Open in new window / Try shogun cloud
--- Log opened Sat Jun 09 00:00:41 2012
-!- zxtx [] has joined #shogun00:09
@sonney2kblackburn, around?00:15
@sonney2kI have a compile error related to %template(IntStdVector) vector<int32_t>;00:16
@sonney2kin java_modular00:16
@sonney2kI think it needs SERIALIZABLE_DUMMY(vector<int32_t>);00:16
blackburnlet me check if we really need that00:16
@sonney2kbut adding that doesn't work00:16
@sonney2kI mean it doesn't change a thing00:16
blackburnI merged chris' code w/o many changes00:16
blackburnsonney2k: we can remove that!00:17
blackburnI'll do00:17
@sonney2kblackburn, you mean we no longer need std::vector in interfaces?00:18
blackburnI thought we need but we don't00:18
@sonney2kdie die std::vector :P00:18
@sonney2klet me see if that fixes the problem here...00:18
blackburnI realize I need to move recently added liblinear to transfer00:19
@sonney2kyou added chris' MTL one?00:19
blackburnhowever patches have gone flatten :(00:20
@sonney2ki guess it is currently in cliblinear?00:20
blackburnno different class00:20
blackburnyes he send me patches to merge00:20
@sonney2kahh ok00:20
blackburnI'll undo not required changes now00:21
@sonney2kI have no idea what you are talking about00:22
blackburnstl added stuff00:22
@sonney2kblackburn, btw can you please merge gsomix' stuff00:23
blackburnargh sure - I thought it will be a new PR and was waiting for mail00:23
@sonney2kgsomix never mentions that a patch is final or ready ... everything is always WIP00:24
@sonney2kblackburn, today I enjoyed swimming in the sea00:28
@sonney2k12-13 C though00:29
blackburndid you visit england?00:29
blackburnyou tend to swim far00:29
blackburn12-13C? argh not any warm00:29
@sonney2kit is so damn cold that I could not swim with face in water00:30
@sonney2kit did hurt like needles put in your head00:30
blackburnwhy didn't you visit some +20C place?00:30
emreceliktene.g. Turkey00:30
@sonney2knot really a car ride away :)00:32
@sonney2k2-3hr car ride I mean :)00:32
@sonney2kalright I better sleep now - long journey back tomorrow!00:32
CIA-9shogun: Soeren Sonnenburg master * rc29a9b2 / (3 files in 2 dirs): fix compile error in java modular -
blackburndid you stay your car in sea?00:32
blackburnthanks! I did the same change :D00:33
CIA-9shogun: Sergey Lisitsyn master * ra85ee05 / (9 files in 4 dirs): Moved liblinear MTL to transfer folder and undo unnecessary stl includes -
blackburnhmm that went ok00:34
CIA-9shogun: Sergey Lisitsyn master * rc15f5a5 / (10 files in 5 dirs): Merge branch 'director_classes' of git:// (+6 more commits...) -
CIA-9shogun: Soeren Sonnenburg master * redcd31c / src/interfaces/modular/modshogun_ignores.i : fix compile error in csharp modular -
-!- blackburn [~blackburn@] has quit [Quit: Leaving.]01:04
-!- zxtx [] has quit [Ping timeout: 265 seconds]01:24
-!- zxtx [] has joined #shogun03:11
-!- wiking [] has joined #shogun04:05
-!- wiking [] has quit [Changing host]04:05
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun04:05
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking]04:12
-!- emrecelikten [~emrecelik@] has quit [Quit: Leaving.]04:41
-!- pluskid [] has joined #shogun07:30
-!- gsomix [~gsomix@] has joined #shogun07:54
-!- wiking [] has joined #shogun08:16
-!- wiking [] has quit [Changing host]08:16
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun08:16
-!- wiking [~wiking@huwico/staff/wiking] has quit [Client Quit]08:17
-!- wiking [] has joined #shogun08:33
-!- wiking [] has quit [Changing host]08:33
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun08:33
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking]08:56
-!- gsomix_ [~gsomix@] has joined #shogun08:57
-!- gsomix [~gsomix@] has quit [Read error: Connection reset by peer]08:57
-!- wiking [] has joined #shogun09:01
-!- wiking [] has quit [Changing host]09:01
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun09:01
-!- n4nd0 [] has joined #shogun10:50
-!- pluskid [] has quit [Ping timeout: 252 seconds]11:32
-!- blackburn [~blackburn@] has joined #shogun11:41
-!- pluskid [~pluskid@] has joined #shogun11:50
-!- emrecelikten [~emrecelik@] has joined #shogun12:22
-!- wiking_ [~wiking@] has joined #shogun12:22
-!- wiking_ [~wiking@] has quit [Changing host]12:22
-!- wiking_ [~wiking@huwico/staff/wiking] has joined #shogun12:22
-!- wiking [~wiking@huwico/staff/wiking] has quit [Ping timeout: 260 seconds]12:24
-!- wiking_ is now known as wiking12:25
-!- n4nd0 [] has quit [Quit: leaving]12:50
-!- ckwidmer [] has joined #shogun12:53
-!- emrecelikten1 [~emrecelik@] has joined #shogun12:57
-!- emrecelikten [~emrecelik@] has quit [Quit: Leaving.]12:57
-!- emrecelikten1 is now known as emrecelikten12:58
blackburnpluskid: you have something wrond with your productivity - it is too high :D13:52
pluskidblackburn: I have long time no commits13:52
pluskidso those are accumulated...13:52
blackburnheh yeah13:52
blackburnlet me try to review that13:53
pluskidactually, I'm having some trouble with the shogun Vw, so have to use LibLinear13:53
pluskidmaybe I could try to figure out the problem (if any) with shogun Vw some time later13:54
blackburnyou changed much things in liblinear13:54
blackburnwhy so?13:54
pluskidblackburn: just refactoring13:54
pluskidblackburn: to allow training example by example13:54
-!- pluskid [~pluskid@] has quit [Quit: Leaving]13:57
-!- wiking [~wiking@huwico/staff/wiking] has quit [Ping timeout: 245 seconds]14:10
-!- wiking [~wiking@] has joined #shogun14:11
-!- wiking [~wiking@] has quit [Changing host]14:11
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun14:11
-!- gsomix_ [~gsomix@] has quit [Ping timeout: 244 seconds]14:29
-!- gsomix_ [~gsomix@] has joined #shogun14:30
-!- emrecelikten [~emrecelik@] has quit [Quit: Leaving.]14:40
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking]14:46
-!- n4nd0 [] has joined #shogun14:58
gsomix_Nietzsche, Nietzsche .____.15:06
CIA-9shogun: Sergey Lisitsyn master * rc660f04 / data : Updated data -
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun15:14
CIA-9shogun: Sergey Lisitsyn master * r765d0dd / (28 files in 6 dirs): Merge branch 'multiclass' of git:// (+29 more commits...) -
-!- pluskid [~pluskid@] has joined #shogun15:23
-!- wiking_ [] has joined #shogun15:30
-!- wiking_ [] has quit [Changing host]15:30
-!- wiking_ [~wiking@huwico/staff/wiking] has joined #shogun15:30
-!- wiking [~wiking@huwico/staff/wiking] has quit [Ping timeout: 244 seconds]15:30
-!- wiking_ is now known as wiking15:30
-!- n4nd0 [] has quit [Ping timeout: 252 seconds]15:34
-!- n4nd0 [] has joined #shogun15:35
CIA-9shogun: Sergey Lisitsyn master * r7b35a82 / (2 files): Some fixes for Cplex -
-!- pluskid [~pluskid@] has quit [Ping timeout: 240 seconds]16:21
-!- pluskid [~pluskid@VPN-18-101-16-183.MIT.EDU] has joined #shogun16:21
CIA-9shogun: Sergey Lisitsyn master * r06aee69 / src/shogun/mathematics/Cplex.cpp : Added missed close paren to Cplex -
-!- n4nd0 [] has quit [Ping timeout: 244 seconds]16:50
-!- blackburn [~blackburn@] has quit [Ping timeout: 244 seconds]17:28
-!- pluskid [~pluskid@VPN-18-101-16-183.MIT.EDU] has quit [Quit: Leaving]17:40
-!- n4nd0 [] has joined #shogun18:06
n4nd0sonney2k: around?19:04
@sonney2kn4nd0, yes19:09
@sonney2kwhats up?19:09
n4nd0I have been stuck here with a problem the whole day and I don't manage to solve it :S19:09
n4nd0I cannot compile any of the examples under libshogun19:10
@sonney2kwhich is?19:10
n4nd0is a problem of mine, locally19:10
n4nd0when I try to compile anything in libshogun I get19:10
n4nd0/usr/local/etc/mosek/6/tools/platform/linux32x86/bin/ undefined reference to `pthread_atfork'19:10
n4nd0collect2: ld returned 1 exit status19:10
@sonney2kmaybe not19:10
n4nd0make: *** [basic_minimal] Error 119:10
n4nd0I have tried with *a lot* of different things but I have not succeeded19:11
@sonney2kn4nd0, does mosek come with its own set of standard system libraries?19:11
n4nd0this is the way one compiles programs with mosek19:11
n4nd0I hope it answers your question19:11
n4nd0gcc examp/lo1.c -o lo1 -I h/ -L platform/linux32x86/bin/ -lmosek -pthread -lc -ldl -lm19:11
n4nd0both h/ and platform/linux32x86/bin/ are mosek directories19:12
@sonney2kwhat I mean is there might be system libraries in the mosek dirs conflicting with the ones installed to your system19:14
n4nd0it can be a possibility19:15
n4nd0but the one that gives the problem, is not in my system I think19:15
n4nd0it's just from mosek19:15
@sonney2kcertainly it is not limiomp's fault19:18
@sonney2kbut maybe your have a different libpthread in mosek dirs somewhere?19:18
n4nd0in my mosek directory19:19
n4nd0find | grep thread returns nothing19:20
@sonney2kn4nd0, other idea is to use -lpthread19:21
@sonney2kat the end19:21
n4nd0sonney2k: yeah19:21
n4nd0that's what I have been trying during the day19:21
n4nd0lpthread compiling with shogun19:21
n4nd0compiling the examples19:21
n4nd0at the end you mean like the last thing in the line I showed you?19:22
n4nd0c++ -O0 -g -I/usr/include -I/usr/include/atlas basic_minimal.cpp -lshogun -L/usr/lib  -lpthread19:22
n4nd0fails in the same way19:23
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking]19:24
@sonney2kn4nd0, and if you put the -L platform/linux32x86/bin/ -lmosek to the end?19:24
n4nd0c++ -O0 -g -I/usr/include -I/usr/include/atlas basic_minimal.cpp -lshogun -L/usr/lib -lpthread -L/usr/local/etc/mosek/6/tools/platform/linux32x86/bin/ -lmosek19:26
n4nd0fails too19:26
-!- wiking [] has joined #shogun19:26
-!- wiking [] has quit [Changing host]19:26
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun19:26
@sonney2kn4nd0, in the same way or what?19:26
n4nd0note that basic_minimal.cpp doesn't use anything from mosek at all19:27
n4nd0sonney2k: yeah, same error19:27
@sonney2kn4nd0, I am lost - you attempt to compile the shogun example right?19:27
@sonney2kso w/o the mosek stuff it compiles right?19:28
n4nd0the error happens with all the examples19:28
n4nd0the mosek stuff I am using is to build shogun19:28
n4nd0I use a USE_MOSEK and include the libraries it needs19:28
@sonney2kn4nd0, no I mean if you don't use mosek link flags19:28
@sonney2kcompile flags19:28
n4nd0I don't know right now since I have not tried19:29
n4nd0but probably the error would disappear19:29
@sonney2kthen please try if  c++ -O0 -g -I/usr/include -I/usr/include/atlas basic_minimal.cpp -lshogun -L/usr/lib -lpthread works19:29
n4nd0since it comes from a file in mosek19:29
n4nd0same error with this last one19:30
n4nd0/usr/local/etc/mosek/6/tools/platform/linux32x86/bin/ undefined reference to `pthread_atfork'19:30
n4nd0collect2: ld returned 1 exit status19:30
@sonney2kwith which command?19:31
n4nd0with the last one you told me19:31
n4nd0I paste again19:31
n4nd0c++ -O0 -g -I/usr/include -I/usr/include/atlas basic_minimal.cpp -lshogun -L/usr/lib -lpthread19:31
n4nd0/usr/local/etc/mosek/6/tools/platform/linux32x86/bin/ undefined reference to `pthread_atfork'19:31
n4nd0collect2: ld returned 1 exit status19:31
@sonney2kn4nd0, what is the output of which c++ ?19:32
@sonney2kand c++ -v19:32
n4nd0which c++19:32
n4nd0c++ -v19:32
n4nd0Using built-in specs.19:32
n4nd0Target: i686-linux-gnu19:33
n4nd0Configured with: ../src/configure -v --with-pkgversion='Ubuntu/Linaro 4.6.1-9ubuntu3' --with-bugurl=file:///usr/share/doc/gcc-4.6/README.Bugs --enable-languages=c,c++,fortran,objc,obj-c++,go --prefix=/usr --program-suffix=-4.6 --enable-shared --enable-linker-build-id --with-system-zlib --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --with-gxx-include-dir=/usr/include/c++/4.6 --libdir=/usr/lib --enable-nls --with-sysroot=/19:33
n4nd0Thread model: posix19:33
n4nd0gcc version 4.6.1 (Ubuntu/Linaro 4.6.1-9ubuntu3)19:33
@sonney2kn4nd0, so you did link libshogun with libiomp5 right?19:34
@sonney2kn4nd0, what is the output of ldd path/to/libshogun.so19:34
n4nd0I did not include any -liomp5 or the like to compile shogun because it doesn't appear in mosek's compilation instructions19:35
n4nd0ldd =>  (0xb76ff000) => /usr/local/etc/mosek/6/tools/platform/linux32x86/bin/ (0xb6980000) => /lib/i386-linux-gnu/ (0xb6942000) => /usr/lib/ (0xb65aa000) => /usr/lib/ (0xb645d000) => /usr/lib/ (0xb5b85000)19:35
n4nd0 => /usr/li/b/ (0xb5b3b000) => /usr/lib/ (0xb5aea000) => /lib/i386-linux-gnu/ (0xb5ad4000) => /lib/ (0xb5ac3000) => /usr/lib/i386-linux-gnu/ (0xb59d8000) => /lib/i386-linux-gnu/ (0xb59ae000) => /lib/i386-linux-gnu/ (0xb5830000)19:35
n4nd0 libgcc_sso.1 => /lib/i386-linux-g/nu/ (0xb5812000) => /usr/local/etc/mosek/6/tools/platform/linux32x86/bin/ (0xb5793000) => /lib/i386-linux-gnu/ (0xb578e000) /lib/ (0xb7700000) => /usr/lib/ (0xb54b9000) => /usr/lib/i386-linux-gnu/ (0xb53b7000)19:35
n4nd0> /usr/lib/i386-linux-gnu/ (0xbb5342000)19:35
n4nd0wow sorry that's quite unreadable19:36
@sonney2kn4nd0, please paste as raw somewhere..19:36
@sonney2kn4nd0, what does grep pthread_atfork lib/i386-linux-gnu/ say?19:38
@sonney2kn4nd0, and next - what does ldd /usr/local/etc/mosek/6/tools/platform/linux32x86/bin/ say?19:39
n4nd0grep pthread_atfork /lib/i386-linux-gnu/
n4nd0Binary file /lib/i386-linux-gnu/ matches19:40
n4nd0and the second one19:40
n4nd0ups, it looks normal to me19:42
@sonney2kvery weird19:42
@sonney2kn4nd0, you are on a 32bit system right?19:42
@sonney2knot that we have some weird 64bit / 32bit mixup19:43
n4nd0yes, I am in 32 bits19:43
n4nd0 uname -a19:43
n4nd0Linux nando-Aspire-5732Z 3.0.0-20-generic-pae #34-Ubuntu SMP Tue May 1 18:35:40 UTC 2012 i686 i686 i386 GNU/Linux19:43
n4nd0uname -m would have been more appropriate :D19:44
n4nd0sonney2k: any other idea?19:55
@sonney2kn4nd0, can you compile mosek examples?19:57
n4nd0sonney2k: yeah, no problem with those19:57
@sonney2kn4nd0, what is the exact cmdline to compile them?19:58
n4nd0gcc lo1.c -o lo1 -I ../../platform/linux32x86/h/ -L ../../platform/linux32x86/bin/ -lmosek -pthread -lc -ldl -lm20:00
n4nd0platform/ is a directory inside my mosek instalation20:00
n4nd0I got the error if I do it with g++!20:01
n4nd0g++ lo1.c -o lo1 -I ../../platform/linux32x86/h/ -L ../../platform/linux32x86/bin/ -lmosek -pthread -lc -ldl -lm20:01
n4nd0/usr/local/etc/mosek/6/tools/platform/linux32x86/bin/ undefined reference to `pthread_atfork'20:01
n4nd0collect2: ld returned 1 exit status20:01
-!- wiking_ [~wiking@] has joined #shogun20:01
-!- wiking_ [~wiking@] has quit [Changing host]20:01
-!- wiking_ [~wiking@huwico/staff/wiking] has joined #shogun20:01
@sonney2kn4nd0, what if you drop -lc -ldl ?20:02
n4nd0with gcc compiles20:02
n4nd0with g++ I get the error20:02
n4nd0g++ lo1.c -o lo1 -I ../../platform/linux32x86/h/ -L ../../platform/linux32x86/bin/ -lmosek -pthread -lm20:02
n4nd0/usr/local/etc/mosek/6/tools/platform/linux32x86/bin/ undefined reference to `pthread_atfork'20:02
n4nd0collect2: ld returned 1 exit status20:02
-!- wiking [~wiking@huwico/staff/wiking] has quit [Ping timeout: 245 seconds]20:02
-!- wiking_ is now known as wiking20:02
@sonney2kn4nd0, please use in lol.c when mosek headers are included:20:03
@sonney2kextern "C" { }20:03
@sonney2karound these includes20:04
n4nd0lol.c is just a mosek example not the code I have written using mosek20:04
-!- gsomix_ is now known as gsomix20:05
@sonney2kn4nd0, do it nevertheless20:05
@sonney2kand try to compile it20:05
n4nd0so there is just one #include "mosek.h"20:05
n4nd0I am on it20:05
-!- blackburn [~blackburn@] has joined #shogun20:05
n4nd0with g++ the same error20:06
n4nd0and with gcc now I get a bunch of new errors20:06
@sonney2kn4nd0, can you paste the modified example somewhere?20:09
n4nd0extern "C" {20:09
n4nd0#include "mosek.h"20:09
@sonney2kn4nd0, no other includes?20:09
n4nd0there is a #include <stdio.h>20:09
n4nd0I didn't surround that one20:09
@sonney2kthat's it?20:10
n4nd0I guess that the file mosek.h contains a bunch of other includes20:10
@sonney2kn4nd0, and if you compile with g++ lo1.c -o lo1 -I ../../platform/linux32x86/h/ -L ../../platform/linux32x86/bin/ -lmosek -lm20:11
n4nd0still with the extern there right?20:12
n4nd0I have tried it both ways, with and without extern20:12
n4nd0still /usr/local/etc/mosek/6/tools/platform/linux32x86/bin/ undefined reference to `pthread_atfork'20:13
n4nd0collect2: ld returned 1 exit status20:13
@sonney2kn4nd0, just to make sure that this is not a c++ problem on your machine - can you install clang?20:14
@sonney2kn4nd0, and then use clang++ to compile the same thing20:14
n4nd0apt-get install clang20:14
n4nd0same flags?20:14
n4nd0clang++ lo1.c -o lo1 -I ../../platform/linux32x86/h/ -L ../../platform/linux32x86/bin/ -lmosek -lm20:15
n4nd0clang: warning: treating 'c' input as 'c++' when in C++ mode, this behavior is deprecated20:16
n4nd0/usr/local/etc/mosek/6/tools/platform/linux32x86/bin/ undefined reference to `pthread_atfork'20:16
n4nd0clang: error: linker command failed with exit code 1 (use -v to see invocation)20:16
n4nd0wow! it compiled using lpthread at the end20:16
n4nd0clang++ lo1.c -o lo1 -I ../../platform/linux32x86/h/ -L ../../platform/linux32x86/bin/ -lmosek -lm -lpthread20:16
n4nd0clang: warning: treating 'c' input as 'c++' when in C++ mode, this behavior is deprecated20:16
n4nd0this last succeeded20:16
@sonney2kn4nd0, and if you do this with g++ instead?20:16
n4nd0and the executable runs fine20:16
n4nd0with g++ it fails20:17
n4nd0what does this mean?20:17
@sonney2kn4nd0, can you install a different version of g++20:17
n4nd0g++ --version20:18
n4nd0g++ (Ubuntu/Linaro 4.6.1-9ubuntu3) 4.6.120:18
n4nd0sure, how should I do it?20:18
@sonney2kn4nd0, apt-cache search 'g++' | grep ^g++20:20
n4nd0should I move to an older version?20:21
@sonney2kn4nd0, which ones are available?20:21
n4nd0and what I have is20:22
n4nd0dpkg -l | grep g++20:22
n4nd0ii  g++                                            4:4.6.1-2ubuntu5                           GNU C++ compiler20:22
n4nd0ii  g++-4.5                                        4.5.3-9ubuntu1                             The GNU C++ compiler20:22
n4nd0ii  g++-4.6                                        4.6.1-9ubuntu3                             GNU C++ compiler20:22
@sonney2kwell then install g++-4.520:23
@sonney2kahh you have it alrady20:23
@sonney2kthen g++-4.5 <cmdline above>20:23
n4nd0is it going to be any problem wiht the version I have in my system right now?20:24
n4nd0I mean, is the installation going to overwrite the current?20:24
@sonney2kyou have it installed already20:24
@sonney2kso just run the cmd20:24
n4nd0ah all right20:25
n4nd0g++-4.5 lo1.c -o lo1 -I ../../platform/linux32x86/h/ -L ../../platform/linux32x86/bin/ -lmosek -lm -lpthread20:25
n4nd0/usr/local/etc/mosek/6/tools/platform/linux32x86/bin/ undefined reference to `pthread_atfork'20:25
n4nd0collect2: ld returned 1 exit status20:25
n4nd0and with another version, g++-4.6, it happens the same20:25
@sonney2kn4nd0, what happens if you rename lol.c -> lol.cpp20:26
@sonney2kany change?20:26
n4nd0same error, using all the different versions, g++, g++-4.5 and g++-4.620:27
@sonney2kwell no idea then - continue with clang++20:28
@sonney2kshogun works with that too so no problem20:28
n4nd0but do you think this could be a problem later?20:29
@sonney2kn4nd0, do we have a choice?20:30
n4nd0I don't know if it pays off trying to solve the problem with g++20:32
n4nd0sonney2k: in any case, thank you very much for your help!20:42
n4nd0hopefully, I will be able to continue with the multiclass example for so using clang++ :)20:42
blackburnwhat is problemyou have?20:45
n4nd0blackburn: sorry man, I am having dinner now20:48
n4nd0blackburn: I have an undefined reference that shows up when compiling my examples in libshogun when I configure shogun to use with mosek20:49
n4nd0I will be back later20:49
gsomixsonney2k, moin. how are you?20:49
@sonney2kgsomix, back home :)21:02
@sonney2kgsomix, so what was the problem our first directorkernel user was having?21:02
gsomixsonney2k, python don't know about polymorphism21:03
gsomixget_lhs returns CFeatures.21:03
gsomixand we haven't methods to covert CFeatures to CDenseFeatures for example21:04
@sonney2kgsomix, that's all?21:04
gsomixthat's all21:04
-!- emrecelikten [~Anubis@] has joined #shogun21:04
@sonney2kI mean the simple fix is to add some conversion methods that do dyn_casts21:05
gsomixI forgot about same virtual method21:05
@sonney2kgsomix, did you fix the virtual method stuff?21:05
gsomixsonney2k, yep21:05
gsomixsonney2k, I think conversion by swig side is better21:06
@sonney2kthe other idea yes21:07
@sonney2kin the same way I did the hack / fix for apply()21:07
@sonney2kgsomix, so with %ignore / an casts21:08
@sonney2kgsomix, mind doing that?21:08
@sonney2kgsomix, btw how far are you with the other director classes?21:09
gsomixsonney2k, mmm, director linear machines are not ready yet.21:09
@sonney2kbecause once these are done I would suggest continuing with dotfeatures21:10
@sonney2kgsomix, whats the problem with linear machines?21:11
blackburnbloody stalin^W soeren21:11
gsomixsonney2k, no problems.21:11
@sonney2kblackburn, well look at how much pluskid does :D21:12
blackburnmuch more than we all in total21:13
@sonney2kblackburn, he is working full time :)21:15
blackburnstop these insinuations :D21:16
@sonney2kblackburn, just stop sleeping and you will be like pluskid :D21:16
blackburnI do not sleep I slep21:18
@sonney2kblackburn, slep faster :)21:19
blackburnone more word and I'll start commiting some useless code to look productive :D21:20
blackburnsonney2k: btw it took a while to detect an error with mkl and linadd21:24
blackburnhmm sonney2k did you ever try to pull from branch without pull request?21:29
CIA-9shogun: Sergey Lisitsyn master * rf209569 / (21 files in 7 dirs): Added SLEP machines -
CIA-9shogun: Sergey Lisitsyn master * r92cd89a / (5 files in 3 dirs): Updated IndicesTree -
CIA-9shogun: Sergey Lisitsyn master * rcc95465 / (3 files in 3 dirs): Added num of nodes counting in indices tree -
CIA-9shogun: Sergey Lisitsyn master * rbac8cc0 / (21 files in 7 dirs): Merge branch 'slep' of git:// -
blackburnsonney2k: enough?21:32
CIA-9shogun: Soeren Sonnenburg master * r062b1f4 / (28 files): convert examples to use new CMath/SGVector/SGMatrix split -
@sonney2kblackburn, what error with mkl/linadd?21:37
@sonney2k(i am the author of both...)21:37
blackburnsonney2k: didn't you read mailing list?21:38
blackburnmkl regression with linadd doesn't work21:38
@sonney2kI only read things very superficially21:40
blackburnsonney2k: I can point to 'wrong' method21:40
@sonney2kdid you fix things?21:41
blackburnI did not managed to21:42
blackburnsuggested Kyle to turn off linadd21:42
@sonney2kblackburn, basically the problem is that svr is translated to svc21:43
@sonney2kwith twice the number of examples21:43
blackburnyes yes and I reduced it by 221:43
blackburnthat helped to avoid segfaults21:43
blackburnbut not to fix things21:43
@sonney2kwhere the 'copy' is just the same examples with label sign flipped to keep things in the eps-tube21:44
@sonney2kshogun's kernel function has some regression_hack flag iirc21:44
@sonney2kblackburn, what do you mean by  I reduced it by 221:45
blackburnsonney2k: a sec21:45
blackburnsonney2k: shogun/regression/svr/SVRLight.cpp:57621:46
@sonney2kblackburn, actually the right fix would be to do a index modulo num_real_examples21:48
blackburncan you do that?21:49
@sonney2kactually the funxtion to call is regression_fix_index(i)21:52
@sonney2kto everything where the error occurs21:53
blackburnbad criterion21:54
@sonney2kblackburn, I don't understand the num=totdoc/2;21:55
@sonney2kisn't num the nubmer of kernels?21:55
blackburnsonney2k: was number of examples twice21:55
@sonney2kanyway in which line is the error?21:55
@sonney2kI mean valgrind error/segfault21:55
blackburnsonney2k: I can tell you how to reproduce21:58
blackburn580 remove /221:58
blackburnand run in python modular21:58
blackburnshould be a nuclear blast21:58
@sonney2kblackburn, did we have this example before?21:59
blackburnno we don't, yes I have added it for that21:59
blackburnintention is to remove that example once the problem is fixed22:00
@sonney2kbugs in the crappiest code ever are sooo hard to fix22:02
blackburnsonney2k: it was pretty hard to detect it22:03
blackburnsonney2k: have cited you 3 times :D22:11
gsomixblackburn, where is my topology? :)22:15
blackburnwait a min22:15
blackburngsomix: page 1322:19
CIA-9shogun: Soeren Sonnenburg master * r1f37d99 / src/shogun/regression/svr/SVRLight.cpp : another attempt to fix linadd + mkl + svr -
gsomixblackburn, :}22:21
blackburngermany scored to 1-022:22
@sonney2kblackburn, can you test if the 'fix' works22:23
blackburnyou can22:23
@sonney2kI am having problems to test anything currently22:23
@sonney2ksomehow atlas is b0rken here so I have 100% segfault guarantee22:24
n4nd0blackburn: congrats for the match yesterday btw!22:27
blackburnhah thanks that was pretty cool22:27
n4nd0blackburn: I didn't watch the match actually22:27
n4nd0I am half-watching Germany now22:27
n4nd0sonney2k: do you like football?22:27
blackburnyeah same shit22:27
n4nd0gsomix: congrats for you too :D22:28
n4nd0blackburn: tomorrow is the day for Spain, we'll see what they do22:28
blackburnvs ?22:28
n4nd0I support this year Sweden though :P22:28
gsomixn4nd0, wut? I don't like a football. :]22:29
gsomixbut tnx22:29
n4nd0yeah ...22:29
n4nd0you know he played in Barcelona, I don't like him that much haha22:30
blackburnthat was badd22:32
n4nd0I have done my first test for the so framework22:40
n4nd0MOSEK Error 1011     : The problem has 10000 variable(s) but the license allows only 300 variable(s) for feature 'PTS'.22:40
@sonney2kblackburn, did it work?22:44
blackburnyes I think so22:45
@sonney2kyay :)22:47
@sonney2khurray atlas compiled22:48
@sonney2know I can finally run things again22:48
@sonney2kn4nd0, welcome to the world of proprietary software22:50
@sonney2kblackburn, btw do you have more time now?22:50
blackburnmore than?22:50
@sonney2kI mean your thesis is done - thanks for the citations btw22:50
@sonney2k-inf :D22:50
blackburngetting into faster pace22:50
@sonney2kit would be nice to have gmm back in shogun ...22:50
@sonney2kblackburn, well ask alesis :D22:51
@sonney2kmaybe he cares22:51
@sonney2khmmhh all multiclass stuff is broken in current git22:52
blackburnagain? argh22:52
blackburnI'll fix22:52
blackburnsonney2k: pluskid's commit ;)22:53
blackburnsonney2k: MulticlassMachine -> DecompositionMulticlassMachine?22:55
blackburnah nevermind too complex22:56
CIA-9shogun: Soeren Sonnenburg master * r89d623e / examples/undocumented/python_modular/ : fix dynprog example -
@sonney2kblackburn, what did he do?23:00
blackburnsonney2k: introduced new class23:00
blackburnbetween multiclassmachine and machine23:00
@sonney2kand we have a basemulticlassmachien now...23:00
gsomixgood night guys23:01
@sonney2kwhat is this for?23:02
blackburnsonney2k: just to store machines23:03
blackburnthings we have in multiclassmachine are too specialized to decomposition23:03
blackburntree stuff works in a slightly different way23:04
@sonney2knite gsomix23:04
@sonney2kn4nd0, which license do you use btw23:05
@sonney2kn4nd0, for mosek I mean23:05
-!- gsomix [~gsomix@] has quit [Quit: Ex-Chat]23:05
CIA-9shogun: Soeren Sonnenburg master * r068eef5 / (2 files): add missing include to fix multiclass machine swig interfaces -
blackburnsonney2k: do not do fixes I am working on :)23:07
@sonney2kblackburn, btw have you seen the CMath split up?23:10
@sonney2kcan you live with it?23:10
@sonney2kI know that you didn't want vector functions to be in SGVector23:10
blackburnhah yeah I can but I do not really like having operations in SGVector23:10
blackburnespecially that you have to23:10
CIA-9shogun: Soeren Sonnenburg master * rb979aae / (3 files in 2 dirs): fix fisher2x3 example and include statistics in maths module -
blackburnI'd like LinearAlgebra or so more23:12
@sonney2kblackburn, well - either we have templated fucntions or templated classes23:12
blackburnmy main concern23:13
@sonney2kI am much more concerned about a proper naming scheme23:13
blackburnSGVector functions work with float64_t* and etc23:13
blackburnnot with vectors23:13
blackburnnaming scheme of?23:13
blackburnsonney2k: ^23:16
@sonney2k<blackburn> naming scheme of?23:19
@sonney2k<sonney2k> for example when we have23:19
@sonney2k<sonney2k> x.add(y)23:19
@sonney2k<sonney2k> (x and y being SGVectors)23:19
@sonney2k<sonney2k> for x + y23:19
@sonney2k<sonney2k> and add(doulbe* x, int x_len, double* y, int y_len)23:19
@sonney2k<sonney2k> should one be called just 'add'23:19
@sonney2k<sonney2k> and the other add_vector or add_<what???>23:19
@sonney2k(got disconnected...)23:19
blackburnbut first is23:19
blackburnand second is23:19
blackburndifferent methods23:19
@sonney2ksure one is static one is not23:20
@sonney2kstill I want to avoid the confusion by giving it a proper name23:20
blackburnI have no opinion on that actually23:20
@sonney2kthen I will wait for heiko :)23:21
n4nd0sonney2k: I have a license that wiking gave me, I don't know of what type it is23:29
@sonney2klet me ask some contact in mosek if we can have a 'deal'23:32
n4nd0sonney2k: thank you :)23:36
blackburnI do not crack matlab23:58
--- Log closed Sun Jun 10 00:00:41 2012