Open in new window / Try shogun cloud
--- Log opened Tue Mar 20 00:00:19 2012
-!- aang [0e8be4d2@gateway/web/freenode/ip.14.139.228.210] has quit [Ping timeout: 245 seconds]02:16
-!- PhilTillet [9d9f0a0e@gateway/web/freenode/ip.157.159.10.14] has quit [Quit: Page closed]02:24
-!- l0nr4n [~l0nr4n@unaffiliated/l0nr4n] has quit [Quit: This computer (or maybe me) has gone to sleep]03:36
-!- vikram360 [~vikram360@117.192.171.204] has joined #shogun05:32
-!- gsomix [~gsomix@188.168.2.63] has joined #shogun05:38
-!- vikram360 [~vikram360@117.192.171.204] has quit [Ping timeout: 264 seconds]05:40
-!- vikram360 [~vikram360@117.192.173.229] has joined #shogun05:40
-!- harshit_ [~harshit@182.68.120.246] has joined #shogun06:07
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun07:53
-!- xiangwang [~chatzilla@159.226.60.224] has joined #shogun08:58
-!- in3xes [~in3xes@180.149.49.227] has joined #shogun09:24
-!- harshit_ [~harshit@182.68.120.246] has quit [Remote host closed the connection]09:25
-!- in3xes [~in3xes@180.149.49.227] has quit [Ping timeout: 245 seconds]09:33
-!- xiangwang [~chatzilla@159.226.60.224] has quit [Quit: ChatZilla 0.9.88.1 [Firefox 11.0/20120312181643]]09:38
-!- gsomix [~gsomix@188.168.2.63] has quit [Read error: Operation timed out]09:40
-!- vikram360 [~vikram360@117.192.173.229] has quit [Ping timeout: 246 seconds]09:42
-!- Traxex [~vinaygarg@14.139.82.6] has joined #shogun10:20
-!- Traxex [~vinaygarg@14.139.82.6] has left #shogun []10:21
-!- vikram360 [~vikram360@117.192.168.142] has joined #shogun10:53
-!- vikram360 [~vikram360@117.192.168.142] has quit [Ping timeout: 245 seconds]12:06
-!- GenX [~Sambhav@14.139.82.6] has quit [Read error: Connection reset by peer]12:28
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun12:29
-!- GenX [~Sambhav@14.139.82.6] has joined #shogun12:29
-!- nicococo [~nico@lacedcoffee.ml.tu-berlin.de] has joined #shogun12:33
n4nd0hi all, can someone lend me a hand wiht a mosek license file?12:50
n4nd0the links on the official page for students/research licenses are broken :(12:50
-!- OgreMagi [~vinaygarg@14.139.82.6] has joined #shogun12:53
OgreMagiHi12:53
n4nd0OgreMagi: hey!12:53
OgreMagiCan you help me with application procedure for GSoC. I am interested in the project "Integrate C5.0"12:54
OgreMagiWhat should be my 1st step?12:54
n4nd0well I am not a mentor, just another student as you probably are :)12:55
n4nd0but maybe I can give you any tip12:55
n4nd0what first step are you looking for? to get into contact with shogun or to prepare your application more specifically?12:55
OgreMagiBoth may be12:56
OgreMagiIs this your 1st time ?12:56
n4nd0applying for gsoc with shogun?12:56
OgreMagiYes12:56
n4nd0yes, it is12:56
OgreMagiand what about GSoC in general ?12:56
n4nd0I've not been a GSoC student before12:57
OgreMagiOkay.12:57
OgreMagiSo, which project are you interested in?12:58
n4nd0SO-SVM12:58
OgreMagiOkay.12:58
OgreMagiSo how are you proceeding?12:59
OgreMagi?13:00
wikingn4nd0: yep13:01
n4nd0so I am making some contributions to the project in general, and working on the topic of the GSoC project I want to apply for13:01
wikingn4nd0: let me dugout for what do you need it?13:01
wikingsince libqp will more or less do the same for you and it's gpl licensed...13:02
wikingn4nd0: gimme an email addr13:02
wikingso that i can fwd you the whole email with the license13:02
n4nd0n4nd0: so I am taking a look to this project http://mloss.org/software/view/250/13:02
wikingah yeah13:02
wikingi wanted to test that as well13:02
wikingit depends as well on mosek?13:03
n4nd0either mosek or cplex13:03
wikingsent13:03
n4nd0thanks :)13:04
n4nd0OgreMagi: so have you tried to get a bit familiarized with the code?13:09
n4nd0forking the repo and getting the source locally13:09
-!- romi_ [~mizobe@187.74.11.203] has quit [Ping timeout: 246 seconds]13:10
wikingso i was saying13:19
wikingthat for doing SO-SVM, e.g. hm-svm13:19
wikingit'd be great to have Interface and develop the general quadratic solver library libqp13:19
n4nd0like the one provided in mosek?13:19
wikingbecause libqp would be great solver for the structuralsvm13:20
wikingso that you don't rely on cplex or mosek13:21
wikingthat is actually proprietary13:21
n4nd0that was exactly a question I have13:21
wikingthere's already an initial interfacing between libqp and shogun13:21
n4nd0do these things exist for C++ in open source?13:21
n4nd0aham13:21
wikingbut it should be interfaced in a nicer way13:21
wikingi'm trying to do the latent svm thingy13:22
wikingand it's the same story13:22
wikinghaving libqp integrated into shogun13:22
wikingwould be great13:22
wikingso now i'm like doing the both at once13:22
wikingbut i had some paperworks shit i had to finish13:22
wikingso i couldnt start it just until recently13:22
wikingso it's still nowhere13:22
n4nd0is it lot of work? the interface of libqp in shogun?13:23
-!- romi_ [~mizobe@187.101.60.135] has joined #shogun13:23
wikingbut i'm hoping to get something out soon, at least in my own repo in a separate branch...13:23
wikingwell as far as i can see no... but of course it should be done with a little bit of care of the api13:23
wikingbecause if you check now13:23
wikinglibqp.h13:24
wikingunder src/shogun/classifier/svm13:24
n4nd0I taking a look to it right nwo13:24
wikingso it's kind of like just a straight plugging in of libqp13:25
wikingand only one part of libqp is integrated...13:26
wikinglibqp_splx.c13:26
wikinglibqp_gsmo.c is still missing13:26
n4nd0I don't really know that is the difference between them13:27
wikinghttp://cmp.felk.cvut.cz/~xfrancv/libqp/html/ as a short description13:28
-!- OgreMagi [~vinaygarg@14.139.82.6] has quit [Ping timeout: 244 seconds]13:32
-!- OgreMagi [~vinaygarg@14.139.82.6] has joined #shogun13:32
-!- l0nr4n [~l0nr4n@hugogascon.ml.tu-berlin.de] has joined #shogun13:33
-!- l0nr4n [~l0nr4n@hugogascon.ml.tu-berlin.de] has quit [Changing host]13:33
-!- l0nr4n [~l0nr4n@unaffiliated/l0nr4n] has joined #shogun13:33
n4nd0wiking: I got the first mail now :D13:40
wikinghahahah13:40
wikingtook some time13:40
n4nd0yeah13:41
n4nd0I have in my mailbox first one that was sent 30 minutes ago and after one sent 36 minutes ago :D13:42
n4nd0btw, didn't ask you before13:43
n4nd0wiking: what do you want to try with HM-SVM?13:43
wikingimages13:43
wiking:)13:43
n4nd0more in particular :P?13:44
wikingheheh13:47
wikingi have images that needs to be categorized based on a given annotation13:48
n4nd0ok13:49
-!- vikram360 [~vikram360@117.192.163.218] has joined #shogun13:52
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving]13:58
-!- vikram360 [~vikram360@117.192.163.218] has quit [Ping timeout: 252 seconds]13:59
-!- vikram360 [~vikram360@117.192.185.106] has joined #shogun13:59
-!- OgreMagi [~vinaygarg@14.139.82.6] has quit [Read error: Connection reset by peer]14:06
-!- romi_ [~mizobe@187.101.60.135] has quit [Ping timeout: 246 seconds]14:18
-!- vikram360 [~vikram360@117.192.185.106] has quit [Ping timeout: 245 seconds]14:21
-!- romi_ [~mizobe@187.101.138.102] has joined #shogun14:30
-!- harshit_ [~harshit@182.68.120.246] has joined #shogun14:33
-!- gsomix [~gsomix@83.234.54.205] has joined #shogun14:33
gsomixhi14:34
-!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has joined #shogun15:24
-!- PSmitAalto [82e9b263@gateway/web/freenode/ip.130.233.178.99] has joined #shogun15:31
PSmitAaltoHello15:32
PSmitAaltoI have a small problem building shogun and I was wondering if you could help15:32
PSmitAaltoI am building shogun with CPLEX support and the python_modular interface15:33
PSmitAaltoNow compiling fails at classifier/LPBoost.cpp15:33
PSmitAaltoclassifier/LPBoost.cpp: In member function 'float64_t shogun::CLPBoost::find_max_violator(int32_t&)': classifier/LPBoost.cpp:82: error: 'labels' was not declared in this scope classifier/LPBoost.cpp: In member function 'virtual bool shogun::CLPBoost::train_machine(shogun::CFeatures*)': classifier/LPBoost.cpp:106: error: 'labels' was not declared in this scope classifier/LPBoost.cpp:108: error: 'labels' was not15:33
PSmitAaltoDoes anyone have an idea what the problem could be?15:34
blackburnPSmitAalto: yes, I know the error15:36
blackburnPSmitAalto: just change labels -> m_labels there15:36
blackburnI am not able to fix it in master branch right now15:37
PSmitAaltoOk, thanks, I'll try that15:38
blackburnPSmitAalto: it is pretty good you found it out cause we are in pre-release state15:38
-!- aang [0e8be4d2@gateway/web/freenode/ip.14.139.228.210] has joined #shogun15:42
PSmitAaltoThe same problem is in LPM.cpp, SubGradientLPM.cpp and svm/CPLEXSVM.cpp15:45
blackburnoh damn thanks for reporting15:46
@sonney2kgsomix, hi...16:04
@sonney2ksounds like you make good progress16:04
blackburnsonney2k: if you have time could you please labels->m_labels in classifier/LPM.cpp; classifier/SubGradientLPM.cpp; classifier/svm/CPLEXSVM.cpp and classifier/LPBoost.cpp?16:05
@sonney2kgsomix, you should expect only one python version to be installed, i.e. what is specified to configure via ./configure --python=python316:05
blackburnsonney2k: gsomix can you enlighten me on this progess too? ;)16:06
@sonney2kblackburn, ok done16:08
CIA-64shogun: Soeren Sonnenburg master * r4064deb / (4 files in 2 dirs): change labels -> m_labels - http://git.io/yxDNjg16:08
blackburnthanks!16:08
blackburnPSmitAalto: you may use latest git now16:08
blackburnsonney2k: btw have you ever used ctags with vim?16:09
blackburnworks like a charm16:09
PSmitAaltoThanks for updating so quickly!16:10
-!- PSmitAalto [82e9b263@gateway/web/freenode/ip.130.233.178.99] has quit [Quit: Page closed]16:11
@sonney2kblackburn, there is a makefile target for that - so yes16:12
@sonney2kbut I forgot what it is - didn't use it for years...16:13
blackburnsonney2k: e.g. with cursor on classname and Ctrl+] you may jump right to its definition16:13
blackburnor even16:13
blackburn:ta CMKL16:13
blackburnand voi-la you are in classifier/mkl/MKL.cpp16:13
@sonney2kohh yeah I need this :)16:14
@sonney2kthe make ctags target needs an update then16:14
blackburnI just got tired with using 5% of vim's potential16:15
blackburnand tried some new things16:15
blackburnsonney2k: I can't see any ctags target16:15
@sonney2kblackburn, in which dir do ctags have to be?16:16
blackburnsonney2k: src would be sufficient16:16
@sonney2kblackburn, cd shogun ; make ctags16:16
blackburnsonney2k: does it work for you? I have no shogun there but there is no ctags mention in makefile..16:17
@sonney2kI guess we need ctags only for libshogun so we can as well have another ctags target in the upper dir16:17
@sonney2kalright... I have to leave now16:18
@sonney2kcu16:18
blackburnsonney2k: see ya16:18
-!- nicococo [~nico@lacedcoffee.ml.tu-berlin.de] has left #shogun []16:24
blackburnaha works in shogun16:24
blackburnstrange no mention in makefile16:24
gsomixsonney2k, okay16:26
gsomixI missed these settings when I read --help. =__=16:27
blackburngsomix: what are the things you are talking about?16:27
-!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has quit [Quit: Page closed]16:33
gsomixblackburn, I did not notice that the version of Python is configured.16:33
gsomixin "./configure".16:34
gsomixsonney2k, thanks a lot16:41
-!- harshit_ [~harshit@182.68.120.246] has quit [Quit: Leaving]17:01
-!- wiking [~wiking@huwico/staff/wiking] has quit [Ping timeout: 245 seconds]17:05
-!- surtani|brb [0e8b5206@gateway/web/freenode/ip.14.139.82.6] has joined #shogun17:08
-!- vikram360 [~vikram360@117.192.172.252] has joined #shogun17:13
-!- l0nr4n [~l0nr4n@unaffiliated/l0nr4n] has quit [Quit: This computer (or maybe me) has gone to sleep]17:56
-!- l0nr4n [~l0nr4n@g231201109.adsl.alicedsl.de] has joined #shogun18:21
-!- l0nr4n [~l0nr4n@g231201109.adsl.alicedsl.de] has quit [Changing host]18:21
-!- l0nr4n [~l0nr4n@unaffiliated/l0nr4n] has joined #shogun18:21
-!- aang [0e8be4d2@gateway/web/freenode/ip.14.139.228.210] has quit [Ping timeout: 245 seconds]18:38
-!- PhilTillet [~Philippe@npasserelle10.minet.net] has joined #shogun18:40
PhilTilletHey18:41
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun18:47
-!- gsomix [~gsomix@83.234.54.205] has quit [Ping timeout: 240 seconds]18:56
-!- vikram360 [~vikram360@117.192.172.252] has quit [Read error: Connection reset by peer]18:59
-!- aparna_ [3b5cc226@gateway/web/freenode/ip.59.92.194.38] has joined #shogun19:24
-!- aparna__ [3b5cc226@gateway/web/freenode/ip.59.92.194.38] has joined #shogun19:26
-!- aparna__ [3b5cc226@gateway/web/freenode/ip.59.92.194.38] has quit [Client Quit]19:27
-!- gsomix [~gsomix@83.234.54.204] has joined #shogun19:53
-!- Varagrawal_ [0e8b7937@gateway/web/freenode/ip.14.139.121.55] has joined #shogun19:56
-!- surtani|brb [0e8b5206@gateway/web/freenode/ip.14.139.82.6] has quit [Quit: Page closed]20:07
Varagrawal_Hey20:10
Varagrawal_I have installed Shogun from source y cloning the repo20:10
Varagrawal_But when I try to run msplicer in the applications folder, I get the error "ERROR IMPORTING MODULES. PLEASE INSTALL SHOGUN"20:11
Varagrawal_Any help?20:11
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun20:17
gsomixVaragrawal_, try to remove a try-except block in msplicer's code. And look where the error occurs.20:20
gsomixn4nd0, hi20:20
Varagrawal_dyn = shogun.Structure.DynProg()20:23
Varagrawal_shogun not defined20:23
gsomixVaragrawal_, sorry.20:23
gsomixI mean, you have to do only import modules. Without catching exceptions.20:23
Varagrawal_ah20:24
Varagrawal_ok20:24
Varagrawal_importerror20:25
Varagrawal_cannot import name SVM20:25
Varagrawal_also when I try to run the minimal.cpp program as given in the tuts20:26
Varagrawal_gsomix: I get the error "Undefined reference to shogun..."20:26
n4nd0Varagrawal_: that undegined reference happens when you run or when you compile it?20:29
n4nd0undefined sorry20:29
Varagrawal_compile20:29
gsomixVaragrawal_, what keys/flags do you use for compiling and linking?20:29
gsomixVaragrawal_, you can view the makefile in the examples.20:29
n4nd0Varagrawal_: probably you should link to the libraries20:30
n4nd0Varagrawal_: try to use -lshogun20:30
Varagrawal_g++ -lshogun minimal.cpp -o minimal20:30
n4nd0I have just tested it with20:31
n4nd0g++ basic_minimal.cpp -o test -lshogun20:31
n4nd0and it worked for me20:31
gsomixsimilary20:31
n4nd0Varagrawal_: so, better now?20:33
Varagrawal_a min20:34
n4nd0sorry20:35
Varagrawal_I got what's wrong20:35
Varagrawal_got it working20:35
n4nd0great20:35
n4nd0in any case20:36
n4nd0you should have a makefile in the directory examples/undocumented/libshogun20:36
n4nd0just go there and run make20:36
Varagrawal_n4nd0: ok20:36
n4nd0no worries about compile flags20:36
Varagrawal_n4nd0: Thanks a bunch mate!! :D20:38
Varagrawal_Ha, found some errors in the code20:38
Varagrawal_:P20:39
Varagrawal_missing ';'s20:39
n4nd0:P20:39
n4nd0your first patch there :)20:39
Varagrawal_Seriously :D20:40
-!- blackburn [~qdrgsm@188.168.2.231] has joined #shogun20:40
gsomixblackburn, yo20:40
blackburnhey20:40
Varagrawal_n4nd0: 'CConjugateIndex' was not declared in this scope20:42
Varagrawal_Is this because of some configure error?20:43
-!- blackburn [~qdrgsm@188.168.2.231] has quit [Ping timeout: 244 seconds]20:45
-!- blackburn [~qdrgsm@188.168.4.147] has joined #shogun20:45
blackburnsonney2k: ping20:46
CIA-64shogun: Sergey Lisitsyn master * reec65f5 / (3 files in 2 dirs): Added missed parameters map handling - http://git.io/mqvTeQ20:56
CIA-64shogun: Sergey Lisitsyn master * rd4f4322 / (9 files): Fixes for python examples - http://git.io/mvQmaA20:56
CIA-64shogun: Sergey Lisitsyn master * r9fbbb9c / testsuite/python_modular/tester.py : Added message if test file is not present - http://git.io/fAnRTw20:56
Varagrawal_n4nd0: I am checking out the CConjugateIndex.h file and everything seems to be in order20:59
Varagrawal_what could be the possible reasons for giving a "not declared in scope" error?20:59
blackburnVaragrawal_: install lapack to make it work21:01
-!- romi_ [~mizobe@187.101.138.102] has quit [Ping timeout: 246 seconds]21:05
Varagrawal_liblapack3gf is already installed21:06
Varagrawal_blackburn: liblapack3gf is already installed21:06
blackburnVaragrawal_: is it detected when you run configure21:06
Varagrawal_blackburn: it looks for jblas21:08
blackburnjblas is required if you want to use java modular21:08
Varagrawal_blackburn: There is no detection as such for lapack21:09
blackburnyes there is21:09
Varagrawal_Ah yes21:09
Varagrawal_blackburn: Vanilla Lapack21:09
Varagrawal_blackburn: My bad21:09
blackburnso, is it Yes or No?21:11
-!- romi_ [~mizobe@187.74.2.1] has joined #shogun21:18
-!- aparna_ [3b5cc226@gateway/web/freenode/ip.59.92.194.38] has quit [Quit: Page closed]21:19
n4nd0Varagrawal_: sorry, I was having dinner21:19
-!- Varagrawal_ [0e8b7937@gateway/web/freenode/ip.14.139.121.55] has quit [Ping timeout: 245 seconds]21:22
blackburnn4nd0: back to kernel subsetting21:23
blackburnyou probably would have to check21:23
blackburnwhether kernel is custom21:23
blackburnand act according to that21:23
blackburni.e. with normal kernels you just subset features21:24
blackburnand with custom you use set_col_subset and set_row_subset21:24
n4nd0blackburn: I have to check those methods21:29
blackburnoops no power left21:29
n4nd0blackburn: but is the idea to leave those to be implemented in the custom kernel?21:30
blackburnsomething like that21:30
-!- Varagrawal_ [0e8b7952@gateway/web/freenode/ip.14.139.121.82] has joined #shogun21:30
Varagrawal_blackburn: Making :P21:31
Varagrawal_it worked21:31
Varagrawal_Just lost my pach :P21:32
Varagrawal_You know21:33
Varagrawal_I could help improve the installation documentation21:34
Varagrawal_I did something similar for TinyOS 2.1.1 on Ubuntu 11.10, but it's still undergoing a review by Mr. Eric Decker21:34
Varagrawal_n4nd0: Sorry for disturbing during your dinner21:35
n4nd0Varagrawal_: no worries at all! you didn't disturb, I apologize you waited for an answer :S21:36
Varagrawal_n4nd0: :D21:36
-!- blackburn [~qdrgsm@188.168.4.147] has quit [Ping timeout: 260 seconds]21:36
Varagrawal_n4nd0: I am still getting a import error for the msplicer code21:36
n4nd0n4nd0: let me check21:37
-!- in3xes [~in3xes@210.212.58.111] has joined #shogun21:38
Varagrawal_n4nd0: any luck?21:39
Varagrawal_I get "cannot import name SVM"21:39
n4nd0Varagrawal_: so you are trying with the executable that comes directly?21:39
Varagrawal_n4nd0: yep21:39
n4nd0Varagrawal_: trying to execute that I get ERROR IMPORTING MODULES, MAKE SURE YOU HAVE SHOGUN INSTALLED21:40
Varagrawal_n4nd0: I got the exact same thing earlier!!! :D21:40
n4nd0Varagrawal_: so I guess there must be sth weird with that executable, the makefile looks a bit weird too21:40
Varagrawal_n4nd0: Then blackburn told me to get rid of the exception block in the beginning21:40
Varagrawal_n4nd0: The import error is basically in the SVM I guess21:41
n4nd0Varagrawal_: I am getting an import error in another place now so I don't know ...21:46
n4nd0Varagrawal_: but we can try to fix it, let's see21:46
Varagrawal_n4nd0: Yeah lets!!21:47
-!- PhilTillet_ [~Philippe@tillet-p42154.maisel.int-evry.fr] has joined #shogun21:47
PhilTillet_Hello :)21:47
Varagrawal_n4nd0: Hey, it's 2 in the morn over here and I have classes at 821:47
Varagrawal_n4nd0: When can I get ahold of you next?21:48
-!- l0nr4n [~l0nr4n@unaffiliated/l0nr4n] has quit [Ping timeout: 265 seconds]21:48
-!- l0nr4n [~l0nr4n@g225120250.adsl.alicedsl.de] has joined #shogun21:49
-!- l0nr4n [~l0nr4n@g225120250.adsl.alicedsl.de] has quit [Changing host]21:49
-!- l0nr4n [~l0nr4n@unaffiliated/l0nr4n] has joined #shogun21:49
-!- PhilTillet [~Philippe@npasserelle10.minet.net] has quit [Ping timeout: 272 seconds]21:49
n4nd0Varagrawal_: I don't know, I normally spend some time around here every day21:50
n4nd0Varagrawal_: just show up and try, blackburn or any other may help you as well ;)21:50
-!- Varagrawal_ [0e8b7952@gateway/web/freenode/ip.14.139.121.82] has quit [Ping timeout: 245 seconds]21:52
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking]21:55
-!- Varagrawal [6f5d43e2@gateway/web/freenode/ip.111.93.67.226] has joined #shogun22:01
Varagrawaln4nd0: You there?22:03
-!- blackburn [5bde8018@gateway/web/freenode/ip.91.222.128.24] has joined #shogun22:06
blackburnsonney2k: we should remove sailfish application - nobody will fix it22:08
blackburnI am currently fixing others22:08
PhilTillet_hi blackburn :)22:08
Varagrawalblackburn: n4nd0 and I both seem to be having the same prob in the msplicer program22:08
blackburnPhilTillet_: hi22:09
-!- PhilTillet_ is now known as PhilTillet22:10
blackburnVaragrawal: although I have no idea how would you use msplicer I'm goint to commit fix soon22:10
blackburngoing*22:10
Varagrawalblackburn: Anyway I can help?22:10
blackburnno, already done22:11
n4nd0blackburn: are you fixing msplicer too?22:15
n4nd0I was on that as well :P22:15
blackburnn4nd0: yes already fixed22:16
n4nd0blackburn: ook :)22:17
blackburnbut I've got some wi-fi connection and can't push :(22:17
blackburnn4nd0: well believe me there is no way we can use msplicer22:17
blackburnI don't give a fuck what a splicing is :D22:17
n4nd0blackburn: me neither :P22:17
Varagrawallol22:18
n4nd0while I was fixing it I was like ... *I have no idea what is this about*22:18
n4nd0haha22:18
Varagrawaln4nd0: You're cracking me up22:18
blackburnok no way to push my commits right now22:19
n4nd0Varagrawal: :O22:19
blackburnyou lost nothing with broken msplicer though :D22:19
n4nd0Varagrawal: did I say something bad?22:19
-!- OgreMagi [~vinaygarg@14.139.82.6] has joined #shogun22:19
OgreMagihi22:20
Varagrawaln4nd0: No man, it was hilarious the way you said it22:20
blackburnhi22:20
n4nd0Varagrawal: haha ok22:20
OgreMagiThis is my 1st time I am applying for GSoC, can anybody help me how to proceed ?22:20
n4nd0Varagrawal: phrasal verbs and I are not very good friends :P22:20
Varagrawaln4nd0: You're not alone22:21
blackburnOgreMagi: you should choice idea you find interesting22:21
blackburnand contact your mentor22:21
blackburnpossible one I mean22:21
OgreMagiblackburn: Yes I am interested in the project named "Integrate C5.0" of Shogun MLT22:22
blackburnOgreMagi: so contact Soeren via sonne at debian.org22:23
n4nd0blackburn: maybe you came across it while fixing msplicer, but just in case there's an indenting issue in StringFeatures.h:38422:23
OgreMagiblackburn: Are you a mentor?22:24
blackburnOgreMagi: no, I'm not, I help Soeren with admin'ing a little22:24
blackburnand active developer :)22:25
blackburnn4nd0: huh we've got a lot of such improper indenting22:26
n4nd0n4nd0: :(22:26
blackburnn4nd0: no worries at all shogun doesn't get slower because of bad indenting ;)22:27
n4nd0blackburn: :P22:27
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun22:28
n4nd0blackburn: I wonder man, is it because of changes in the interfaces that these errors show up?22:28
n4nd0blackburn: e.g. before it was set_string_features and now set_features in CStringFeatures22:28
blackburnn4nd0: errors with msplicer etc?22:29
n4nd0blackburn: yeah, like these small things in the names of the methods and the like22:29
blackburnexactly22:29
blackburnnothing more22:29
blackburnPreProc -> Preprocessor22:29
n4nd0blackburn: I would be really cool if there was something to fix them automatically22:30
n4nd0I bet it is already invented somewhere22:30
blackburnalso22:30
n4nd0have you heard of it before?22:30
blackburnand even some forgotten (float64_t* w, int32_t len) -> SGVector<float64_t> w conversion22:30
blackburnn4nd0: I can't mind how can it be automatic22:31
n4nd0blackburn: well I don't really know either but for example the autocomplete function of terminador or eclipse22:31
blackburnwell for better compatibility we should rather not change names22:32
n4nd0blackburn: I guess that something similar to that could be used22:32
blackburnit is a overkill for me22:32
n4nd0I meant terminator before, sorry22:33
blackburnI mean that is not really need to automate 2 line fixes22:33
n4nd0mmm maybe22:34
n4nd0blackburn: what as the set_col_subset and set_row_subset for CustomKernel you tell me about22:35
n4nd0blackburn: why doesn't a normal set_subset suffice?22:35
blackburnn4nd0: basically same as set_subset22:35
blackburnwell it doesn't assume features are equal22:35
blackburn(while training they are for sure)22:35
n4nd0blackburn: ok, I see the idea but probably I should take a closer look to CustomKernel to understand it better22:37
blackburnn4nd0: it is just a matrix22:37
blackburnmain thing is normal kernels22:37
n4nd0blackburn: one more thing, do you remember that yesterday you were talking about that there were no problems with github? :P22:38
blackburnyeah22:38
n4nd0all the commits I have done for the multiclass stuff have gone into the qda pull request22:38
n4nd0not such a big problem maybe ... but it kind of sucks22:38
blackburnwell probably you have commited it into qda branch?22:39
n4nd0I merged qda branch into master before doing the pull request22:39
blackburnn4nd0: why did you do that?22:40
blackburnyou may do pull request from different branches22:40
n4nd0blackburn: aham22:40
n4nd0it looked reasonable for me :O22:40
blackburnoops22:41
n4nd0ok I think I get it now ... different branches -> different pull requests22:41
blackburnn4nd0: scikits guys use pull request to discuss things22:42
n4nd0blackburn: I will try to fix it22:43
blackburnwould be pain in the ass22:44
n4nd0although I don't know if I can change the pull request from master to another branch22:44
n4nd0probably22:44
n4nd0I guess that to remove the unwanted commits from the pull requests would be easier22:44
blackburnlooks very interesting sometimes22:44
n4nd0blackburn: I'll try to do that then, remove the commits related to multiclass from the pull request and do then in another multiclass branch22:45
blackburnn4nd0: so you have all the qda things in the qda branch?22:45
n4nd0n4nd0: yes, qda things are in my qda branch22:46
n4nd0but I did the pull request from master :S22:46
blackburnthen you could close master pull request22:47
blackburnand create one from qda22:47
n4nd0looks good22:47
n4nd0I've homework for tomorrow then :P22:48
blackburnthen you may reset your master back to its normal state22:48
blackburnand apply multiclass things on top of it22:48
n4nd0do you think it doesn't need to be done in another branch the multiclass things?22:48
blackburnnot really necessary, it is up to you22:49
blackburnI mean main intention of using branches here is to make possible to work on two things22:49
n4nd0I think I'd probably do it in another branch22:50
n4nd0anyhow, time to sleep for me today22:50
n4nd0good night people22:50
n4nd0blackburn: and thank you very much for all your help again :)22:51
blackburnconnection lost again huh22:51
blackburnn4nd0: ok see you22:51
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving]22:51
-!- Varagrawal [6f5d43e2@gateway/web/freenode/ip.111.93.67.226] has quit [Ping timeout: 245 seconds]22:55
PhilTilletblackburn, I guess it's ok if my opencl code is not fully optimized? :p23:05
blackburnPhilTillet: which code?23:05
PhilTilleti'm supposed to write some code for SVM on OpenCL for the patch! ^^23:06
blackburnah sure23:06
-!- gsomix [~gsomix@83.234.54.204] has quit [Ping timeout: 245 seconds]23:06
blackburnI think it is ok23:07
blackburnjust show the proof of concept ;)23:07
PhilTilletwell, I'm almost done implementing the SVM formula with a gaussian kernel, but integrating it will be more difficult :p23:08
PhilTilletdataset rarely go over 100 000 examples right?23:09
blackburndepends on domain..23:10
PhilTillethaha ok :p23:10
PhilTilletactually I am chosing which algorithm i'll use to sum the components of a vector, and some are more efficient from size 100 000 :p23:10
blackburnPhilTillet: do not pay attention on such details, just integrate code which sucessfully uses opencl when applying svm23:11
PhilTilletI probably ask myself too much questions indeed23:11
blackburnjust load support feature vectors and compute score23:12
PhilTilletok :p23:12
PhilTilletand when is the "deadline" by the way?23:12
blackburndeadline for?23:12
PhilTilletthe patch23:12
blackburnwell there is no such deadlines..23:13
blackburndeadline*23:13
PhilTilletok :p23:14
blackburnMarch 26 is students proposals opening IIRC23:14
PhilTilletyes23:14
blackburnand something like 6, April is proposal deadline23:14
blackburnso probably it would be nice to have any contributions before proposal submit deadline23:15
blackburncause you may refer that and it is advantage for sure23:15
PhilTilletyes23:16
PhilTilletI'll try to :p I have a tough week because I have my first conference presentation ever next week23:16
PhilTilletbut i'll try to find the time23:16
blackburnPhilTillet: as for your 'custom' idea23:16
blackburnI think if it is not the only thing you want to work on you should contact mentor of any other idea23:17
blackburnand send at least two proposals23:18
blackburnI mean if it wasn't possible to select one slot for OpenCL you may compete in some other area23:19
PhilTilletOkay :)23:19
blackburnI am afraid the number of slots will be unknown by the deadline23:19
PhilTilletyes, like last year23:19
blackburnin this case I guess Soeren won't be able to make any decision on OpenCL23:20
blackburnok late already23:21
blackburnPhilTillet: see you23:21
PhilTilletsee you :)23:21
-!- blackburn [5bde8018@gateway/web/freenode/ip.91.222.128.24] has quit [Quit: Page closed]23:21
-!- OgreMagi [~vinaygarg@14.139.82.6] has left #shogun []23:25
-!- PhilTillet [~Philippe@tillet-p42154.maisel.int-evry.fr] has quit [Quit: Leaving]23:38
--- Log closed Wed Mar 21 00:00:19 2012