IRC logs of #shogun for Friday, 2013-04-05

« Thursday, 2013-04-04 Index Saturday, 2013-04-06 »
--- Log opened Fri Apr 05 00:00:21 2013
-!- ZeThomas [~thomas@109.133.206.114] has joined #shogun01:27
-!- shogun-buildbot_ [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]03:45
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun03:45
@sonney2kshogun-buildbot, force build libshogun03:46
shogun-buildbotno such builder 'libshogun'03:46
@sonney2kshogun-buildbot, force build "deb1 - libshogun"03:47
shogun-buildbotbuild forced [ETA 7m40s]03:47
shogun-buildbotI'll give a shout when the build finishes03:47
@sonney2kshogun-buildbot, force build "bsd1 - libshogun"03:48
shogun-buildbotbuild forced [ETA 3m09s]03:48
shogun-buildbotI'll give a shout when the build finishes03:48
@sonney2kshogun-buildbot, force build "cyg1 - libshogun"03:48
shogun-buildbotbuild forced [ETA 14m51s]03:48
shogun-buildbotI'll give a shout when the build finishes03:48
@sonney2kshogun-buildbot, force build "deb2 - static_interfaces"03:48
shogun-buildbotThe build has been queued, I'll give a shout when it starts03:48
@sonney2kshogun-buildbot, force build "deb3 - modular_interfaces"03:48
shogun-buildbotThe build has been queued, I'll give a shout when it starts03:48
@sonney2kshogun-buildbot, force build "rpm1 - libshogun"03:49
shogun-buildbotbuild forced [ETA 8m10s]03:49
shogun-buildbotI'll give a shout when the build finishes03:49
@sonney2kshogun-buildbot, force build "ubu1 - libshogun"03:49
shogun-buildbotbuild forced [ETA 32m06s]03:49
shogun-buildbotI'll give a shout when the build finishes03:49
@sonney2kshogun-buildbot, force build "nighlty_all"03:49
shogun-buildbotno such builder 'nighlty_all'03:49
@sonney2kshogun-buildbot, force build "nightlty_all"03:49
shogun-buildbotno such builder 'nightlty_all'03:49
@sonney2kshogun-buildbot, force build "nightly_all"03:49
shogun-buildbotThe build has been queued, I'll give a shout when it starts03:49
@sonney2kshogun-buildbot, force build "nightly_none"03:50
shogun-buildbotThe build has been queued, I'll give a shout when it starts03:50
shogun-buildbotHey! build bsd1 - libshogun #871 is complete: Success [build successful]03:51
shogun-buildbotBuild details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/87103:51
shogun-buildbotHey! build deb1 - libshogun #1061 is complete: Success [build successful]03:56
shogun-buildbotBuild details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/106103:56
shogun-buildbotbuild forced [ETA 10m45s]03:56
shogun-buildbotI'll give a shout when the build finishes03:56
shogun-buildbotHey! build rpm1 - libshogun #421 is complete: Success [build successful]03:56
shogun-buildbotBuild details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/42103:56
shogun-buildbotHey! build cyg1 - libshogun #707 is complete: Success [build successful]04:03
shogun-buildbotBuild details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/70704:03
shogun-buildbotHey! build deb2 - static_interfaces #876 is complete: Success [build successful]04:07
shogun-buildbotBuild details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/87604:07
shogun-buildbotbuild forced [ETA 40m11s]04:07
shogun-buildbotI'll give a shout when the build finishes04:07
shogun-buildbotHey! build ubu1 - libshogun #402 is complete: Success [build successful]04:21
shogun-buildbotBuild details are at http://www.shogun-toolbox.org/buildbot/builders/ubu1%20-%20libshogun/builds/40204:21
shogun-buildbotbuild forced [ETA 40m00s]04:30
shogun-buildbotI'll give a shout when the build finishes04:30
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]04:32
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun04:32
@sonney2kshogun-buildbot, force build "deb1 - libshogun"04:32
shogun-buildbotThe build has been queued, I'll give a shout when it starts04:32
@sonney2kshogun-buildbot, force build "bsd1 - libshogun"04:33
shogun-buildbotbuild forced [ETA 3m18s]04:33
shogun-buildbotI'll give a shout when the build finishes04:33
@sonney2kshogun-buildbot, force build "nightly_none"04:33
@sonney2kshogun-buildbot, force build "nightly_all"04:33
shogun-buildbotThe build has been queued, I'll give a shout when it starts04:34
@sonney2kshogun-buildbot, force build "ubu1 - libshogun"04:34
shogun-buildbotThe build has been queued, I'll give a shout when it starts04:34
shogun-buildbotbuild forced [ETA 32m02s]04:34
shogun-buildbotI'll give a shout when the build finishes04:34
@sonney2kshogun-buildbot, force build "rpm1 - libshogun"04:34
shogun-buildbotbuild forced [ETA 7m46s]04:34
shogun-buildbotI'll give a shout when the build finishes04:34
@sonney2kshogun-buildbot, force build "deb3 - modular_interfaces"04:34
shogun-buildbotThe build has been queued, I'll give a shout when it starts04:34
@sonney2kshogun-buildbot, force build "deb2 - static_interfaces"04:34
shogun-buildbotThe build has been queued, I'll give a shout when it starts04:34
@sonney2kshogun-buildbot, force build "cyg1 - libshogun"04:34
shogun-buildbotbuild forced [ETA 15m01s]04:34
shogun-buildbotI'll give a shout when the build finishes04:34
@sonney2kshogun-buildbot, force build "nightly_default"04:34
shogun-buildbotThe build has been queued, I'll give a shout when it starts04:35
shogun-buildbotbuild #708 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/70804:35
shogun-buildbotHey! build bsd1 - libshogun #872 is complete: Success [build successful]04:36
shogun-buildbotBuild details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/87204:36
shogun-buildbotHey! build rpm1 - libshogun #422 is complete: Success [build successful]04:41
shogun-buildbotBuild details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/42204:41
shogun-buildbotbuild forced [ETA 14m17s]04:52
shogun-buildbotI'll give a shout when the build finishes04:52
shogun-buildbotbuild forced [ETA 8m17s]05:00
shogun-buildbotI'll give a shout when the build finishes05:00
shogun-buildbotHey! build deb1 - libshogun #1062 is complete: Success [build successful]05:02
shogun-buildbotBuild details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/106205:02
shogun-buildbotbuild forced [ETA 34m59s]05:02
shogun-buildbotI'll give a shout when the build finishes05:02
shogun-buildbotHey! build ubu1 - libshogun #403 is complete: Success [build successful]05:06
shogun-buildbotBuild details are at http://www.shogun-toolbox.org/buildbot/builders/ubu1%20-%20libshogun/builds/40305:06
shogun-buildbotHey! build nightly_none #301 is complete: Success [build successful]05:07
shogun-buildbotBuild details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_none/builds/30105:07
shogun-buildbotbuild forced [ETA 42m32s]05:07
shogun-buildbotI'll give a shout when the build finishes05:07
shogun-buildbotHey! build nightly_all #291 is complete: Success [build successful]05:31
shogun-buildbotBuild details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_all/builds/29105:31
shogun-buildbotbuild forced [ETA 11m18s]05:31
shogun-buildbotI'll give a shout when the build finishes05:31
shogun-buildbotHey! build deb2 - static_interfaces #877 is complete: Success [build successful]05:37
shogun-buildbotBuild details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/87705:37
shogun-buildbotbuild forced [ETA 46m34s]05:37
shogun-buildbotI'll give a shout when the build finishes05:37
shogun-buildbotHey! build deb3 - modular_interfaces #1005 is complete: Success [build successful]05:53
shogun-buildbotBuild details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/100505:53
shogun-buildbotHey! build nightly_default #345 is complete: Success [build successful]06:13
shogun-buildbotBuild details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/34506:13
-!- ZeThomas [~thomas@109.133.206.114] has quit [Quit: Leaving]07:34
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun08:16
@sonney2kshogun-buildbot, force build "cyg1 - libshogun"08:18
shogun-buildbotThe build has been queued, I'll give a shout when it starts08:19
shogun-buildbotbuild forced [ETA 15m01s]08:19
shogun-buildbotI'll give a shout when the build finishes08:19
shogun-buildbotbuild #709 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/70908:19
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving]08:43
-!- n4nd0 [~nando@n186-p150.kthopen.kth.se] has joined #shogun09:18
n4nd0sonne|work: hi!09:43
n4nd0I checked how this history of changes for the webpage could be added09:57
n4nd0It looks like this could be a good star: https://bitbucket.org/dwatkinsweb/django-simple-history09:58
n4nd0I reached it from this post: http://stackoverflow.com/questions/987669/tying-in-to-django-admins-model-history09:58
-!- ZeThomas [~thomaspr@2a02:2c40:100:a000:fce1:5064:9d1c:71c2] has joined #shogun10:01
sonne|workn4nd0: hey!10:08
sonne|workn4nd0: looks quite promising10:08
sonne|workn4nd0: is this then already integrated in the admin view somehow?10:08
n4nd0sonne|work: I don't think so10:09
sonne|workI mean what I don't understand is how one can put an old revision live10:09
n4nd0the admin view is customizable10:09
n4nd0so I think the idea would be to use this django-simple-history to automatically create revisions of the articles every time they are updated10:10
n4nd0or created10:10
n4nd0and then modify the admin view as we like to show these old versions10:10
sonne|workn4nd0: ok - the only thing that would not work with this is that we cannot select the version to be shown on the website10:11
sonne|workit would be nice if we just edit10:11
sonne|workand then at some point tag a publish flag10:12
n4nd0sonne|work: aham. Yes, that is an even better idea10:12
n4nd0That could be done adding a new model probably10:13
n4nd0I don't know if it would be the best solution though10:13
sonne|workno idea what is best but certainly adding just an integer for the revision and a boolean flag 'published' would do10:19
sonne|workohh no11:01
sonne|workit is snowing again!11:01
blackburnsonne|work: ha-ha11:03
sonne|workblackburn: btw congrats: http://investors.tomtom.com/releasedetail.cfm?ReleaseID=75422711:06
blackburnsonne|work: oh everybody hates moscow :D11:09
blackburnespecially its citizens11:09
blackburnthey basically live in jams11:09
blackburnsonne|work: one more thing is that it is crazy expensive to live there and people live outside of the city - then they have to say move 2 hours by train and stay 2 hours in jams11:12
blackburnI have absolutely no idea how do they live :D11:12
blackburnbut I know quite a few that live that way11:12
sonne|workdoesn't sound nice..11:17
blackburnsonne|work: but it is believed that there is no life outside moscow so 10% of russia lives there :D11:18
sonne|workhehe11:19
blackburnsonne|work: 20M of 140M totally - crazy isn't it?11:20
sonne|workits big11:23
wikingblackburn: does that mean that we can directly commit to shogun-toolbox/shogun.git?11:23
blackburnwiking: yes I think so11:23
blackburnnot to master though11:23
wikingyeah sure thing11:24
blackburnsonne|work: do you mind commits to develop without branches?11:24
blackburnI don't know if we really should use that hotfix stuff11:24
wikingn4nd0: i saw the emails.. sorry but i'm away atm but i'll answer asap11:24
n4nd0wiking: sure, no problem11:25
n4nd0wiking: I think that he can find what he asked for in the doc of latent svm though11:25
sonne|workblackburn: yes11:25
blackburnsonne|work: yes - use hotfix?11:26
sonne|workthat too yes11:26
-!- heiko [~heiko@nat-174-82.internal.eduroam.ucl.ac.uk] has joined #shogun11:27
blackburnheiko: is the new releaser!11:28
blackburn:D11:28
heikoyeee! :)11:29
* sonne|work relaxes :D11:35
-!- sonne|work is now known as sonne|nowork11:35
heikosonne|nowork, blackburn as I said, I probably need some help. I would be up for creating a howto on this and update on the way11:36
heikolets discuss later, got a huge pile of work lying here :)11:36
blackburnn4nd0: oh we've got vi{c,k}tors11:41
blackburnZeThomas: hey11:43
ZeThomasblackburn: hi11:43
blackburnlet me get back to your thing11:43
blackburnis it still failing?11:43
ZeThomasyes11:43
n4nd0blackburn: hehe yeah :D11:46
blackburnZeThomas: error 139?11:47
ZeThomasindeed11:47
ZeThomashappens in the init step apparently11:49
ZeThomasline 83 in http://pastebin.com/567TrTnx11:49
ZeThomaswithin the init in line 38: sg.DirectorKernel.init(self, l_features, r_features)11:51
ZeThomasI can't dig deeper than that :)11:55
blackburnZeThomas: so 3811:55
ZeThomasyep11:55
blackburnZeThomas: can you please run gdb on it?12:00
ZeThomasok, you might have to talk me through this, how do I gdb a python prog?12:04
blackburnZeThomas: gdb --args python yourscript.py12:04
blackburnthen start12:04
blackburnthen c12:04
ZeThomasok, I get this: http://pastebin.com/4Bc4dfij12:07
blackburnZeThomas: 'back'12:10
ZeThomashttp://pastebin.com/0Dw1msdr12:11
blackburnZeThomas: quite interesting heh12:15
ZeThomasblackburn: I'm not getting very much out of this... :)12:16
blackburnI expect it fails on to->name12:16
ZeThomasyou mean in the typecheck (is that for the (dummy)features?)12:20
blackburnZeThomas: yeah that's crazy12:20
ZeThomascould it be because I have these variables l and r as attributes?12:20
blackburnno shouldnot be12:21
ZeThomas(renamed them all, to no avail)12:22
blackburnohhh12:23
blackburnwhy do you call init(self, l_features, r_features)12:23
blackburnit should be (l_features, r_features) just12:23
blackburnsonne|nowork: another trick of python12:24
blackburnsonne|nowork: method with 2 parameters is called with 3 parameters and no error (??)12:24
blackburn*args, **kwargs probably12:25
blackburnZeThomas: ^12:25
ZeThomasblackburn: because DirectorKernel.init is a method with 3 params12:27
blackburnZeThomas: no thats __init__12:28
blackburnvirtual bool init(CFeatures* l, CFeatures* r)12:28
ZeThomasevery instance of it automatically fills in itself as first argument12:28
ZeThomasno, every class method is like that12:28
blackburnyes I know what you mean but I am not sure it is the case12:28
ZeThomasblackburn: ok yeah, every python class method12:28
blackburnhmm let me think12:29
ZeThomasI don't know how CPython then handles this12:29
blackburnZeThomas: cpython?12:29
ZeThomasblackburn: scrap that last remark12:30
blackburnZeThomas: yes you are right self is needed here12:31
blackburnZeThomas: could you try to name it custom_init or whatever?12:32
blackburnI am worried about name clash12:32
ZeThomasblackburn: I tried this: http://pastebin.com/DsP4ka0z and this works fine12:33
blackburntrue12:33
ZeThomasbut wait12:34
blackburndo you see any difference?12:34
ZeThomasI get this warning with this toy example, which I don't get in the other case12:35
ZeThomasRuntimeWarning: [WARN] Enforcing to use only one thread due to restrictions of directors12:35
blackburnZeThomas: that's ok12:35
ZeThomaswhy is that?12:35
blackburnZeThomas: director kernel can be called only from one thread12:36
ZeThomasyeah, I know, but why is it that in the crashing case I don't get this warning?12:36
blackburnyour code is not getting to that line12:36
blackburnit fails before12:36
ZeThomascould it be because I called init already once, in my __init__ ?12:39
blackburnbut it worked back when?12:40
blackburnthen*12:41
ZeThomasok there's some stupid stuff in there, I don't know why (I think my lacking vim-fu)12:43
ZeThomaslike the real __init__ in another one in the abstract class?12:43
ZeThomasI'm fixing that first12:43
blackburnyou mean that __init__ inside __init__?12:44
ZeThomasyeah, don't really know why that's there...12:45
ZeThomasok, I seems to work now, this was weird12:49
ZeThomastraining that svm seems to take ages though12:51
blackburnthat's strange12:52
-!- n4nd0 [~nando@n186-p150.kthopen.kth.se] has quit [Ping timeout: 246 seconds]12:53
ZeThomasstill on it12:55
blackburnZeThomas: is data big?12:58
blackburnthe one you work on12:58
ZeThomas3500 of these objects I defined my kernel on13:01
ZeThomasok, now it's done training :)13:03
-!- sumit [73f91219@gateway/web/freenode/ip.115.249.18.25] has joined #shogun13:10
-!- n4nd0 [~nando@n186-p150.kthopen.kth.se] has joined #shogun13:14
ZeThomasdammit13:22
ZeThomasafter about two iterations, I get that damned 139 again13:23
ZeThomasI'll gdb it13:23
blackburnehmm13:25
ZeThomasblackburn: ok, here we go: http://pastebin.com/z8fp64VC13:54
blackburnZeThomas: no info at all :)13:54
ZeThomasand http://pastebin.com/GZZa4naM13:55
blackburnZeThomas: what happened after the first iteration?13:56
ZeThomasthis is the code: http://pastebin.com/F9n3GLTy ; iterations are at line 12313:58
ZeThomasso it must segfault before the second iteration starts, seeing the output13:59
ZeThomashowever, that's not a constant, since the previous time it segfaulted after 2 iterations14:00
-!- lambday [3ba2174f@gateway/web/freenode/ip.59.162.23.79] has joined #shogun14:02
lambdayheiko: hi..14:04
-!- sumit [73f91219@gateway/web/freenode/ip.115.249.18.25] has quit [Ping timeout: 245 seconds]14:33
-!- lambday [3ba2174f@gateway/web/freenode/ip.59.162.23.79] has quit [Ping timeout: 245 seconds]14:38
ZeThomasblackburn: it's weird, from outside gdb it crashes inconsistently in location14:39
ZeThomasfrom inside it's always at the same plce14:39
blackburnZeThomas: 139?14:41
ZeThomasyes, still that one14:42
-!- sonne|nowork [~sonnenbu@sams-office-nat.tomtomgroup.com] has quit [Quit: Leaving.]14:43
blackburnZeThomas: can't see anything really wrong14:44
blackburnZeThomas: any idea what is changing?14:44
blackburnZeThomas: which line it fails?14:44
ZeThomasblackburn: is there anything else I can type in gdb to give you more info?15:09
blackburnZeThomas: no not really -  what is the line in your python script it fails?15:10
ZeThomasblackburn: sec, I'm finding it out15:14
-!- sumit [73f91219@gateway/web/freenode/ip.115.249.18.25] has joined #shogun15:19
ZeThomashaha I just added some print statements to localise, and it gets further than before15:25
ZeThomasgotta love computers15:25
n4nd0ZeThomas: are you using all the time the same data?15:30
ZeThomasn4nd0: it's randomised, but I set a fixed seed for consistency in testing... If I'm not mistaken this takes all 'randomness' away15:31
n4nd0ZeThomas: ok, I think for debuggin you should fix the data15:32
n4nd0otherwise it could be completely fine that the code crashes in different parts for each execution15:32
ZeThomaswell, but that's what setting a seed does, right15:33
n4nd0if I were you, I would ensure it :)15:33
ZeThomasyou're making me curious... checking right now15:35
ZeThomasyep, that's good15:40
n4nd0ZeThomas: is there any randomness anywhere else in your method?15:42
ZeThomasblackburn, n4nd0: it's in the loop at line 87 (this time...); http://pastebin.com/F9n3GLTy15:44
ZeThomasand n4nd0: nope, no randomness, as far as I'm aware15:44
n4nd0ZeThomas: my point is:15:44
ZeThomasthe gdb crash location seems now consistent though with how it was outside of it...15:45
n4nd0if there is no randomness in the input to your program or on the task it does15:45
n4nd0then there must be some invalid memory access15:45
ZeThomasthat's what I was thinking as well15:45
n4nd0since I believe you cannot do this in python, I believe it must be in the C++ code15:46
n4nd0valgrind may help you to find it out15:46
n4nd0but if you are going to valgrind some python program, you better use a suppression file. Otherwise you will see lot of junk15:47
ZeThomasok, this sounds very vague to me15:48
n4nd0hehe ok, sorry15:48
n4nd0ZeThomas: so the data you are using, is it big?15:48
n4nd0you could for example input minimal data15:48
n4nd0print what's going on every few steps on your algorithm15:49
n4nd0and make sure it happens as you expect15:49
ZeThomasok, i'll try that15:55
n4nd0ok gtg now, see you later16:01
-!- n4nd0 [~nando@n186-p150.kthopen.kth.se] has quit [Quit: leaving]16:01
ZeThomasblackburn: I tried with a smaller dataset now (like a 100 something), and it fails at line 102...16:04
blackburnlet me check16:05
blackburn102??16:05
blackburnwhat about deeper?16:05
ZeThomasit's very weird, because with other kernels this part of the code never fails, and is very unrelated to any shogun code16:06
-!- votjak [~chatzilla@host-46-241-3-209.bbcustomer.zsttk.net] has joined #shogun16:07
-!- votjak [~chatzilla@host-46-241-3-209.bbcustomer.zsttk.net] has quit [Quit: ChatZilla 0.9.90 [Firefox 20.0/20130326150557]]16:14
ZeThomasblackburn: this makes no sense (well, not to me at least): it fails on the line: copy = self.__class__(self.contents, (), self.path, self.filename)16:15
ZeThomaswhich basically constructs a new object, with 3 strings and a tuple as args...16:16
blackburnyeah makes no sense to me well16:16
blackburnas well*16:16
-!- ahcorde [5813e5d6@gateway/web/freenode/ip.88.19.229.214] has joined #shogun16:20
ZeThomasI'm sick of staring at code, I'm going home16:33
ZeThomascu guys, thanks for the help&patience16:33
ZeThomasI'll be back ;-)16:33
-!- sumit [73f91219@gateway/web/freenode/ip.115.249.18.25] has quit [Ping timeout: 245 seconds]16:44
-!- ZeThomas [~thomaspr@2a02:2c40:100:a000:fce1:5064:9d1c:71c2] has quit [Quit: Leaving]16:48
shogun-buildbotbuild #346 of nightly_default is complete: Failure [failed tutorial]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/34617:07
@sonney2kblackburn, ZeThomas' problem really looks badly like some valgrind errors might have occured17:34
blackburnsonney2k: yeah something crazy17:34
shogun-buildbotbuild #873 of bsd1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/873  blamelist: Soeren Sonnenburg <sonne@debian.org>18:40
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun19:03
n4nd0Unknown parameter: --gtest=/home/wiking/gmock-1.6.0/gtest19:10
n4nd0muahaha19:10
-!- ahcorde [5813e5d6@gateway/web/freenode/ip.88.19.229.214] has quit [Ping timeout: 245 seconds]19:22
-!- lambday [0e8b614f@gateway/web/freenode/ip.14.139.97.79] has joined #shogun19:24
heikolambday: hi19:47
lambdayheiko: hey19:47
lambdaymore git stuffs :(19:48
heikogit?19:48
heikoah the new dev model you mean?19:48
lambdaysergey's mail19:48
lambdayyes...19:48
heikolambday: this is actually gonna help us a lot :)19:48
lambdayheiko: hmmm... i still haven't checked it..19:49
heikobasically you develop new stuff in feature branches19:49
heikowhich are then committed/merged19:49
heikoand eventually merged into thedevelopment branch19:49
lambdayhmm...19:50
lambdayI need to check this gitflow thing...19:51
n4nd0heiko: btw, we still need to do the PR through github, don't we?19:51
lambdayand in my system, all I need to do is to change the upstream to "features", right?19:51
heikon4nd0: I think yet, but I haven't comitted anything since then19:52
n4nd0heiko: ok19:52
heikolambday, n4nd0 I did not yet have time to check out the details, but check19:52
heikohttp://buildamodule.com/video/change-management-and-version-control-deploying-releases-features-and-fixes-with-git-how-to-use-a-scalable-git-branching-model-called-gitflow19:52
heikohttp://labs.grupow.com/blog/2011/07/05/getting-started-with-git-flow19:53
heikoand sergey's mail19:53
heikolambday: yes19:53
heikoand then19:53
heikohttp://nvie.com/posts/a-successful-git-branching-model/19:53
lambdayheiko: hmmm...19:53
n4nd0nice, I like videolectures19:53
lambdayme too me too :D19:53
heikome too :)19:53
lambdayheiko: by the way, on a different topic, for classification using svm, n train and m test points, the train kernel should be nxn, right? then we train the svm, and then we fit the test kernel nxm in it after training, and use it to get labels.. right?19:56
heikolambday:  let me think ...19:58
heikolambday: yes, thats how it works19:58
heikoin shogun we have this lhs rhs thing on the kernel19:58
lambdayheiko: just a sec, be right back19:58
heikolhs is always training, rhs is first training, then test19:58
heikoso exactly training matrix is nxn, test matrix is nxm19:58
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving]20:05
@sonney2kheiko, blackburn: how would students work with this model?20:07
@sonney2kIMHO they just fork the develop branch20:07
blackburnsonney2k: PRs20:07
@sonney2kand create there PRs against that20:07
blackburnsonney2k: not really though20:08
blackburnthey do not fork the branch20:08
blackburnthey work on feature/my-super-feature20:08
blackburnand do PR to the develop20:08
blackburnheiko: sonney2k: http://warm-reaches-4085.herokuapp.com/ website of tapkee runs clojure now :D20:10
heikoblackburn:  nice20:10
blackburnargh I broke equations again20:11
heikosonney2k, blackburn I still gotta do some reading on this, sorry20:11
blackburnheiko: on what?20:11
heikothe new git20:11
blackburnwhy sorry?20:11
blackburn:)20:11
heikoand think how students commit stuff20:11
heikosince too busy to look at it currently20:11
heikogoing home, bye all!20:21
-!- heiko [~heiko@nat-174-82.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.]20:21
shogun-buildbotbuild #1006 of deb3 - modular_interfaces is complete: Failure [failed test libshogun]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1006  blamelist: Soeren Sonnenburg <sonne@debian.org>20:22
@sonney2kblackburn, that's how it would look like https://dl.dropbox.com/u/31480471/shogun-website-ann.png21:41
blackburnsonney2k: yeah much better now21:54
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun21:59
-!- lambday [0e8b614f@gateway/web/freenode/ip.14.139.97.79] has quit [Ping timeout: 245 seconds]22:03
@sonney2kn4nd0, https://dl.dropbox.com/u/31480471/shogun-website-ann.png22:07
@sonney2kblackburn, huh? no difference...22:07
@sonney2konly some nice text22:07
blackburnheh well w/o text it was just an image22:08
@sonney2kblackburn, I just wanted to give you an idea where the text / logo could be22:08
@sonney2kblackburn, *lol* http://www.washingtonpost.com/blogs/worldviews/files/2013/03/2013-03-25T050307Z_01_PYO04_RTRIDSP_3_KOREA-NORTH.jpg22:10
@sonney2kcyberattack of the finest22:10
n4nd0sonney2k: yes, I saw in logs22:10
blackburnsonney2k: oh this guy rocks22:10
n4nd0it is looking nice I think22:10
n4nd0ah wait you updated with text22:11
n4nd0ah no, same last time I looked22:11
@sonney2kn4nd0, so now we need to polish the ws page / announcement a bit and then if we get in we send of calls as much as we can22:11
n4nd0sonney2k: what about the workshop advert I sent you?22:11
@sonney2kblackburn, that is how I imagine you when coding :P22:12
blackburnsonney2k: uh this guy is like 3x of me22:12
@sonney2kcold war computers22:12
blackburnso please do not compare :D22:12
@sonney2kand look at the power switch at the right bottom of the black box22:12
blackburnwell that's military thing22:13
@sonney2kblackburn, aren't you some kind of heavy-weight champion?22:13
blackburnsonney2k: no I am not 150 kg22:13
blackburn:D22:13
blackburnlike that guy22:13
@sonney2kyeah he says world domination and he means it22:13
@sonney2klike myself22:14
@sonney2kshogun world domination22:14
@sonney2khe uses atom bombs22:14
@sonney2kand we use ML22:14
@sonney2kand C++22:14
@sonney2kas explosive I would say22:14
@sonney2k*scnr*22:14
n4nd0we need an hymn22:14
blackburnsonney2k: btw check he always wears black22:15
@sonney2kand a flag22:15
blackburnvery stylish22:15
blackburnI am an korean expert lol22:15
blackburnhe tends to look like his grandfather22:15
blackburnkim ir sen that is22:15
@sonney2khttps://lh3.ggpht.com/_T-_fQZXgzsE/TA_syVg7Y1I/AAAAAAAAACA/3GRJxDwFzi4/s1600/musashi.jpg22:16
blackburnor rather kim il sung in english22:16
blackburnwell we have a flag22:16
* sonney2k is tired... cu...22:16
blackburnhttps://dl.dropbox.com/u/10139213/shogun/sonne.jpg22:17
blackburnsonney2k: ^22:17
@sonney2kblackburn, yeah but as I said I don't wear pants22:17
blackburnsonney2k: we can assume that for a flag22:17
blackburnw/o pants it won't be as official22:17
n4nd0haha22:18
blackburnsonney2k: and these pants are special22:18
blackburnI see the device is getting air circulated22:19
@sonney2kblackburn, heh22:19
@sonney2kblackburn, look anonymous loves kim too http://www.heise.de/newsticker/meldung/Anonymous-attackiert-Online-Angebote-Nordkoreas-1836224.html?view=zoom;zoom=122:19
blackburnsonney2k: ohh that's way too stupid22:20
blackburnhe is not that bad22:20
blackburnit is the game and he kinda has no choice22:20
* sonney2k ZZzzz...22:22
n4nd0oh shit Korea has started launching missiles22:32
n4nd0http://25.media.tumblr.com/ff5157f3ced7988cfacc8c77248fe420/tumblr_mkli36AIDM1qigfjto1_400.gif22:33
blackburnheh22:34
n4nd0ok22:50
n4nd0it didn't work out with 21 either22:51
n4nd0blackburn: oops that was for you :P22:51
-!- cwidmer [8ca3fe9d@gateway/web/freenode/ip.140.163.254.157] has joined #shogun22:51
n4nd0wrong tab22:51
blackburnn4nd0: well may be waffles go wrong22:51
blackburnhey there cwidmer22:51
blackburnjoin #tapkee22:51
--- Log closed Sat Apr 06 00:00:25 2013
« Thursday, 2013-04-04 Index Saturday, 2013-04-06 »

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