Open in new window / Try shogun cloud
--- Log opened Mon Apr 29 00:00:29 2013
-!- gsomix [~Miranda@] has quit [Quit: Miranda IM! Smaller, Faster, Easier.]00:39
-!- iglesiasg [] has quit [Quit: leaving]01:23
-!- iglesiasg [] has joined #shogun01:30
-!- mode/#shogun [+o iglesiasg] by ChanServ01:30
-!- iglesiasg [] has quit [Quit: leaving]02:03
shogun-buildbotbuild #374 of nightly_default is complete: Success [build successful]  Build details are at
-!- kaushalb09 [] has quit [Ping timeout: 245 seconds]08:06
-!- iglesiasg [] has joined #shogun08:20
-!- mode/#shogun [+o iglesiasg] by ChanServ08:20
-!- sijin [~smuxi@] has quit []08:51
-!- sonne|work [] has joined #shogun08:51
-!- vgorbati [c3ee5cb1@gateway/web/freenode/ip.] has joined #shogun09:08
-!- lisitsyn [] has quit [Quit: Leaving.]09:09
-!- sids_aquarius [] has joined #shogun10:25
-!- sids_aquarius [] has quit [Ping timeout: 256 seconds]10:30
-!- foulwall [] has joined #shogun11:06
-!- mikhailBelous [~towelenee@] has joined #shogun11:14
-!- van51 [~van51@] has joined #shogun11:20
-!- iglesiasg [] has quit [Quit: leaving]11:33
-!- deerishi [73f88294@gateway/web/freenode/ip.] has joined #shogun11:33
deerishi@sonney2k : hi11:33
deerishii was just drafting my gsoc proposal ,and i wanted to ask if i could cite from a few papers?11:34
sonne|workdeerishi: sure11:39
-!- gsomix [] has joined #shogun11:41
gsomixgood morning11:41
-!- Dgomzi [b49531e7@gateway/web/freenode/ip.] has joined #shogun12:11
-!- sids_aquarius [~sidi_@] has joined #shogun12:43
-!- foulwall [] has quit [Ping timeout: 272 seconds]13:18
-!- sids_aquarius_ [~sidi_@] has joined #shogun13:20
-!- sids_aquarius [~sidi_@] has quit [Read error: Connection reset by peer]13:20
-!- sids_aquarius_ is now known as sids_aquarius13:22
-!- deerishi_ [73f88294@gateway/web/freenode/ip.] has joined #shogun13:30
-!- Dgomzi [b49531e7@gateway/web/freenode/ip.] has quit [Ping timeout: 245 seconds]13:30
-!- deerishi [73f88294@gateway/web/freenode/ip.] has quit [Quit: Page closed]13:35
-!- iglesiasg [] has joined #shogun13:38
-!- mode/#shogun [+o iglesiasg] by ChanServ13:38
gsomixiglesiasg: yo13:41
gsomixiglesiasg: huh, autocomplete is not working for your nickname in miranda im.13:42
-!- gsomix [] has left #shogun ["I'm a happy Miranda IM user! Get it here:"]13:42
-!- gsomix [] has joined #shogun13:43
@iglesiasggsomix: mm that's weird13:43
@iglesiasgit works for me here13:44
@iglesiasggsomix: I guess you are not but just in case, are you startting with @?13:44
-!- iglesiasg [] has quit [Quit: Reconnecting]13:44
-!- gsomix [] has quit [Read error: Connection reset by peer]13:45
van51@sonney2k: hey, so about the json file issue I made the change to SG_FREE. Would you like me to try and fix the issue in SerializableFile?13:45
-!- iglesiasg [] has joined #shogun13:45
-!- mode/#shogun [+o iglesiasg] by ChanServ13:45
-!- gsomix [] has joined #shogun13:45
gsomixiglesiasg: I think it's just bug. because now autocomplete works proper... after crashing Miranda :)13:46
@iglesiasggsomix: hehe ok13:46
-!- iglesias1 [] has joined #shogun13:49
-!- iglesias1 [] has left #shogun []13:49
-!- gsomix [] has quit [Quit: Miranda IM! Smaller, Faster, Easier.]14:17
@iglesiasgShogun Machine Learning toolbar14:30
@wikingwhere's listing?14:38
@wikinglistyn? :)14:38
@wikingsonne|work: around?14:39
-!- alibezz [~alibezz@2001:12f0:601:a921:7c4c:a0ab:e23e:9c51] has joined #shogun14:39
sonne|workwiking: yes?14:40
@wikingsonne|work: got some minuts?14:41
@wikingsonne|work: 1) freebsd bot with python_modular14:41
@wikingsonne|work: it will fail for sure but let's add it and we'll try to fix it somehow14:41
sonne|workso just install the dependencies and make sure that it compiles locally14:42
sonne|workthen I can enable it14:42
sonne|workwith the settings you give me14:42
@wikingsonne|work: 2) librt (random forest implementation) can only handle binary labelling.... and afaik random forest should be able to do multiclass, or?14:42
@wikingsonne|work: yeah well this is my current problem14:42
@wikingsonne|work: for python_modular + freebsd:
@wikingsonne|work: ideas?14:43
@wikingsonne|work: i mean the first 2 things it's easy to fix...14:43
@wikingsonne|work: but i have no clue what the last error is and how it could be fixed14:43
@wikingsonne|work: that's the output when u try to do ./configure --interfaces=python_modular && make on freebsd bot14:44
sonne|worksth weird is going on14:44
sonne|workyou are on a 64bit machine14:44
sonne|workbut use only 32bit it seems14:44
@wikingsonne|work: i haven't touched anything apart from this in ./configure script14:44
sonne|worksure but this is the reason14:45
@wikingany ideas where i could start fixing this problem?14:45
-!- deerishi_ [73f88294@gateway/web/freenode/ip.] has quit [Ping timeout: 245 seconds]14:45
@wikingi mean all the libraries are 64 bit14:45
@wikingon that fbsd bot14:45
sonne|workit is easy to get it to compile14:46
@wikingso i don't even know why it tries to use 32bit14:46
sonne|workbut that would be more of a workaround14:46
sonne|workit should use 64bit14:46
sonne|workthe error you are getting is basically saying swig's int/long etc data types are just 32bit14:47
sonne|workand I forced this define on on 64bit machines14:47
sonne|workbecause swig doesn't inherit included #defines14:47
@wiking/usr/local/bin/swig2.0: ELF 64-bit LSB executable, x86-64, version 1 (FreeBSD), dynamically linked (uses shared libs), for FreeBSD 9.1, not stripped14:47
-!- kaushalb09 [] has joined #shogun14:48
@wikingsonne|work: btw: am i right about random forest? or it is really just binary?14:49
sonne|workrandom forest == bagged decision tree14:49
sonne|workso you can do anything14:49
sonne|workbinary / mc14:49
@wikingthen librt sucks :(14:49
@wikingas the labeling it accepts atm is only 0,1,-114:50
@wikingcerr << "Incorrect label (only +1, 0, -1 supported)" << endl;14:50
-!- lisitsyn [] has joined #shogun14:51
@wikingsonne|work: brb, if u have a fix idea, let me know14:53
@wikinglisitsyn: ping14:53
lisitsynwiking: pong14:54
@wikinglisitsyn: fucking librt14:54
@wikinglisitsyn: only support cerr << "Incorrect label (only +1, 0, -1 supported)" << endl;14:55
lisitsynwiking: positive, negative, unknown?14:55
@wikingyeah i have no clue14:57
@wikingi'll check what weka uses14:57
lisitsynwiking: iglesiasg: just fyi I am in berlin starting from 9th July, staying in golden tulip hotel15:02
@iglesiasglisitsyn: thanks for the info!15:02
@iglesiasgI will book this week15:02
vgorbatilisitsyn: hi, I have recently submitted an application - could you please take a look if it is ok for you?15:03
vgorbatilisitsyn: when you have time, of course)15:04
lisitsynvgorbati: will take a look yes15:04
lisitsynwiking: I am listing !15:04
-!- kaushalb09 [] has quit [Ping timeout: 260 seconds]15:08
@wikinglisitsyn: sorry :(15:09
lisitsynwiking: haha15:11
-!- deerishi [73f88294@gateway/web/freenode/ip.] has joined #shogun15:14
-!- sijin [~smuxi@] has joined #shogun15:17
-!- Shogo [4e207df3@gateway/web/freenode/ip.] has joined #shogun15:20
-!- Shogo [4e207df3@gateway/web/freenode/ip.] has quit [Client Quit]15:21
-!- gsomix [~Miranda@] has joined #shogun15:40
gsomixlisitsyn: Lee See Tsyn15:41
* gsomix afk15:41
-!- iglesiasg [] has quit [Quit: leaving]15:43
-!- romi__ [~mizobe@] has joined #shogun15:44
-!- iglesiasg [] has joined #shogun15:47
-!- mode/#shogun [+o iglesiasg] by ChanServ15:47
-!- iglesiasg [] has left #shogun []15:47
-!- iglesiasg [] has joined #shogun15:47
-!- mode/#shogun [+o iglesiasg] by ChanServ15:47
-!- sids_aquarius [~sidi_@] has left #shogun ["Off to code | sleep"]15:53
-!- van51 [~van51@] has left #shogun ["QUIT :Leaving."]15:54
-!- deerishi [73f88294@gateway/web/freenode/ip.] has quit [Quit: Page closed]16:08
-!- deerishi [73f88294@gateway/web/freenode/ip.] has joined #shogun16:15
-!- alibezz [~alibezz@2001:12f0:601:a921:7c4c:a0ab:e23e:9c51] has quit [Ping timeout: 264 seconds]16:18
mikhailBelousHow can i contact with georg, he didn't responded for multiply emails for 5 days?16:28
@iglesiasgmikhailBelous: he must be really busy I am afraid16:34
mikhailBelousI understand, but it's only 3 days to create a proposal16:36
@iglesiasgmikhailBelous: insist if you think it is appropriate16:37
mikhailBeloushow i can insist?16:39
@iglesiasgmikhailBelous: sending another mail16:40
@iglesiasgin any case, you can also try asking us around here. It might be something we can help you with16:41
-!- alibezz [~alibezz@2001:12f0:601:a921:7c4c:a0ab:e23e:9c51] has joined #shogun16:42
-!- mikhailBelous [~towelenee@] has quit []16:57
sonne|workwiking: look at this16:59
sonne|work#if (__WORDSIZE == 64) || (LONG_MAX != INT_MAX)16:59
sonne|work# error "SWIG wrapped code invalid in 64 bit architecture, regenarete code using -DSWIGWORDSIZE64"16:59
sonne|workwiking: maybe you can figure out what LONG_MAX and INT_MAX on this bsd machine are17:00
@wikingsonne|work: where do we save __WORDSIZE ?17:03
sonne|workwiking: according to wikipedia LONG_MAX is related to long, it is
sonne|workand INT_MAX to int17:03
-!- deerishi [73f88294@gateway/web/freenode/ip.] has quit [Ping timeout: 245 seconds]17:03
@wikingaccording to ./configure the word size is 64 bits correctly17:03
-!- alibezz [~alibezz@2001:12f0:601:a921:7c4c:a0ab:e23e:9c51] has quit [Quit: Leaving.]17:04
sonne|workohh actually it is this error no?17:04
sonne|work#ifdef SWIGWORDSIZE6417:04
sonne|work#define SWIGWORDSIZE6417:04
sonne|work#ifndef LONG_MAX17:04
sonne|work#include <limits.h>17:04
sonne|work#if (__WORDSIZE == 32) || (LONG_MAX == INT_MAX)17:04
sonne|work# error "SWIG wrapped code invalid in 32 bit architecture, regenarete code using -DSWIGWORDSIZE32"17:04
sonne|workso __WORDSIZE == 32!17:05
@wikingi mean i've just ran ./configure17:05
sonne|workthen LONG_MAX == INT_MAX17:05
@wikingand it's output is 64 bits17:05
sonne|workso sizeof(int) == sizeof(long)17:05
sonne|workcan you check that?17:05
@wikingso imo __WORDSIZE = 6417:06
@wikingyep just a sec17:06
sonne|workoops gtg, will be around in the evening17:06
-!- wencan [] has joined #shogun17:07
@wikingsonne|work: 8 417:07
@wikingsonne|work: printf ("%ld %ld\n", sizeof(long), sizeof(int));17:08
-!- deerishi [73f88294@gateway/web/freenode/ip.] has joined #shogun17:09
-!- iglesiasg [] has quit [Quit: leaving]17:10
@wikingeven better:17:10
@wiking8 417:10
@wiking9223372036854775807 214748364717:10
@wikingsonne|work: the second line is: the result for         printf ("%ld %d\n", LONG_MAX, INT_MAX);17:11
-!- mikhailBelous [~towelenee@] has joined #shogun17:11
@wikingsonne|work: i really wonder what's going on:
mikhailBelousI got a problem instaling project. When i sudo make install. It's write error cannot stat "./settings/*.h" no such file or directory17:19
wencanHi, I'm Wencan. I'm interested in the Gsoc Project (Fast Reading and writing of shogun features / objects in standard file formats).17:19
wencanCould you add me my github id: wencanluo to the repository?17:19
@wikingwencan: you dont need to be added17:21
@wikingwencan: fork a repository for yourself17:21
@wikingand do there the changes17:21
@wikingcommit and push it there17:21
wencanOk. That's fine.17:21
@wikingand then submit a pull request for the develop branch17:22
@wikingmikhailBelous: and?17:22
@wikingmikhailBelous: i mean with that error it still should wokr17:22
mikhailBelousokey, so it's mean I made some other mistake installing17:23
@wikingwell what do you try to test17:24
@wikingi mean what and where does it fail17:24
@wikingdo u try to run an example?17:24
mikhailBelousI try to compile sample from your site17:24
@wikingwhat sample?17:25
-!- deerishi [73f88294@gateway/web/freenode/ip.] has quit [Ping timeout: 245 seconds]17:26
@wikingmikhailBelous: it would be nice if you could actually copy-paste somehow the actual error17:26
mikhailBelousOkay wait a minute please17:28
@wikingthat's ok17:28
-!- mikhailBelous [~towelenee@] has quit []17:29
-!- mikhailBelous [~quassel@] has joined #shogun17:30
@wikingmikhailBelous: plz put it here as other's might encounter the same problem17:33
mikhailBelousI compiled your project then trying to compile simple program, just init and exit(from your site sample).17:35
mikhailBelousAnd it outputs error17:35
mikhailBelousa.cpp:(.text+0x29): undefined reference to `shogun::init_shogun(void (*)(_IO_FILE*, char const*), void (*)(_IO_FILE*, char const*), void (*)(_IO_FILE*, char const*), void (*)(bool&, bool&))'17:35
@wikingok what's a.cpp?17:35
@wikingbut i think you are missing -lshogun flag from your linking17:36
-!- gsomix [~Miranda@] has quit [Quit: Miranda IM! Smaller, Faster, Easier.]17:36
mikhailBelous first example17:43
-!- van51 [] has joined #shogun17:47
@wikingmikhailBelous: yep17:47
@wikingi guess you are running something like17:47
@wikinggcc a.cpp17:47
@wikingbut you'll need to tell that it should link it to libshogun as well17:47
@wikingso something like17:47
@wikinggcc a.cpp -lshogun17:47
mikhailBelousI do it17:47
@wikingthen maybe17:48
@wikinggcc a.cpp -L/usr/local/lib -lshogun17:48
@wikingif your prefix is /usr/local for you ./configure script17:48
mikhailBelousFound mistake17:48
mikhailBelousI had build project without configure for debian17:49
-!- lisitsyn [] has quit [Quit: Leaving.]17:53
@wikingi really don't know how you could have built without ./configure17:55
mikhailBelousI made just configure, not debian configure17:57
-!- thoralf [] has joined #shogun17:57
thoralfHey all.17:57
-!- kaushalb09 [] has joined #shogun17:59
thoralfA small memory management issue in shogun while stream-ascii-file-parsing: When getting a copy of stream->get_vector() and assigning it to a local variable, I get in trouble then SG_UNREF(stream).18:00
thoralfDoes anyone know what to do and/or to read? :)18:00
-!- iglesiasg [] has joined #shogun18:00
-!- mode/#shogun [+o iglesiasg] by ChanServ18:00
-!- alibezz [~alibezz@] has joined #shogun18:09
-!- vgorbati [c3ee5cb1@gateway/web/freenode/ip.] has quit [Ping timeout: 245 seconds]18:25
-!- iglesiasg [] has quit [Quit: leaving]18:29
-!- iglesiasg [] has joined #shogun18:29
-!- mode/#shogun [+o iglesiasg] by ChanServ18:29
van51@sonney2k: when you can please review the PR for the json fix.18:43
van51i think the @ in front is redundant :P18:44
-!- iglesiasg [] has quit [Quit: leaving]18:47
-!- iglesiasg [] has joined #shogun18:51
-!- mode/#shogun [+o iglesiasg] by ChanServ18:51
thoralfsonney2k: Are you available?18:57
-!- kaushalb09 [] has quit [Ping timeout: 252 seconds]19:00
-!- sids_aquarius [] has joined #shogun19:12
-!- sids_aquarius_ [] has joined #shogun19:12
-!- sids_aquarius [] has quit [Client Quit]19:13
-!- sids_aquarius_ is now known as sids_aquarius19:13
@sonney2kthoralf, whats up?19:38
@sonney2kmikhailBelous, pong19:38
@sonney2kwiking, that is indeed weird19:39
@wikingsonney2k: yeah kind of like have no clue wtf19:39
@sonney2kwiking, could you please grep in the swig headers for that message19:42
@sonney2kwiking, and add a #warning macro there19:43
@sonney2kto print out the values?19:43
@wikingsonney2k: lol?!19:44
@sonney2kwiking, but it is this difficult to get the actual values:
@wikingsonney2k: i mean i've just deleted from that macro check19:44
@wiking#if (__WORDSIZE == 32) || (LONG_MAX == INT_MAX)19:44
@wiking__WORDSIZE == 3219:44
@wikingand now it's compiling19:44
@wikingso WTF?19:44
@sonney2kahh WORDSIZE=32?19:44
@wikingbut where does it get it19:44
@wikingi have noooo idea19:44
@wikingbecause as u can see from that dummy example19:45
@wikingthere wordsize == 6419:45
@wikingso somewher it gets redefined19:45
thoralfsonney2k: Just sent a mail to the mailing list.19:47
@wikingwit this19:47
@wiking__WORDSIZE = 3219:47
@wikingso i'm guessing somewhere in those headers __WORDSIZE is redefined19:48
@wikingi'm trying now a grep -r __WORDSIZE /usr/local/*19:49
@wikingbut no help...19:50
@wikingit's not defined anywhere there19:50
@wikingjust used19:50
@sonney2kwiking, but how can it be defined then?19:53
@wikingsonney2k: mmm still wondering19:53
@wikingi think i found it19:54
@sonney2kthere is wordsize in there19:54
@wikingclang++ -o test -> wordsize = 3219:54
@wikingclang test.c -o test -> wordsize = 6419:55
@wiking__wordsize is not a portable macro apparently19:55
@wiking(at least in freebsd stdint.h :P)19:55
@sonney2kwiking, crazy19:55
@wiking /* GNU and Darwin define this and people seem to think it's portable */19:55
@sonney2kso now what19:55
@wikingfrom /usr/include/stdint.h19:55
@sonney2kswig uses it19:55
@sonney2kit is used in limits.h19:55
@sonney2kwiking, is it used in limits.h in fbsd too?19:56
@wikingi mean not __WORDSIZE19:56
@wikingthat's some linux hacko19:57
@sonney2kwiking, so it is undefined?19:57
@wikingwell not undefined19:58
@wikingas u can see it's defined in stdint.h19:58
@wikingbut as it says that bug i've just pasted19:58
@wikingit's not the most reliable macro19:58
@wikingon freebsd19:58
@wikingsee clang++ and clang output19:58
@sonney2kwiking, how about you forcefully define __WORDSIZE=64 ?19:59
@sonney2kwiking, try to add this to the -D defines list19:59
@wikingsonney2k: same behavoir for g++4620:00
@wikingsonney2k: you mean clang++ -D__WORDSIZE=64 -o test20:01
@wikingas it doesn't help unfortunately20:01
@wikingit's still 3220:01
@sonney2kwiking, so where is it defined then?20:04
@wikingsonney2k: as i said20:04
@sonney2kit must be somewhere in /usr/**.h20:04
@wikingin /usr/include/stdint.h20:04
@wikingbut it's aint reliable20:04
@sonney2kwiking, yeah but how is it defined there?20:04
@wiking#if defined(UINTPTR_MAX) && defined(UINT64_MAX) && (UINTPTR_MAX == UINT64_MAX)20:05
@wiking#define __WORDSIZE              6420:05
@wiking#define __WORDSIZE              3220:05
@wikingand one of those is not defined in c++ case20:05
@sonney2kso it is overridden20:05
@sonney2kbut it looks good to me20:06
@wikingyeah but still20:06
@wikingaccording to the bug20:06
@wikingHowever the way it's defined is utterly unsafe: when UINTPTR_MAX or UINT64_MAX are not defined (which is the case for C++, as their definitions in e.g. x86/_stint.h are wrapped in20:06
@wikingif !defined(__cplusplus) || defined(__STDC_LIMIT_MACROS)20:07
@wiking__WORDSIZE is always defined as 32, which is wrong on 64bit systems.20:07
@sonney2kwiking, could you please try -D__LP64__ too20:07
@wikingclang++ -D__WORDSIZE=64 -D__LP64__ -o test20:08
@wiking[wiking@sandbox ~]$ ./test20:08
@wiking8 420:08
@wiking9223372036854775807 214748364720:08
@sonney2kwiking, please print the 3 values above20:10
@sonney2ktwo actually :D20:10 error: use of undeclared identifier 'UINTPTR_MAX'20:11
@wiking        std::cout << UINTPTR_MAX << " " << UINT64_MAX << std::endl;20:11
@wiking                     ^20:11 error: use of undeclared identifier 'UINT64_MAX'20:11
@wiking        std::cout << UINTPTR_MAX << " " << UINT64_MAX << std::endl;20:11
-!- vgorbati [d4029f22@gateway/web/freenode/ip.] has joined #shogun20:14
@sonney2kwiking, cool - so define both to 0xffffffffffffffff20:15
@sonney2kwiking, or alternatively include stdint.h and define __LP64__20:15
@sonney2k#ifdef __LP64__20:16
@sonney2k   85 #define INT64_MIN       (-0x7fffffffffffffff-1)20:16
@sonney2k   86 #define INT64_MAX       0x7fffffffffffffff20:16
@sonney2k   87 #define UINT64_MAX      0xffffffffffffffff20:16
@sonney2k   88 #else20:16
@sonney2kthat is what I found in there20:16
-!- gsomix [~Miranda@] has joined #shogun20:16
gsomixhi, guys20:16
@wikingclang++ -DUINTPTR_MAX=0xffffffffffffffff -DUINT64_MAX=0xffffffffffffffff -o test20:16
@wikingthat works20:16
gsomixehm, and girls :)20:16
@wikingsonney2k: that macro works only on linux20:18
@wikingsonney2k: we'll have to go with -DUINTPTR_MAX=0xffffffffffffffff -DUINT64_MAX=0xffffffffffffffff20:18
@wikingon fbsd20:18
@sonney2kwiking, ok so try if it compiles then...20:19
@sonney2kand even runs sth20:19
gsomixsonney2k: good evening. so, I'm finally writing proposal for I/O project.20:21
@sonney2kgsomix, ok!20:22
-!- lisitsyn [~blackburn@] has joined #shogun20:28
gsomixlisitsyn: hey20:28
-!- mode/#shogun [+o lisitsyn] by ChanServ20:28
gsomixlisitsyn: Ben, I need help20:28
@lisitsynanyone is up to present shogun at euroscipy?20:40
@lisitsynwiking: brussels! you should do that!20:41
vgorbatilisitsyn: hi, I have submitted a PR we discussed yesterday (just informing you:)).20:42
@lisitsynahh yes let me check it then20:43
@lisitsynvgorbati: oops you have updated data folder accidentally I guess20:45
vgorbatilisitsyn: no, I did actually put there a regression data file for newly added test20:45
vgorbatilisitsyn: remove it?20:46
@lisitsynyeah lets keep it old yet20:46
vgorbatilisitsyn: what is a good way to do this?20:47
-!- kaushalb09 [] has joined #shogun20:47
@lisitsynvgorbati: git reset HEAD~120:48
@lisitsyncommit and force push20:48
vgorbatilisitsyn: I should issue reset HEAD~1 in data submodule I guess?20:49
@lisitsynif you commited there yes20:49
vgorbatilisitsyn: updated20:54
@lisitsyngood job20:54
@lisitsynlet me wait for travis and then I'll merge20:55
vgorbatilisitsyn: sure. I am thinking of trying to implement some DR method now. What is your opinion?20:55
@lisitsynvgorbati: if you feel ok porting not the best C++ code I have something20:57
vgorbati lisitsyn: sure20:58
@lisitsynvgorbati: that could be a guide for adding a new method:
vgorbatilisitsyn: yeah, I saw it)21:00
@lisitsynvgorbati: what about SOM?21:00
@lisitsynvgorbati: or manifold sculpting21:01
vgorbatilisitsyn: regarding your words, ' provided in a clean C++ library' is a lie?:) (text on their main page)21:01
vgorbatilisitsyn: SOM sound good for me21:01
@lisitsynvgorbati: no it is pretty clean I think but slow21:01
@lisitsynvgorbati: eigen3 based code is faster21:01
@lisitsynand mike gashler uses ehm strange methods for eigendecompositions21:02
@lisitsynbut other code must be ok21:02
@lisitsynvgorbati: SOM looks pretty big21:02
@lisitsynlets rather go for manifold sculping21:02
vgorbatilisitsyn: at least I have heard about it:) is manifold sculpting smaller?21:03
@lisitsynyes pretty small21:03
vgorbatilisitsyn:got it21:03
@lisitsynvgorbati: SOM in waffles is like >1.5K LoC21:03
vgorbatilisitsyn: and it is present in the 'Waffles', right?21:03
vgorbatilisitsyn: yeah, it will probably require implementing some base class for neural network, etc.21:04
vgorbatilisitsyn: SOM I mean21:04
@lisitsynand it is not like classic dimension reduction21:05
vgorbatilisitsyn: I guess there's no eigendecompositions there)21:05
@lisitsynvgorbati: no not only that21:06
vgorbatilisitsyn: SOM is all-to-all neural network IIIRC?21:07
vgorbatilisitsyn: or no, it is hopfield's net21:08
@lisitsynI do not remember exactly21:08
@lisitsynvgorbati: it has connections from all feature nodes to all mapping space nodes21:09
@lisitsynvgorbati: it is not that typical because it maps to some 10 x 10 space or sth like that21:10
vgorbatilisitsyn: anyway, maybe it will be the next DR. I will start working on sculpting from tomorrow then21:11
@lisitsynsounds like a plan21:11
-!- van51 [] has left #shogun ["QUIT :Leaving."]21:12
vgorbatilisitsyn: well, I definitely want to continue working on shogun in any case21:12
@lisitsynwhat do you mean?21:12
vgorbatilisitsyn: I thought that 'sounds like a plan' meant that I am planning before actually being accepted:) so I decided to clarify21:14
@lisitsynno I mean plan for that week for now21:14
vgorbatilisitsyn: ah, I guess so21:14
@sonney2kwiking, so does it compile now?21:19
-!- kaushalb09 [] has quit [Ping timeout: 264 seconds]21:30
@wikingsonney2k: when? ah with those macros...21:37
@wikingyeah i mean i've compiled it earlier by removing __WORDSIZE == 32 from that macro21:38
@sonney2kwiking, well then so just add this workaround to configure21:45
@sonney2kwiking, we have a freebsd() function there21:46
@sonney2kand if wordsize is 64 just add these defines21:46
-!- romi__ [~mizobe@] has quit [Ping timeout: 258 seconds]21:46
@wikingsonney2k: :DDD21:47
@wikingyes captain!21:47
-!- van51 [] has joined #shogun21:47
-!- romi__ [~mizobe@] has joined #shogun21:49
-!- van51 [] has quit [Client Quit]21:50
-!- van51 [] has joined #shogun21:56
vgorbatilisitsyn: btw, PR passed (except as usual)22:20
-!- shogun-notifier- [] has joined #shogun22:25
shogun-notifier-shogun: Soeren Sonnenburg :develop * d0406f0 / / (3 files):
shogun-notifier-shogun: fix json serialization (Closes: #943)22:25
@sonney2kvan51, sorry but the json fix was much more complex ^22:26
@sonney2kvan51, but still we might have the memory leak with the string for you to fix22:26
-!- vgorbati [d4029f22@gateway/web/freenode/ip.] has quit [Ping timeout: 245 seconds]22:31
van51sonney2k: it needs an entire re-design?22:31
van51and what memory leak are you referring to?22:32
shogun-buildbotbuild #714 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at  blamelist: Soeren Sonnenburg <>22:35
-!- lisitsyn1 [~blackburn@] has joined #shogun22:35
-!- gsomix_ [~Miranda@] has joined #shogun22:36
-!- gsomix [~Miranda@] has quit [Ping timeout: 245 seconds]22:36
-!- lisitsyn [~blackburn@] has quit [Ping timeout: 252 seconds]22:37
-!- travis-ci [] has joined #shogun22:49
travis-ci[travis-ci] it's Soeren Sonnenburg's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun:
-!- travis-ci [] has left #shogun []22:49
-!- alibezz [~alibezz@] has quit [Quit: Leaving.]22:53
-!- alibezz [~alibezz@] has joined #shogun22:54
shogun-notifier-shogun: Soeren Sonnenburg :develop * c7a100a / tests/integration/python_modular/
shogun-notifier-shogun: add fancy difflib based diff to determine failures on travis22:59
@sonney2kvan51, yes - IMHO this has never worked22:59
-!- sumit [73f91219@gateway/web/freenode/ip.] has joined #shogun23:00
@sonney2kvan51, there were lots of checks failing when the result was NULL23:00
@sonney2kbut in fact 0 meant - no error23:00
@sonney2kwiking, I've added some better diff to determine the error we have w/ travis in tester.py23:01
@sonney2kwiking, now it will print the original / expected values23:01
@sonney2kand also a diff23:01
@sonney2klets hope that helps...23:01
mikhailBelousIs it possible to go out from town for 20 july -10 august and still be in GSOC?23:03
van51sonney2k: sorry, not really sure I am following.. when what result was null?23:03
@sonney2kvan51, just look at the diff of my commit above23:03
@sonney2kvan51, but we might still have the leak with the filename23:04
shogun-buildbotbuild #715 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at  blamelist: Soeren Sonnenburg <>23:10
van51sonney2k: so with your changes, SerializableJsonFile is ok?23:13
van51and now it's just a matter of whether the m_filename from the superclass leaks?23:13
-!- sumit [73f91219@gateway/web/freenode/ip.] has quit [Ping timeout: 245 seconds]23:14
@sonney2kvan51, should be yes23:15
van51i'll fetch the new code and see if filename is indeed causing a problem23:16
-!- travis-ci [] has joined #shogun23:37
travis-ci[travis-ci] it's Soeren Sonnenburg's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun:
-!- travis-ci [] has left #shogun []23:37
@sonney2kvan51, thanks23:38
@sonney2kwiking, alright looks like differences are in the 14th digit after the coma23:40
van51hehe what for? haven't done anything yet :p23:40
@sonney2kvan51, we have one test failing on travis-ci but not our own buildbot23:42
@sonney2kdifference is ~1e-1423:42
van51and that error comes from deserialization?23:45
van51also please avoid writing "failing" after my name.. i'm getting nervous :p23:45
@wikingsonney2k: mmm lemme check as well23:46
@wikingsonney2k: any ideas what could be the underlying cause23:54
@wikingbtw: i couldn't find any freebsd specific function/variable in the configure script23:54
--- Log closed Tue Apr 30 00:00:29 2013