Open in new window / Try shogun cloud
--- Log opened Tue Mar 15 00:00:30 2016
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun00:14
shogun-notifier-shogun: Sanuj :develop * 4b6eb6b / / (4 files): https://github.com/shogun-toolbox/shogun/commit/4b6eb6b9d7a6a1de0bd5bc88def2168b29a89b9400:14
shogun-notifier-shogun: add linear ridge regression in cookbook00:14
shogun-notifier-shogun: Heiko Strathmann :develop * 32270be / / (4 files): https://github.com/shogun-toolbox/shogun/commit/32270befedc90f6a944e779aa99d3e6e2171944300:14
shogun-notifier-shogun: Merge pull request #3064 from sanuj/cookbook00:14
shogun-notifier-shogun:00:14
shogun-notifier-shogun: add least squares regression cookbook00:14
shogun-buildbotbuild #521 of trusty - libshogun - viennacl is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/trusty%20-%20libshogun%20-%20viennacl/builds/521  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Sanuj <sanuj.sharma.in@gmail.com>00:35
-!- BurningBear [~zpcxh95@121.237.188.207] has joined #shogun00:53
-!- travis-ci [~travis-ci@ec2-54-224-162-220.compute-1.amazonaws.com] has joined #shogun00:56
travis-ciit's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: https://travis-ci.org/shogun-toolbox/shogun/builds/11600058700:56
-!- travis-ci [~travis-ci@ec2-54-224-162-220.compute-1.amazonaws.com] has left #shogun []00:56
shogun-notifier-shogun: Heiko Strathmann :develop * 271ca53 / doc/cookbook/README.md: https://github.com/shogun-toolbox/shogun/commit/271ca53e9023eb2e3fc2057f1209f9f8c8bf5c3e00:57
shogun-notifier-shogun: add some comments on meta example testing00:57
shogun-buildbotbuild #522 of trusty - libshogun - viennacl is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/trusty%20-%20libshogun%20-%20viennacl/builds/52200:58
shogun-notifier-shogun: Heiko Strathmann :develop * 8a6a9fe / CMakeLists.txt: https://github.com/shogun-toolbox/shogun/commit/8a6a9fedb6013e9052cb985f80d26c9fe453d7da01:03
shogun-notifier-shogun: only run meta integration tests when meta examples are actually built01:03
shogun-buildbotbuild #3641 of deb1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/3641  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>01:19
shogun-buildbotbuild #3642 of deb1 - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/364201:21
shogun-buildbotbuild #351 of CentOS7 - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/CentOS7%20-%20libshogun/builds/351  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>01:21
shogun-buildbotbuild #1143 of precise - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/1143  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>01:21
shogun-buildbotbuild #148 of FC22 - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC22%20-%20libshogun/builds/148  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>01:21
shogun-buildbotbuild #2799 of bsd1 - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/2799  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>01:22
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has quit [Quit: Leaving.]01:35
-!- travis-ci [~travis-ci@ec2-54-227-61-151.compute-1.amazonaws.com] has joined #shogun01:43
travis-ciit's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: https://travis-ci.org/shogun-toolbox/shogun/builds/11600855501:43
-!- travis-ci [~travis-ci@ec2-54-227-61-151.compute-1.amazonaws.com] has left #shogun []01:43
shogun-buildbotbuild #758 of deb4 - python3 is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb4%20-%20python3/builds/758  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>02:06
shogun-buildbotbuild #595 of clang34 - thread analysis is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/595  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>02:09
-!- will__ [cc2db69d@gateway/web/freenode/ip.204.45.182.157] has joined #shogun02:13
will__Hi guys! I have installed shogun and run some examples successfully. Please give me some guide on what to do next.02:14
-!- travis-ci [~travis-ci@ec2-54-224-162-220.compute-1.amazonaws.com] has joined #shogun02:15
travis-ciit's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: https://travis-ci.org/shogun-toolbox/shogun/builds/11600963402:15
-!- travis-ci [~travis-ci@ec2-54-224-162-220.compute-1.amazonaws.com] has left #shogun []02:15
shogun-buildbotbuild #688 of clang34 - undefined behaviour analysis is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20undefined%20behaviour%20analysis/builds/688  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>02:16
will__Oh, btw, I'm a GSoC student hoping to participate in a project of shugun02:21
-!- GandalfTheWizard [~ian@112.10.106.188] has quit [Ping timeout: 268 seconds]02:25
-!- curiousguy13 [~curiousgu@120.59.228.40] has quit [Ping timeout: 252 seconds]02:32
-!- will___ [cc2db69d@gateway/web/freenode/ip.204.45.182.157] has joined #shogun02:36
-!- will__ [cc2db69d@gateway/web/freenode/ip.204.45.182.157] has quit [Ping timeout: 252 seconds]02:38
-!- GandalfTheWizard [~ian@112.10.106.188] has joined #shogun02:45
-!- GandalfTheWizard [~ian@112.10.106.188] has quit [Ping timeout: 248 seconds]02:52
-!- GandalfTheWizard [~ian@112.10.106.188] has joined #shogun03:11
-!- GandalfTheWizard [~ian@112.10.106.188] has quit [Ping timeout: 276 seconds]03:19
shogun-buildbotbuild #604 of debian wheezy - memcheck is complete: Failure [failed memory check]  Build details are at http://buildbot.shogun-toolbox.org/builders/debian%20wheezy%20-%20memcheck/builds/604  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>03:30
-!- BurningBear [~zpcxh95@121.237.188.207] has quit []03:35
-!- GandalfTheWizard [~ian@112.10.106.188] has joined #shogun03:40
-!- GandalfTheWizard [~ian@112.10.106.188] has quit [Ping timeout: 244 seconds]03:47
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]04:03
-!- GandalfTheWizard [~ian@112.10.106.188] has joined #shogun04:07
-!- GandalfTheWizard [~ian@112.10.106.188] has quit [Ping timeout: 250 seconds]04:14
-!- GandalfTheWizard [~ian@112.10.106.188] has joined #shogun04:20
-!- GandalfTheWizard [~ian@112.10.106.188] has quit [Max SendQ exceeded]04:20
-!- GandalfTheWizard [~ian@112.10.106.188] has joined #shogun04:22
-!- GandalfTheWizard [~ian@112.10.106.188] has quit [Max SendQ exceeded]04:23
-!- curiousguy13 [~curiousgu@120.59.102.243] has joined #shogun04:34
-!- GandalfTheWizard [~ian@112.10.106.188] has joined #shogun04:41
shogun-buildbotbuild #1095 of nightly_default is complete: Failure [failed notebooks]  Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/1095  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Sanuj <sanuj.sharma.in@gmail.com>04:42
-!- sanuj [0e8bc409@gateway/web/freenode/ip.14.139.196.9] has joined #shogun05:19
-!- sanuj [0e8bc409@gateway/web/freenode/ip.14.139.196.9] has quit [Client Quit]05:21
-!- sanuj [0e8bc409@gateway/web/freenode/ip.14.139.196.9] has joined #shogun05:23
-!- sanuj [0e8bc409@gateway/web/freenode/ip.14.139.196.9] has quit [Ping timeout: 252 seconds]05:29
-!- lambday [6a33854e@gateway/web/freenode/ip.106.51.133.78] has quit [Ping timeout: 252 seconds]06:03
-!- GandalfTheWizard [~ian@112.10.106.188] has quit [Ping timeout: 260 seconds]06:26
-!- will___ [cc2db69d@gateway/web/freenode/ip.204.45.182.157] has quit [Ping timeout: 252 seconds]06:29
@wikingshogun-buildbot: force build --branch=develop 'CentOS7 - libshogun'06:42
shogun-buildbotbuild forced [ETA 2m45s]06:42
shogun-buildbotI'll give a shout when the build finishes06:42
shogun-buildbotbuild #352 of CentOS7 - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/CentOS7%20-%20libshogun/builds/35206:42
@wikingshogun-buildbot: force build --branch=develop 'CentOS7 - libshogun'06:43
shogun-buildbotbuild forced [ETA 2m45s]06:43
shogun-buildbotI'll give a shout when the build finishes06:43
-!- GandalfTheWizard [~ian@112.10.106.188] has joined #shogun06:43
shogun-buildbotbuild #353 of CentOS7 - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/CentOS7%20-%20libshogun/builds/35306:46
@wikingshogun-buildbot: force build --branch=develop 'FC22 - libshogun'06:48
shogun-buildbotbuild forced [ETA 2m58s]06:48
shogun-buildbotI'll give a shout when the build finishes06:48
-!- GandalfTheWizard [~ian@112.10.106.188] has quit [Ping timeout: 276 seconds]06:49
-!- sanuj [0e8bc409@gateway/web/freenode/ip.14.139.196.9] has joined #shogun06:51
sanujwiking: all tests are passing on my local with -DENABLE_TESTING=ON06:51
shogun-buildbotbuild #149 of FC22 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC22%20-%20libshogun/builds/14906:53
@wikingshogun-buildbot: force build --branch=develop 'bsd1 - libshogun'06:59
shogun-buildbotbuild forced [ETA 2m40s]06:59
shogun-buildbotI'll give a shout when the build finishes06:59
shogun-buildbotbuild #2800 of bsd1 - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/280007:01
@wikingshogun-buildbot: force build --branch=develop 'precise - libshogun'07:16
shogun-buildbotbuild forced [ETA 9m48s]07:16
shogun-buildbotI'll give a shout when the build finishes07:16
shogun-buildbotbuild #1144 of precise - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/114407:24
@wikingshogun-buildbot: force build --branch=develop 'deb4 - python3'07:26
shogun-buildbotbuild forced [ETA 5m23s]07:26
shogun-buildbotI'll give a shout when the build finishes07:26
shogun-buildbotbuild #759 of deb4 - python3 is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb4%20-%20python3/builds/75907:26
-!- sanuj [0e8bc409@gateway/web/freenode/ip.14.139.196.9] has quit [Quit: Page closed]07:42
-!- sanuj [0e8bc409@gateway/web/freenode/ip.14.139.196.9] has joined #shogun07:43
-!- GandalfTheWizard [~ian@112.10.106.188] has joined #shogun08:09
-!- Izaya [~Eejya@103.225.100.51] has joined #shogun08:25
-!- besser82 [~besser82@fedora/besser82] has joined #shogun08:33
-!- mode/#shogun [+o besser82] by ChanServ08:33
IzayaHi, what would be a good entrance task for the easy shogun installation project ?08:40
-!- mizari [~mizari@95-174-213-100.nts.su] has quit [Ping timeout: 276 seconds]08:54
-!- sanuj [0e8bc409@gateway/web/freenode/ip.14.139.196.9] has quit [Ping timeout: 252 seconds]09:06
-!- sanuj [0e8bc409@gateway/web/freenode/ip.14.139.196.9] has joined #shogun09:12
-!- sanuj [0e8bc409@gateway/web/freenode/ip.14.139.196.9] has quit [Ping timeout: 252 seconds]09:23
-!- Izaya [~Eejya@103.225.100.51] has quit [Quit: Leaving]09:36
-!- besser82 [~besser82@fedora/besser82] has quit [Ping timeout: 268 seconds]10:06
-!- HeikoS [~heiko@5751934b.skybroadband.com] has joined #shogun10:44
-!- mode/#shogun [+o HeikoS] by ChanServ10:44
@HeikoSwiking: jojo10:45
@wikinghey10:52
@wikingwhatsup10:52
@HeikoSwiking: this time out problem is weird10:57
@HeikoSwiking: but wanted to ask about this10:57
@HeikoShttps://github.com/shogun-toolbox/shogun/pull/306310:57
@HeikoSany objections?10:57
@HeikoSa bit cleaner for the cmake option10:57
@HeikoSlisitsyn: hi11:05
@HeikoSwiking: this PR did not cause a travis build: https://github.com/shogun-toolbox/shogun/pull/307311:19
@HeikoSany idea why?11:19
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun11:20
shogun-notifier-shogun: Saurabh7 :develop * 1a0e910 / / (4 files): https://github.com/shogun-toolbox/shogun/commit/1a0e910add23dba2d9d8482cf1ab53ff0c1773a911:20
shogun-notifier-shogun: add precompute to base Distance11:20
shogun-notifier-shogun: Heiko Strathmann :develop * 93ae619 / / (4 files): https://github.com/shogun-toolbox/shogun/commit/93ae619a2cb727d5175952bf4991aefb6865f26011:20
shogun-notifier-shogun: Merge pull request #3073 from Saurabh7/prcomp_base11:20
shogun-notifier-shogun:11:20
shogun-notifier-shogun: add precompute to base Distance11:20
curiousguy13HeikoS, hey11:23
@HeikoScuriousguy13: hi11:24
curiousguy13i wanted to talk to you regarding this : https://github.com/shogun-toolbox/shogun/pull/307111:24
curiousguy13so are we completely removing the NATIVE interface from shogun?11:24
@HeikoScuriousguy13: yes, we now have eigen3 always enabled11:25
curiousguy13ohh, ok11:25
@HeikoScuriousguy13: but obviously only delete if the function is implemented elsewhere11:25
@HeikoSand if not, port it to eigen311:25
curiousguy13ok, so range fill is implemented as part of the SGVector class11:26
shogun-buildbotbuild #2801 of bsd1 - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/2801  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Saurabh7 <saurabh.mahindre@gmail.com>11:26
curiousguy13so, I should just port my NATIVE implementation to Eigen311:27
curiousguy13?11:27
shogun-buildbotbuild #150 of FC22 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC22%20-%20libshogun/builds/150  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Saurabh7 <saurabh.mahindre@gmail.com>11:28
@HeikoScuriousguy13: no first check whether that is necessary11:29
@HeikoScuriousguy13: if there is no such case, just remove the native11:29
curiousguy13okay11:32
curiousguy13so, if we add a new method to linalg like cholesky we need to add it's eigen3 implemetation, right?11:32
curiousguy13HeikoS, ^11:33
@HeikoScuriousguy13: yes11:35
@HeikoSsee  https://github.com/shogun-toolbox/shogun/issues/307411:35
curiousguy13HeikoS, nice. I'll give this one a shot.11:38
@HeikoScuriousguy13: yes this will be really really useful11:38
curiousguy13HeikoS, I am interested in the shogun-detox GSoC project btw.11:38
@HeikoScuriousguy13: make sure to send small patches11:38
@HeikoSand aim for high unit test coverage11:38
@HeikoScuriousguy13: cool good to hear11:38
@HeikoScuriousguy13: you are the second person to be interested in it iirc11:39
lisitsynh e l l o11:39
@HeikoSlisitsyn: SERGEY11:39
@HeikoSNO EXCUSES11:39
curiousguy13HeikoS, cool. I'll ask if I have any doubts. Thanks :)11:39
lisitsynHeikoS: SORRY11:39
@HeikoSlisitsyn: NOT ACCEPTE11:40
@HeikoSD11:40
@HeikoS!!!11:40
lisitsynBUT I AM SORRY11:40
@HeikoSlisitsyn: ok then11:42
@HeikoSyou shall be forgiven11:42
lisitsynlol11:42
lisitsynlets review some prs11:42
@HeikoSlisitsyn: I did a few11:42
@HeikoSlisitsyn: there is a weird time out11:42
@HeikoSon travis11:42
@HeikoSno idea why11:43
@HeikoScannot reprouce locally11:43
lisitsynmean computation +1236 lines11:43
@HeikoSlisitsyn: I told her  to re-do in small bits11:43
@HeikoSlisitsyn: hey how can I update the website at beta?11:43
@HeikoSjust push to web2?11:43
lisitsynHeikoS: not yet11:43
lisitsynbut soon11:43
@HeikoSokok11:44
@HeikoSlooking forward11:44
@HeikoSthe time out is caused by shogun-unit-test11:44
lisitsynmeh that sucks11:45
@HeikoSwiking: ^11:45
@HeikoSjust managed to get this locally11:45
lisitsynHeikoS: any idea what happened?11:45
@HeikoSlisitsyn: no idea11:45
@HeikoSit just doesnt terminate11:45
@HeikoSIll check recently added unit tests11:45
lisitsynHeikoS: can you reproducei t?11:46
lisitsynlets use gdb11:46
@HeikoSprobably some joker did an infinite loop that only appears random11:46
@HeikoSlisitsyn: I compiled with -DENABLE_TESTING=On and everything else default11:46
lisitsynHeikoS: just run it and attach gdb11:46
@HeikoSlisitsyn: I leave that to you :)11:46
lisitsynmeeeh11:46
lisitsynok let me try11:46
@HeikoSIll update the cookbook style11:46
lisitsynHeikoS: ah actually11:48
lisitsynI tried read the docs theme11:48
lisitsynbut our tabs break with that11:48
lisitsynbecause it has a different layout11:48
shogun-buildbotbuild #354 of trusty - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/trusty%20-%20libshogun/builds/354  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Saurabh7 <saurabh.mahindre@gmail.com>11:49
lisitsynyou'd need some repair for that11:49
@HeikoSlisitsyn: ah11:49
@HeikoSok then I wont do it in that case11:49
@HeikoSlisitsyn: and also leave it to you11:49
@HeikoSas you already started ;)11:49
lisitsynHeikoS: it was in 2015 I think11:50
lisitsyn:D11:50
@HeikoSI see11:50
@HeikoSwill try again now11:50
lisitsynHeikoS: yeah just try and lets see11:50
lisitsynI bet tabs would look ugly11:50
lisitsynbut we can try to repair it11:50
@HeikoSlisitsyn: it looks nice11:54
@HeikoSbut it doesnt work :)11:54
lisitsynHeikoS: ok11:54
lisitsynshould be possible to repair11:54
lisitsynHeikoS: can you do a branch so I fix it?11:54
@HeikoSlisitsyn: sure, but travis wont build it11:55
@HeikoSlisitsyn: so not sure a branch makes sense11:55
@HeikoSI think better do locally11:55
@HeikoSits just adding 3 lines to the conf.py to change style11:55
shogun-buildbotbuild #1145 of precise - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/1145  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Saurabh7 <saurabh.mahindre@gmail.com>11:55
lisitsynHeikoS: yeah give me diff then ;)11:55
@HeikoS# use fancy style11:55
@HeikoSimport sphinx_rtd_theme11:55
@HeikoShtml_theme = "sphinx_rtd_theme"11:55
@HeikoShtml_theme_path = [sphinx_rtd_theme.get_html_theme_path()]11:55
@HeikoSlast three lines of conf.py11:55
@HeikoSlisitsyn: but the unit test freeze is more important11:56
@wikingHeikoS: sorry didn't see12:06
@wikingwas out12:06
@HeikoSwiking: no worries12:06
@wikingwhat's happening with the stupid tests again12:06
@HeikoSwiking: unit tests infinite loop12:06
shogun-buildbotbuild #760 of deb4 - python3 is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb4%20-%20python3/builds/760  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Saurabh7 <saurabh.mahindre@gmail.com>12:06
@HeikoScurrently trying to isolate, it happens random12:06
@wikingHeikoS: yeah maybe that should be debugged12:06
@wikingshogun-buildbot: force build --branch=develop 'precise - libshogun'12:07
shogun-buildbotbuild forced [ETA 9m02s]12:07
shogun-buildbotI'll give a shout when the build finishes12:07
@wikingwe need to fix ctags and ply detection12:07
@wikingas they are not complete12:07
@wikingsee python3 and fbsd builds12:07
@HeikoSwiking: will have a look12:07
@HeikoSwiking: btw the doxygen upload is broken12:07
@HeikoSonly class list is there12:08
@HeikoSI am tending to just leave it, the doxygen docs were horrible anyways12:08
@wikingno they weren't horrible12:09
@wikingi was using the first time i used that12:09
@wikingi mean shogun12:09
@HeikoSwiking: by horrible I mean outdated12:09
shogun-buildbotbuild #596 of clang34 - thread analysis is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/596  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Saurabh7 <saurabh.mahindre@gmail.com>12:09
@HeikoSunmaintained12:09
@wikingHeikoS: so fucking not true :)12:09
@HeikoSwhich of the pages you want to keep?12:10
@wikingall12:10
@wiking?12:10
@HeikoSexamples are just pasted code12:10
@HeikoSeventually these are replaced by meta12:10
@HeikoSthats the biggest chunk12:10
@HeikoSmany other bits are just inserted from our readmes12:10
@wikingyes12:10
@HeikoSthat can stay12:10
@wikingbtw12:11
@wikingi'm just checkong now doxygen12:11
@wikingwhat's the actual problem?12:11
@HeikoSwell anyways, no need to decide now, just saying it is broken12:11
@HeikoSwiking: I moved it into a subfolder12:11
@HeikoSwiking: not sure what exactly goes wrong, but I think its just the upload12:11
@wikinghave you actually12:12
@wikingchecked12:12
@wikingsince you are talking12:12
@wiking?:)12:12
@wikinghttp://shogun.ml/doc/en/latest/namespaceshogun.html12:12
@HeikoSyes, class list is there12:12
@HeikoSthe rest is not12:12
@wikingcan you show me12:12
@wikingso i can easily debug12:12
@HeikoShttp://shogun.ml/doc/en/latest/index.html12:12
@wikingok12:12
@wikingchecking and fixing12:13
@HeikoSvs http://shogun.ml/doc/en/3.0.0/index.html12:13
@HeikoSok gotta run, see you later12:13
shogun-buildbotbuild #1146 of precise - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/114612:16
shogun-buildbotbuild #689 of clang34 - undefined behaviour analysis is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20undefined%20behaviour%20analysis/builds/689  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Saurabh7 <saurabh.mahindre@gmail.com>12:17
-!- travis-ci [~travis-ci@ec2-54-224-162-220.compute-1.amazonaws.com] has joined #shogun12:18
travis-ciit's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: https://travis-ci.org/shogun-toolbox/shogun/builds/11608370112:18
-!- travis-ci [~travis-ci@ec2-54-224-162-220.compute-1.amazonaws.com] has left #shogun []12:18
-!- HeikoS [~heiko@5751934b.skybroadband.com] has quit [Ping timeout: 244 seconds]12:18
-!- mizari [~mizari@isc.tsu.ru] has joined #shogun12:30
-!- HeikoS [~heiko@untrust-out.swc.ucl.ac.uk] has joined #shogun12:45
-!- mode/#shogun [+o HeikoS] by ChanServ12:45
-!- sanuj [0e8bc402@gateway/web/freenode/ip.14.139.196.2] has joined #shogun12:47
@HeikoSwiking: btw when I said remove doxygen docs, I meant everything but the class list.12:55
@HeikoSwiking: did we talk about the same thing?12:55
@HeikoSlisitsyn: any results regarding the style thing?13:01
lisitsynHeikoS: well I'm working ;)13:17
sanujHeikoS: hi13:21
sanujHeikoS: dunno what to do about the buildbot freezing due to knn. doesn't happen on my laptop.13:24
@wikingHeikoS: btw i'm just checkoing13:25
@wikingthat thing has been broken13:25
@wikingfor a while :D13:25
shogun-buildbotbuild #605 of debian wheezy - memcheck is complete: Failure [failed memory check]  Build details are at http://buildbot.shogun-toolbox.org/builders/debian%20wheezy%20-%20memcheck/builds/605  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Saurabh7 <saurabh.mahindre@gmail.com>13:26
@wikingi guess there was again13:26
@wikinga none backward compatibility change13:26
@wikingin the mds13:26
@wikingman we cannot scale like this13:28
@wikingthat every time there's an unilateral change13:28
@wikingand somebody else have to clear up the mess13:28
@HeikoSsanuj: its not knn13:30
@HeikoSsanuj: its the unit test that freezes13:30
@HeikoSwiking: details?13:30
@HeikoSlisitsyn: do you know about shogun's group lasso solver?13:31
@wikingHeikoS: details: 1) doxygen changing from one dir to another 2) md changes 3) adding ply and other dependencies13:31
@wikingif you do these things13:31
@wikingyou do all the way to the end13:31
@wikingfixing buildbot13:31
@wikingfixing doxygen generation13:31
@wikingetc.13:31
@wikingif you cannot13:31
@HeikoSwiking: I wish I could13:31
@wikingthen before merging13:31
@wikingyou take care13:31
@HeikoSbut I dont have buildbot access13:31
@wikingof communicating13:31
@wikingand not merge13:32
@wikingand then wait others to fix it13:32
@wikingand pls13:32
@wikingwhen you squash a commit13:32
@wikingdo not13:32
@wikingever put there your commit history13:32
@wikingtry to come up with a real commit history log13:33
lisitsynHeikoS: I do13:33
@HeikoSwiking: will avoid next time, thanks13:33
lisitsynHeikoS: why?13:33
@HeikoSwiking: the thing with the merges is that nothing ever moves if I wait for poeple to react13:33
@wikingyes13:33
@wikingi dont think i dont react13:33
@wikingin a reasonable time13:33
@HeikoSwiking: but how can I help now with the fix?13:33
@wikingdunnno beacuse i have no idea13:34
@wikinghow this got broken13:34
@wikingand it's not only because13:34
@wikingcookbook changes13:34
@HeikoSwiking: the md change was long time ago13:35
@HeikoSdont think it was broken since then13:35
@HeikoSlisitsyn: I need a group lasso13:36
-!- mizari [~mizari@isc.tsu.ru] has quit [Ping timeout: 240 seconds]13:36
@HeikoSlisitsyn: can I just give it objective function that I implement myself?13:36
@HeikoSand then define groups in variable?13:36
@HeikoSor do I need to give it in matrix form?13:36
lisitsynHeikoS: you'd need to define groups as objects13:36
lisitsynhttps://github.com/shogun-toolbox/shogun/blob/develop/examples/undocumented/python_modular/classifier_featureblock_logistic_regression.py13:37
@HeikoSlisitsyn: the thing is I have an objective that I want to avoid formulating it as a matrix thing13:37
-!- mizari [~mizari@isc.tsu.ru] has joined #shogun13:37
lisitsynHeikoS: our solver is heavily specialized13:37
@HeikoSlisitsyn: this is logistic regression13:37
@HeikoSI just want squared loss13:38
@HeikoSi.e. regression13:38
@HeikoSbut it is not quite regression13:38
@HeikoSI have my own objective13:38
lisitsynHeikoS: https://github.com/shogun-toolbox/shogun/blob/develop/src/shogun/lib/slep/slep_mc_plain_lr.cpp13:40
lisitsynthat's where it is implemented13:41
@HeikoSlisitsyn: so it has to be matrix form13:41
@HeikoSgna13:41
@HeikoSpitty13:41
@HeikoSmy matrix is sparse13:41
@HeikoSdont want to store it13:41
lisitsynHeikoS: it uses dense_dot_range13:41
@HeikoSin the form of features13:41
lisitsyndot features13:42
@HeikoSlisitsyn: ah so I can overload13:42
lisitsynI think so13:42
@HeikoScool that helped13:42
@HeikoSwill have a look13:42
@HeikoSwiking: it is LaRank unit test13:42
@HeikoSthat freezes13:42
@HeikoSat least locally on my machine13:43
sanujHeikoS: LaRank freezes on my machine also sometimes13:43
sanujHeikoS: i'm using sinc for krr cookbook page13:44
@HeikoSsanuj: thats fine, how big is training and test data? are there test labels?13:44
@HeikoSor are you adding a new?13:45
sanujHeikoS: no i added in the previous update only13:45
sanujHeikoS: training has 100 samples and testing has 1013:45
@HeikoSgood13:48
shogun-notifier-shogun: Heiko Strathmann :develop * 6588827 / tests/unit/multiclass/LaRank_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/65888277a0a167424a9d5ed9cc353d669b1bbada13:52
shogun-notifier-shogun: disable la rank unit test as sometimes freezes13:52
@HeikoSwiking: ^ lets hope that does it13:54
@HeikoSshogun-buildbot: force build --branch=develop 'precise - libshogun'13:54
shogun-buildbotThe build has been queued, I'll give a shout when it starts13:54
shogun-buildbotbuild #2802 of bsd1 - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/2802  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>13:54
@HeikoSlisitsyn: can you check this, ctags http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/2802/steps/configure/logs/stdio13:55
@HeikoSdont know how exactly you added it13:55
shogun-buildbotbuild #151 of FC22 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC22%20-%20libshogun/builds/151  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>13:56
shogun-buildbotbuild #355 of trusty - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/trusty%20-%20libshogun/builds/35513:56
lisitsynHeikoS: argh13:56
@HeikoSbuildbot madness13:57
@HeikoSlisitsyn:  check this: http://buildbot.shogun-toolbox.org/builders/FC22%20-%20libshogun/builds/150/steps/test/logs/stdio13:58
@HeikoSlisitsyn: looks like we need bettter string handling in cmake13:58
@HeikoSreged replace is not reliable13:59
@HeikoSregex13:59
shogun-buildbotbuild forced [ETA 9m32s]14:03
shogun-buildbotI'll give a shout when the build finishes14:03
shogun-buildbotHey! build precise - libshogun #1148 is complete: Success [build successful]14:11
shogun-buildbotBuild details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/114814:11
-!- mizari [~mizari@isc.tsu.ru] has quit [Ping timeout: 244 seconds]14:15
-!- travis-ci [~travis-ci@ec2-54-167-124-247.compute-1.amazonaws.com] has joined #shogun14:23
travis-ciit's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: https://travis-ci.org/shogun-toolbox/shogun/builds/11608370114:23
-!- travis-ci [~travis-ci@ec2-54-167-124-247.compute-1.amazonaws.com] has left #shogun []14:23
shogun-buildbotbuild #761 of deb4 - python3 is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb4%20-%20python3/builds/761  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>14:25
shogun-buildbotbuild #597 of clang34 - thread analysis is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/597  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>14:29
@HeikoSwiking: I attempted some fixes ^   you have to be patient with me ;)14:31
@HeikoSwiking: btw can we do something about benchmarks? some ppl tell me they ran things in mlpacks benchmarks -- it would be cool to collect them somewhere. Is that a lot of work?14:32
-!- bollu [~Adium@1.186.2.226] has joined #shogun14:32
@HeikoSwiking: can you check the kmeans update Pr and merge if it is fine? From my side, only squash is needed14:33
-!- travis-ci [~travis-ci@ec2-54-167-124-247.compute-1.amazonaws.com] has joined #shogun14:34
travis-ciit's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: https://travis-ci.org/shogun-toolbox/shogun/builds/11611182214:34
-!- travis-ci [~travis-ci@ec2-54-167-124-247.compute-1.amazonaws.com] has left #shogun []14:34
sanujHeikoS: what kind of kernels can be passed to krr?14:35
@HeikoSsanuj: all I guess14:35
@HeikoSwhy?14:35
@HeikoSany CKernel14:36
@HeikoSread the code to confirm ;)14:36
@HeikoSI dont know14:36
shogun-buildbotbuild #690 of clang34 - undefined behaviour analysis is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20undefined%20behaviour%20analysis/builds/690  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>14:36
sanujgaussian is used everywhere which is a CDotkernel14:36
sanujHeikoS: will see the code14:36
bolluhey guys, I'd come here asking for bugs to squash in shogun, and with help writing a project proposal related to cleaning up shogun. Where do I start?14:42
bolluS14:42
-!- bollu [~Adium@1.186.2.226] has quit [Quit: Leaving.]14:45
-!- bollu [~Adium@1.186.2.226] has joined #shogun14:49
-!- bollu [~Adium@1.186.2.226] has quit [Quit: Leaving.]14:56
sanujHeikoS: there are 2 training methods in krr15:04
sanujshall i talk about that in the cookbook page?15:04
-!- besser82 [~besser82@fedora/besser82] has joined #shogun15:09
-!- mode/#shogun [+o besser82] by ChanServ15:09
-!- bollu [~Adium@1.186.2.226] has joined #shogun15:17
-!- bollu [~Adium@1.186.2.226] has quit [Max SendQ exceeded]15:17
-!- bollu [~Adium@1.186.2.226] has joined #shogun15:18
-!- bollu [~Adium@1.186.2.226] has quit [Client Quit]15:19
@HeikoSsanuj: hi15:26
@HeikoSsanuj: no, in fact, can you make it use eigen3, and remove this gauss seidel shit? :)15:26
@HeikoSsanuj: no need for that15:27
@HeikoSthe matrix is pdf15:27
@HeikoSpsd15:27
@HeikoSjust use Eigen's LLT solve15:27
sanujHeikoS: yeah, i just realized that15:27
@HeikoSno hand implemented solver in KRR class thats crazy15:27
shogun-notifier-shogun: Heiko Strathmann :develop * e13b939 / / (9 files): https://github.com/shogun-toolbox/shogun/commit/e13b93907c3b81563bab2fac1b6085eab86b295615:27
shogun-notifier-shogun: avoid using REGEXP for untangling filenames in meta examples15:27
shogun-notifier-shogun: Heiko Strathmann :develop * ac14e1b / / (9 files): https://github.com/shogun-toolbox/shogun/commit/ac14e1b11a993e2d1de29ace6e149dcc768eaa1b15:27
shogun-notifier-shogun: Merge pull request #3076 from karlnapf/feature/meta_example_cmake_fix15:27
shogun-notifier-shogun:15:27
shogun-notifier-shogun: fix FC22 buildbot (hopefully)15:27
-!- mizari [~mizari@95-174-213-100.nts.su] has joined #shogun15:28
@HeikoSshogun-buildbot: force build --branch=develop 'FC22 - libshogun'15:28
shogun-buildbotbuild forced [ETA 2m58s]15:28
shogun-buildbotI'll give a shout when the build finishes15:28
sanujHeikoS: i can try to make krr use eigen3 llt15:28
@HeikoSsanuj: yes definitely15:28
@HeikoSalso get rid of lapack calls in there15:28
@HeikoSsanuj: and please put the method where (K+I)^-1y is computed in a method that is virtual15:29
shogun-buildbotbuild #152 of FC22 - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/FC22%20-%20libshogun/builds/15215:29
@HeikoSthen we can overload this for the approx KRR versions15:29
sanujHeikoS: sounds good15:29
@HeikoSwiking: http://buildbot.shogun-toolbox.org/builders/FC22%20-%20libshogun15:29
@HeikoShappy? :)15:29
-!- bollu [~Adium@1.186.2.226] has joined #shogun15:31
-!- bollu [~Adium@1.186.2.226] has quit [Client Quit]15:33
sanujHeikoS: i wanted to reference a bibtex for krr15:34
@HeikoSsanuj: good idea15:35
sanujdo i need to make a different reference.bib15:35
@HeikoSput in landmark papers or book chapters15:35
@HeikoSsanuj: no just put it in the existing15:35
@HeikoSsanuj: please check the readme, if it is not clear, please update it :)15:35
sanujHeikoS: i tried that15:35
sanujand this in rst => .. bibliography:: ../../references.bib15:36
@HeikoSdidnt work?15:36
sanujHeikoS: got this citation David Barber. Bayesian reasoning and machine learning. Cambridge University Press, 2012.15:37
sanujthis is for knn15:37
@HeikoSbibtex has tags15:39
@HeikoS:cite:`barber2012bayesian`15:39
@HeikoSso barber2012bayesian needs to be in the bibtex15:39
sanujcool15:39
@HeikoSyou can put more entried in the bibtex file15:40
sanujlet me try15:40
-!- bollu [~Adium@1.186.2.226] has joined #shogun15:40
@HeikoSand then do cite:`yourNewTag2012`15:40
-!- bollu [~Adium@1.186.2.226] has quit [Client Quit]15:41
sanujHeikoS: i'll read the sphinx bibtex doc once ;)15:41
shogun-buildbotbuild #606 of debian wheezy - memcheck is complete: Failure [failed memory check]  Build details are at http://buildbot.shogun-toolbox.org/builders/debian%20wheezy%20-%20memcheck/builds/606  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>15:43
shogun-buildbotbuild #356 of CentOS7 - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/CentOS7%20-%20libshogun/builds/356  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>15:46
shogun-buildbotbuild #2803 of bsd1 - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/2803  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>15:46
@wikingHeikoS: much better15:52
@HeikoSwiking: lisitsyn has to do the ctags, he added that, I dont know what he exactly did15:53
@HeikoSwiking: btw did you see my eigen3 change ?15:53
@HeikoSwiking: in the open PR15:53
-!- besser82 [~besser82@fedora/besser82] has quit [Ping timeout: 250 seconds]15:54
sanujHeikoS: it's working now but16:01
sanujWARNING: duplicate label for keys barber2012bayesian and barber2012bayesian16:01
@HeikoSsanuj: can you gist your .bib file?16:02
sanujyeah16:02
sanujHeikoS: https://gist.github.com/sanuj/95ebbc1fd08ed5a7877316:02
@HeikoSsanme here16:06
@HeikoSmmmh16:06
@wikingHeikoS: i think you broke centos716:08
@wikinghttp://buildbot.shogun-toolbox.org/builders/CentOS7%20-%20libshogun/builds/356/steps/configure/logs/stdio16:08
@wiking;)16:08
-!- arasu [daf8068c@gateway/web/freenode/ip.218.248.6.140] has joined #shogun16:08
@HeikoSchecking in a bit16:09
@HeikoSwiking: this might be cmake version?16:09
@wikingdunn16:09
@wikingdunno16:09
@HeikoSsince it complains about unknown components of a function call I make16:09
@wikingwhich is?16:09
@HeikoSget_filename_component16:10
@HeikoSCMake Error at examples/meta/cpp/CMakeLists.txt:6 (get_filename_component):   get_filename_component unknown component DIRECTORY16:10
@HeikoSonly happensin centos16:10
@wikingis it available since 2.8.8 ?:)16:10
@wikingbecause if not16:10
@wiking(centos has 2.8.11)16:10
@wikingbut we require16:10
@wiking2.8.816:10
@wikingbackward compatibility16:10
@wikingis the mantra16:10
@HeikoSyeah I get it16:11
@wikingwe cannot just use everything YOLO16:11
@HeikoSbut how should I know before man?16:11
@HeikoSstop pretending that I dont care about this16:11
@wikingyou check16:11
@wikingyou dont care enough16:11
@wikingyou should check it16:11
@HeikoSfor 15 OS?16:11
@wikingno16:11
@wikingfor cmake version16:11
@HeikoSwhat do we have buildbot for?16:11
@wikingyou know that we require16:11
@wiking2.8.816:11
@HeikoSfor this16:11
@wikingit's in the cmake16:11
@HeikoSso I let the bot check and then fix16:11
@wikingno16:11
@wikingthis is bullshit driven development16:11
@wikingsoryr but this is the truth16:12
@HeikoStsss16:12
@wikingthe cmake requested version16:12
@wikingis in the cmakefile16:12
@wikingif you take a function16:12
@wikingthen check prior adding16:12
@wikingi did this16:12
@wikingfor all the Find*bullshit.cmake16:12
@wikingthat's why we have those there16:12
@wikinga lot of them available in 3.x16:12
@HeikoSpoint here is: it seems there16:12
@HeikoSbut has changed API16:12
@HeikoSanyways I will fix it16:13
@wikingthnx16:13
@wikingmaybe16:14
@wikingyou should try using this16:14
@wikinghttp://docs.buildbot.net/0.8.12/manual/cmdline.html?highlight=try#try16:14
@wikingthat requires setting up this: http://docs.buildbot.net/0.8.12/manual/cfg-schedulers.html?highlight=try#try-schedulers16:14
@wikinginfra has the configs for buildbot16:14
-!- arasu_ [daf8068c@gateway/web/freenode/ip.218.248.6.140] has joined #shogun16:16
-!- arasu [daf8068c@gateway/web/freenode/ip.218.248.6.140] has quit [Ping timeout: 252 seconds]16:18
-!- arasu [daf8068c@gateway/web/freenode/ip.218.248.6.140] has joined #shogun16:19
-!- curiousguy13 [~curiousgu@120.59.102.243] has quit [Remote host closed the connection]16:20
shogun-buildbotbuild #598 of clang34 - thread analysis is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/598  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>16:21
-!- arasu_ [daf8068c@gateway/web/freenode/ip.218.248.6.140] has quit [Ping timeout: 252 seconds]16:22
@HeikoSwiking: checking, thanks for that16:23
@HeikoSsanuj: found a fix16:23
shogun-notifier-shogun: Heiko Strathmann :develop * e6808db / doc/cookbook/source/examples/classifier/knn.rst: https://github.com/shogun-toolbox/shogun/commit/e6808db3ce7189f69af852a28bb642adc226919c16:23
shogun-notifier-shogun: only put references that were cited16:23
@HeikoSsanuj: see the new knn.rst16:23
@HeikoSsanuj: btw please insert proper bibtex references, yours was messy. Respect the classes, put in urls in the\url flag, etc16:24
@HeikoSsanuj: only cite properly published things, not max wellings course notes16:24
shogun-buildbotbuild #691 of clang34 - undefined behaviour analysis is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20undefined%20behaviour%20analysis/builds/691  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>16:28
sanujHeikoS: haha, sure16:33
sanujHeikoS: proper citation from now :)16:33
arasuHello everyone. I'm submitting a GSoC proposal for Shogun and we've been asked to introduce ourselves on the IRC.... which to be honest, is quite cryptic to me.16:34
arasuAnyways, I'm Arasu, a 2nd year UG student from IIT Hyderabad, in India.16:34
arasuI'd like to try out for the Hip Deep Learning project.16:34
sanujHeikoS: that bibtex was autogenerated by google scholar16:36
@HeikoSsanuj: yes and its crap ;)16:36
@HeikoSarasu: nice to meet you16:36
@HeikoSlooking forward to your tonributions16:36
sanujHeikoS: :)16:36
@HeikoSsanuj: google rarely produces good bibtex entries16:36
arasuI'm the guy who embarrassed himself by not reading the first paragraph of the "GSoC guidelines" on Shogun's page :P16:36
arasuRegarding my work, I've made a few PRs regarding some stochastic optimizers16:37
@HeikoSarasu: that is fine ... there are loads of you ;)16:37
@HeikoSarasu: I see so that is you16:37
@HeikoSarasu: try to ping lisitsyn about them16:37
@HeikoSor yorkerlin16:37
arasuand I've made some layers (Tanh, HardTanh) which are almost done. And a few constructors and typo fixes which were pushed.16:38
arasuYup, I did. I mailed him to remind him.16:38
arasuAlthough it's probably very hectic for him now anyways.16:39
sanujHeikoS: not sure what to cite for krr16:43
sanujso i'm leaving this for now16:44
sanujand making a pr16:44
@HeikoSsanuj: the david barber book has a cahpter on this fot sure16:44
@HeikoSarasu: please dont write private emails16:45
@HeikoSbut ping in chat16:45
@HeikoSwiking: anything to complain here? https://github.com/shogun-toolbox/shogun/pull/307716:45
lisitsynhey arasu16:45
arasuYes, I realized my mistake. I didn't know how exactly to remind him, that was all.16:45
lisitsynI've seen your mail no worries16:45
arasuCoincidentally, I replied back to that just 20 seconds ago.16:46
shogun-buildbotbuild #762 of deb4 - python3 is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb4%20-%20python3/builds/762  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>16:46
arasuI had a doubt regarding the Hip Deep Learning project that I wanted to discuss with you guys about.16:47
sanujHeikoS: the book is loading, will check it out16:47
arasuWhat is Shogun's future plan regarding Deep Learning?16:48
@HeikoSlisitsyn: http://buildbot.shogun-toolbox.org/builders/deb4%20-%20python3/builds/762/steps/configure/logs/stdio16:48
@HeikoSIll leave this one for you16:48
@HeikoSarasu: get all the basic algorithms in, and them make them run fast16:48
lisitsynyeah ^16:48
arasuYes, one of my proposals is to refactor the code and the unit tests with linalg and OpenMP.16:48
@HeikoSarasu: shogun will never be the state of the art deep learning platform, but it will contain competitive implementations of all standard algos16:48
@HeikoSthat is a good idea16:49
@HeikoSespecially parallelising16:49
@HeikoSbut more importantly we want things to run on graphic cards16:49
@HeikoSmuch more gains than for multicore16:49
sanujHeikoS: are we going to continue using viennacl?16:49
arasuViennaCL takes care of the GPU support, right? It was clarified to me on an issue discussion.16:49
@HeikoSsanuj: I guess so16:50
arasuBy sanuj16:50
@HeikoSlisitsyn: ^16:50
sanujHeikoS: cool16:50
arasuBy extending to other algorithms, I'm in a dilemma of proposing between16:50
arasuRNNs and LSTMs versus DBNs, DBMs16:51
lisitsynI am fancy of having recurrent networks as well16:52
@HeikoSwiking: are the infra script automatically updated on buildbot ?16:52
arasuContemporary work is most on the former. RNNs are very hot topics these days.16:52
sanujHeikoS: leaving now, will send the PR soon16:52
@HeikoSarasu: start with simplest16:52
@HeikoSsanuj: nice looking forward16:52
sanujbye16:52
@HeikoSarasu: we dont need many, we need *well implemented*16:52
-!- sanuj [0e8bc402@gateway/web/freenode/ip.14.139.196.2] has quit [Quit: Page closed]16:52
lisitsynarasu: if you're particularly interested in rnns that's ok16:53
arasui just wanted to know what your focus was.16:54
@wikingHeikoS: no16:55
@HeikoSwiking: about the python3 build, how to install python-ply there? Is this something I can do via the infra?16:55
@wikingHeikoS: it's installed16:55
@wikingas i said previously16:55
@wikingthe detection of it16:55
@wikingis not good16:55
@wikingor not complete16:55
@HeikoSwiking: ah sorry didnt get that16:55
@wikingsame goes for ctags16:55
arasuconsidering hip deep learning was a non-priority project already, i wasn't sure about my freedom of choice in proposal.16:55
@HeikoSlisitsyn:  ^16:56
@HeikoSwiking: how would you try to find python modules in cmake?16:57
@wikingdunno16:57
@wikingtry github16:57
lisitsynarasu: ah no worries, if you have some suggestions we're pretty much open16:57
@wikingto see for examples16:57
lisitsynarasu: I'd focus on you + shogun, not very concrete details of particular projects16:58
@HeikoSyeah thats where I got mine from16:58
@wikingHeikoS: well that's not good enough then :)))16:59
@HeikoSgotta take a break now, will fix later17:00
shogun-buildbotbuild #599 of clang34 - thread analysis is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/599  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>17:03
shogun-buildbotbuild #692 of clang34 - undefined behaviour analysis is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20undefined%20behaviour%20analysis/builds/692  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>17:10
@HeikoSwiking: ah man, I broke them all17:10
@HeikoShttp://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/599/steps/compile/logs/stdio17:10
@HeikoSI dont understand this here17:10
@HeikoSany ideas?17:11
arasulisitsyn: okay, that sounds great. i'll send a proposal draft around rnn's by 2 days. thanks :)17:11
@HeikoSlisitsyn: ^17:11
-!- Bestony [~Bestony@171.43.9.157] has joined #shogun17:12
BestonyHello, i am student of gsoc.17:12
@HeikoShi17:12
BestonyHi, Heikos, i want to know ,if i only have other language experience ,can i submit proposal for student?17:13
Bestonyi do some frontend and php in free time .17:14
Bestonyi know that Shogun need student for documenting.Maybe i am not  clever for Shogun.17:15
@HeikoSBestony: Shogun is c++17:16
@HeikoSand this year in particular, we want to do software eingineering17:17
@HeikoSengineering17:17
@HeikoSso better find something else .... or have a look what would suit you17:17
Bestonyok.Thanks for your reply!17:17
@wiking:)))))))))))))))))17:17
-!- Bestony [~Bestony@171.43.9.157] has left #shogun []17:17
@wikingHeikoS: some of them were broken17:18
@wikingsome not17:18
-!- arasu [daf8068c@gateway/web/freenode/ip.218.248.6.140] has quit [Quit: Page closed]17:19
@wikingHeikoS: mmm how did you end up changing compiler flags?17:25
@HeikoSwiking: if I do this, does it overwrite things in other places?17:56
@HeikoSSET_TARGET_PROPERTIES(${GENERATED_CPP_BINARY} PROPERTIES COMPILE_FLAGS "-Wno-unused-variable")17:56
@HeikoSI do this for the cpp meta examples, as they have unused variables17:56
@wiking:>17:58
@wikingwhere did u do it?17:58
@wikingwhich cmakefile?17:58
@wikingbut yeah i suppose17:58
@wikingas b4 that17:58
@wikingthere wasn't a problem17:58
@HeikoSit is the tester though18:00
@HeikoSADD_EXECUTABLE(meta_example_integration_tester ${CMAKE_CURRENT_SOURCE_DIR}/tester.cpp)18:00
@HeikoSADD_DEPENDENCIES(meta_example_integration_tester shogun)18:00
@HeikoStarget_link_libraries(meta_example_integration_tester shogun ${SANITIZER_LIBRARY})18:00
@HeikoSshould I add this here:18:00
@HeikoSIF(SANITIZER_FLAGS)18:00
@HeikoSset_target_properties(meta_example_integration_tester PROPERTIES COMPILE_FLAGS ${SANITIZER_FLAGS})18:00
@HeikoSENDIF()18:00
@HeikoShave the feeling that is it18:00
@HeikoSit is done to the libshogun examples18:00
@HeikoSand the tester is same kind of thing18:01
@wikingwhich cmake file18:01
@wikingexactly18:01
@wikingare we talking about18:01
@wikingwhere u do this18:01
@HeikoStwo18:01
@HeikoSbut the last one is18:02
@HeikoStests/meta18:02
@HeikoSline 2218:02
@HeikoSwhere I build the c++ program that compares two dumped outputs18:02
@HeikoSthis is where something might be missing18:02
@HeikoSand the other file is: examples/meta/cpp/CMakeLists.txt18:03
@HeikoSthere I surpress the unused variable warnings, but I think that should not override other things18:03
@wikingtwo18:03
@wiking?18:03
@wikingmmm18:04
@HeikoStwo?18:04
@HeikoSI think first file is the reason18:04
@wikingthe sanitizers18:04
@wikingare compiling18:04
@wikingtests18:04
@HeikoSthe integration tester program is very much like a libshogun text18:05
@HeikoSthat is c++ example that performs shogun code18:05
shogun-buildbotbuild #607 of debian wheezy - memcheck is complete: Failure [failed memory check]  Build details are at http://buildbot.shogun-toolbox.org/builders/debian%20wheezy%20-%20memcheck/builds/607  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>18:07
shogun-buildbotbuild #357 of CentOS7 - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/CentOS7%20-%20libshogun/builds/357  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>18:09
-!- arasu [daf8068c@gateway/web/freenode/ip.218.248.6.140] has joined #shogun18:09
shogun-buildbotbuild #2804 of bsd1 - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/2804  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>18:09
-!- bollu [~Adium@1.186.5.127] has joined #shogun18:13
-!- bollu [~Adium@1.186.5.127] has quit [Client Quit]18:17
-!- bollu [~Adium@1.186.5.127] has joined #shogun18:19
-!- arasu [daf8068c@gateway/web/freenode/ip.218.248.6.140] has quit [Quit: Page closed]18:19
shogun-notifier-shogun: Heiko Strathmann :develop * 1be22a6 / tests/meta/CMakeLists.txt: https://github.com/shogun-toolbox/shogun/commit/1be22a6e4d721a8e5e8f60826010e9a87f404e0e18:21
shogun-notifier-shogun: try to fix buildbot via treating integration tester just like a libshogun example18:21
@HeikoSshogun-buildbot: force build --branch=develop 'clang34 - thread analysis'18:21
shogun-buildbotThe build has been queued, I'll give a shout when it starts18:21
shogun-notifier-shogun: Heiko Strathmann :develop * 1a4ea16 / / (9 files): https://github.com/shogun-toolbox/shogun/commit/1a4ea167434941c71ce8ee1a912a9b57777cfd5b18:22
shogun-notifier-shogun: fix cmake build for version <= 2.8.1118:22
shogun-notifier-shogun: Heiko Strathmann :develop * 5fac3dd / / (9 files): https://github.com/shogun-toolbox/shogun/commit/5fac3dd37009f2eb4555b73877bbeb3c9afd3d1e18:22
shogun-notifier-shogun: cleaner version check18:22
shogun-buildbotbuild #600 forced18:22
shogun-buildbotI'll give a shout when the build finishes18:22
shogun-notifier-shogun: Heiko Strathmann :develop * 5049704 / / (9 files): https://github.com/shogun-toolbox/shogun/commit/50497042a6490c3925df75a3f3d365cea3c33fff18:22
shogun-notifier-shogun: Merge pull request #3077 from karlnapf/feature/cmake18:22
shogun-notifier-shogun:18:22
shogun-notifier-shogun: fix cmake build for version <= 2.8.1118:22
@HeikoSshogun-buildbot: force build --branch=develop 'CentOS7 - libshogun'18:22
shogun-buildbotbuild forced [ETA 2m58s]18:22
shogun-buildbotI'll give a shout when the build finishes18:22
@HeikoSwiking: centos is fixed :)18:23
shogun-buildbotbuild #358 of CentOS7 - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/CentOS7%20-%20libshogun/builds/35818:23
@wikinggreat18:23
-!- bollu [~Adium@1.186.5.127] has quit [Ping timeout: 276 seconds]18:27
shogun-buildbotbuild #600 of clang34 - thread analysis is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/60018:28
-!- lambday [6a3384f0@gateway/web/freenode/ip.106.51.132.240] has joined #shogun18:38
-!- mode/#shogun [+o lambday] by ChanServ18:38
@wikingmmm doxygen can generate svg :)18:39
shogun-buildbotbuild #763 of deb4 - python3 is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb4%20-%20python3/builds/763  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>18:39
shogun-buildbotbuild #2805 of bsd1 - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/2805  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>18:42
shogun-buildbotbuild #601 of clang34 - thread analysis is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/601  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>18:45
-!- curiousguy13 [~curiousgu@120.59.102.243] has joined #shogun18:50
shogun-buildbotbuild #693 of clang34 - undefined behaviour analysis is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20undefined%20behaviour%20analysis/builds/693  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>18:52
-!- sanuj [0e8bc402@gateway/web/freenode/ip.14.139.196.2] has joined #shogun19:03
-!- bollu [~Adium@1.186.0.15] has joined #shogun19:11
shogun-notifier-shogun: Viktor Gal :develop * 639088c / doc/doxygen/ (4 files): https://github.com/shogun-toolbox/shogun/commit/639088c197f6012c952e58268991fea7d721bbdb19:12
shogun-notifier-shogun: Update and fix doxygen config files19:12
@wikingshogun-buildbot: force build --branch=develop 'nightly_default'19:12
shogun-buildbotThe build has been queued, I'll give a shout when it starts19:12
-!- bollu [~Adium@1.186.0.15] has quit [Client Quit]19:14
-!- sonne|osx [~sonne@f052250177.adsl.alicedsl.de] has joined #shogun19:22
-!- lambday [6a3384f0@gateway/web/freenode/ip.106.51.132.240] has quit [Ping timeout: 252 seconds]19:26
sanujHeikoS: i don't know why buy 209 tests are failing on my machine19:33
sanujpython-modular and generated19:33
-!- bollu [~Adium@1.186.130.24] has joined #shogun19:37
-!- bollu1 [~Adium@1.186.130.24] has joined #shogun19:38
-!- bollu [~Adium@1.186.130.24] has quit [Ping timeout: 244 seconds]19:42
-!- sanuj [0e8bc402@gateway/web/freenode/ip.14.139.196.2] has quit [Ping timeout: 252 seconds]19:47
-!- sanuj [0e8bc402@gateway/web/freenode/ip.14.139.196.2] has joined #shogun19:52
shogun-buildbotbuild #608 of debian wheezy - memcheck is complete: Failure [failed memory check]  Build details are at http://buildbot.shogun-toolbox.org/builders/debian%20wheezy%20-%20memcheck/builds/608  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>19:59
-!- besser82 [~besser82@fedora/besser82] has joined #shogun20:20
-!- mode/#shogun [+o besser82] by ChanServ20:20
shogun-buildbotbuild #764 of deb4 - python3 is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb4%20-%20python3/builds/764  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>20:30
shogun-buildbotbuild #1096 forced20:32
shogun-buildbotI'll give a shout when the build finishes20:32
shogun-buildbotbuild #2806 of bsd1 - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/2806  blamelist: Viktor Gal <viktor.gal@maeth.com>20:32
-!- lambday [6a3384f0@gateway/web/freenode/ip.106.51.132.240] has joined #shogun20:34
-!- mode/#shogun [+o lambday] by ChanServ20:34
@lambdayHeikoS: hey20:35
shogun-buildbotbuild #602 of clang34 - thread analysis is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/602  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>20:36
shogun-buildbotbuild #694 of clang34 - undefined behaviour analysis is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20undefined%20behaviour%20analysis/builds/694  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>20:43
sanujlambday: hi20:47
sanujlambday: https://github.com/shogun-toolbox/shogun/issues/307220:47
@wikingHeikoS: http://shogun.ml/doc/en/latest/index.html20:48
@wikingdocs fixed20:48
@wikingas a positive sideeffect20:48
@wikinghttp://shogun.ml/doc/en/latest/classshogun_1_1CDotFeatures.html20:48
@wikingsvg is being used to generate the dependency graph20:48
@wikingand you can zoom and pan around :P20:49
lisitsynwiking: do you happen to know whether doxygen could generate better links?20:52
lisitsynclassshogun_1_1CDotFeatures20:53
lisitsynhell20:53
@lambdaysanuj: hey21:02
@lambdaychecking21:02
sanujlambday: cool :)21:03
sanujlambday: also this https://github.com/shogun-toolbox/shogun/issues/307421:04
@wikinglisitsyn: mmm only shorter :)21:04
-!- besser82 [~besser82@fedora/besser82] has quit [Ping timeout: 240 seconds]21:04
@lambdaysanuj: okay.. so let's do the native removal first21:06
sanujlambday: okay, i have worked with linalg before....almost a year back, so i have forgotten most of the stuff21:06
@lambdayI was also thinking of using a better notation for rowwise and colwise things that we have.. I'll add a couple of issues regarding that later :)21:06
@lambdayhaha don't worry.. I myself forgot almost all of it21:07
sanujlambday: cool :)21:07
@lambdaysanuj: get your hands dirty with viennacl also...21:08
sanujlambday: shall i get rid of the native implementation and the backend as well?21:08
@lambdaywe need some gpu stuffs21:08
sanujlambday: building viennacl with shogun is tricky... i was having some problems which i don't remember anymore :P21:09
@lambdaysanuj: yeah that would make sense.. but I'd suggest you send a small patch first.. remove native for one particular operation.. make sure that the tests still work..21:09
@lambdaysanuj: cmake just have to find your opencl and viennacl shared libraries.. shouldn't be too hard :)21:10
@lambdayanyway it would be worth investing a couple of hours for.. cause we need to make this work :)21:10
sanujlambday: sure, i'll start with the most easy one - add.h21:11
@lambdaysounds cool21:11
@lambdaylet's do this iteratively21:11
sanujlambday: agreed21:11
@lambdaywhen all of NATIVEs are gone, we'll kill the backend21:11
sanujsounds cool21:12
@HeikoSwiking: http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/600/steps/compile/logs/stdio21:12
@HeikoSthis is what I got with the sanitizer thing21:12
@HeikoSmmmh21:12
@HeikoSah wait21:13
@HeikoSthats different now21:13
@HeikoSso gotta change it for the cpp as well21:13
@HeikoSsanuj: jojo21:14
@HeikoSlambday: jo21:14
@lambdayHeikoS: hola21:14
@HeikoSlambday: how are thingz?21:14
sanujHeikoS: hey21:14
@lambdayHeikoS: good.. redesigned the hell out of it :/ this is what I have so far https://github.com/lambday/flash/blob/develop/src/flash/statistics/MMD.cpp#L5421:15
@wikingHeikoS: there are 2 sanitizers ;)21:15
@wikingjust for the record21:15
@HeikoSwiking: ok thanks21:15
@HeikoShad no idea about any of this tbh ;)21:15
@HeikoSlambday: cool checking21:16
@lambdayHeikoS: so the idea is - for each burst, we precompute the kernels on each block - then send each precomputed kernel to a separate thread (either cpu or gpu)21:16
@lambdaythey do their job.. when they are done, we get the statistic21:16
@lambdayif we can't allocate enough memory for all the blocks in the burst, we remind the user to use small number of blocks per burst21:17
@lambdaythis is still WIP.. other cases have to be added21:18
@HeikoSlambday: I think this is pretty neat21:19
@HeikoSespecially since it automatically send things to other thread21:19
@lambdayyeah the computation manager takes care of that21:19
@lambdaycm.use_gpu().compute(....)21:19
@lambdayor cpu21:19
@lambdayplus this approach separate the equation itself from the computation.. the equation is same for all the mmds21:20
@lambdayso they are kept separate21:20
@lambdaythey are just functors21:20
@lambdayq-time, b-test, linear time - all will use this21:20
@lambdayfor q-time, the block is the whole data set21:20
@lambdayfor b-test, we ask the user about the blocksize and for linear, we set it to 2/321:21
@lambdayor so21:21
shogun-notifier-shogun: Heiko Strathmann :develop * ee728ef / examples/meta/cpp/CMakeLists.txt: https://github.com/shogun-toolbox/shogun/commit/ee728efda76cac31f53121c7189bc01b7069897021:21
shogun-notifier-shogun: sanitize meta cpp examples if needed21:21
@HeikoSlambday:  and then each of these seperate things can also do things in parallel if we cant21:21
@lambdayeach thread has its own copy - so they can run in parallel21:21
@HeikoSand yes computation separated from equation is good21:21
@HeikoSsince yes, computation is always the same structure21:22
@HeikoSshogun-buildbot: force build --branch=develop 'clang34 - thread analysis'21:22
shogun-buildbotbuild #603 forced21:22
shogun-buildbotI'll give a shout when the build finishes21:22
@lambdayfor hsic too, similar things can be done21:22
@HeikoSlambday: I think for the one sample test, the story will be the same21:23
sanujlambday: removed native implementation and related unittests in add.h21:23
sanujlambday: everything passing21:23
@lambdayyeah the code would look pretty similar for all the cases - easier to read21:23
@lambdaysanuj: awesome! send the patch21:23
sanujlambday: BTW it was me who wrote these native implementations :P21:23
sanujlast year21:24
sanujyeah21:24
@lambdaysanuj: haha that's ironic :D21:24
@lambdayHeikoS: I totally screwed up the code last time I can see.. I myself can't understand how things are working/where the pieces are..21:24
@lambdayit's good that I am spending time on it to fix it21:25
@lambday:(21:25
@HeikoShehe21:25
@HeikoSok21:25
@HeikoSyes definitely21:25
@HeikoSlambday: what is really important is also to actually test things from a statistical qualitative perspective in notebooks later21:25
@HeikoScheck that results make sense21:25
@lambdayHeikoS: yeah that would be the exciting part.. all of this is just plain coding21:26
@lambdayHeikoS: trying to be done with this by the end of this week..21:26
@lambdayHeikoS: BTW shall I spend some time to read nonparametric statistics?21:27
@lambdaybasics21:27
-!- bollu1 [~Adium@1.186.130.24] has quit [Quit: Leaving.]21:27
@HeikoSlambday: no dont worry fornow21:27
-!- besser82 [~besser82@fedora/besser82] has joined #shogun21:28
-!- mode/#shogun [+o besser82] by ChanServ21:28
@HeikoSlambday: for now we want to get everything working so that we can do expeirment21:28
sanujlambday: https://github.com/shogun-toolbox/shogun/pull/307921:28
shogun-buildbotbuild #603 of clang34 - thread analysis is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/60321:28
@HeikoSlambday: keep in mind we dont want to stay too long in this initial phase21:28
@HeikoSlambday: otherwise its hard to finish/polish things21:28
@lambdayHeikoS: yeah.. maybe we can discuss a possible timeline for this thing..21:30
@HeikoSlambday: would be good I guess21:30
@HeikoSlambday: how much longer do you think this will take21:30
@lambdayonce I finish this, next step would be to write examples for cookbook and make the notebook work again..21:31
@HeikoSuntil we can start writing API examples and tick boxes of which methods/papers we can do21:31
@HeikoSlambday: but it needs to be integrated with a feature branch as well -- doesnt that take time?21:31
@HeikoSlambday: document interfaces etc?21:31
@HeikoSon the other hand we can reuse old docs21:31
@lambdayHeikoS: yeah that too is gonna take time.. but if the code is there, I can reuse old docs and do it parallely..21:32
-!- sonne|osx [~sonne@f052250177.adsl.alicedsl.de] has quit [Quit: sonne|osx]21:32
@wikinghohohoho21:32
@wikingi just found this21:32
@lambdayI guess by the end of this week I can finish as much of it as it's there21:32
@wikingor actually was sent to me21:33
@wikinghttps://www.python.org/dev/peps/pep-0020/21:33
@wiking:DDDDDDDDDDDDDDDDDDDDDD21:33
@wiking"Explicit is better than implicit"21:33
@wikingnow i know why python programmers never use variables or configuration files21:33
@wiking:))))21:33
@HeikoSwiking: haha you never saw this before? :D21:33
-!- sanuj [0e8bc402@gateway/web/freenode/ip.14.139.196.2] has quit [Quit: Page closed]21:33
@wikingnever21:33
@wikingthis is the first time21:33
@wikingthis is the biggest bullshit21:33
@wikingi ever seen in my life21:33
@HeikoSlambday: yeah I can also help you writing the cookbook examples21:33
@wiking:D21:33
@HeikoSlambday: so maybe lets do this: once the standard MMD stuff  works, you put it in afeature brach and start iterating there21:34
@HeikoSin the meantime, I get the API examples going21:34
-!- sanuj [0e8bc409@gateway/web/freenode/ip.14.139.196.9] has joined #shogun21:34
@HeikoSwiking: haha21:34
@HeikoSah man, I love it when you make fun of them ;)21:34
@HeikoSFlat is better than nested.21:35
shogun-notifier-shogun: Heiko Strathmann :develop * 21b3cd8 / examples/meta/cpp/CMakeLists.txt: https://github.com/shogun-toolbox/shogun/commit/21b3cd8fa6d53fa9a31fa70ceea1b0a241a4b39921:36
shogun-notifier-shogun: give meta examples *exactly* the same treatment as libshogun examples21:36
@HeikoSshogun-buildbot: force build --branch=develop 'clang34 - thread analysis'21:36
shogun-buildbotThe build has been queued, I'll give a shout when it starts21:36
@wikingHeikoS: :DDDD21:36
shogun-buildbotbuild #604 forced21:37
shogun-buildbotI'll give a shout when the build finishes21:37
sanujlambday: shall i send more PRs like that or we wait for travis first and you merge?21:39
shogun-buildbotbuild #1096 of nightly_default is complete: Failure [failed notebooks]  Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/109621:39
@lambdayHeikoS: okay I'll do that.. by the end of this week I'll put this thing in a feature branch.. so next week I can give to write the examples.. and some actual testing.. alongwith debugging21:41
@lambdaysanuj: I'm waiting for travis.. you can send more PRs :D21:41
@HeikoSlambday: yes, maybe even start putting things now. As soon as the quadratic time MMD works, you can put it in :)21:41
@HeikoSits doesnt have to be perfect in feature branch21:41
@lambdayHeikoS: okay..21:41
sanujlambday: one implementation per PR?21:42
@HeikoSjust compile and then I can already start trying things out.... in case something needs change, we save ourselves a bit of effort21:42
@lambdayoh one question - for quadratic time mmd, we need api for the variance under null and alternative both right?21:42
@lambdayor just the null is fine?21:42
@HeikoSlambday: both would be good21:42
@HeikoSbut alternative is just variance of the mmd samples itself right?21:42
@HeikoSlambday: the core interface doesnt need alt, since we dont need it for the test21:43
@HeikoSbut it would be good to have access to this for experimenting21:43
@HeikoSlambday: I like your gpu switch :)21:43
@lambdayHeikoS: haha need some time to actually implement the opencl kernel :D21:43
@HeikoSlambday: how much faster was summing all elements of kernel matrix on gpu?21:43
@HeikoSlambday: linalg should do it no?21:43
@lambdayHeikoS: yeah but linalg is again a mess.. it doesn't have block-wise matrix sum using viennacl21:44
@HeikoSwiking: whoooo http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/60421:44
@lambdayshould be faster21:44
@HeikoSlambday: then that shold be added I guess21:44
@lambdaythat's what I need to check21:44
@HeikoSlambday: for quadratic time mmd, this is not needed21:44
@lambdaynah21:44
@HeikoSand there the speedups are most crucial21:44
@HeikoSlambday: I think we might have to re-write one of the paper statements, which is "permutation test is slow"21:45
@lambdaybut for b-test and linear-time this can make significant difference21:45
@HeikoSthis was only because of their crappy implementation21:45
@HeikoSlambday: yeah ok then21:45
@lambdayHeikoS: another question21:45
@lambdayHeikoS: so what we do for single kernel is --> compute all kernel matrices for a bunch of blocks --> send them to the compute engine to execute parallely21:46
@lambdayfor multiple kernels, this way won't work21:46
@lambdayso I was thinking to invoke the parallelism per kernel21:46
@lambdaybut that depends - how many kernels do we use for kernel selection?21:46
@lambdayI am just trying to get an idea - where to put the parallelism21:47
shogun-notifier-shogun: Heiko Strathmann :develop * a6036fd / tests/meta/CMakeLists.txt: https://github.com/shogun-toolbox/shogun/commit/a6036fd590d1631322993e9873d2a681d5232ad321:47
shogun-notifier-shogun: give meta integration tester same treatment as the libshogun examples21:47
@lambdaysanuj: you can send a PR for a bunch of them.. just make sure it's not too large21:47
@HeikoSI see21:47
sanujlambday: sure21:47
@HeikoSlambday: so the costs are m^2 for m kernels21:48
@HeikoSthat is the MMD has to be computed (and stored) that many times21:48
@HeikoSbut what exactly do you mean?21:48
@HeikoSshogun-buildbot: force build --branch=develop 'clang34 - thread analysis'21:48
shogun-buildbotThe build has been queued, I'll give a shout when it starts21:48
shogun-buildbotHey! build clang34 - thread analysis #604 is complete: Warnings [warnings test]21:48
shogun-buildbotBuild details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/60421:48
shogun-buildbotbuild #605 forced21:50
shogun-buildbotI'll give a shout when the build finishes21:50
@lambdayI mean, say, if we have 10 kernels and 1000 blocks per burst - all of these kernels AND all of these blocks can be computed parallely.. but to make things simple, I want to use parallel computation (a) either per *kernel* or (b) per block..21:50
@HeikoSlambday: I see what you mean21:50
@lambdayfor single kernel, it's always *per* block21:50
@lambday1000 threads running in parallel21:50
@HeikoSusually I think there are more blocks than kernels right?21:50
@lambdaythat's what I think - so it makes sense to invoke the parallelism block-wise, rather than kernel-wise21:51
@HeikoSlambday: yes absolutely21:51
@HeikoSas long as the cpu is busy that is fine21:51
@lambdayfor multiple kernel, we internally reduce the number of blocks we are processing per burst..21:51
@HeikoSyes21:51
@lambdayokay21:52
@wikinglambday: how do distribute it?21:53
@lambdaywiking: yeah that's the question21:53
@wikingmmm21:54
@wikingbut within machine21:54
@wikingor among machines?21:54
@lambdaywiking: as of now, say, gpu21:54
@lambdaywiking: but maybe hadoop job later21:54
@lambday:D21:54
@wikingok21:54
@wikingeasy21:54
@wikinghttp://actor-framework.org/21:55
@lambdaywiking: this is cool stuff21:55
@wikingyep21:56
@wikingatm i'm writing my simple example21:56
@wikingcaf + libhdfs21:56
@wikingand then adding shogun later21:56
@lambdaysexy21:57
@lambdayokay I am definitely trying this - once the basic things work..21:57
@wikingi'd like this to be the distributed backend of shogun21:57
@lambdaywiking: how is it different from spark?21:58
@wikingwell spark is first of all jvm21:58
@wikingand spark is an abstraction over akka21:58
@wikingwhich is a jvm actor framework21:59
@wikingso yeah21:59
@wikingthis is c++ actor framework21:59
@lambdayhave to check it out22:00
shogun-buildbotHey! build clang34 - thread analysis #605 is complete: Warnings [warnings test]22:01
shogun-buildbotBuild details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/60522:01
@HeikoSlambday: no hadoop or multi machine please :)22:05
@HeikoSthats a second step22:05
@HeikoSfirst is to make it actually work at all, then multicore, then maybe (if the science requests it) we can distribute it22:05
@HeikoSfor now, there is not need for that22:06
@HeikoSlambday: woul dbe more interesting for other things22:06
lisitsynI heard hadoop22:06
@HeikoSthe ozone one was nice to distribute22:06
@HeikoSlambday: no22:06
@HeikoSlisitsyn: no22:06
@HeikoSlisitsyn: your ears tricked you22:06
@lambdayHeikoS: yeah none now :)22:06
@lambdayHeikoS: just the single machine gpu would be nice22:06
@HeikoSlambday: yeaj22:07
@HeikoSand multicore22:07
@HeikoSI canm always get a machine with 100 cores and a few GB of ram if we need power22:07
@HeikoSa good multicore implementation is quite useful for such things already22:07
@HeikoSshogun-buildbot: force build --branch=develop 'clang34 - undefined behaviour analysis'22:08
shogun-buildbotbuild #695 forced22:08
shogun-buildbotI'll give a shout when the build finishes22:08
@lambdayHeikoS: yeah.. maybe I'll have to tune it a bit later.. but not thinking of such things as of now..22:09
@HeikoSlambday: Im keen to write API examples22:09
@HeikoSlambday: so let me know when I can start with one22:09
@lambdayHeikoS: okay by the end of this week this would be API example ready22:10
@lambdayat least the mmds22:10
@lambdayI'll feature branch it22:10
@HeikoSlambday: great22:15
@HeikoSlet me know!22:15
shogun-buildbotHey! build clang34 - undefined behaviour analysis #695 is complete: Warnings [warnings test]22:16
shogun-buildbotBuild details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20undefined%20behaviour%20analysis/builds/69522:16
@HeikoSlambday: going home now22:17
@HeikoSlambday: see you later!22:17
sanujlambday: getting error in this22:17
@lambdayHeikoS: okay see you :)22:17
@HeikoSwiking: the clang analysis now are compiling again22:17
sanujlambday: https://github.com/shogun-toolbox/shogun/blob/d8eb73dd89f47e0da28f07163c4f635b96d0ec00/src/shogun/kernel/ExponentialARDKernel.cpp#L9522:17
@HeikoSwiking: all thats missing is lisitsyn for the ctags fix and me the python3 ply fix22:18
@lambdayHeikoS: for element-wise, we need NATIVE :(22:18
@lambdaysanuj:22:18
sanujlambday: after removing native from Elementwiseunaryoperation.h22:18
@lambdaysanuj: this lambda expression won't be supported by eigen3 or anything else.. only native can support thta22:19
sanujlambday: HeikoS ohoh22:19
@lambdaysanuj: eigen3 has only a few NATIVE element-wise things22:19
sanujlambday: i see22:20
@lambdaysanuj: you can do one thing - remove the specialization.. make it the default22:20
@lambdayoh but then too I am afraid it won't work22:20
sanujlambday: then how will we get rid of native backend22:21
@lambdaylet me think22:21
sanujHeikoS: there?22:22
-!- HeikoS [~heiko@untrust-out.swc.ucl.ac.uk] has quit [Ping timeout: 276 seconds]22:23
sanujlambday: i'll leave elementwise op for now22:23
sanujand remove other implementations22:24
@lambdaysanuj: for these elementwise things, eigen3 is slow from our perspective... suppose you do an element-wise sin, it returns the result in a newly created matrix..22:24
@lambdaythere is a lot of copying involed..22:24
@lambdayyeah that might be better idea for now22:24
sanujokay22:24
sanujlambday: i don't think there is any harm in keeping native implementations wherever we need them22:25
-!- besser82 [~besser82@fedora/besser82] has quit [Ping timeout: 250 seconds]22:38
sanujlambday: https://github.com/shogun-toolbox/shogun/pull/307922:38
sanujall checks are passing22:38
@lambdaysanuj: cool.. thanks22:39
shogun-notifier-shogun: Sanuj :develop * 3551260 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/3551260aa2108c01f884873f6ee807bd171243fe22:39
shogun-notifier-shogun: remove native add in linalg22:39
shogun-notifier-shogun: Soumyajit De :develop * 018971a / / (3 files): https://github.com/shogun-toolbox/shogun/commit/018971a489e293a2dbe3af5667ea79206d9b7ca922:39
shogun-notifier-shogun: Merge pull request #3079 from sanuj/linalg_cleanup22:39
shogun-notifier-shogun:22:39
shogun-notifier-shogun: remove native add in linalg22:39
@lambdaysanuj: looking forward to your next patch22:40
sanujlambday: https://github.com/shogun-toolbox/shogun/pull/308022:40
-!- besser82 [~besser82@fedora/besser82] has joined #shogun22:51
-!- mode/#shogun [+o besser82] by ChanServ22:51
sanujlambday: this is the last one22:55
sanujlambday: https://github.com/shogun-toolbox/shogun/pull/308122:55
sanujlambday: now there is native implementatino left in ElementwiseOperations.h, ElementwiseUnaryOperation.h and sin.h22:57
sanujwhich we are not going to remove i guess22:57
shogun-buildbotbuild #609 of debian wheezy - memcheck is complete: Failure [failed memory check]  Build details are at http://buildbot.shogun-toolbox.org/builders/debian%20wheezy%20-%20memcheck/builds/609  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>22:58
-!- besser82 [~besser82@fedora/besser82] has quit [Ping timeout: 250 seconds]23:16
sanujlambday: going to sleep23:26
sanujgoodnight23:26
-!- sanuj [0e8bc409@gateway/web/freenode/ip.14.139.196.9] has quit [Quit: Page closed]23:26
-!- besser82 [~besser82@fedora/besser82] has joined #shogun23:28
-!- mode/#shogun [+o besser82] by ChanServ23:28
shogun-buildbotbuild #765 of deb4 - python3 is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb4%20-%20python3/builds/765  blamelist: Viktor Gal <viktor.gal@maeth.com>23:29
shogun-buildbotbuild #2807 of bsd1 - libshogun is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/2807  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Sanuj <sanuj.sharma.in@gmail.com>, Soumyajit De <heavensdevil6909@gmail.com>23:32
shogun-buildbotbuild #606 of clang34 - thread analysis is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/606  blamelist: Viktor Gal <viktor.gal@maeth.com>23:35
shogun-buildbotbuild #696 of clang34 - undefined behaviour analysis is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20undefined%20behaviour%20analysis/builds/696  blamelist: Viktor Gal <viktor.gal@maeth.com>23:42
-!- besser82 [~besser82@fedora/besser82] has quit [Ping timeout: 250 seconds]23:59
--- Log closed Wed Mar 16 00:00:32 2016