20:39 | |
shogun-buildbot | build #87 of osx1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/87 blamelist: Soeren Sonnenburg <sonne@debian.org>, Thoralf Klein <thoralf.klein@zib.de>, jey1401 <jey1401@gmail.com>, Kevin <kevinhughes27@gmail.com>, Fernando Iglesias <fernando.iglesiasg@gmail.com>, van51 | 20:46 |
shogun-buildbot | <vangelis_51@hotmail.com> | 20:46 |
shogun-buildbot | build #1698 of deb3 - modular_interfaces is complete: Failure [failed configure] Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1698 | 20:59 |
-!- pickle27 [~Kevin@199.119.128.114] has joined #shogun | 21:00 | |
@sonney2k | pickle27, hey... | 21:00 |
pickle27 | sonney2k, hey | 21:00 |
@sonney2k | I wish wiking was here to help with this | 21:00 |
@sonney2k | pickle27, the nmatrix install instructions are pretty clear but hmhh | 21:01 |
pickle27 | sonney2k, oh for the build bot | 21:01 |
pickle27 | sonney2k, so we have a ruby modular build bot? I wasn't sure | 21:01 |
@sonney2k | pickle27, sure | 21:01 |
@sonney2k | pickle27, we are getting this far now http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/1698 | 21:02 |
@sonney2k | fail on configure | 21:02 |
pickle27 | sonney2k, how did you put nmatrix on the bot? | 21:04 |
@sonney2k | pickle27, not at all | 21:09 |
@sonney2k | pickle27, question is whether we should implement that in cmake downloading rather | 21:09 |
pickle27 | sonney2k, ah i see, that is a wiking question | 21:10 |
pickle27 | I think we should leave it, I mean we don't install NumPy if they are missing it do we? | 21:10 |
@sonney2k | pickle27, no but we install json-c, eigen3 etc | 21:13 |
@sonney2k | but yes it would be an option to not do it | 21:13 |
@sonney2k | OMFG | 21:13 |
@sonney2k | I just realize libshogun.so* is 250MB in size now? | 21:13 |
@sonney2k | wtf happened | 21:13 |
@sonney2k | IIRC we were at two digits before gsoc | 21:14 |
pickle27 | sonney2k, no idea | 21:14 |
pickle27 | i didn't do it :0 | 21:14 |
@sonney2k | pickle27, that is what you say :D | 21:17 |
pickle27 | haha | 21:17 |
pickle27 | alright I have to head out now, looking at apartments today | 21:18 |
* sonney2k tries to compile shogun 2.1 for reference | 21:18 | |
pickle27 | Im going to read the logs for the result, Im curious too | 21:19 |
@sonney2k | it will take just 1 min :D | 21:20 |
@sonney2k | 44M! | 21:21 |
@sonney2k | compile time with cold caches 1m17 | 21:21 |
@sonney2k | stripped 12M | 21:22 |
pickle27 | woah | 21:23 |
pickle27 | compile times have gotten pretty long | 21:23 |
@sonney2k | stripped libshogun now is 20M | 21:24 |
@sonney2k | pickle27, I guess due to using more templates than before. maybe eigen3 not sure | 21:26 |
@sonney2k | we will see | 21:26 |
@sonney2k | alright on may 13 everything was still ok | 21:27 |
-!- van51 [~van51@athedsl-410351.home.otenet.gr] has quit [Quit: Leaving.] | 21:28 | |
@sonney2k | foulwall, around? | 21:29 |
@sonney2k | hmmhh jul18 had already 210MB | 21:35 |
@sonney2k | compile time w/ ccache up to 2:20 | 21:36 |
@sonney2k | Jun1 138MB | 21:44 |
@sonney2k | May 19 135MB | 21:49 |
@sonney2k | May 14 128MB | 21:53 |
-!- vgorbati [~vgorbati@91.216.173.29] has quit [Quit: vgorbati] | 22:08 | |
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has joined #shogun | 22:18 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 22:18 | |
@sonney2k | iglesiasg, did you know that our libshogun.so is now 250MB in size? | 22:27 |
@iglesiasg | sonney2k, no, I didn't notice it actually | 22:27 |
@iglesiasg | sonney2k, is that good or bad? :) | 22:27 |
@sonney2k | iglesiasg, well we had sth like 44 | 22:28 |
@sonney2k | so well that explains compile time and everthing | 22:28 |
@iglesiasg | I see | 22:29 |
@iglesiasg | sonney2k, do you think we should try to reduce it somehow? | 22:29 |
@iglesiasg | sonney2k, btw since I am using my office computer basically for everything, I don't notice any more that it takes some time to compile | 22:29 |
@iglesiasg | it is so fast | 22:29 |
@sonney2k | heh | 22:30 |
@sonney2k | no idea | 22:30 |
@sonney2k | I don't know what I measured | 22:30 |
@sonney2k | but now I have 122 MB for shogun 2.1 | 22:30 |
@sonney2k | I guess we need some shell script that tests all versions | 22:32 |
@sonney2k | alrighty script is running | 22:43 |
@iglesiasg | sonney2k, checking dimension of libshogun.so in different versions? | 22:50 |
@sonney2k | iglesiasg, yeah every version is now being built and size determined | 22:53 |
@sonney2k | shocking how many versions don't build at all | 22:53 |
@iglesiasg | lol | 22:53 |
-!- besser82 is now known as besser82|afk | 22:54 | |
@sonney2k | 98380b322004796b058d74d6c24b9f0dad1bff80 251506947shogun/libshogun.so.14.0 | 22:54 |
@sonney2k | <commit> <size in byte> <libshogun name> | 22:54 |
@sonney2k | iglesiasg, I think we should have a plot of this and update it done by our buildbot | 22:55 |
@iglesiasg | sounds interesting | 23:01 |
@iglesiasg | sonney2k, so does the size increase commit by commit? | 23:01 |
@sonney2k | iglesiasg, I am doing the other way round | 23:16 |
@sonney2k | ce4ee057ce3980725275ced9416c889f5cdf8b98 250093795shogun/libshogun.so.14.0 | 23:16 |
@sonney2k | it will take all night and maybe tomorrow even | 23:16 |
@iglesiasg | I see | 23:20 |
--- Log closed Fri Sep 06 00:00:47 2013 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!