| --- Log opened Fri Apr 20 00:00:19 2012 | ||
| -!- emrecelikten [~emre@176.40.238.20] has joined #shogun | 00:08 | |
| -!- wiking [~wiking@huwico/staff/wiking] has quit [Ping timeout: 248 seconds] | 01:04 | |
| -!- emrecelikten [~emre@176.40.238.20] has quit [Quit: Leaving.] | 01:16 | |
| -!- emrecelikten [~emre@176.40.238.20] has joined #shogun | 01:17 | |
| -!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 01:30 | |
| -!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 02:11 | |
| -!- av3ngr [av3ngr@nat/redhat/x-svsxtaryupzxrgvv] has joined #shogun | 02:13 | |
| -!- PhilTillet [~Philippe@157.159.42.154] has quit [Quit: Leaving] | 02:42 | |
| -!- emrecelikten [~emre@176.40.238.20] has quit [Quit: Leaving.] | 03:13 | |
| -!- xiangwang [~chatzilla@159.226.60.224] has joined #shogun | 05:15 | |
| -!- xiangwang [~chatzilla@159.226.60.224] has quit [Ping timeout: 244 seconds] | 05:21 | |
| -!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 07:05 | |
| -!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 07:10 | |
| CIA-64 | shogun: Evgeniy Andreev master * rb919bef / examples/undocumented/python_modular/regression_gaussian_process_modular.py : fixed regression gaussian process example for python3 support - http://git.io/f08Skg | 07:59 |
|---|---|---|
| CIA-64 | shogun: Soeren Sonnenburg master * rd466587 / examples/undocumented/python_modular/regression_gaussian_process_modular.py : | 07:59 |
| CIA-64 | shogun: Merge pull request #473 from gsomix/python3_interface | 07:59 |
| CIA-64 | shogun: Fixed regression gaussian process example for python3 support - http://git.io/Vr85dg | 07:59 |
| CIA-64 | shogun: Soeren Sonnenburg master * r8b0a9d7 / (74 files in 11 dirs): | 08:05 |
| CIA-64 | shogun: Merge pull request #472 from gsomix/big_deal_with_sgvector | 08:05 |
| CIA-64 | shogun: Start with SGVector -> SGVector& migration (+12 more commits...) - http://git.io/aGuOHQ | 08:05 |
| CIA-64 | shogun: Soeren Sonnenburg master * r39c82e6 / (src/shogun/regression/LARS.cpp src/shogun/regression/LARS.h): fix compile error: make LARS unavailable w/o lapack - http://git.io/9e1viA | 08:12 |
| -!- wiking [~wiking@78-23-189-112.access.telenet.be] has joined #shogun | 08:55 | |
| -!- wiking [~wiking@78-23-189-112.access.telenet.be] has quit [Changing host] | 08:55 | |
| -!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 08:55 | |
| -!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 09:02 | |
| -!- xiangwang [~chatzilla@159.226.60.224] has joined #shogun | 09:05 | |
| -!- uricamic [9320543b@gateway/web/freenode/ip.147.32.84.59] has joined #shogun | 09:06 | |
| -!- xiangwang [~chatzilla@159.226.60.224] has quit [Client Quit] | 09:06 | |
| -!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 09:15 | |
| -!- xiangwang [~chatzilla@159.226.60.224] has joined #shogun | 09:28 | |
| -!- Jiuding [d2198538@gateway/web/freenode/ip.210.25.133.56] has joined #shogun | 09:30 | |
| -!- Jiuding [d2198538@gateway/web/freenode/ip.210.25.133.56] has quit [Client Quit] | 09:31 | |
| xiangwang | hi wiking, pull request made succesfully. Thanks:) | 09:36 |
| -!- harshit_ [~harshit@182.68.43.52] has joined #shogun | 09:47 | |
| wiking | xiangwang: no worries | 09:51 |
| wiking | gsomix: yeeeey cool patches! | 09:51 |
| wiking | gsomix: around | 09:51 |
| wiking | ? | 09:51 |
| -!- harshit_ [~harshit@182.68.43.52] has quit [Ping timeout: 260 seconds] | 09:56 | |
| xiangwang | ok, just keep on working++ | 09:56 |
| -!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 10:01 | |
| -!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has joined #shogun | 10:06 | |
| blackburn | ok I am finally finished with that shit | 10:07 |
| blackburn | who is around? wiking? gsomix? sonne|work? | 10:08 |
| wiking | here | 10:08 |
| wiking | finished with what shit? | 10:08 |
| blackburn | wiking: passed optics ungraded exam | 10:09 |
| wiking | hahahah congrats | 10:09 |
| blackburn | that was the most awful exam ever | 10:09 |
| blackburn | I guess | 10:10 |
| wiking | btw: don't we want to expose the primal objective value in svm? | 10:10 |
| wiking | i would need it... | 10:10 |
| blackburn | comatic aberrations, chromatic aberrations | 10:10 |
| blackburn | hrhrrh | 10:10 |
| wiking | hahahaha | 10:10 |
| blackburn | seidel aberrations | 10:10 |
| blackburn | holography | 10:11 |
| blackburn | denisyuk holography | 10:11 |
| n4nd0 | blackburn: I am here around too, I am glad you finish you exam ;) | 10:11 |
| n4nd0 | finished* | 10:11 |
| blackburn | n4nd0: uh sorry I missed your nickname in the list | 10:11 |
| blackburn | thanks | 10:11 |
| blackburn | wiking: I thought it is available | 10:11 |
| blackburn | wiking: do you mean multiclass? | 10:11 |
| wiking | blackburn: noun... i'm talking about svmocas | 10:11 |
| blackburn | wiking: binary? | 10:12 |
| wiking | yeps | 10:12 |
| blackburn | svm base class has primal objective method IIRC | 10:12 |
| blackburn | so it doesn't depend on implementation | 10:12 |
| wiking | mmm where? :) | 10:13 |
| wiking | man i was looking cmachine and clinearmachine api | 10:13 |
| wiking | haven't seen one... | 10:13 |
| -!- av3ngr [av3ngr@nat/redhat/x-svsxtaryupzxrgvv] has quit [Quit: That's all folks!] | 10:14 | |
| blackburn | wiking: SVM | 10:14 |
| wiking | checking now CSVM | 10:14 |
| blackburn | how can that be linearmachine? ;) | 10:14 |
| blackburn | what is the objective | 10:14 |
| wiking | get_objective? | 10:15 |
| blackburn | float64_t compute_svm_dual_objective(); | 10:15 |
| blackburn | this one | 10:15 |
| wiking | mmm | 10:15 |
| blackburn | or primal it is up to you | 10:15 |
| blackburn | ;) | 10:15 |
| wiking | ok then there's only one thing i think | 10:15 |
| wiking | class CSVMOcas : public CLinearMachine | 10:15 |
| wiking | shogun/classifier/svm/SVMOcas.h | 10:16 |
| wiking | this needs to be fixed afaik | 10:16 |
| wiking | and this is why i was checking linear machine ;P | 10:16 |
| wiking | since i'd like to use ocas for latent | 10:16 |
| blackburn | ohohoho | 10:17 |
| blackburn | I see | 10:17 |
| wiking | so i guess before anything | 10:17 |
| wiking | i should do the porting of svmocas to csvm | 10:17 |
| blackburn | no way! | 10:17 |
| wiking | or use libocas straight away | 10:17 |
| blackburn | wiking: it shouldn't be csvm | 10:17 |
| blackburn | csvm is a kernel machine | 10:18 |
| wiking | ah so there's no linear svm api | 10:18 |
| blackburn | yes exactly | 10:18 |
| wiking | ok i see now | 10:18 |
| blackburn | I am afraid we would need that | 10:18 |
| wiking | yep | 10:18 |
| wiking | so we would need CSVM and CKERNELSVM | 10:18 |
| wiking | or seomthing like this | 10:18 |
| wiking | or CSVM and from that have a CKERNELSVM inherited | 10:18 |
| blackburn | just insert CLinearSVM between CLinearMachine and CSVMOcas | 10:18 |
| blackburn | no I don't think we need to rename basic svm | 10:20 |
| wiking | ok | 10:20 |
| wiking | anyways we need a standard api for linear svm | 10:20 |
| wiking | where one can get access to the primal/dual objective value | 10:20 |
| blackburn | things getting tougher w/o multiple inheritance | 10:20 |
| blackburn | wiking: actually even computing routines are different | 10:20 |
| blackburn | computing primal of linear svm costs nothing | 10:20 |
| blackburn | would be really fast I mean | 10:20 |
| wiking | yeah it's only that currently you do not have access to those values ;P | 10:20 |
| blackburn | wiking: just add it is easy ;) | 10:20 |
| wiking | i mean i could have just hacked now CSVMOcas and expose the values in it | 10:20 |
| wiking | but i thought it's better to do something more general | 10:21 |
| wiking | and since now i've realized that we don't have a standard api for linear svm | 10:21 |
| blackburn | yes CLinearSVM sounds legal for me | 10:21 |
| wiking | ok i'll make one | 10:21 |
| blackburn | sonne|work: CLinearSVM | 10:21 |
| wiking | heheh again a branch switch | 10:21 |
| wiking | i looooove git stash :D | 10:22 |
| blackburn | #optimization, #shogun, #svms | 10:22 |
| blackburn | that should take his attention | 10:22 |
| wiking | hahahaha | 10:22 |
| wiking | blackburn: we've agreed yesterday | 10:23 |
| wiking | to have a shogun/optimizer folder | 10:23 |
| wiking | with all the qp solvers being there | 10:23 |
| blackburn | yay | 10:23 |
| blackburn | I shall read the logs | 10:23 |
| wiking | only thing that sonne|work isn't sure yet what should we do about tron and pr_loqo | 10:23 |
| n4nd0 | wiking: what is tron exactly? I have not found that much doc within the code | 10:28 |
| n4nd0 | and google search gets other hits by tron as you can imagine :) | 10:28 |
| wiking | :D | 10:28 |
| blackburn | n4nd0: some optimizer | 10:28 |
| wiking | blackburn: ah yeah forgot to mention. would u mind if i change the doxygen config that the inherited methods are as well present in the documentation of a class? | 10:33 |
| blackburn | I don't but no idea what sonne|work would say | 10:33 |
| wiking | ok well u guys will decide about the pull request :P | 10:33 |
| blackburn | I am actually thinking | 10:34 |
| blackburn | we should spend some week in may | 10:34 |
| blackburn | on doc | 10:34 |
| blackburn | kind of doc sprint heh | 10:34 |
| wiking | yeyey | 10:34 |
| -!- Marty28 [~marty@74-21.vpn.rz.uni-potsdam.de] has joined #shogun | 10:38 | |
| blackburn | n4nd0: do you like it? | 10:42 |
| blackburn | wiking: and you? | 10:42 |
| n4nd0 | doc improvement? | 10:42 |
| n4nd0 | I might be kind of boring but it is probably necessary... so yes I like it | 10:43 |
| blackburn | no we can make it fun | 10:45 |
| n4nd0 | how? | 10:45 |
| blackburn | http://priceofoil.org/wp-content/uploads/2009/10/barack-obama-yes-we-can.jpg | 10:45 |
| blackburn | n4nd0: barack obama says we can | 10:45 |
| n4nd0 | haha all right | 10:45 |
| n4nd0 | but not how, looks fishy | 10:46 |
| blackburn | n4nd0: be a mathematician | 10:46 |
| blackburn | existence is everything | 10:46 |
| wiking | blackburn: most definitely would be up for it | 10:47 |
| -!- Marty28 [~marty@74-21.vpn.rz.uni-potsdam.de] has quit [Quit: Colloquy for iPad - http://colloquy.mobi] | 10:49 | |
| blackburn | what about middle of may? | 10:49 |
| n4nd0 | after the 7th it is good for me | 10:50 |
| wiking | mm i'm away between 13-20 and 24->31 | 10:50 |
| wiking | otherwise it's good | 10:50 |
| -!- gsomix [~gsomix@188.168.13.79] has quit [Read error: Operation timed out] | 10:51 | |
| blackburn | wiking: is it serious you will be in iceland or so? | 10:52 |
| wiking | blackburn: in july yes | 10:52 |
| wiking | blackburn: but there i'll be 24/7 available | 10:52 |
| wiking | in may i have conferences at those dates... that's why i cannot really work | 10:52 |
| -!- xiangwang [~chatzilla@159.226.60.224] has quit [Ping timeout: 265 seconds] | 10:52 | |
| blackburn | I see | 10:52 |
| shogun-buildbot | build #500 of ruby_modular is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/ruby_modular/builds/500 blamelist: sonne@debian.org | 10:53 |
| -!- gsomix [~gsomix@188.168.13.79] has joined #shogun | 10:55 | |
| sonne|work | wiking, blackburn please no CLinearSVM | 10:56 |
| blackburn | sonne|work: why? | 10:56 |
| wiking | sonne|work: ok what then? | 10:56 |
| sonne|work | nothing | 10:56 |
| blackburn | sonne|work: where to place helpers like objective stuff? | 10:57 |
| sonne|work | blackburn: in LinearMachine directly | 10:57 |
| blackburn | sonne|work: any motivation why we should avoid CLinearSVM? | 10:57 |
| wiking | sonne|work: so i should then reimplement ocas optimizer if i want to get access to objective value? | 10:57 |
| sonne|work | blackburn: what would you derive liblinear from? | 10:57 |
| blackburn | sonne|work: ehm CLinearSVM? | 10:58 |
| sonne|work | haha | 10:58 |
| sonne|work | it is logistic regression, linear programming machine, support vector regression *and* linear svm | 10:58 |
| sonne|work | all in one thing | 10:58 |
| wiking | yep true that | 10:59 |
| blackburn | ah you're right | 10:59 |
| blackburn | sonne|work: do you like doc sprint idea? | 10:59 |
| sonne|work | I am certainly happy if you all write doc :D | 11:00 |
| wiking | ehehhe so ok | 11:00 |
| wiking | should we then just add a new function to linear machine | 11:00 |
| blackburn | sonne|work: it would need your help too ;) | 11:00 |
| wiking | that is 'optional'? | 11:00 |
| blackburn | especially for legacy stuff | 11:01 |
| sonne|work | wiking: yes - even implement it directly in there | 11:01 |
| sonne|work | blackburn: why that? | 11:01 |
| sonne|work | then it is not good :P | 11:01 |
| blackburn | sonne|work: I am not able to doc some POIMs (wtf it is actually?), alignment kernels blabla | 11:01 |
| * sonne|work hates writing docs - and yes lots if not most of the doc was written by me | 11:01 | |
| wiking | sonne|work: directly there? i mean it depends on the solver whether i have access to it at all... | 11:02 |
| sonne|work | wiking: I thought you wanted to compute objective? | 11:02 |
| sonne|work | svm objective is pretty clear if you have training data, w and b | 11:02 |
| sonne|work | and C | 11:02 |
| wiking | yes but only svm has it | 11:02 |
| blackburn | wiking: hm but why not to compute dual objective of LDA's w? | 11:03 |
| blackburn | would be interesting haha | 11:03 |
| sonne|work | ok then - float64_t compute_primal_objective() { SG_NOTIMPLEMENTED; } | 11:03 |
| wiking | sonne|work: yep that's what i thought | 11:03 |
| wiking | that have an pure virtual function | 11:03 |
| wiking | in linear machine | 11:03 |
| wiking | and if we can have such | 11:03 |
| wiking | we supply one | 11:04 |
| wiking | if not then it's notimplemented | 11:04 |
| sonne|work | not pure virtual but virtual yes | 11:04 |
| wiking | hehe ok | 11:04 |
| blackburn | okay lets schedule some doc madness in the second half of may | 11:07 |
| wiking | ok can i ask a naive question | 11:09 |
| wiking | like in the case of CSVMOcas | 11:10 |
| wiking | why the svmocas related variables are protected? | 11:10 |
| wiking | are we planning ahead that maybe one day somebody will do an inheritance of it where he/she wants access to those variables | 11:11 |
| wiking | but i mean this is true for a lot of machines.. | 11:11 |
| blackburn | wiking: but they are protected so what is the problem? | 11:12 |
| wiking | ok other way to formulate my question: why aren't they private | 11:12 |
| blackburn | sorry I misread your question as statement :) | 11:13 |
| blackburn | yes it makes sense to provide such extensibility I think | 11:13 |
| -!- PhilTillet [~Philippe@npasserelle10.minet.net] has joined #shogun | 11:14 | |
| PhilTillet | Hello everybody | 11:14 |
| PhilTillet | :) | 11:14 |
| blackburn | hi | 11:14 |
| PhilTillet | I'm finally in spring break for 1 week yeee | 11:15 |
| wiking | PhilTillet: good on you | 11:15 |
| wiking | enjoy it ;) | 11:15 |
| PhilTillet | haha | 11:15 |
| PhilTillet | I won't get bored :p a lot of work | 11:16 |
| PhilTillet | i'm kind of disappointed by what I do at school so I end up having to do a lot of work on my own :p | 11:18 |
| -!- xiangwang [~chatzilla@159.226.60.224] has joined #shogun | 11:20 | |
| -!- xiangwang [~chatzilla@159.226.60.224] has quit [Ping timeout: 276 seconds] | 11:26 | |
| shogun-buildbot | build #520 of csharp_modular is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/csharp_modular/builds/520 blamelist: sonne@debian.org, gsomix@gmail.com | 11:27 |
| blackburn | nice patch | 11:28 |
| shogun-buildbot | build #533 of octave_modular is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/octave_modular/builds/533 blamelist: sonne@debian.org, gsomix@gmail.com | 11:34 |
| -!- xiangwang [~chatzilla@159.226.60.224] has joined #shogun | 11:35 | |
| -!- PhilTillet [~Philippe@npasserelle10.minet.net] has quit [Ping timeout: 248 seconds] | 11:40 | |
| shogun-buildbot | build #509 of lua_modular is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/lua_modular/builds/509 blamelist: sonne@debian.org, gsomix@gmail.com | 11:40 |
| shogun-buildbot | build #510 of java_modular is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/java_modular/builds/510 blamelist: sonne@debian.org, gsomix@gmail.com | 11:51 |
| -!- PhilTillet [~Philippe@157.159.42.154] has joined #shogun | 11:52 | |
| shogun-buildbot | build #506 of python_modular is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/python_modular/builds/506 blamelist: sonne@debian.org, gsomix@gmail.com | 11:58 |
| sonne|work | gsomix: you need to adjust all typemaps to work with SGVector& too | 12:02 |
| gsomix | sonne|work, ok. I need a little time. I have just returned. | 12:06 |
| sonne|work | gsomix: and of course changing typemaps when there are still some SGVector's without & around will still cause breakage - so maybe better just copy typemaps for now to have a SGVector& and SGVector variant and then slowly proceed migrating them | 12:11 |
| gsomix | sonne|work, got it. | 12:17 |
| sonne|work | gsomix: all the other things I have in mind are piece of cake compared to this transition | 12:22 |
| sonne|work | most intrusive so should be done first and will trigger lots of bugs | 12:22 |
| -!- sonne|work [~sonnenbu@194.78.35.195] has left #shogun [] | 13:17 | |
| -!- xiangwang_ [~chatzilla@210.73.4.120] has joined #shogun | 13:21 | |
| -!- xiangwang_ [~chatzilla@210.73.4.120] has quit [Quit: ChatZilla 0.9.88.2 [Firefox 9.0.1/20111220165912]] | 13:21 | |
| -!- xiangwang [~chatzilla@159.226.60.224] has quit [Ping timeout: 272 seconds] | 13:36 | |
| -!- sonne|work [~sonnenbu@194.78.35.195] has joined #shogun | 13:49 | |
| blackburn | *pleeease turn on conditioners* | 13:50 |
| PhilTillet | haha | 13:50 |
| PhilTillet | blackburn, what about turning on the preconditioners? | 13:50 |
| PhilTillet | :D | 13:50 |
| blackburn | PhilTillet: I don't mind if any preconditioner is able to cool the room heh | 13:51 |
| PhilTillet | well | 13:52 |
| PhilTillet | preconditioners are cool | 13:52 |
| PhilTillet | :D | 13:52 |
| PhilTillet | or maybe you can put a preconditioner on your conditioner | 13:54 |
| PhilTillet | so that the temperature converges faster | 13:54 |
| PhilTillet | o_o | 13:54 |
| blackburn | I hope it would converge to some minus this time | 13:55 |
| PhilTillet | haha | 13:56 |
| PhilTillet | Better if the problem is convex too | 13:56 |
| PhilTillet | XD | 13:56 |
| n4nd0 | :O | 13:57 |
| blackburn | I'll become convex if temperature in this room stays the same or it becomes hotter | 13:57 |
| PhilTillet | lol | 13:57 |
| PhilTillet | n4nd0, yes we're a bit crazy | 13:59 |
| n4nd0 | what the hell man, today it shows in Sweden and you in Russia are warm | 13:59 |
| n4nd0 | no way | 13:59 |
| n4nd0 | snows | 13:59 |
| n4nd0 | not shows | 13:59 |
| n4nd0 | haha | 13:59 |
| blackburn | snows? | 14:00 |
| n4nd0 | yeah | 14:00 |
| blackburn | I'm dying because of the hell here | 14:00 |
| n4nd0 | it has snowed today here | 14:00 |
| blackburn | it should be +30 | 14:00 |
| PhilTillet | weird | 14:00 |
| PhilTillet | is sweden that far from Russia? | 14:01 |
| blackburn | russia is far from russia | 14:01 |
| PhilTillet | lol | 14:01 |
| blackburn | hmm +24 | 14:01 |
| n4nd0 | haha | 14:01 |
| blackburn | but feels like hell | 14:01 |
| blackburn | they broke conditioners in the office | 14:02 |
| -!- xiangwang [~chatzilla@159.226.60.224] has joined #shogun | 14:03 | |
| n4nd0 | see you later guys | 14:04 |
| gsomix | sonney2k, sonne|work around? | 14:06 |
| -!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 248 seconds] | 14:08 | |
| -!- xiangwang [~chatzilla@159.226.60.224] has quit [Ping timeout: 245 seconds] | 14:09 | |
| -!- xiangwang [~chatzilla@159.226.60.224] has joined #shogun | 14:23 | |
| -!- xiangwang [~chatzilla@159.226.60.224] has quit [Ping timeout: 260 seconds] | 14:29 | |
| -!- n4nd0 [~nando@n129-p107.kthopen.kth.se] has joined #shogun | 14:30 | |
| gsomix | blackburn, wut? today is cold, desu | 14:31 |
| blackburn | gsomix: cold? you should visit netcracker office with turned off conditioners | 14:31 |
| -!- xiangwang [~chatzilla@159.226.60.224] has joined #shogun | 14:32 | |
| blackburn | however it is way too warm outside as well | 14:35 |
| gsomix | blackburn, hmm. I'm in my room and I'm cold, hehe | 14:36 |
| -!- xiangwang [~chatzilla@159.226.60.224] has quit [Ping timeout: 245 seconds] | 14:37 | |
| blackburn | are you cold? | 14:37 |
| PhilTillet | blackburn, I'm kind of cold too | 14:38 |
| -!- xiangwang [~chatzilla@159.226.60.224] has joined #shogun | 14:39 | |
| blackburn | as for me I want to sleeeep :) | 14:40 |
| PhilTillet | haha | 14:40 |
| PhilTillet | me too, but i'm reading some svm stuffs :p | 14:41 |
| PhilTillet | does Shogun use SMO for training SVMs? | 14:41 |
| blackburn | PhilTillet: at some point | 14:42 |
| blackburn | libsvm is some kind of | 14:42 |
| PhilTillet | yes I read it was based on that | 14:42 |
| PhilTillet | but not exactly the same | 14:42 |
| blackburn | in the shogun we just incorporate different solver | 14:44 |
| blackburn | with some improvement along the way to merge | 14:44 |
| PhilTillet | okay | 14:47 |
| PhilTillet | cause i've read several papers on parallelizing svm, and most of the time they just use SMO for learning | 14:47 |
| -!- pluskid [~pluskid@173.254.214.60] has joined #shogun | 14:52 | |
| -!- xiangwang [~chatzilla@159.226.60.224] has quit [Ping timeout: 265 seconds] | 14:52 | |
| -!- n4nd0 [~nando@n129-p107.kthopen.kth.se] has quit [Quit: leaving] | 15:00 | |
| -!- harshit_ [~harshit@182.68.43.52] has joined #shogun | 15:04 | |
| -!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has quit [Quit: Page closed] | 15:04 | |
| gsomix | time to compile | 15:05 |
| * gsomix afk | 15:05 | |
| -!- Marina_ [55b38cc8@gateway/web/freenode/ip.85.179.140.200] has joined #shogun | 15:16 | |
| -!- xiangwang [~chatzilla@159.226.60.224] has joined #shogun | 15:17 | |
| Marina_ | @sonney2k: if i try to train a svm on the real spice data, I get the error: km = kernel.get_kernel_matrix(); SystemError: Out of memory error, tried to allocate 20000000000000000 bytes using malloc. | 15:18 |
| Marina_ | isn't it possible to train with so many datas? | 15:19 |
| PhilTillet | Marina_, isn't that 2 e16 bytes? :p | 15:22 |
| -!- harshit_ [~harshit@182.68.43.52] has quit [Remote host closed the connection] | 15:22 | |
| Marina_ | the dna_train.dat is 9.814.454 KB | 15:23 |
| -!- xiangwang [~chatzilla@159.226.60.224] has quit [Ping timeout: 260 seconds] | 15:23 | |
| Marina_ | or what do you mean? | 15:23 |
| sonne|work | Marina_: use a small subset of the data | 15:30 |
| Marina_ | @sonney2k how many data should I use for the subset? and should I load them first all into the programm or is it possible to load only a subset from the file? | 15:36 |
| sonne|work | Marina_: start with very few (say 1000) and then gradually take more | 15:41 |
| sonne|work | 100k is probably enough to show anything of interest | 15:42 |
| Marina_ | @sonne2k so i load them all with lm.load_dna('real_data/dna_train.dat') and then take the first 1000 | 15:43 |
| Marina_ | right? | 15:43 |
| sonne|work | no | 15:45 |
| sonne|work | load just a few | 15:45 |
| Marina_ | but how do I do so? | 15:45 |
| sonne|work | x=file('my_file.dna').readlines()[:100] ? | 15:46 |
| sonne|work | do it in whatever language you progam in | 15:47 |
| -!- harshit_ [~harshit@182.68.43.52] has joined #shogun | 15:47 | |
| harshit_ | hello everyone :) | 15:48 |
| harshit_ | sonney2k, here ? | 15:48 |
| -!- n4nd0 [~nando@n129-p107.kthopen.kth.se] has joined #shogun | 15:50 | |
| n4nd0 | sonne|work: hey! I got some time to be in shogun mode :) | 15:54 |
| n4nd0 | sonne|work: do you want me to try first if MulticlassKernel is ok now after Heiko's work? | 15:55 |
| n4nd0 | sonne|work: or should I continue with cover tree?? | 15:55 |
| sonne|work | n4nd0: I am most interested in the timing experiment | 15:55 |
| sonne|work | harshit_: yes | 15:55 |
| harshit_ | sonne|work, got to know form logs that svr is released ..! Are you working on it ? | 15:56 |
| harshit_ | I am asking coz I want to do that | 15:56 |
| gsomix | sonne|work, ok, it compiled. python_modular, ruby_modular, octave_modular are fine now. | 15:57 |
| n4nd0 | sonne|work: using CBlas function then right? I don't have that clear if we concluded whether CBlas or the decomposition (x-y)^2 = x^2 + y^2 -2?x?y is preferable | 15:57 |
| gsomix | sonne|work, i need to go. cu little later | 15:57 |
| sonne|work | n4nd0: cblas has no norm(x,y) function | 15:58 |
| sonne|work | n4nd0: so try the decomposition using blas | 15:58 |
| n4nd0 | sonne|work: let me look in the logs, I remember blackburn suggested to use a cblas function | 16:00 |
| sonne|work | n4nd0: dnrm2 | 16:00 |
| n4nd0 | sonne|work: yeah exactly | 16:00 |
| sonne|work | n4nd0: that doesn't help, will only do ||x|| | 16:01 |
| sonne|work | not ||x-y|| | 16:01 |
| sonne|work | harshit_: I started already - but what you are trying to do (SVMLin) is also nice! | 16:01 |
| n4nd0 | sonne|work: I could comptue the difference vector first | 16:02 |
| sonne|work | which is already slower than doing x^T y in the decomposition :) | 16:02 |
| sonne|work | and blas | 16:02 |
| PhilTillet | Ha, now I understand why GaussianKernel is a DotKernel and not a distance kernel :p | 16:05 |
| sonne|work | PhilTillet: yeah nice trick eh? | 16:05 |
| PhilTillet | yes :D | 16:06 |
| PhilTillet | on GPU, norm2(x-y) is faster :p this was why I was facing a little design issue for my patch | 16:06 |
| sonne|work | PhilTillet: why that? | 16:07 |
| sonne|work | I mean if you already hav ||x||^2 and ||y||^2 precomputed | 16:07 |
| sonne|work | x^T y cannot possible slower or? | 16:07 |
| PhilTillet | yes, if you have these two precomputed | 16:07 |
| PhilTillet | indeed we have | 16:07 |
| PhilTillet | :p | 16:07 |
| sonne|work | yeah of course you need to precompute these | 16:08 |
| sonne|work | otherwise -> slow | 16:08 |
| PhilTillet | the point is that i was still with these matrix stuff, I did some benchmark replacing dot product with dot on GPU | 16:08 |
| PhilTillet | but well, it is pointless I think | 16:08 |
| PhilTillet | gpu starts to benefit for blas1,2 for vector size > 100 000 | 16:09 |
| PhilTillet | and I guess we rarely have features of dimension more than 100 000 :p | 16:09 |
| PhilTillet | well, and I was thinking, maybe the cleanest thing is to make some GPUKernels, like GPUDotKernels, etc... | 16:12 |
| sonne|work | PhilTillet: maybe it is really only possible if this is done for a specific kernel, specific features + kernel machine | 16:15 |
| sonne|work | so GaussianKernelMachine | 16:16 |
| sonne|work | working only on 'simple' features | 16:16 |
| PhilTillet | Hmmm | 16:16 |
| PhilTillet | Could be possible with sparse features too | 16:16 |
| PhilTillet | but I agree that for online learning, it's probably not efficient | 16:17 |
| PhilTillet | :p | 16:17 |
| sonne|work | how could that work with sparse features? | 16:17 |
| sonne|work | isn't that random memory access then again? | 16:17 |
| PhilTillet | Hmm, it's possible to do sparse linear algebra on GPU :p | 16:18 |
| PhilTillet | I guess what's important is to have all the features at the beginning so you can cache them efficiently | 16:19 |
| PhilTillet | I've seen a paper with SMO on GPU where they use chunking because the kernel matrix is too large to fit in memory | 16:20 |
| PhilTillet | they had 50x speedup on learning compared to libsvm >< | 16:25 |
| -!- harshit_ [~harshit@182.68.43.52] has quit [Ping timeout: 276 seconds] | 16:25 | |
| -!- uricamic [9320543b@gateway/web/freenode/ip.147.32.84.59] has quit [Quit: Page closed] | 16:34 | |
| sonne|work | PhilTillet: yeah sure that makes sense | 16:38 |
| sonne|work | PhilTillet: actually chunking methods like svmlight precompute a part of the kernel matrix in each iteration (about 40 kernel rows) | 16:39 |
| PhilTillet | hmm, interesting | 16:39 |
| -!- Marina_ [55b38cc8@gateway/web/freenode/ip.85.179.140.200] has quit [Ping timeout: 245 seconds] | 16:42 | |
| -!- blackburn [5bde8018@gateway/web/freenode/ip.91.222.128.24] has joined #shogun | 16:47 | |
| -!- blackburn [5bde8018@gateway/web/freenode/ip.91.222.128.24] has quit [Quit: Page closed] | 16:56 | |
| gsomix | sonney2k, sonne|work java_modular is ok now. | 17:22 |
| -!- harshit_ [~harshit@182.68.43.52] has joined #shogun | 17:23 | |
| -!- PhilTillet [~Philippe@157.159.42.154] has left #shogun ["Leaving"] | 17:31 | |
| -!- pluskid [~pluskid@173.254.214.60] has quit [Quit: Leaving] | 17:40 | |
| -!- harshit_ [~harshit@182.68.43.52] has quit [Remote host closed the connection] | 17:46 | |
| -!- xiangwang [~xiangwang@210.73.4.120] has joined #shogun | 18:03 | |
| n4nd0 | sonne|work: there is no huge speedup using the decomposition :( | 18:33 |
| n4nd0 | this is the way I am computing it, just in case I got the idea wrong | 18:33 |
| n4nd0 | http://snipt.org/uhigh2#expand | 18:33 |
| -!- emrecelikten [~emre@176.40.238.20] has joined #shogun | 19:08 | |
| -!- xiangwang [~xiangwang@210.73.4.120] has quit [Remote host closed the connection] | 19:14 | |
| @sonney2k | gsomix, ok then I better merge things :) | 19:43 |
| @sonney2k | n4nd0, no wonder | 19:44 |
| @sonney2k | n4nd0, you should precompute x^T x | 19:44 |
| @sonney2k | and store these constants | 19:44 |
| CIA-64 | shogun: Soeren Sonnenburg master * r48b3ffd / (42 files in 7 dirs): | 19:46 |
| CIA-64 | shogun: Merge pull request #458 from pluskid/multiclass-refactor2 | 19:46 |
| CIA-64 | shogun: Refactoring of multiclass (+11 more commits...) - http://git.io/Zjyx0Q | 19:46 |
| CIA-64 | shogun: Soeren Sonnenburg master * rd6cc3eb / (7 files in 7 dirs): | 19:47 |
| CIA-64 | shogun: Merge pull request #474 from gsomix/big_deal_with_sgvector | 19:47 |
| CIA-64 | shogun: add SGVector& typemaps for 'smooth' transition (+7 more commits...) - http://git.io/dbyMcQ | 19:47 |
| n4nd0 | sonney2k: to compute x^T x the one I am doing it right now with cblas_ddot(xlen, x, 1, x, 1) is fine or not? | 19:50 |
| @sonney2k | n4nd0, just use CMath::dot() (which effectively does this) | 19:50 |
| n4nd0 | the way instead of the one ^ | 19:50 |
| n4nd0 | all right | 19:51 |
| n4nd0 | do you expect a major improvement storing it? | 19:51 |
| @sonney2k | n4nd0, of course | 19:51 |
| @sonney2k | factor 3 faster then | 19:51 |
| n4nd0 | wait a moment | 19:52 |
| @sonney2k | n4nd0, btw I think you should compare this to JL's covertree distance | 19:52 |
| shogun-buildbot | build #754 of libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/libshogun/builds/754 blamelist: pluskid@gmail.com | 19:52 |
| @sonney2k | ours should be faster then | 19:52 |
| n4nd0 | with precomputing x^T x | 19:53 |
| @sonney2k | yes | 19:53 |
| n4nd0 | this is for the whole training data right? | 19:53 |
| @sonney2k | yes, going once over all data | 19:53 |
| n4nd0 | I mean for each training vector x store x^T x | 19:54 |
| @sonney2k | yes | 19:54 |
| @sonney2k | have a look at CGaussianKernel | 19:54 |
| @sonney2k | it does the same | 19:54 |
| n4nd0 | ok, I will check it | 19:54 |
| @sonney2k | I guess you can even just copy the code | 19:54 |
| n4nd0 | but I thought in any case that in shogun we normally assume that these type of matrices do not fit into memory | 19:54 |
| @sonney2k | dammed buildbot :D | 19:54 |
| @sonney2k | n4nd0, most algorithms assume that data fits in memory | 19:55 |
| @sonney2k | but storing a scalar per vector is not really a lot even if not | 19:55 |
| n4nd0 | in the same way it is done for the training data | 19:57 |
| @sonney2k | yes | 19:57 |
| n4nd0 | it is good to do it for the query points | 19:57 |
| @sonney2k | you have to do it for lhs and rhs | 19:58 |
| n4nd0 | got it | 19:58 |
| n4nd0 | thank you for the pointer to CGaussianKernel :) | 19:58 |
| CIA-64 | shogun: Soeren Sonnenburg master * r732f288 / src/shogun/classifier/svm/SVM.h : add missing friend 'class' Foo to fix compile error - http://git.io/rs1rSg | 19:58 |
| @sonney2k | n4nd0, let me know whether JLs distance is faster or ours now - and if ours is faster our KNN+covertree should be too | 19:59 |
| @sonney2k | :P | 19:59 |
| @sonney2k | shogun-buildbot, be happy! | 20:04 |
| shogun-buildbot | What you say! | 20:04 |
| emrecelikten | someone set up us the bomb | 20:04 |
| @sonney2k | shogun-buildbot, you are the bomb! | 20:05 |
| shogun-buildbot | build #756 of libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/libshogun/builds/756 | 20:05 |
| shogun-buildbot | What you say! | 20:05 |
| @sonney2k | yay! | 20:05 |
| gsomix | sonney2k, ok. I'm waiting for shogun-buildbot judgement. | 20:06 |
| shogun-buildbot | build #686 of r_static is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/r_static/builds/686 blamelist: sonne@debian.org | 20:09 |
| CIA-64 | shogun: Soeren Sonnenburg master * ra6d10e0 / src/shogun/machine/MulticlassMachine.cpp : fix compile warnings - http://git.io/cdw-Rg | 20:09 |
| shogun-buildbot | build #708 of cmdline_static is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cmdline_static/builds/708 blamelist: sonne@debian.org | 20:13 |
| CIA-64 | shogun: Soeren Sonnenburg master * r36b4b6e / examples/undocumented/libshogun/classifier_mklmulticlass.cpp : fix include MultiClass -> Multiclass - http://git.io/E7bhbA | 20:14 |
| @sonney2k | gsomix, it will take some time - have to cleanup pluskids bugs :) | 20:14 |
| @sonney2k | gsomix, I really hope we can make some fast progress on this sgvector business | 20:15 |
| @sonney2k | this and some reshuffling into lib/external / multitask / multiclass are the big changes | 20:15 |
| @sonney2k | all the rest are just incremental feature updates | 20:15 |
| @sonney2k | not so likely to break the whole build | 20:15 |
| CIA-64 | shogun: Soeren Sonnenburg master * r53cc701 / examples/undocumented/libshogun/classifier_libsvmmulticlass.cpp : another MultiClass -> Multiclass include fix - http://git.io/4HG5DA | 20:17 |
| shogun-buildbot | build #687 of octave_static is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/octave_static/builds/687 blamelist: sonne@debian.org | 20:17 |
| gsomix | sonney2k, I also hope so. | 20:18 |
| @sonney2k | shogun-buildbot, stinky-finger! | 20:18 |
| shogun-buildbot | What you say! | 20:18 |
| gsomix | sonney2k, I have some misunderstanding with reference counting. can we discuss these issues a little later? | 20:20 |
| @sonney2k | sure but first the SGVector& business | 20:20 |
| shogun-buildbot | build #674 of python_static is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/python_static/builds/674 blamelist: sonne@debian.org | 20:21 |
| gsomix | sonney2k, ok. I need to go now. cu, later for discuss future plan. | 20:23 |
| @sonney2k | gsomix, cu | 20:23 |
| * sonney2k waits for the buildbot to send out love and happiness again | 20:24 | |
| shogun-buildbot | build #215 of nightly_none is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_none/builds/215 | 21:10 |
| -!- genix [~gsomix@83.234.54.246] has joined #shogun | 21:26 | |
| -!- gsomix [~gsomix@188.168.13.79] has quit [Ping timeout: 276 seconds] | 21:30 | |
| -!- genix is now known as gsomix | 21:38 | |
| gsomix | haha | 21:38 |
| gsomix | >:] | 21:38 |
| gsomix | wiking, around | 21:38 |
| wiking | ye | 21:44 |
| gsomix | wiking, did you wanted to tell me smth? I'm slowpoke - missed your message. | 21:46 |
| wiking | ah yeah | 21:46 |
| wiking | i think maybe blackburn already told you about a little change we'll need to do in Features.h | 21:46 |
| wiking | and in all the inherited classes | 21:46 |
| gsomix | wiking, const methods? | 21:47 |
| wiking | yes | 21:47 |
| gsomix | aha | 21:47 |
| wiking | ok cool so you know about it \o/ | 21:47 |
| -!- n4nd0 [~nando@n129-p107.kthopen.kth.se] has quit [Quit: leaving] | 21:48 | |
| gsomix | :3 | 21:48 |
| * gsomix afk | 21:48 | |
| wiking | thnx in advance for the changes! | 21:49 |
| wiking | now it's time for the night session developing... | 21:53 |
| gsomix | it's time to homework for me. | 22:02 |
| @sonney2k | alright the deduplication meeting is over and everyone of you who was in before is still in :D | 22:33 |
| wiking | hahahaha | 22:34 |
| wiking | now the only question is | 22:34 |
| wiking | who's in | 22:34 |
| wiking | but i guess we'll only know this next week | 22:34 |
| CIA-64 | shogun: Soeren Sonnenburg master * rac7a60e / (2 files): fix examples MultiClass -> Multiclass - http://git.io/eoanrA | 22:35 |
| wiking | nyiha | 22:35 |
| wiking | the we'll only know it next monday between 21:00 and 22:00 CEST | 22:36 |
| CIA-64 | shogun: Soeren Sonnenburg master * rfa27ebf / (16 files in 9 dirs): MultiClass -> Multiclass replacement in all remaining examples - http://git.io/Bzc30Q | 22:37 |
| wiking | sonney2k: so there's gonna be a structural-svm gsoc project? | 22:37 |
| @sonney2k | yeah | 22:37 |
| wiking | ok cool | 22:37 |
| @sonney2k | hmmhh pluskid really did lots of changes causing all MC examples to fail | 22:38 |
| wiking | :( | 22:39 |
| @sonney2k | simple changes are missing but I guess I never said that one has to compile for all interfaces and run make check-examples | 22:39 |
| @sonney2k | anyway it will be over soon | 22:39 |
| @sonney2k | wiking, anyway 8 students is a lot and I hope you will all have fun :) | 22:40 |
| wiking | hahaha so should i understand the implied meaning here ? ;) | 22:40 |
| @sonney2k | I am not implying anything | 22:41 |
| wiking | ah ok | 22:41 |
| wiking | :D | 22:42 |
| emrecelikten | I sadly could not find time for doing the patch requirement for this project | 22:46 |
| emrecelikten | Maybe I will work on something on my own if none of my proposals get accepted in general | 22:47 |
| emrecelikten | I still have that HMM project in my mind :D | 22:47 |
| @sonney2k | emrecelikten, you mean continous output hmms? | 22:49 |
| emrecelikten | Yep, Gaussian mixtures | 22:49 |
| emrecelikten | I am a text-to-speech guy | 22:49 |
| emrecelikten | It's used a lot | 22:49 |
| @sonney2k | yeah | 22:49 |
| @sonney2k | would be a very welcome contribution... maybe you even find C/C++ code for that already | 22:50 |
| @sonney2k | huh? someone has withdrawn his proposal?! | 22:50 |
| emrecelikten | I think there are various tools, but performance is quite important I think | 22:50 |
| emrecelikten | Baum-Welch takes a lot of time | 22:50 |
| @sonney2k | got to leave train | 22:50 |
| @sonney2k | cu | 22:50 |
| emrecelikten | Bye! | 22:50 |
| wiking | sonney2k: mmm wasn't it 9 slots? | 22:51 |
| emrecelikten | What are the projects of you guys? | 22:55 |
| emrecelikten | I know that Fernando applied to structured SVM outputs | 22:55 |
| gsomix | emrecelikten, "Many changes, improvements, bugs, lolz and vodka for Good!" | 22:57 |
| wiking | gsomix: the best project \o/ | 22:58 |
| wiking | emrecelikten: i've applied for latent svm | 22:58 |
| emrecelikten | Vodking all the time then huh? | 22:58 |
| emrecelikten | Too bad that I'm a straight edge :P | 22:59 |
| emrecelikten | wiking: Nice. I applied for structured SVM output and C5.0 | 22:59 |
| wiking | yeah structured svm would be cool to have | 22:59 |
| emrecelikten | But since I haven't sent a patch, I don't think I have a chance | 22:59 |
| wiking | well i've send a lot of patches | 23:00 |
| wiking | but not much related to latent svm yet :( | 23:00 |
| wiking | that is still under progress | 23:00 |
| emrecelikten | I think it's fine, you're supposed to work on it later | 23:00 |
| wiking | i might have something tonight | 23:00 |
| emrecelikten | If you could do something serious by now, there wouldn't be a need for GSoC, would there? | 23:00 |
| emrecelikten | :D | 23:00 |
| wiking | just cooked myself a turkish coffee so i should have the power ;) | 23:01 |
| wiking | but yeah the latent svm would be in big need for having structural svm | 23:01 |
| emrecelikten | Should I send you some? :P | 23:01 |
| wiking | emrecelikten: i have some here with me ;P | 23:01 |
| wiking | not the structural svm | 23:01 |
| wiking | but the turkish coffee ;) | 23:01 |
| emrecelikten | Haha | 23:01 |
| wiking | anyhow | 23:02 |
| wiking | i want to finish this | 23:02 |
| wiking | i'm aaalmost there | 23:02 |
| wiking | the minimization is done | 23:02 |
| wiking | now i just have to figure out the handle the latent variable handling in a nice way | 23:03 |
| wiking | and then i might already be able to run the first dummy test | 23:03 |
| wiking | and actually it'd be great if it really works out | 23:03 |
| wiking | since it's gonna use libocas | 23:03 |
| emrecelikten | Best of luck to you then :) | 23:03 |
| wiking | which suppose to be very efficient | 23:03 |
| wiking | and latent svm will need that | 23:04 |
| wiking | since each iteration is basically consist of a cutting plane finding.... | 23:05 |
| wiking | *search | 23:05 |
| wiking | but i even have to now construct a dummy test for the whole thing | 23:05 |
| wiking | so that i can see that actually it works :) | 23:05 |
| wiking | not just calculating infinity | 23:05 |
| wiking | and till now i haven't managed to find a simple train/test set .... so it seem i'll have to construct it myself | 23:06 |
| emrecelikten | You shouldn't worry about being accepted though, regarding the conversation at top | 23:09 |
| emrecelikten | 8 slots is a lot | 23:09 |
| emrecelikten | Unless someone is ninjaing your project, you should be fine I guess | 23:09 |
| emrecelikten | My "primary" project had 2 slots last year | 23:10 |
| emrecelikten | It's Spartan-like competition to get accepted there | 23:10 |
| emrecelikten | Not saying that it's easy here or something, but there wasn't any "friendly competition", that's what I meant :D | 23:16 |
| @sonney2k | emrecelikten, wiking - the funny thing is that last year no one applied for the structured output task and the latent svm! | 23:27 |
| wiking | :( | 23:27 |
| wiking | too bad | 23:27 |
| emrecelikten | sonney2k: I sort of feel bad about not applying to Shogun last year | 23:36 |
| emrecelikten | I discovered it too late | 23:36 |
| @sonney2k | emrecelikten, well it was much tougher to get in last year | 23:37 |
| @sonney2k | 5 slots for 73 proposals | 23:37 |
| emrecelikten | It's like 8 slots for 50 proposals now? | 23:37 |
| @sonney2k | yeah | 23:37 |
| wiking | mmm i think i have a problem with git svn rebase ;) | 23:38 |
| @sonney2k | but decision is easier this year too - what counts is # and quality of patches | 23:38 |
| emrecelikten | Zero of which are mine :P | 23:38 |
| wiking | oh i was meaning to ask | 23:39 |
| wiking | don't we want to do the documentation in http://docutils.sourceforge.net/rst.html ? | 23:39 |
| wiking | easy to use and nice html can be generated as well | 23:40 |
| @sonney2k | emrecelikten, yeah well ... contributing not just directly before gsoc is also an option... even for next year (if we apply & get in...) | 23:43 |
| @sonney2k | wiking, but it won't work with doxygen | 23:43 |
| @sonney2k | or are you thinking of writing a new separate doc? | 23:44 |
| wiking | well | 23:46 |
| wiking | maybe a bit more of documentation would be good | 23:46 |
| wiking | aaah btw maybe u haven't seen it | 23:46 |
| wiking | about doxygen | 23:46 |
| wiking | could we turn on the flag that it actually adds the inherited functions as well to a class interface? | 23:47 |
| wiking | or u have strong objections about that | 23:47 |
| emrecelikten | sonney2k: I know, I know :) | 23:48 |
| wiking | INLINE_INHERITED_MEMB = YES | 23:51 |
| wiking | i think this is the doxygen option | 23:51 |
| @sonney2k | too bad that pluskid is not here :( | 23:52 |
| @sonney2k | wiking, I know that flag - do you really think that helps? | 23:52 |
| wiking | now that i'm doing latent machine | 23:53 |
| wiking | it's a direct inherited from linear machine | 23:53 |
| wiking | t | 23:53 |
| wiking | h | 23:53 |
| wiking | which is a machine | 23:53 |
| wiking | so i was like ok let's see what are the funcitons | 23:53 |
| wiking | so i was checking two headers | 23:54 |
| wiking | looking for what are the pure virtual functions etc | 23:54 |
| @sonney2k | they apply method of multiclass svm is gone?!? | 23:54 |
| @sonney2k | cannot find it in multiclasskernelmachine either | 23:54 |
| wiking | so it'd be great if for example i could see from the class that i'm inheriting from that ok these are the function i'll need to implement and actually i have by default | 23:55 |
| @sonney2k | wiking, I mean we can try it | 23:55 |
| wiking | so in my case it's helpful | 23:55 |
| wiking | of course now that clang has this amazing error and warning messages | 23:55 |
| wiking | it's actually making a lot easier to see where did i do something wrong with the inheritance and missing some pure virtual function implementation | 23:55 |
| wiking | :P | 23:56 |
| @sonney2k | wiking, send a pull request :) | 23:56 |
| wiking | heheh ok | 23:56 |
| wiking | interfaces/modular/modshogun.doxy is the config file right? | 23:56 |
| wiking | oh no | 23:58 |
| wiking | there's Doxyfile ;) | 23:58 |
| --- Log closed Sat Apr 21 00:00:15 2012 | ||
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!
