IRC logs of #shogun for Wednesday, 2016-06-15

« Tuesday, 2016-06-14 Index Thursday, 2016-06-16 »
--- Log opened Wed Jun 15 00:00:40 2016
-!- lambday [0551433c@gateway/web/freenode/ip.5.81.67.60] has joined #shogun00:29
-!- mode/#shogun [+o lambday] by ChanServ00:29
@lambdayOXPHOS: yo00:29
lisitsynbazdmeg!00:30
lisitsyndaily bazdmeg time00:30
@lambdaylisitsyn: what's that?00:31
lisitsynlambday: wiking taught me some curses00:32
@lambdaylisitsyn: haha00:32
OXPHOSlambday: whaaaaaat00:57
OXPHOSlambday: who are you00:57
@lambdayOXPHOS: Kal-El00:58
@lambdayOXPHOS: checking your patch.. looks pretty good now :)00:59
@lambdayjust a few minor issues, commenting00:59
OXPHOSlambday: thx..late night work?01:00
@lambdayOXPHOS: won't be awake for that long01:01
@lambdaylisitsyn: awake?01:02
@lambdaywiking: ?01:03
lisitsynlambday: on my way to bed01:03
lisitsynjust falling doooownn01:03
OXPHOSlambday: Does coding reading promotes sleeping XD01:03
@lambdayokay01:03
lisitsynslloowww01:03
lisitsynwhat'su p01:03
@lambdaylisitsyn: maybe have a quick look at OXPHOS's patch? https://github.com/shogun-toolbox/shogun/pull/327701:03
lisitsynlambday: can check morning01:03
lisitsynif get some time01:03
@lambdaylisitsyn: coding reading will help you sleep better, as OXPHOS says :D01:03
@lambdaylisitsyn: okay thanks :)01:04
OXPHOShaha. turned down01:04
@lambdayOXPHOS: yeah, story of my life! :D01:04
OXPHOSlambday: ohhhhhhhhh so sad01:04
@lambdayOXPHOS: BTW did I answer your question about the transfer to cpu thingi?01:05
OXPHOSlambday: sort of..so we have to assume the user won't switch backend/or we need a restriction?01:06
OXPHOSI mean the dot onGPU() can be changed easily01:06
OXPHOSI was also worried, one user transfer data to GPU with ViennaCL, and then he switched to CUDA and try to transfer it back to CPU with CUDA01:07
@lambdayOXPHOS: well, we have to assume something! if data is on GPU and user registered a GPU backend, then it is a fair assumption that operations are performed on GPU data01:07
@lambdayOXPHOS: haha let's not worry about that now01:07
OXPHOSkk..01:08
@lambdayOXPHOS: I am not entirely sure whether multiple gpu backend support will happen/is required anytime soon01:08
OXPHOSlambday: also currently we cannot destroy GPU without destroying the whole Vector. So if the user decides to go GPU, he has to stick to it or construct a new Vector01:08
OXPHOSjust add a method to reset GPUptr?01:09
@lambdayOXPHOS: well, transfer to cpu can do that, no?01:09
OXPHOSlambday: I removed to reset part..01:09
OXPHOSin case someone just wants to get a SGVector but also leave this Vector intact?01:10
OXPHOSfeel like they should be separate01:10
@lambdayOXPHOS: okay makes sense..01:10
@lambdaya release gpu memory method01:11
OXPHOSyep01:11
@lambdaycool!01:11
OXPHOSand actually SG_FREE doesn't work in shared_ptr(SG_MALLOC(), )01:11
OXPHOSbecause it has to take arguments01:11
OXPHOSbut let me pull out the error message01:12
@lambdayOXPHOS: you can pass it a lambda01:12
@lambdayso it would be something like shared_ptr(SG_MALLOC(....), [](T* ptr) { SG_FREE(ptr); });01:12
OXPHOSI can try01:13
OXPHOSBTW until last week I thought your name was lamb day for some reason01:13
@lambdayOXPHOS: I've been called worse :D01:13
OXPHOSsad night ;(01:14
@lambdayOXPHOS: so it's not that bad :D01:14
OXPHOShaha01:14
@lambdayOXPHOS: actually my first mail to shogun community in my first GSoC was just this : hey people, I am lambda y (not lamb day)01:14
OXPHOSlambday: so once data is onGPU, CPU part is "masked"01:15
@lambdayOXPHOS: I think it should be fine..01:15
OXPHOSlambday: I must learn from the mail01:15
OXPHOSkk01:15
OXPHOSlambday: oh you mean your first GSoC lol. Guess you'll have to state that everytime01:16
@lambdayhaha01:16
@lambdayOXPHOS: alright.. off to sleep01:17
@lambdayOXPHOS: please update the PR.. we want to merge it soon01:18
@lambdayOXPHOS: and make sure you get it reviewed by wiking or lisitsyn or HeikoS also01:18
OXPHOSlambday: sure I'll chase them. good night!01:18
@lambdayOXPHOS: night night01:19
-!- lambday [0551433c@gateway/web/freenode/ip.5.81.67.60] has quit [Quit: Leaving.]01:19
-!- OXPHOS [9d8b131c@gateway/web/freenode/ip.157.139.19.28] has quit [Quit: Page closed]03:11
-!- GandalfTheWizard [~Eva@112.10.170.90] has joined #shogun03:12
-!- arianepaola [~ariane@unaffiliated/arianepaola] has left #shogun []03:19
-!- sanuj [~sanuj@117.203.14.130] has joined #shogun06:06
sanujwhat happened to buildbot?06:07
Saurabh7ye something was wrong last night too06:10
sanujbuildbot be like: The service is not available. Please try again later.06:11
Saurabh7its on aws now right , dunno wasup06:12
-!- sanuj [~sanuj@117.203.14.130] has quit [Ping timeout: 276 seconds]06:31
@wikinghehe no buldbot for some time now07:44
@wiking(under progress to fix it)07:44
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun08:01
-!- sanuj [~sanuj@117.203.14.130] has joined #shogun08:19
shogun-buildbotbuild #54 of xenial - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/xenial%20-%20libshogun/builds/54  blamelist: Viktor Gal <vigsterkr@gmail.com>, Saurabh7 <saurabh.mahindre@gmail.com>08:36
shogun-buildbotbuild #17 of clang - undefined behaviour analysis is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang%20-%20undefined%20behaviour%20analysis/builds/17  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Viktor Gal <vigsterkr@gmail.com>, Ariane Paola Gomes <arianepaola@users.noreply.github.com>, Saurabh708:46
shogun-buildbot<saurabh.mahindre@gmail.com>08:46
-!- sanuj [~sanuj@117.203.14.130] has quit [Remote host closed the connection]08:48
shogun-buildbotbuild #0 of FC23 - libshogun - aarch64 is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC23%20-%20libshogun%20-%20aarch64/builds/0  blamelist: Viktor Gal <vigsterkr@gmail.com>, Saurabh7 <saurabh.mahindre@gmail.com>08:50
-!- GandalfTheWizard [~Eva@112.10.170.90] has quit [Quit: Leaving.]09:02
@wikingshogun-buildbot: force build --branch=develop 'CentOS7 - libshogun'09:32
shogun-buildbotThe build has been queued, I'll give a shout when it starts09:32
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Remote host closed the connection]09:32
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun09:33
shogun-buildbotbuild #1021 of nightly_none is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_none/builds/1021  blamelist: Viktor Gal <vigsterkr@gmail.com>, Saurabh7 <saurabh.mahindre@gmail.com>09:42
-!- arianepaola [~ariane@unaffiliated/arianepaola] has joined #shogun09:48
@wikingshogun-buildbot: dance10:07
shogun-buildbot<(^.^<)10:07
shogun-buildbot<(^.^)>10:07
shogun-buildbot(>^.^)>10:07
shogun-buildbot(7^.^)710:07
shogun-buildbot(>^.^<)10:07
-!- besser82_ [~besser82@fedora/besser82] has joined #shogun10:16
-!- mode/#shogun [+o besser82_] by ChanServ10:16
shogun-buildbotbuild #1151 of nightly_default is complete: Failure [failed test notebooks]  Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/1151  blamelist: Viktor Gal <vigsterkr@gmail.com>, Saurabh7 <saurabh.mahindre@gmail.com>10:32
@wikingSaurabh7: ping10:36
@wikingSaurabh7: https://github.com/shogun-toolbox/shogun/pull/328810:37
@wikingsee the comment of Heiko10:37
Saurabh7hm no test for it atm10:59
Saurabh7but I will have to compare against some constant values then11:00
Saurabh7let me add one11:00
Saurabh7wiking: should I push the test in PR or ?11:18
Saurabh7in the same PR i mean11:22
-!- besser82_ [~besser82@fedora/besser82] has quit [Ping timeout: 250 seconds]11:25
Saurabh7brb11:30
@wikingSaurabh7: same same :)11:44
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has joined #shogun12:15
-!- mode/#shogun [+o HeikoS] by ChanServ12:15
@wikingHeikoS: so we got buildbot back :)12:15
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has quit [Quit: Leaving.]12:44
-!- arianepaola [~ariane@unaffiliated/arianepaola] has quit [Excess Flood]13:14
-!- arianepaola [~ariane@unaffiliated/arianepaola] has joined #shogun13:16
@wikingarianepaola: lets merge that pr13:36
-!- sanuj [~sanuj@117.203.14.130] has joined #shogun13:36
sanujlisitsyn, hey13:36
lisitsynsanuj: hey13:36
@wikingarianepaola: oh wait, can you just amend that commit with the description that Heiko referred to: https://github.com/shogun-toolbox/shogun/wiki/Project-description13:36
@wikingplz13:36
@wikingping me when its done13:37
@wikingthnx13:37
sanujlisitsyn, did you see my email?13:37
lisitsynsanuj: yeah but got no time yet to check13:37
sanujokay13:37
sanuji'll start working on the unit-tests for now13:37
lisitsynokay13:37
sanujwiking, you were talking about something we have to check coverage of code....13:49
@wikingpeeeeeeeeeeeeeeople13:55
@wikingwe have a bug13:55
@wikinghttp://buildbot.shogun-toolbox.org/builders/deb4%20-%20python3/builds/882/steps/install/logs/stdio13:55
sanujooooooh13:55
@wikingit's from Heiko actually :)13:56
lisitsynOLD PyTHon!13:56
@wikingyep yep13:57
-!- HeikoS [~heiko@nat-201-233.internal.eduroam.ucl.ac.uk] has joined #shogun14:06
-!- mode/#shogun [+o HeikoS] by ChanServ14:06
@wikingHeikoS: bugzi14:09
@HeikoSwiking: jo14:09
@HeikoSwiking: nice one on the buildbot14:09
@HeikoSso I can merge more cookbooks later14:09
@wikingyeah and he found you the bug you've created for python3 ;)14:09
@HeikoSwiking: ah nice14:09
@HeikoSchecking14:10
@HeikoSarianepaola: ping14:11
@HeikoSarianepaola: can you pls fix this: http://buildbot.shogun-toolbox.org/builders/deb4%20-%20python3/builds/881/steps/install/logs/stdio14:11
@HeikoSarianepaola: i explicitly asked whether this is python 2 and 3 compatible, remember? :)14:11
@HeikoSwiking: ok lets hope that is solved soon, I have to do some work now, will check back later14:13
-!- [Chris] [~Chris]@interferon.am10.uni-tuebingen.de] has joined #shogun14:15
[Chris]hi, I have a question regarding combined kernels and cross validation. How should I specify the features for CrossValidation if I'm using a SVM with a combined kernel consisting of two rbf kernels with different feature encodings? Is there some example code on that ?14:17
arianepaolahello everyone14:26
arianepaolahi wiking, I saw your message in the logs14:32
arianepaolawiking: I will add the description14:32
arianepaolaHeikoS: Python3... I will update it14:33
@HeikoSarianepaola: hi14:33
@HeikoSthanks14:33
@HeikoSwhenever you write python code for shogun, it needs to work under both 2/3, like I said in the PR14:33
arianepaolaI must have changed the interpreter14:34
arianepaoladoes it need to be < Python 2.614:34
@HeikoSno 2.7 is fine14:34
arianepaolaor is Python >= 2.6 ok14:34
arianepaolaok14:34
sanujwiking, i built shogun with ENABLE_COVERAGE=ON14:38
sanujnow how to see coverage?14:38
-!- c4goldsw [5da420e6@gateway/web/cgi-irc/kiwiirc.com/ip.93.164.32.230] has joined #shogun14:40
c4goldswHello14:40
c4goldswHeikoS Hey, how are you?14:43
@HeikoSc4goldsw: jo"!14:44
@HeikoSpretty good14:44
c4goldswI saw your comments on my PR - aside from what you listed, I think it's fine.14:44
c4goldswBut, its not being included b/c Pan is doing linalg, right?14:44
c4goldswIf that's the case, I'll just close it.14:45
c4goldswHeikoS ^14:51
@HeikoSc4goldsw: sorry was in a call14:52
@HeikoSok cool14:52
@HeikoSI mean pans stuff will refactor the very basics of linalg14:52
@HeikoSbut the operation will still be needed14:52
@HeikoSas well as the unit tests etc14:52
@HeikoSthey will just be refactored a bit14:52
c4goldswFair enough - I'll still close it anyways.14:53
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun15:07
shogun-notifier-shogun: Ariane Paola Gomes :develop * 5a6f847 / examples/meta/generator/generate.py: https://github.com/shogun-toolbox/shogun/commit/5a6f8474b3c09d52911ecc9a5a520923738cbe7c15:07
shogun-notifier-shogun: Fix Python 2 to 3 exception conversion.15:07
shogun-notifier-shogun: Viktor Gal :develop * cd345d1 / examples/meta/generator/generate.py: https://github.com/shogun-toolbox/shogun/commit/cd345d15ca03819c8a62c013d51fb66d9f556e7e15:07
shogun-notifier-shogun: Merge pull request #3291 from arianepaola/fix_2to315:07
shogun-notifier-shogun:15:07
shogun-notifier-shogun: Fix Python 2 to 3 exception conversion.15:07
arianepaola^^ HeikoS15:07
@wikingHeikoS: puuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuub15:07
arianepaola^^ wiking setup.py is ready to merge15:07
shogun-buildbotbuild #718 of trusty - libshogun - viennacl is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/trusty%20-%20libshogun%20-%20viennacl/builds/718  blamelist: Ariane Paola Gomes <arianepaola@users.noreply.github.com>15:07
@wikingwhy trusty faaaails fuck15:08
@HeikoSthx :)15:08
arianepaolashogun-buildbot: dance15:08
shogun-buildbot<(^.^<)15:08
shogun-buildbot<(^.^)>15:08
shogun-buildbot(>^.^)>15:08
shogun-buildbot(7^.^)715:08
shogun-buildbot(>^.^<)15:08
arianepaolahahaha15:08
shogun-buildbotbuild #719 of trusty - libshogun - viennacl is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/trusty%20-%20libshogun%20-%20viennacl/builds/719  blamelist: Viktor Gal <vigsterkr@gmail.com>15:08
@wikingmmmm this is weirdo http://buildbot.shogun-toolbox.org/builders/trusty%20-%20libshogun%20-%20viennacl/builds/719/steps/compile/logs/stdio15:10
@wikinglars_unittest.cc15:11
@wikingwtf?15:11
@wikingwhy did it started to error now?!15:11
arianepaolait had a bad day15:11
arianepaolalol15:11
@wikingshogun-buildbot: force build --branch=develop 'deb4 - python3'15:11
shogun-buildbotbuild #883 forced15:11
shogun-buildbotI'll give a shout when the build finishes15:11
arianepaolaprobably some teenager growing up stuff15:11
arianepaolarebellion, etc.15:12
arianepaolastarting to break everything on buildbot now15:12
@wikingshogun-buildbot: force build --branch=develop 'trusty - libshogun - viennacl'15:12
shogun-buildbotbuild #720 forced15:12
shogun-buildbotI'll give a shout when the build finishes15:12
@wikingfuck15:12
@wikinglets see now15:12
arianepaolawiking: when using -DPythonModular=ON and release build it crashes15:13
shogun-buildbotbuild #720 of trusty - libshogun - viennacl is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/trusty%20-%20libshogun%20-%20viennacl/builds/72015:13
@wiking:D15:13
@wikingnice15:13
@wikingwhen was this unit test changed?15:17
shogun-buildbotbuild #883 of deb4 - python3 is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb4%20-%20python3/builds/88315:19
@wikingok cool arianepaola \o/ fixed15:20
@wikingHeikoS: ping ping15:20
arianepaolawiking: good15:20
@wikingmmm how do we add a package to pypi15:20
@wiking?15:20
arianepaolawiking: setup.py PR is updated15:20
@wikingarianepaola: yeah i'm just wanna check if HeikoS is ok with it15:21
@wikingarghfuck15:21
@wikingHeikoS: LETS FUCKING START DOING WHAT WE AGREED ON15:21
arianepaolawiking: python setup.;why register15:21
@wikingHeikoS: https://github.com/shogun-toolbox/shogun/pull/3240#issuecomment-22618068315:21
@wikingso i added my +115:21
@HeikoSwiking: whats the matter?15:21
arianepaolahttp://python-packaging.readthedocs.io/en/latest/minimal.html -> Publishing On PyPI15:21
@wikingremember? :D15:21
@wikingonly merge when we have 2 +1s15:21
@wiking:>15:22
@HeikoSI am lost as in a call15:22
@HeikoSgive me some context15:22
@wikingheheh so15:22
@wikingremember weve agreed to only merge15:22
@wikinga gsoc pr15:22
@HeikoSwhich one?15:22
@wikingwhen there are 2 +1s on15:22
@wikingthe pr from core contributers15:22
@wikingthis pr https://github.com/shogun-toolbox/shogun/pull/3240#issuecomment-22618068315:22
@wikingso i just need your +1 :D15:22
@HeikoSok checking15:22
@wikingHeikoS: my question is about description only15:23
@wikingif you are fine15:23
@wikingwith that15:23
@HeikoSdescription15:23
@HeikoShttps://github.com/shogun-toolbox/shogun/pull/3240/files#diff-2eeaed663bd0d25b7e608891384b7298R29615:23
@HeikoS=?15:23
@wikingReadme.md15:23
@wiking +with open(os.path.join(setup_py_location, 'README.md')) as f:15:24
@wiking+  readme = f.read()15:24
@HeikoSah yes15:24
@HeikoSgood15:24
@HeikoSBUT15:24
arianepaolawiking: I think HeikoS means https://github.com/shogun-toolbox/shogun/pull/3240/files#diff-2eeaed663bd0d25b7e608891384b7298R31715:24
@HeikoSwe should update the readme15:24
@HeikoSguys I am lost sorry15:24
@HeikoShalf in a call15:24
@HeikoSwhat exactly is the question?15:24
@wikingah but ok15:24
@wikingso you wanna change some stuff in README.md15:24
@HeikoSwiking: yes15:25
@wikingthat's ok but that should be changed15:25
@wikingin another pr15:25
@wikingnot here15:25
@HeikoSI want to put in the description from the wiki15:25
@HeikoSyes15:25
@HeikoS+1 from my side for the pr15:25
arianepaolais this line ok? https://github.com/shogun-toolbox/shogun/pull/3240/files#diff-2eeaed663bd0d25b7e608891384b7298R31715:25
@HeikoSyes15:25
@wikingput the +115:26
@wiking:D15:26
@HeikoSdone15:26
@HeikoSin text form :)15:26
@wikingahahha15:26
arianepaolabtw, we can add Operating System :: Microsoft :: Windows :: Windows 3.1 or Earlier15:26
arianepaolaand Operating System :: Microsoft :: MS-DOS15:26
arianepaolarofl15:26
@wikingheheh no15:26
@HeikoSis that on our buildbot? ;)15:26
@wikingwe dont support hat15:26
@wiking*that15:26
@wikinganyhow merging15:26
@wikingand lets fucking register ourselves into pypi15:26
arianepaolathanks wiking15:26
shogun-notifier-shogun: Ariane Paola Gomes :develop * 2151b9c / setup.py: https://github.com/shogun-toolbox/shogun/commit/2151b9c8cce2aca068a421460e580d8c5d39829315:27
shogun-notifier-shogun: Added Shogun Python package generation using setup.py.15:27
shogun-notifier-shogun: Viktor Gal :develop * 3998da8 / setup.py: https://github.com/shogun-toolbox/shogun/commit/3998da842f988a6cc816ef4b225241c81da1859115:27
shogun-notifier-shogun: Merge pull request #3240 from arianepaola/feature/setup_py15:27
shogun-notifier-shogun:15:27
shogun-notifier-shogun: Shogun Python package using setup.py15:27
@HeikoSwhoooo :)15:27
@HeikoScongrats15:27
@HeikoSso now lets install on all machine on buildbot :)15:27
@wiking:D15:27
@wikingah that will not happen15:27
@wiking:>15:27
sanujHeikoS, i had updated this https://github.com/shogun-toolbox/shogun/pull/328115:27
@wikingregistering15:27
@wiking:)15:27
@HeikoSsanuj: thanks!15:27
@HeikoSI will check a bit later15:28
shogun-buildbotbuild #721 of trusty - libshogun - viennacl is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/trusty%20-%20libshogun%20-%20viennacl/builds/721  blamelist: Viktor Gal <vigsterkr@gmail.com>, Ariane Paola Gomes <arianepaola@users.noreply.github.com>15:28
sanujHeikoS, the math compiles fine on my local but not on buildbot15:28
sanujfor svm_cookbook15:28
sanujwiking, i built shogun with ENABLE_COVERAGE=ON,now how to see coverage?15:29
@HeikoSsanuj: I sent you this tool a while ago15:29
@HeikoSthat can just do it for the unit test binary15:29
sanujokay15:29
@HeikoSsanuj: or you can check how the buildbot uploads the coverage15:29
-!- besser82_ [~besser82@fedora/besser82] has joined #shogun15:31
-!- mode/#shogun [+o besser82_] by ChanServ15:31
-!- besser82_ [~besser82@fedora/besser82] has quit [Ping timeout: 260 seconds]15:39
-!- sanuj [~sanuj@117.203.14.130] has quit [Ping timeout: 240 seconds]15:43
-!- besser82_ [~besser82@fedora/besser82] has joined #shogun15:48
-!- mode/#shogun [+o besser82_] by ChanServ15:48
-!- besser82_ [~besser82@fedora/besser82] has quit [Remote host closed the connection]15:58
-!- besser82 [~besser82@fedora/besser82] has joined #shogun16:01
-!- mode/#shogun [+o besser82] by ChanServ16:01
-!- sanuj [~sanuj@117.203.14.130] has joined #shogun16:11
sanujHeikoS, gcov wants optimization off16:11
sanujdo i need to rebuild the tests16:12
@HeikoSsanuj: yes thats ok16:12
@HeikoSyes16:12
sanujhow to set optimization off16:12
@HeikoSgood for developing16:12
@HeikoSCMAKE_BUILD_MODE16:12
@HeikoSShould be "Debug"16:12
sanujit is already debug16:12
sanujHeikoS, i had build with ==> cmake -DENABLE_TESTING=on -DPythonModular=on -DCMAKE_BUILD_TYPE=Debug ..16:13
@HeikoSthat has optimisation off16:13
@wikingarianepaola: Server response (403): You are not allowed to store 'shogun' package information16:13
@wiking:S16:13
arianepaolalol16:13
arianepaolahmmm16:14
@wikingweird16:14
@wikingtrying to google why16:14
arianepaolahttps://wiki.python.org/moin/TestPyPI16:15
arianepaoladid you try to run: python setup.py register before upload?16:16
arianepaola^^ wiking16:16
@wiking?16:18
@wikingi ran16:18
@wikingpython setup.py register16:18
@wikingand that was the output :)16:18
@wikingah ok16:18
arianepaolathere is also no shogun package on pypi16:18
@wikingpython setup.py register -r https://testpypi.python.org/pypi16:18
@wikingso i shoudl try this first?16:19
arianepaolait should work on pypi directly16:19
@wikingi get 40316:19
sanujHeikoS, how to run unit-test with gcov?16:22
sanujgcov ./shogun-unit-test16:23
@HeikoSsanuj: I dont know, read the manual, should be in there16:23
@HeikoSthat command doesnt work?16:23
sanujno16:23
sanuj./shogun-unit-test.gcno:cannot open notes file16:23
@HeikoSok then manual :)16:24
-!- OXPHOS [9d8b131c@gateway/web/freenode/ip.157.139.19.28] has joined #shogun16:27
-!- GandalfTheWizard [~Eva@112.10.170.90] has joined #shogun16:33
sanujHeikoS, it says that to use gcov, the code should be compiled with gcc options: -fprofile-arcs and -ftestcoverage16:34
sanujdoes that happen in debug mode?16:35
@wikingoh for god sakes16:36
@wikinghttps://pypi.python.org/pypi/shogun/16:36
@wiking:D16:36
arianepaolayay16:36
@wikingso this guy16:36
@wikingcreated a python package16:36
@wikingthat fucking does nothing16:36
@wikingand took our name16:36
@wiking:)16:36
@wikingtime to message16:37
@wikinghttps://github.com/portantier/shogun/issues/116:40
@wiking:)16:40
sanujwiking, have you used shogun with gcov?16:40
sanujhaha16:41
c4goldswwiking pong16:57
@wikingc4goldsw: hey17:01
@wikingsanuj: mmmm lemme recall this part i wrote like 3 years ago17:01
sanujwiking, thanks17:01
c4goldswI just want to clarify that this is essentially what I should be doing: https://gist.github.com/c4goldsw/fd325503b9b81a2a7f434291732e19e017:02
c4goldswas compared to https://github.com/shogun-toolbox/shogun/blob/develop/src/shogun/regression/LeastAngleRegression.cpp17:02
@wikingc4goldsw: does it compile? :)17:03
c4goldswNo, because it's superclass would also have to be parametereized / tempalted :(17:03
@wikingbecause this is a problem for example17:04
@wikinghttps://gist.github.com/c4goldsw/fd325503b9b81a2a7f434291732e19e0#file-templated-lars-L14117:04
@wikingwhich super class?17:04
@wikingso while Xorig.matrix is now SGMatrix<T> that could be any of those args17:04
@wikingMatrixXd17:04
c4goldswCLinearMachine17:04
c4goldswCLeastAngleRegession inherits from it17:05
@wikingis actually typedef Matrix<double, Dynamic, Dynamic> MatrixXd;17:05
@wikingso that's not a good mapping :(17:05
c4goldswhmm, okay17:05
@wikingbecause you could have int32_t or any other int type there17:05
@wikingor float32_t for that matter17:05
@wikingbut you'd be doing a double map17:05
@wikingso that for sure has to be figured out17:06
@wikingand as well wondering17:06
c4goldswBut to be clear, am I taking the right approach?17:07
@wikingwhether one can do/or makes sense to do a LARS on CLeastAngleRegression<bool> or CLeastAngleRegression<char>;17:07
@wikingmmm17:07
@wikingi'm wondering now a bit17:07
c4goldswAh, but I had that question last week - which types make sense?17:07
@wikingbecause you know there are other means to figure out the type of the CFeatures17:07
c4goldswWhat do you mean by that?17:08
@wikinghttps://github.com/shogun-toolbox/shogun/blob/develop/src/shogun/features/FeatureTypes.h17:09
@wikingso if you call CFeatures->get_feature_type()17:09
@wikingyou'll get the type17:09
@wikingsame for EFeatureClass17:10
@wikingi mean the thing is17:10
@wikingthat LARS could even get a C_COMBINED17:10
@wiking:)17:10
@wikingso maybe instead of templating17:11
@wikingyou could like do an actual type detection within the train_machine part17:11
@wikingand do the rigth call17:12
@wikingbecuase i wonder how would you solve the issue of type mapping with eigen17:12
@wikingthis way btw you dont need to support templating in the partent class17:12
@wiking*parent17:12
c4goldswAlright then17:13
@wikingagain this might need a bit more thinking17:13
@wikingbecause actually whatever you'll write here17:13
@wikingmost of the code17:13
@wikingshoudl be reusable17:13
@wikingin other parts of shogun17:13
@wikingsince this is a problem as you saw in many palces17:14
c4goldswIn what sense?17:15
c4goldswI understand your type detection approach17:15
c4goldswBut I'm a little unclear of what problems could arise, as  I haven't thoroughly read through train_machine17:15
@wikingin sense thwat this type detection thing should be available in other parts17:15
@wikingand how to map an SGMatrix<int8_t> to an eigen map17:16
c4goldswWhat possible values can eigen maps take?17:16
c4goldswI though they were parameterized.17:16
c4goldswpossible types*17:16
@wikingi guess whatever u give it :)17:17
c4goldswAlright then - wouldn't this simply become a problem of parameterizing everything else that uses it?17:18
c4goldswWhich could be quite a big problem.17:18
@wiking?17:19
OXPHOSwiking: hey!17:20
OXPHOSwiking: can you have a look at the linalg when you got a sec? https://github.com/shogun-toolbox/shogun/pull/327717:20
OXPHOSthx!17:20
c4goldswWell, you asked "how to map an SGMatrix<int8_t>" to an eigen map.  Wouldn't  that just involve doing : Matrix<int8_t, Dynamic, Dynamic> MatrixXd; ?17:21
c4goldswOr, let's use a more sensible type17:21
@wikingno17:21
@wikingMatrixXd is a reserved already defined type in eigen17:22
@wikingthat is Matrix<double, Dynamic, Dynamic>17:22
sanujlisitsyn, there?17:23
c4goldswThen let's take a step back - instead of templating / parameterizing all of LeastAngleRegression, you suggest that I can just check the type of CFeatures* data, then act accordingly.17:24
arianepaolawiking: did you find any info regarding launchpad packaging format requirements?17:24
c4goldswcheck inside of train_machine*.  That's what you were saying, right?17:24
@HeikoSsanuj: jo17:24
@HeikoSnow checking PRs17:24
-!- ptizoom [~christian@host-92-21-196-52.as13285.net] has quit [Ping timeout: 258 seconds]17:25
@wikingc4goldsw: yep17:25
c4goldswGood.  So, you're suggesting that the outstanding problem is that I would have to find ways of making CFeatures* of various types compatible with eigen maps, right?17:26
@wikingarianepaola: https://developer.ubuntu.com/en/publish/other-forms-of-submitting-apps/ppa/17:26
@wikingc4goldsw: yep17:27
arianepaolathanks wiking17:27
c4goldswSo, should I continue in that direction?17:27
arianepaolawiking: do you know https://build.opensuse.org/ ?17:27
@wikingc4goldsw: nou17:27
@wikingsorry17:27
@wikingarianepaola: noup17:27
@wikingc4goldsw: yeap since this templating what you've started with17:27
c4goldswwiking that's fine.17:27
@wikingi'm not sure will lead to a solution17:28
arianepaolawiking: seems quite large -> openSUSE Build Service hosts 44,424 projects, with 364,974 packages, in 68,967repositories and is used by 44,475 confirmed developers.17:28
c4goldswwiking haha with the way things are going I don't think they will17:28
@wikingarianepaola: how do we apply/register?17:28
arianepaolaI think it is free17:28
arianepaolaI will check17:28
@wikingmmm i mean the thing is that we wanna ha ppa17:29
@wikingbecause it's an easy way for ubuntu/debian17:29
@HeikoS+1 for ppa17:30
@HeikoSnightly build17:30
@HeikoSand every release triggers an automatic one17:30
c4goldswwiking don't think it will *17:31
@wikingc4goldsw: wil will17:31
@wikingjust takes time17:31
shogun-notifier-shogun: OXPHOS :develop * 9950533 / / (7 files): https://github.com/shogun-toolbox/shogun/commit/9950533524e3d393a0bd7eb94393524b419a9dd617:31
shogun-notifier-shogun: means cookbook - meta examples and integration test17:31
shogun-notifier-shogun: Heiko Strathmann :develop * 59c003f / / (7 files): https://github.com/shogun-toolbox/shogun/commit/59c003fdbb01a5f2af680c5a99de7e331195e7cb17:31
shogun-notifier-shogun: Merge pull request #3183 from OXPHOS/cookbook17:31
shogun-notifier-shogun:17:31
shogun-notifier-shogun: Add kmeans page to cookbook17:31
@wiking\o/17:31
@wikingcookering17:31
arianepaolaalthough it probably out of date: http://openlife.cc/blogs/2011/february/looking-opensuse-build-service-and-launchpad-ppa-aka-how-build-packages-mepsql17:31
arianepaolathe opensuse buildservice could be useful for rpm though17:32
shogun-buildbotbuild #722 of trusty - libshogun - viennacl is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/trusty%20-%20libshogun%20-%20viennacl/builds/722  blamelist: OXPHOS <engelzora@gmail.com>17:32
arianepaolappa is easier to add17:32
c4goldswwiking No, I mean templating everything - you just said that wouldn't work.  So, I'll focus on making train_machine compatible with different types (Sorry for making you talk to two people at once).  Another question: which types should I make it compatible with?17:33
shogun-buildbotbuild #723 of trusty - libshogun - viennacl is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/trusty%20-%20libshogun%20-%20viennacl/builds/723  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>17:33
c4goldswwiking would it even make sense for non float types?17:34
@wikingc4goldsw: good question17:34
@wikingit's really up to the algorithm17:35
@HeikoSc4goldsw, wiking couldnt we just have some template methods that do the conversions for us?17:35
c4goldswI'm looking at it:https://en.wikipedia.org/wiki/Least-angle_regression17:35
sanujHeikoS, for svm cookbook, the math compiles on my local but not on buildbot17:35
@HeikoSI also think the change between int and float is not so interesting17:35
@HeikoSbut rather float32 vs float6417:35
@HeikoSand int32 vs int6417:35
@HeikoSthese 4 are like 80% for the use cases where one would be interested17:35
@HeikoSin changing representation17:35
sanuji tried some online thing to translate math and it worked on that too17:35
@HeikoSsanuj:17:36
@HeikoShttp://buildbot.shogun-toolbox.org/cookbook_pr/033f9c8fa9e07d040d0c94e6c16092a1090983c4/examples/classifier/kernel_svm.html17:36
sanujyeah i saw that17:36
@wikingHeikoS: could17:37
c4goldswHeikoS I've been doing that, but it could potentially involve templating all superclasses and sublcasses of those superclasses.  CLeastAngleRegression has CLinearMachine as a parent and I've have to template that if I'm to template LARs17:37
@HeikoSc4goldsw: the algorithm classes would not change17:37
@HeikoSonly the features17:37
@HeikoSand then there is a template method that gives corresponding eigen vector/matrix17:37
-!- ptizoom [~christian@host-92-21-231-167.as13285.net] has joined #shogun17:37
@HeikoSand we can use MatrixBase in the algorithm if eigen is used17:37
@HeikoSor even auto17:38
@HeikoSin the algo you then do17:38
@HeikoSfeatures->to_eigen_bla()17:38
@HeikoSor actually17:38
@HeikoSa wrapper for map would be good17:38
@HeikoSfeatures->to_eigen_map(num_rows, num_cols)17:38
@HeikoSoffset maybe17:38
@HeikoSand that method is templated17:39
@HeikoSsanuj: saw my comment in data?17:39
c4goldswIt's a little hard for me to grasp this, could you possibly make a Gist of it?17:39
c4goldswHeikoS ^17:39
sanujHeikoS, yeah17:39
sanujshall i try changing some parameters for the svm17:40
@HeikoSc4goldsw: I see, sorry17:41
@HeikoSc4goldsw: I am currently a bit overbusy, so no time to draft it out17:41
@HeikoSbut my point is this:17:41
@HeikoSthe feature class is templated right?17:41
c4goldswYes17:41
@HeikoSok17:42
@HeikoSso why not add a method to DenseFeeatures17:42
@HeikoSthat returns an eigen map to the feature matrix?17:42
@HeikoSof the correct type17:42
@HeikoSLike VectorXd or VectorXf17:42
@HeikoSlike for example get_feature_matrix17:43
@HeikoSreturns SGMatrix<T>17:43
c4goldswThat makes sense17:43
@HeikoSbut for eigen its a bit different17:43
@HeikoSwe cannot just pass on the T17:43
@HeikoSbut for T=float64_t, we return MatrixXd, and for T=float32_t we return MatrixXf17:44
c4goldswAh, that's clearer now.17:44
c4goldswHeikoS Thanks17:46
@HeikoSc4goldsw: are you looking at a concrete example?17:47
c4goldswHeikoS yes: https://gist.github.com/c4goldsw/fd325503b9b81a2a7f434291732e19e0#file-templated-lars-L13517:48
OXPHOSHeikoS: hey! got time to check the PR? https://github.com/shogun-toolbox/shogun/pull/327717:50
c4goldswHeiko: In this file, I've templated everything so far, which has lead to errors being thrown, since superclass members are not parameterized.  wiking pointed out that this line is an issue.  So, as you suggested,  I'll get onto making a get_feature_matrix() which can return eigen maps of differing types17:50
@HeikoSOXPHOS: hi17:50
@HeikoSgoing through things right now17:50
@HeikoScoming back to you soon :)17:50
OXPHOSHeikoS: thx!17:50
@HeikoSc4goldsw: why is this line a problem?17:50
@HeikoSlines like Map<VectorXd> are a problem17:51
@HeikoSbecause there, you have a float64_t17:51
@HeikoSwhereas SGMatrix<ST> is ok, it is template17:51
@HeikoSjust need a way to create the eigen map from the SG structures in a templated way17:51
c4goldswHeikoS ah, excuse me17:51
@HeikoSso I think the way to go is to have some factory where one can create eigen maps for existing SG data types17:52
@HeikoSI think there should be something like this already somewhere17:52
@HeikoSjust need to template it17:52
@HeikoSlisitsyn:  ^ ?17:52
c4goldswHeikoS should it be templated for all types?17:52
@HeikoSif we use eigen317:52
@HeikoSwe can only template for those that are supported by eigen17:52
@HeikoSi dont know how many that are17:53
@HeikoSgotta check their manual17:53
c4goldswI'll find out then.17:53
@HeikoSnice one that you are digging on this btw!17:53
@HeikoSreally useful17:53
@HeikoSto change lars accuracy to 32bit without doing anything17:53
c4goldswAh, you're welcome.  I'm off for now, I'll be back on tomorrow.17:54
@HeikoSsee you! :)17:55
-!- c4goldsw [5da420e6@gateway/web/cgi-irc/kiwiirc.com/ip.93.164.32.230] has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]17:55
@HeikoSOXPHOS: ok17:58
@HeikoSOXPHOS: can you give me a quick update what you recently did there and what I should look out for?17:59
@HeikoSOXPHOS: whats CPU_CACHE_LINE SIZE?17:59
@HeikoSOXPHOS: actually just talking to lambday18:00
@HeikoSwill get back after18:00
sanujHeikoS, do i need to check mem leak in python for this https://github.com/shogun-toolbox/shogun/pull/328318:00
@HeikoSsanuj: will get back soon18:01
sanujokay18:01
-!- GandalfTheWizard [~Eva@112.10.170.90] has quit [Quit: Leaving.]18:13
OXPHOSHeikoS: hey thanks for the comments18:20
OXPHOSHeikoS: so basically we don't allow [] visit18:21
@HeikoSOXPHOS: yep18:25
@HeikoSfound a problem18:25
@HeikoSthinking and commenting18:25
@HeikoSOXPHOS: ok put final comment18:35
@HeikoSOXPHOS: can you read the final comment first?18:38
@HeikoSand othen go backwards from there18:38
OXPHOSHeikoS: sure18:38
@HeikoSOXPHOS: btw, you should aim for 100% unit test coverage of this code18:38
@HeikoSplease read everything carefully and let me know if you feel you udnerstand it18:39
sanujHeikoS, see https://github.com/shogun-toolbox/shogun-data/pull/10218:39
sanuji updated the data18:39
sanujshould be good now18:39
@HeikoSwhat was wrong?18:39
OXPHOSSo if I have to transfer everything back to SGVector first then compare the element value by value - I have to assume everything associate with operator SGVector() works18:39
OXPHOSHeikoS^18:39
shogun-notifier-shogun-data: Sanuj :master * baf949d / testsuite/meta/classifier/kernel_svm.dat: https://github.com/shogun-toolbox/shogun-data/commit/baf949d1d1449c82e63f1a39ad12d38f4a82723818:39
shogun-notifier-shogun-data: add data for kernel svm cookbook18:39
sanujHeikoS, increased the width of gaussian kernel18:39
shogun-notifier-shogun-data: Heiko Strathmann :master * 038f080 / testsuite/meta/classifier/kernel_svm.dat: https://github.com/shogun-toolbox/shogun-data/commit/038f080e9eb720097900b34e51372798a385d59418:39
shogun-notifier-shogun-data: Merge pull request #102 from sanuj/master18:39
shogun-notifier-shogun-data:18:39
shogun-notifier-shogun-data: add data for kernel svm cookbook18:39
@HeikoSOXPHOS: I am not sure what you mean18:40
@HeikoSsanuj: always a good idea :)18:40
@HeikoSOXPHOS: can you ask again?18:40
sanuj:D18:40
@HeikoSsanuj: any idea about the math?18:41
sanuji'm updating it18:41
sanujlet's see if that works18:41
OXPHOSHeikoS: by 100% unit test coverage - does it include every single method in the class?18:41
@HeikoSOXPHOS: every single line of code18:41
@HeikoSOXPHOS: there is a tool to check that18:41
@HeikoSwiking: still awake?18:42
@HeikoSOXPHOS: but dont worry about that for now18:42
@HeikoSthe main thing we need to change for now is the copying18:42
@HeikoSand as a result the reference counting18:42
@HeikoSand as a result the alignment18:42
sanujHeikoS, When using gcov, you must first compile your program with two special GCC options: ‘-fprofile-arcs -ftest-coverage’.18:43
sanujhow to do this ^18:43
OXPHOSHeikoS: okay...but i don't feel like it's a big problem copying SGVector. They'll be done just once by user18:43
@HeikoSsanuj: ah ok, lemme check18:43
@HeikoSOXPHOS: it is a big problem18:43
@HeikoSimagine you have two SGVectors18:43
@HeikoSand you want to compute their dot product18:43
@HeikoSvia linalg::dot18:43
@HeikoSlinalg:dot accepts Vector at this point18:44
@HeikoSupon casting SGVector to Vector, the data is copied18:44
@HeikoSso the dot product is preceeded by copying the two vectors18:44
@HeikoSsee?18:44
OXPHOSBut this is based on the assumptions that the vector will be used (for calculation) for only once18:45
OXPHOSEvery SGVector can have an associated Vector18:46
OXPHOSSpeaking of which, since Vector is just a shallow copy of SGVector, what is the difference of having only SGVector?18:47
OXPHOSHeikoS^18:47
@HeikoSsanuj: can you check CMakeLists.txt in the main dir18:47
@HeikoSline 192, 19318:47
@HeikoStry adding them there18:47
@HeikoS(re-run cmake afterwards)18:47
sanujokay18:47
@HeikoSsanuj: otherwise, you could also compile and link the unit tests manually18:48
@HeikoSah no18:48
@HeikoSsanuj: I dont know18:48
@HeikoSsanuj: what about checking our covarage build?18:49
@HeikoSOXPHOS: you are right, if memory lives on cpu, Vector wraps the *same* memory as SGVector used to construct it18:49
@HeikoSwhy we should have another class for it, in my opinion is separation of concerns18:50
@HeikoSe.g. no element-wise access18:50
@HeikoSsanuj: we can probably also add them to the debug build, but need to ask besser82 or wiking for that18:51
@HeikoSsanuj: http://buildbot.shogun-toolbox.org/builders/coverity%20analysis/builds/10218:51
@HeikoSever checked this one?18:51
sanujno18:52
sanujthis is the coverage build?18:52
@HeikoSI think only static analysis18:53
sanujHeikoS, the alignment looks weird now http://buildbot.shogun-toolbox.org/cookbook_pr/708245f57c6f5a0d5385277303ff714587bc3ab6/examples/classifier/kernel_svm.html18:53
sanujit is again perfectly fine on my local18:53
sanujdo we have old version of sphinx18:54
@HeikoScan you send the pr link again?18:55
@HeikoSalignment weird indeed18:55
@HeikoSbut maybe its just some latex command that should not be used18:55
@HeikoSlemme check18:55
OXPHOSHeikoS: i see. the current idea is, Vector holds two vectors, one on GPU, one on CPU. If onGPU is true, CPUvec part is masked. As the calculation is done on GPU and CPU won't be updated18:55
sanujHeikoS, PR is here https://github.com/shogun-toolbox/shogun/pull/328118:55
OXPHOSHiekoS: that's why we transfer_to_CPU everytime we try to copy it to a new Vector18:55
@HeikoSOXPHOS: lets have a private chat18:56
@HeikoSI get confused otherwise ;)18:56
OXPHOSHeikoS: Can't use this command in this window..can you receive me?18:57
OXPHOS"Can't use this command in this window"18:57
@HeikoSok then :)18:57
@HeikoSnevermind18:57
@HeikoSVector holds two pointers18:57
@HeikoSCPU one points to the data that was handed in on construction18:58
@HeikoSi.e. the one in the SGVector18:58
sanujHeikoS, normally when do you get time to review/merge PRs and chat on irc?18:58
@HeikoSsanuj: nowish18:58
OXPHOSHeikoS: yes18:58
sanujHeikoS, earlier than this not possible?18:59
@HeikoSOXPHOS: I did not get what you mean by this transfer thing you said18:59
@HeikoSsanuj: sometimes :)18:59
sanujjust asking because my sleeping time starts :D18:59
@HeikoSsanuj: haha18:59
@HeikoSyeah I am usually around earlier as well18:59
@HeikoSbut at some point I also need to work18:59
sanujyeah18:59
@HeikoSso sometimes I turn off irc18:59
sanuji understand19:00
@HeikoS(remember I am not paid for this, but for my normal job ;) )19:00
sanuj:D19:00
sanujHeikoS, you are the best :D19:00
OXPHOSHeikoS: I was trying to justify that transfer stuff - and also the copy problem - let me start over again19:00
@HeikoSOXPHOS: ok19:00
@HeikoSso a copy should only ever happen19:01
@HeikoSwhen data is put on gpu19:01
@HeikoSsanuj: its the &&19:01
@HeikoSremove it and all will be good19:01
sanujokay19:01
OXPHOSHeikoS: so CPU and GPUptr of Vector should be synchronized some time, because the calculation is done on one end19:01
OXPHOSHeikoS: if Vector is a shallow copy of SGVector, there's chance that the vector on CPU is modified, like SGVector[i] = 019:02
OXPHOSand the *Vector.CPUptr is also changed19:03
@HeikoSno need for that19:03
OXPHOSHeikoS: but there is no way for Vector to mark this change as it is done in SGVector19:03
@HeikoSfor the syncronisation I mean19:03
@HeikoSyou copy data to GPU, then you can forget about the data in CPU (it can even be freed, i.e. ref-counter decreased)19:04
@HeikoSyou move data to GPU, then you do stuff with it, and at some point you transfer it back to main memory, then you have to allocate memory for that again (or use previously allocated)19:04
OXPHOSIf CPUptr is not freed - re-use previous allocated memory - SGVector is changed; if the CPUptr is freed - new memory - new SGVector19:05
OXPHOSThis should be something user decides? or we should?19:06
OXPHOSHeikoS^19:07
@HeikoSuser19:07
@HeikoSoption a) allocate new block for transfering back19:07
@HeikoSwith fresh ref-counter as it was just created19:07
@HeikoSoption b) user provides pointer (or SGVector) that has pre-allocated memory19:08
OXPHOSHeikoS: i see19:09
@HeikoSabout this thing of re-using the memory that was used to copy things to GPU, i need to think19:10
@HeikoSbut thats not a big difference19:10
@HeikoSand easy to change19:10
@HeikoSthe main point is: if data is never transferred to GPU, it is never copied19:10
OXPHOSHeikoS: cool. make sense19:11
sanujHeikoS, is this fine http://buildbot.shogun-toolbox.org/cookbook_pr/c29762cbcd29c568b084a32be454a1b7c5bf5319/examples/classifier/kernel_svm.html19:11
sanujHeikoS, how did you know && was causing the problem?19:12
@HeikoScan you add a new line for the constraints and say "subject to" in the text?19:12
@HeikoSsanuj: intuition19:12
sanujcan i include "subject to" in the math area19:12
@HeikoSsanuj:19:12
@HeikoSno19:12
@HeikoSdo two math sections19:12
@HeikoSand normal text in between19:12
sanujcool19:12
OXPHOSHeikoS: honestly now I prefer the old idea again - have separate CPUVector and GPUVector. you see if we're doing calculation on GPU, we have to go through a CPUptr step19:12
@HeikoSOXPHOS: so what makes this bad is19:13
@HeikoSthat things are decided at compile time them19:13
@HeikoSand we dont want that19:13
@HeikoSthink: what would be the signature of linalg::dot?19:13
@HeikoSwould it accept CPUVector or GPUVector?19:13
@HeikoSand how to change this at runtime19:14
@HeikoSimpossible19:14
@HeikoSthats the idea of not having separate classes for CPU/GPU19:14
@HeikoSOXPHOS: see that?19:14
OXPHOSHeikoS sry im not familiar with compile/runtime stuff..it won't work even we have a factory method and base class, and everything was passed as base class ptr(also linalg::dot takes baseptr)?19:16
@HeikoSnope19:16
OXPHOS;(19:16
@HeikoSOXPHOS: it is very easy to understand19:16
@HeikoSthink about it this way:19:17
@HeikoSthe difference is WHEN you take the decision which method is called19:17
@HeikoSlinalg:dot(CPUVec) or linalg:dot(GPUVec)19:17
@HeikoScompile time, this is decided when you compile shogun19:17
@HeikoSand cannot be changed afterwards19:17
@HeikoSruntime: you can change the behaviour at runtume19:18
@HeikoSnow with a base class pointer, you would think that can be avoided19:18
@HeikoSas one can pass both CPU and GPU19:18
@HeikoSBUT19:18
@HeikoSthe factory, which one does it return? this is also decided at compile time19:18
@HeikoSwhich hard-wires the backend into shogun binary19:18
@HeikoSand cannot be changed anymore after it has been compiled19:19
OXPHOSokay so it's not like we can switch backend anytime after compiling19:19
OXPHOScan i try, in Vector constructor, if HAVE_VIENNACL, everything goes to GPU directly? skipping that CPUptr step?19:21
@HeikoSyes19:22
sanujHeikoS, what do you think about this http://buildbot.shogun-toolbox.org/cookbook_pr/c4ab6ffb384ea563327f64612b659033e9e21f13/examples/classifier/kernel_svm.html19:22
@HeikoSOXPHOS: nope19:22
@HeikoSevery copy to GPU is expensive, right?19:22
@HeikoSso it should never happen implicitly19:22
@HeikoSbut always require the user to call a method19:23
@HeikoSthat has a warning sign on it19:23
@HeikoSsanuj: thats it :)19:23
@HeikoSsanuj: check whether travis executes you integration test, then ping me and I wil lmerge19:23
OXPHOSHeikoS: okay i see. thanks so much!19:24
sanujHeikoS, travis is going to take time19:24
sanujand i'm going to sleep :D19:24
OXPHOSHeikoS: hope wiking won't be upset if i remove that aligned stuff XD19:24
@HeikoSsanuj: good night :)19:24
@HeikoSOXPHOS: yeah align....19:24
@HeikoSmmh19:25
OXPHOSsanuj: night night19:25
@HeikoSI dont know much about it tbh19:25
sanujbut one more thing before i sleep19:25
sanujhttps://github.com/shogun-toolbox/shogun/pull/328319:25
sanujHeikoS, do i need to check for mem leaks ^19:25
@HeikoSOXPHOS: but I dont see how we can align the data itself (!= the Vector class), without having to copy19:25
@HeikoSOXPHOS: we can put that back in later19:26
@HeikoSbut for sure, wiking will be angry19:26
OXPHOSHeikoS: yeah I was thinking the same..but he told me to use align before I have the deepcopy? but anyway it's easy to add back19:26
@HeikoSyeah19:26
@HeikoSkeep the snippets19:26
@HeikoSwe cant copy, thats for sure19:26
@HeikoSI mean copy in cpu case19:26
@HeikoSbefore the move_to_gpu() call19:26
OXPHOSokay19:26
@HeikoSOXPHOS: ok, Ill head home now19:27
@HeikoSOXPHOS: thanks for the discussion :)19:27
@HeikoSOXPHOS: and pls let me know if there are more questions19:27
sanujhome sweet home19:27
@HeikoSOXPHOS: important to getthis right19:27
@HeikoSsanuj: indeed :)19:27
@HeikoSclimbing wall is my home :D19:27
OXPHOSHeikoS it's a nice one. thanks!19:27
sanujclimbing wall ?19:28
sanujHeikoS, mem leaks?19:29
sanujokay sleeping19:30
sanujHeikoS, OXPHOS goodnight :)19:31
-!- sanuj [~sanuj@117.203.14.130] has quit [Quit: Leaving]19:31
-!- HeikoS [~heiko@nat-201-233.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.]19:39
-!- HeikoS [~heiko@nat-201-233.internal.eduroam.ucl.ac.uk] has joined #shogun19:40
-!- mode/#shogun [+o HeikoS] by ChanServ19:40
-!- HeikoS [~heiko@nat-201-233.internal.eduroam.ucl.ac.uk] has quit [Client Quit]19:41
-!- besser82 [~besser82@fedora/besser82] has quit [Ping timeout: 258 seconds]20:32
arianepaolabye everyone, see you tomorrow21:21
OXPHOSarianepaola: good night21:23
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]21:39
@wikingOXPHOS: u wrote kmeans cookbook? http://buildbot.shogun-toolbox.org/cookbook_pr/cec96964e162b2442e7d7515d2610b883f00fba6/examples/clustering/kmeans.html ?22:35
@wikingsee argmin22:35
OXPHOSwiking: yes?22:36
@wikingthat argmin has a problem right?22:37
@wikinglol22:37
@wikingnot22:37
@wiking(my browser/internet)22:37
@wikingsorry22:37
OXPHOSwiking haha22:37
OXPHOSso it's fine?22:37
@wikingyeo22:39
-!- tctara [~quassel@128.199.61.169] has joined #shogun23:53
OXPHOSwiking: there?23:58
--- Log closed Thu Jun 16 00:00:42 2016
« Tuesday, 2016-06-14 Index Thursday, 2016-06-16 »

Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!