IRC logs of #shogun for Saturday, 2012-03-10

« Friday, 2012-03-09 Index Sunday, 2012-03-11 »
--- Log opened Sat Mar 10 00:00:19 2012
-!- in3xes [~in3xes@180.149.49.230] has joined #shogun00:34
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun01:23
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Read error: Operation timed out]02:56
-!- in3xes_ [~in3xes@180.149.49.230] has joined #shogun04:44
-!- in3xes__ [~in3xes@180.149.49.230] has joined #shogun04:46
-!- in3xes [~in3xes@180.149.49.230] has quit [Ping timeout: 248 seconds]04:48
-!- in3xes_ [~in3xes@180.149.49.230] has quit [Ping timeout: 252 seconds]04:50
-!- in3xes_ [~in3xes@180.149.49.230] has joined #shogun06:28
-!- in3xes__ [~in3xes@180.149.49.230] has quit [Ping timeout: 246 seconds]06:32
-!- in3xes__ [~in3xes@180.149.49.230] has joined #shogun06:38
-!- in3xes_ [~in3xes@180.149.49.230] has quit [Ping timeout: 260 seconds]06:41
-!- in3xes__ [~in3xes@180.149.49.230] has quit [Ping timeout: 244 seconds]06:59
-!- in3xes__ [~in3xes@180.149.49.230] has joined #shogun07:12
-!- in3xes_ [~in3xes@180.149.49.230] has joined #shogun07:44
-!- in3xes__ [~in3xes@180.149.49.230] has quit [Ping timeout: 245 seconds]07:47
-!- blackburn [~qdrgsm@109.226.78.202] has joined #shogun08:25
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking]09:45
-!- in3xes__ [~in3xes@180.149.49.230] has joined #shogun09:45
-!- in3xes_ [~in3xes@180.149.49.230] has quit [Ping timeout: 244 seconds]09:47
-!- vikram360 [~vikram360@117.192.164.156] has quit [Ping timeout: 265 seconds]09:48
-!- vikram360 [~vikram360@117.192.174.187] has joined #shogun09:49
CIA-64shogun: Sergey Lisitsyn master * r0e176ef / (4 files in 2 dirs): Improvements for MC classifiers - http://git.io/vlgSIA09:51
-!- in3xes_ [~in3xes@180.149.49.230] has joined #shogun09:52
-!- cronor [~cronor@e178169201.adsl.alicedsl.de] has quit [Ping timeout: 260 seconds]09:54
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun09:55
-!- in3xes__ [~in3xes@180.149.49.230] has quit [Ping timeout: 246 seconds]09:55
-!- cronor [~cronor@e178176255.adsl.alicedsl.de] has joined #shogun09:55
-!- cronor [~cronor@e178176255.adsl.alicedsl.de] has quit [Read error: Connection reset by peer]10:02
-!- flxb [~flxb@e178176255.adsl.alicedsl.de] has joined #shogun10:10
-!- flxb [~flxb@e178176255.adsl.alicedsl.de] has left #shogun []10:12
-!- vikram360 [~vikram360@117.192.174.187] has quit [Ping timeout: 245 seconds]10:13
-!- vikram360 [~vikram360@117.192.174.163] has joined #shogun10:15
-!- vikram360 [~vikram360@117.192.174.163] has quit [Ping timeout: 252 seconds]10:19
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking]10:28
-!- in3xes_ is now known as in3xes10:40
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun10:44
n4nd0blackburn: good morning!10:51
n4nd0blackburn: how is it going?10:51
blackburnn4nd0: guten tag ;)10:51
blackburnfine, and you?10:51
n4nd0blackburn: fine too :)10:52
n4nd0blackburn: I have a couple of questions10:54
blackburnshoot10:54
n4nd0blackburn: yesterday I was checking if the QDA stuff I have done so far had some memory leaks10:54
blackburnyeah painful shit10:54
n4nd0so I used valgrind and target an example done with python modular10:54
blackburnright10:55
n4nd0and the thing is that I got stuff as "still reachable" or "possibly lost"10:55
n4nd0I was a bit concerned about that10:55
n4nd0but actually it also appeared when I run the example with the lines where QDA is used commented out10:56
blackburnn4nd0: valgrind always reports some errors on python itself10:56
blackburnuse valgrind --leak-check=full --log-file=somelog python yourexample10:56
blackburnand then just find 'shogun' in your log file :)10:57
n4nd0blackburn: ok cool10:57
n4nd0I used a suppressions file for the python stuff10:57
blackburnthere is a way to ignore python things with valgrind10:57
n4nd0but I get it doesn't remove everything10:57
blackburnyeah, you'd have to recompile python for that IIRC10:58
n4nd0I read something about that10:58
n4nd0does it pay off? I mean is it sth good to do?10:58
blackburnC++ example would be simplier to check for leaks etc10:58
blackburnhmm I did not it10:58
n4nd0ok10:59
n4nd0and another thing10:59
n4nd0I am having a hard time trying to find the svd :S10:59
n4nd0I am using this wrap_dgesvd function11:00
blackburnaha11:00
blackburnwhat's the problem?11:00
n4nd0so when I run it with valgrind it looks to work fine11:01
n4nd0but without it fails on an assert11:01
n4nd0python: malloc.c:4631: _int_malloc: Assertion `(unsigned long)(size) >= (unsigned long)(nb)' failed.11:02
blackburnhmm11:02
n4nd0but I think this one is related to the parameters work and lwork that I thought I shouldn't worry about when using the wrap function11:02
n4nd0they are in dgesvd but the wrap seems to take care of them11:02
blackburntry to insert SG_SPRINT("lworka=%d",lworka); at lapack.cpp:26011:03
blackburnand check whether it is positive11:03
n4nd0in any case I found it really weird that it runs with valgrind but not without it11:03
blackburnno, not really weird, it is sometimes11:04
n4nd0really, why so? any clue?11:05
blackburnsome sandbox or so11:05
blackburnvalgrind takes more care :)11:05
n4nd0what is this thing about that sometimes sg_print doesn't work but sg_sprint does?11:07
n4nd0(this looks like a tv show, many questions :P)11:07
blackburnn4nd0: sg_print uses m_sgio field of class derived from SGObject11:08
n4nd0blackburn: lworka is positive, >>>>>>> lworka=1411:08
blackburnand sg_sprint uses global SGIO11:08
blackburnhm..11:08
blackburnwhen something could be wrong with your params11:08
n4nd0blackburn: I am checking changing the ones that were not example the same as they are used in pinv, let's see11:09
n4nd0blackburn: it seems there was a problem using 'N' in jobu11:11
n4nd0blackburn: it worked after changing it to 'A'11:11
n4nd0the thing is that I am not interested in U in the svd decomposition A = U * S * V^T so I just wanted to ignore it11:12
n4nd0I thought 'N' was for that11:12
blackburnn4nd0: is it a QDA?11:13
n4nd0blackburn: yes, I am doing this for QDA11:13
blackburnwhat is the matrix you decompose?11:13
n4nd0but I said nothing before ... it does still fail11:13
blackburnfeature?11:13
n4nd0features of one class - mean of the features of that class11:14
n4nd0and that thing is to be done for every class11:14
n4nd0scikit-learn's code11:14
blackburnah I see11:14
blackburnn4nd0: I just mean you could compute gram matrices instead of it11:15
n4nd0http://snipt.org/ufgp611:16
n4nd0blackburn: I have no idea what is that :S11:16
blackburngram matrix?11:17
blackburnXX'11:17
blackburnerr11:17
blackburnX'X11:17
blackburneigvectors(X'X) = V, where UEVT = svd(X)11:17
n4nd0and you mean to get the eigenvalues of that?11:17
blackburnUEV' I mean11:17
n4nd0ok I see11:17
blackburnI don't know if it is better11:18
blackburndinner time11:18
blackburn:)11:18
n4nd0the matrix will be bigger I think11:18
blackburnbrb11:18
n4nd0dinner?11:18
n4nd0what time is it in Russian man? :P11:18
blackburn14.1811:18
n4nd0early dinner then11:18
blackburnprobably :)11:18
blackburnlunch is a better word may be? ;)11:28
blackburnn4nd0: ok did you trace the error?11:29
n4nd0blackburn: haha oh lunch! that's better now yes11:52
n4nd0I think I had some trouble with the dimensions :S11:52
n4nd0my bad11:53
-!- in3xes [~in3xes@180.149.49.230] has quit [Remote host closed the connection]12:08
CIA-64shogun: Sergey Lisitsyn master * r835eb0b / (2 files in 2 dirs): Rejection strategies fixes - http://git.io/tKwikQ12:20
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun12:29
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking]12:58
CIA-64shogun: Sergey Lisitsyn master * r47abf3f / src/shogun/classifier/svm/MulticlassOCAS.cpp : Warning fix for mc ocas - http://git.io/5u2KkQ13:26
CIA-64shogun: Sergey Lisitsyn master * rd123ca6 / (3 files in 2 dirs): Proper rejection handling - http://git.io/sk3ZfA13:26
-!- vikram360 [~vikram360@117.192.166.231] has joined #shogun13:30
blackburndamn i'm an idiot13:30
-!- vikram360 [~vikram360@117.192.166.231] has quit [Ping timeout: 272 seconds]13:37
-!- vikram360 [~vikram360@117.192.162.199] has joined #shogun13:38
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 252 seconds]13:50
CIA-64shogun: Sergey Lisitsyn master * reee5670 / (2 files in 2 dirs): Finally proper rejects handling - http://git.io/ZtZ0RA14:00
CIA-64shogun: Sergey Lisitsyn master * rdf59536 / src/shogun/machine/MulticlassMachine.cpp : Quick fix - http://git.io/AJ0CCQ14:02
CIA-64shogun: Sergey Lisitsyn master * rf9907d2 / (2 files): Added reject number counting - http://git.io/ejr_2A14:10
-!- harshit_ [~harshit@182.68.130.241] has joined #shogun14:10
harshit_hello everyone14:14
vikram360hello harshit_14:19
harshit_so there are some ppl out there i thought i hav entered some ghostcity !14:20
harshit_btw i was planning to implement C5.0 in shogun14:20
harshit_need some starting advise14:20
vikram360harshit_ : haha, I was planning on doing the same thing for my gsoc proposal14:21
harshit_that seems to be the most easiest one14:22
-!- blackburn [~qdrgsm@109.226.78.202] has left #shogun []14:39
vikram360harshit_ : not really, I chose it because I've worked with DTrees before.14:42
harshit_thats exactly what i am saying decision trees are easiest to implement at least for me14:44
shogun-buildbotbuild #586 of libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/libshogun/builds/586  blamelist: blackburn91@gmail.com14:50
shogun-buildbotbuild #587 of libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/libshogun/builds/58714:53
-!- harshit_ [~harshit@182.68.130.241] has left #shogun ["Leaving"]15:01
-!- vikram360 [~vikram360@117.192.162.199] has quit [Ping timeout: 248 seconds]15:11
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun15:12
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun15:18
-!- vikram360 [~vikram360@117.192.162.199] has joined #shogun15:22
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 260 seconds]15:43
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking]15:44
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun16:37
-!- vikram360 [~vikram360@117.192.162.199] has quit [Read error: Connection reset by peer]17:01
-!- vikram360 [~vikram360@117.192.162.199] has joined #shogun17:01
-!- in3xes [~in3xes@180.149.49.227] has joined #shogun17:04
-!- flxb [~flxb@e178168097.adsl.alicedsl.de] has joined #shogun17:09
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun17:13
-!- blackburn [~qdrgsm@83.234.54.35] has joined #shogun17:22
-!- flxb [~flxb@e178168097.adsl.alicedsl.de] has quit [Ping timeout: 244 seconds]17:26
-!- vikram360 [~vikram360@117.192.162.199] has quit [Ping timeout: 272 seconds]17:28
-!- vikram360 [~vikram360@117.192.188.21] has joined #shogun17:32
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking]17:33
-!- in3xes_ [~in3xes@180.149.49.227] has joined #shogun17:38
-!- in3xes [~in3xes@180.149.49.227] has quit [Ping timeout: 252 seconds]17:41
-!- blackburn [~qdrgsm@83.234.54.35] has quit [Ping timeout: 252 seconds]17:45
-!- in3xes_ [~in3xes@180.149.49.227] has quit [Ping timeout: 244 seconds]17:46
-!- blackburn [~qdrgsm@188.168.13.202] has joined #shogun18:02
-!- vikram360 [~vikram360@117.192.188.21] has quit [Ping timeout: 264 seconds]18:07
-!- blackburn [~qdrgsm@188.168.13.202] has quit [Quit: Leaving.]18:29
-!- vikram360 [~vikram360@117.192.188.21] has joined #shogun18:40
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun19:24
-!- asm_ [4e352848@gateway/web/freenode/ip.78.53.40.72] has joined #shogun19:30
-!- vikram360 [~vikram360@117.192.188.21] has quit [Ping timeout: 260 seconds]20:11
-!- vikram360 [~vikram360@117.192.188.21] has joined #shogun20:12
-!- asm_ [4e352848@gateway/web/freenode/ip.78.53.40.72] has quit [Quit: Page closed]20:51
-!- wiking [~wiking@huwico/staff/wiking] has quit [Ping timeout: 272 seconds]20:55
-!- vikram360 [~vikram360@117.192.188.21] has quit [Read error: Connection reset by peer]20:57
-!- vikram360 [~vikram360@117.192.188.21] has joined #shogun20:57
-!- harshit_ [~harshit@182.68.130.241] has joined #shogun20:59
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun21:21
-!- flxb [~flxb@g225027199.adsl.alicedsl.de] has joined #shogun21:30
-!- flxb [~flxb@g225027199.adsl.alicedsl.de] has quit [Quit: flxb]22:38
-!- harshit_ [~harshit@182.68.130.241] has quit [Quit: Leaving]22:57
CIA-64shogun: Sergey Lisitsyn master * rfa1f8cd / (57 files in 9 dirs): Some rearrangements again - http://git.io/VHnYpQ23:22
CIA-64shogun: Sergey Lisitsyn master * rcb7c871 / (4 files in 2 dirs): Added supervised KLTSA - http://git.io/lsrBUQ23:22
shogun-buildbotbuild #588 of libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/libshogun/builds/588  blamelist: blackburn91@gmail.com23:27
--- Log closed Sun Mar 11 00:00:19 2012
« Friday, 2012-03-09 Index Sunday, 2012-03-11 »

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