Open in new window / Try shogun cloud
--- Log opened Tue Jun 05 00:00:41 2012
n4nd0heiko: good night, a bit late though :S00:12
blackburnhah! I managed to introduce some group theory in thesis00:13
n4nd0group theory?00:17
blackburnyes :D00:18
blackburncyclic group of gradient operators00:18
n4nd0I think I had never heard about group theory before00:19
blackburnheh funny00:20
n4nd0take a look at this00:33
n4nd0at the messages with SG_PRINT in particular00:33
n4nd0I don't really like the idea of printing these messages everywhere, it makes the code kind of unreadable00:33
n4nd0but they make the checks everywhere in MOSEK examples00:34
n4nd0blackburn: what do you think?00:34
blackburnwho did that?00:34
n4nd0that code is mine00:34
blackburnwell having informative messages is cool00:35
n4nd0I don't really like to have them everywhere in the code00:35
blackburnwhat is alternative?00:37
n4nd0I was looking for some cool idea of yours :D00:38
blackburnREQUIRE(condition, message)00:39
n4nd0pse I don't know, maybe00:39
n4nd0looks better than what I do now to tell the truth00:39
blackburnREQUIRE(mosek->get_rescode()==MSK_RES_OK, "Result code is not OK");00:40
n4nd0all right, I had enough for today00:51
n4nd0too tired :S00:51
n4nd0good night!00:51
-!- n4nd0 [] has quit [Quit: leaving]00:51
blackburngood night00:51
-!- cronor [] has quit [Read error: Connection reset by peer]00:57
-!- cronor [] has joined #shogun00:57
-!- romi_ [~mizobe@] has quit [Ping timeout: 252 seconds]01:12
-!- heiko [] has quit [Ping timeout: 260 seconds]01:15
-!- av3ngr [av3ngr@nat/redhat/x-cdloprclgxvognxz] has joined #shogun02:21
-!- av3ngr [av3ngr@nat/redhat/x-cdloprclgxvognxz] has quit [Remote host closed the connection]02:21
-!- blackburn [~blackburn@] has left #shogun []02:27
-!- cronor_ [] has joined #shogun02:40
-!- cronor [] has quit [Ping timeout: 245 seconds]02:41
-!- cronor_ is now known as cronor02:41
-!- cronor_ [] has joined #shogun03:40
-!- cronor [] has quit [Ping timeout: 260 seconds]03:42
-!- cronor_ is now known as cronor03:42
-!- cronor_ [] has joined #shogun04:21
-!- cronor [] has quit [Ping timeout: 265 seconds]04:23
-!- cronor_ is now known as cronor04:23
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun04:42
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking]04:57
-!- zxtx [] has quit [Quit: Leaving]06:03
-!- zxtx [] has joined #shogun06:03
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun06:06
-!- n4nd0 [] has joined #shogun07:32
n4nd0good morning wiking, how is it going?07:52
wikingn4nd0: heheh great07:53
wikingalready awake for 3 hours now :)07:53
n4nd0wow you started early today then!07:53
n4nd0same time there? 07:53 here07:53
wikingsame time07:53
n4nd0jet lag?07:53
n4nd0I just wondered if you took a look to the so framework that is ready so far07:55
n4nd0if everything is fitting well for your latent-svm07:55
-!- cronor [] has quit [Quit: cronor]07:58
-!- cronor [] has joined #shogun07:59
-!- cronor [] has quit [Read error: Connection reset by peer]08:55
-!- cronor [] has joined #shogun08:55
-!- uricamic [~uricamic@2001:718:2:1634:40ea:f122:b876:bdd1] has joined #shogun08:56
-!- n4nd0 [] has quit [Ping timeout: 250 seconds]09:42
-!- cronor [] has quit [Quit: cronor]09:55
-!- cronor [] has joined #shogun09:56
-!- cronor [] has quit [Quit: cronor]10:01
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking]10:27
-!- cronor [] has joined #shogun10:32
-!- heiko1 [] has joined #shogun10:37
-!- foo__ [2e1fd566@gateway/web/freenode/ip.] has joined #shogun10:58
-!- n4nd0 [] has joined #shogun11:17
@sonney2kheiko1, did you notice that all tests fail in11:46
@sonney2k#0  0x00002b7330c4a5b0 in shogun::CStatistics::incomplete_gamma () from //usr/local/stow/shogun-2.0/lib/
@sonney2k#1  0x00002b7330d7aa29 in shogun::CStatistics::inverse_incomplete_gamma (a=0, y0=0.050000000000000044) at mathematics/Statistics.cpp:122311:46
@sonney2k#2  0x0000000000401034 in test_inverse_incomplete_gamma () at statistics.cpp:4511:46
@sonney2k#3  0x00000000004011a5 in main (argc=1, argv=0x7fffd9c09058) at statistics.cpp:5911:46
heiko1sonney2k, no I did not, though I checked them, thanks for noticing11:50
@sonney2kheiko1, they all hang on buildbot (and on my machine too)11:53
heiko1I for some reason did not see that11:53
heiko1I already spotted the bad line11:53
CIA-9shogun: Soeren Sonnenburg master * rc4402d5 / examples/undocumented/libshogun/modelselection_parameter_combination_test.cpp : don't use SGVector in modsel example (fixes compile error) -
CIA-9shogun: Heiko Strathmann master * r62ab975 / examples/undocumented/libshogun/statistics.cpp : fixed a non-terminating test case -
CIA-9shogun: Heiko Strathmann master * r90df188 / examples/undocumented/libshogun/statistics.cpp : Merge pull request #568 from karlnapf/master -
CIA-9shogun: Soeren Sonnenburg master * r3177332 / examples/undocumented/libshogun/library_dyn_int.cpp : fix compile error in dynint example and reduce array size to speed it up -
CIA-9shogun: Heiko Strathmann master * r2b1ff82 / examples/undocumented/libshogun/statistics.cpp : temprary fix for the test (there is somethign wrong with the ALGOLIB routines) -
CIA-9shogun: Heiko Strathmann master * rc699c87 / examples/undocumented/libshogun/statistics.cpp : Merge pull request #569 from karlnapf/master -
heiko1sonney2k, something strange is wrong there12:01
heiko1I removed the test for now12:02
heiko1I will have a look a bit later, have to study now (last one tomorrow)12:02
-!- wiking [] has joined #shogun12:41
-!- wiking [] has quit [Changing host]12:41
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun12:41
-!- n4nd0 [] has quit [Read error: Operation timed out]13:38
-!- heiko1 [] has quit [Ping timeout: 260 seconds]13:40
-!- puffin444 [472e31fb@gateway/web/freenode/ip.] has joined #shogun13:55
foo__I have questions about "StringFeatures::obtain_from_char_features" method.14:05
foo__I get this error : SG_ERROR( "symbol does not fit into datatype \"%c\" (%d)\n", (char) max_val, (int) max_val);14:06
foo__I use StringFeatures<uint16_t> and obtain_from_char_features to get 2-mer frequence histogram features, is that correct ?14:07
-!- puffin444 [472e31fb@gateway/web/freenode/ip.] has quit [Quit: Page closed]14:32
-!- heiko [] has joined #shogun14:33
-!- gsomix [~gsomix@] has joined #shogun14:43
foo__could someone explain me the use of StringFeatures::obtain_from_char_features plz14:53
shogun-buildbotbuild #923 of r_static is complete: Success [build successful]  Build details are at
shogun-buildbotbuild #833 of octave_static is complete: Success [build successful]  Build details are at
shogun-buildbotbuild #910 of python_static is complete: Success [build successful]  Build details are at
-!- n4nd0 [] has joined #shogun15:09
n4nd0hey gsomix15:09
n4nd0how is it going?15:09
n4nd0gsomix: I read you were working with the director classes15:11
n4nd0I have a wish related to them :)15:11
n4nd0I've to go now but I'll try to catch you later and tell you about it15:15
-!- n4nd0 [] has quit [Quit: leaving]15:15
-!- oliver_ [c07c1afb@gateway/web/freenode/ip.] has joined #shogun15:27
-!- heiko1 [] has joined #shogun15:48
-!- heiko [] has quit [Ping timeout: 260 seconds]15:49
-!- uricamic [~uricamic@2001:718:2:1634:40ea:f122:b876:bdd1] has quit [Quit: Leaving.]15:58
-!- uricamic [] has joined #shogun16:03
-!- gsomix [~gsomix@] has quit [Ping timeout: 244 seconds]16:03
-!- cronor_ [~cronor@] has joined #shogun16:10
-!- cronor [] has quit [Read error: Operation timed out]16:12
-!- cronor [~cronor@] has joined #shogun16:14
-!- cronor_ [~cronor@] has quit [Ping timeout: 245 seconds]16:14
-!- cronor [~cronor@] has left #shogun []16:15
-!- oliver_ [c07c1afb@gateway/web/freenode/ip.] has quit [Quit: Page closed]17:26
-!- n4nd0 [] has joined #shogun17:31
-!- foo__ [2e1fd566@gateway/web/freenode/ip.] has quit [Ping timeout: 245 seconds]17:51
-!- gsomix [~gsomix@] has joined #shogun17:54
-!- uricamic [] has quit [Quit: Leaving.]17:58
-!- n4nd0 [] has quit [Ping timeout: 252 seconds]18:00
-!- uricamic [] has joined #shogun18:03
-!- uricamic [] has quit [Client Quit]18:05
@sonney2kheiko1, ok - good luck for tomorrow18:50
heiko1sonney2k, thanks! :)18:57
gsomixsonney2k, hey. how are you?18:59
@sonney2kgsomix, fine :)19:00
@sonney2kgsomix, I hope you made some big progress while I am absent :D19:00
@sonney2k...and passed all your exams of course !19:00
gsomixsonney2k, unfortunately, only exams.19:02
gsomixbut I'm working now19:02
-!- zxtx [] has quit [Remote host closed the connection]19:06
-!- zxtx [] has joined #shogun19:10
@sonney2kheiko1, you are the owner of shogun/statistics basically19:13
heiko1sonney2k, yes, why?19:14
@sonney2kheiko1, I am cleaning up CMath - where would you want to see entropy() / relative entropy() / fishers exact test for whatever ?19:14
heiko1Yeah these should go to statistics in my eyes19:14
heiko1and keep a wrapper in CMath but put a deprecated or so19:15
@sonney2kheiko1, where in statistics?19:15
@sonney2kI will just remove them19:15
heiko1perhaps one can split up the statistics later19:15
heiko1oh well there is only one file for the CStatistics class19:15
heiko1the stuff in the statistics folder is for the new tests19:15
heiko1But I want to merge later19:15
@sonney2kheiko1, shogun/statistics?19:17
heiko1no in math19:17
heiko1there are all the basic functions19:17
@sonney2khmmhh why is it not in shogun/statistics/Statistics.h etc?19:18
heiko1because I added the statistics folder recently (fraft)19:18
heiko1I want to do that19:18
@sonney2kso lgamma/ tgamma etc should also go into CStatistcs right?19:19
heiko1also there is a difference between these advanced stat methods and the basic functions like mean/variance/fisher etc19:19
@sonney2kheiko, ok thanks - then I will put it in there19:20
heiko1sonney2k, alright!19:21
-!- blackburn [d5578aee@gateway/web/freenode/ip.] has joined #shogun19:25
blackburnhey there19:25
gsomixblackburn, yo19:29
@sonney2kblackburn, about GMM - if you have time to fix it - please do.19:29
blackburn-inf time19:30
blackburnthese days19:30
blackburn2 more I guess19:30
blackburnfinally preparing thesis19:30
@sonney2kthen its destiny is sealed19:30
blackburnI just was surprised you removed whole GSoC project :)19:30
-!- heiko1 [] has quit [Ping timeout: 260 seconds]19:40
-!- n4nd0 [] has joined #shogun19:52
n4nd0sonney2k, gsomix: about the director classes, I think it could be a good idea to give support for the argmax function used in SO19:54
n4nd0for the other ones (delta loss, joint feature) it could be useful too19:54
n4nd0sonney2k: what do you think?19:54
-!- romi_ [~mizobe@] has joined #shogun20:37
-!- puffin444 [472e31fb@gateway/web/freenode/ip.] has joined #shogun21:43
n4nd0wiking: about Psi(x,y) and Psi(x,y,h), sonney2k suggested to use a implicit representation for the joint features21:45
n4nd0following the idea of COFFIN21:45
n4nd0something similar to what is done with CDotFeatures I believe21:45
@sonney2kn4nd0, agreed21:50
@sonney2kn4nd0, gsomix - the earlier/more we can use the new director stuff the better21:51
n4nd0sonney2k: I will devote time to the COFFIN issue once I have tested and debugged the current implementation with multiclass classification21:51
@sonney2kn4nd0, sure get the beast to work first :D21:53
@sonney2kwiking, thanks for the update22:08
n4nd0can someone please help me with this error?22:08
wikingsonney2k: no worries22:08
n4nd0../../shogun/mathematics/Mosek.h:66: Error: Syntax error in input(3).22:08
n4nd0make[1]: *** [modshogun_wrap.cxx] Error 122:08
@sonney2kwiking, good to see you being back on track now22:08
n4nd0and the line it is referring to is22:09
wikingsonney2k: yep just i'm still waking up at 4am :D22:09
n4nd0static void MSKAPI print(void* handle, char str[]);22:09
@sonney2kwiking, and for the future - please send us incremental pull requests22:09
wikingother than that i'm good :P22:09
wikingsonney2k: well i won't until i don't have something really useful22:09
@sonney2kno need to do full rebases then22:09
wikingsonney2k: but yeah i'll first have to do the full reforking22:09
@sonney2kwiking, well we expect everyone to succeed anyways22:09
wikingsonney2k: so that i'll be able to do a pull request one day22:10
@sonney2kso there is no reason in not submitting patches (that don't break other stuff) one by one22:10
wikingyeah sure22:10
@sonney2kwiking, I think all you need to do is git format-patch on the old clone and then apply it to the new one22:11
@sonney2kwiking, n4nd0 went through this btw22:11
@sonney2kand all other too :)22:11
wikingsonney2k: yep i was hoping for that22:11
wikingyeah i guessed :)22:11
@sonney2kn4nd0, my suggestion is to ignore the whole mosek class22:12
n4nd0sonney2k: ok, I'll try with that22:12
@sonney2kn4nd0, %ignore CMosek; is the syntax22:13
n4nd0sonney2k: do you plan that we'll include this use of Mosek in the "final" version?22:13
@sonney2kn4nd0, we can keep it but as I said almost no one can use it22:14
@sonney2kcertainly not me, not the buildbot, ...22:14
n4nd0n4nd0: yeah, we really need something to solve this QP22:14
@sonney2kn4nd0, well if there is time we can create some kind of optimizer interface that wraps the common tasks22:15
@sonney2kn4nd0, maybe this is post GSoC though...22:15
@sonney2kwe have cplex as interface too22:15
n4nd0sonney2k: Nico told me you discussed with him about it22:16
n4nd0sonney2k: do you think that will imply lot of work?22:16
@sonney2kn4nd0, well if it is 'just' one certain variant - say general QP with linear inequality and equality constraints + box constraints - no22:17
@sonney2kbut there is lots of stuff underneath22:17
@sonney2k...type of optimizer (dual simplex, interior point,...)22:18
@sonney2kall that can make a huge difference in speed22:18
@sonney2kwiking, I hope you agree it is better to not have a 100MB crap data file in shogun's src code repo... so this repository recreation was really needed and worth the pain22:19
* sonney2k has ping turn around times of 3600ms - yay22:21
blackburnsonney2k: are you still at hiddensee?22:22
@sonney2kof course22:23
@sonney2ktotally hidden in the dune22:23
@sonney2k...sunset is just over22:23
-!- gsomix [~gsomix@] has quit [Quit: Ex-Chat]23:33
-!- heiko [] has joined #shogun23:47
-!- puffin444 [472e31fb@gateway/web/freenode/ip.] has quit [Quit: Page closed]23:55
--- Log closed Wed Jun 06 00:00:41 2012