Open in new window / Try shogun cloud
--- Log opened Wed May 01 00:00:29 2013
alibezzlisitsyn: iglesiasg: how do the embed methods at tapkee_methods.hpp are supposed to work? I mean, in the bigger picture of shogun's architectute?00:30
-!- Asix3 [~asix3@d47-69-144-148.nap.wideopenwest.com] has joined #shogun01:19
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]01:42
Asix3Hi! I'm going to try an get an application together for GSOC. I'm wondering if there's any way to search the shogun mailing list01:54
Asix3I've subscribed already, but I'm not sure if there's an archive of past messages01:54
Asix3...and I just found it. Nevermind!01:56
shogun-buildbotbuild #325 of nightly_none is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_none/builds/32503:03
alibezzlisitsyn: hey, when you specified on step 7 "Add a class CSammonMapping that calls the tapkee library code for sammon mapping (check PR #1015 for example).Add a class CSammonMapping that calls the tapkee library code for sammon mapping (check PR #1015 for example)."03:35
alibezzlisitsyn: you mean a class in src/shogun/converter, right?  i'm assuming so as of now03:36
alibezz=)03:36
-!- wencan [~wencan@c-71-61-182-121.hsd1.pa.comcast.net] has joined #shogun03:38
-!- cwidmer [8ca3fe9d@gateway/web/freenode/ip.140.163.254.157] has joined #shogun03:54
-!- cwidmer [8ca3fe9d@gateway/web/freenode/ip.140.163.254.157] has quit [Client Quit]03:56
Asix3ok, I've forked shogun from git and cloned it to my system. I've ran ./configure, but 'make' isn't doing anything04:17
Asix3well, it's giving me: make: *** No rule to make target '.config'.  Stop.04:20
Asix3...and it looks like I figured that problem out, too -- sorry!04:34
-!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has joined #shogun04:40
-!- foulwall_ [~foulwall@li379-21.members.linode.com] has joined #shogun04:58
-!- alibezz [~alibezz@177.182.164.8] has quit [Quit: Leaving.]05:27
-!- alibezz [~alibezz@177.182.164.8] has joined #shogun05:46
-!- alibezz [~alibezz@177.182.164.8] has quit [Client Quit]05:47
-!- van51 [~van51@athedsl-401908.home.otenet.gr] has left #shogun ["PING 1367380762"]05:59
-!- wencan [~wencan@c-71-61-182-121.hsd1.pa.comcast.net] has quit [Quit: wencan]06:13
-!- gsomix [~Miranda@109.169.243.24] has left #shogun ["I'm a happy Miranda IM user! Get it here: http://miranda-im.org"]08:07
-!- gsomix [~Miranda@109.169.243.24] has joined #shogun08:07
@sonney2klisitsyn, how many proposal do you expect?08:24
@lisitsynsonney2k: at least 30 then08:25
@lisitsynsonney2k: have you seen eveybody gets less proposals this yeat?08:25
@lisitsynyear*08:25
@sonney2kno where should I have seen this?08:26
@lisitsynsonney2k: mentors list08:26
@lisitsynsonney2k: [GSoC Mentors] Low(er) number of applications this year?08:27
@sonney2klisitsyn, ohh yes replied to the thread08:40
gsomixgood morning08:48
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun09:30
-!- gsomix [~Miranda@109.169.243.24] has quit [Read error: Connection reset by peer]09:37
-!- gsomix [~gsomix@109.169.243.24] has joined #shogun09:42
-!- poojits [75d35d4a@gateway/web/freenode/ip.117.211.93.74] has joined #shogun09:59
poojitsHi lisitsyn, I would like to discuss issue 338. The links you have provided there are broken.10:05
-!- van51 [~van51@athedsl-401908.home.otenet.gr] has joined #shogun10:17
@lisitsynpoojits: oh let me try to find out where they are now10:20
@lisitsynpoojits: okay the paper link updated but code is kind of lost10:23
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds]10:32
-!- hushell [43bd6474@gateway/web/freenode/ip.67.189.100.116] has joined #shogun10:33
poojitslisitsyn: thanks, I will have a look.10:37
-!- poojits [75d35d4a@gateway/web/freenode/ip.117.211.93.74] has quit [Quit: Page closed]10:57
hushellHello guys I am Shell11:03
hushellHi Sergey I see your reply I come here before going to bed11:04
@lisitsynhushell: hey11:04
@lisitsynHeikoS: nice to see you here11:04
@lisitsynoops11:04
@lisitsyn:D11:05
@lisitsynok nice to see you both here11:05
@lisitsynhah11:05
hushellnice to meet you guys11:05
@lisitsynhushell: alright just hang there when you are available11:06
hushellI have a short question regarding the Binarylabels::score_to_probabilities()11:06
@lisitsynalthough you are in a totally different timezone11:06
@lisitsynhushell: what is the question?11:06
hushellyeah I am in UtC-811:07
@lisitsynhaha 12 hours between us11:07
hushellwhy this function doesn't output anything expect changing the confidences11:08
@HeikoSlisitsyn: Hi!11:08
@lisitsynHeikoS: ^ could you please answer?11:08
@HeikoShushell: why should it?11:09
@HeikoSthe labels have a vector for the scores and one for the confidences11:09
@HeikoSthis translates scores to confidences11:09
@HeikoSwhy should it do something else?11:09
@lisitsynhave to go now11:10
hushellCould you take a look at Eq.(7) in this paper: http://hal.archives-ouvertes.fr/docs/00/10/39/55/PDF/cr102875872670.pdf11:11
hushellit needs the learned parameters11:11
hushellI don't know much about Platt's rescoring, it seems like a calibration11:12
hushellusing a sigmoid function11:13
@HeikoSlooking ...11:13
@HeikoShushell: I see11:14
@HeikoSyou need the A and B11:14
@HeikoSlet me quickly check the code11:14
hushellexactly11:15
@HeikoSok, my suggestion then is:11:15
@HeikoSto put the sigmoid fitting method into CStatistics11:15
@HeikoSas fit_sigmoid(SGVector<float64_t>)11:16
@HeikoSthis method returns A and B11:16
@HeikoSand is called from the binary labels11:16
@HeikoShushell: what do you think of this?11:16
hushellWhere is the CStatistics?11:17
hushellSo you mean do it in another way11:17
@HeikoShushell: no just make it modular11:18
@HeikoSor wait11:18
hushellOk I found it11:18
hushelldocumentation is great!11:18
@HeikoSsince you need the scores *and* the parameters11:18
@HeikoSand sigmoid fitting is not likely to happen somewhere else11:18
@HeikoSmaybe put it into the scores_to_probabilities method11:19
@HeikoSdont really know whats best11:19
@HeikoSwhat do you think?11:19
hushellcould you explain a bit how the scores_to_probabilities being implemented?11:21
hushellwhat's the target to fit?11:21
hushellsorry I haven't read this piece of code11:22
@HeikoSno worries11:22
@HeikoSit does two things: 1) fit a sigmoid via newton search, result is parameters a and b. 2) uses this sigmoid to compute all the probabilities given the scores11:22
@HeikoScode is taken from a paper btw11:22
@HeikoShave a look at the last lines of the method11:23
@HeikoSthere are the important parts11:23
@HeikoSand also a and b11:23
hushellthe algorithm is very clear11:26
hushellso why not just store the a and b11:26
@HeikoSnot sure about this11:26
@HeikoSscores might change11:26
@HeikoSand without calling the method again they outdate11:26
@HeikoSmaybe make a and b optional parameters11:27
@HeikoSif they are provided, they are not coimputed11:27
@HeikoSvia method in CStatistics11:27
@HeikoSbut you can also precompute yourself and then pass them11:27
hushellthat's true it's being call during testing11:27
@HeikoSI think thats good11:27
@HeikoSdo you want to move the code or should I do it?11:28
-!- poojits [75d35d4a@gateway/web/freenode/ip.117.211.93.74] has joined #shogun11:28
hushellI am not very clear how to do it11:29
@HeikoShushell: dont worry11:29
hushellSeems a good chance to learn11:29
@HeikoSI will do it now, give me 10 mins11:29
hushellwhich function should I call from CStatistics11:30
@HeikoSyou already have lots of stuff in your PR, lets not make it even more complex :)11:30
@HeikoSI will tell you in a bit11:30
hushellOkay let's see how you make it :)11:30
hushellThanks so much. That's true, many files are involved11:31
hushellIt11:31
hushellIt11:31
hushellIt's my first time to code in such a regular way11:31
hushellI just went through the code, it looks like the targets come from the counting of positives and negatives11:37
-!- ozan [~ozan@ee-167-250.eee.metu.edu.tr] has joined #shogun11:37
hushellinteresting, in Lin11:42
hushell's note he included everything11:43
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun11:44
-!- Viksvampire19 [~AndChat49@223.185.83.159] has joined #shogun11:45
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds]12:05
@HeikoShushell: ok done, I will quickly compile/test/merge and then let you know12:08
hushellAwesome!12:09
hushellThanks for your help!12:09
hushellI have to go to bed now, talk you later! Have a nice day!12:10
@HeikoShushell: ok good night!12:10
@HeikoShave a look at the score_to_prob method12:11
@HeikoSthere you will see how you can use CStatistics to get the sigmoid parameters12:11
-!- hushell [43bd6474@gateway/web/freenode/ip.67.189.100.116] has quit []12:12
-!- abinash [75ef5e6e@gateway/web/freenode/ip.117.239.94.110] has joined #shogun12:15
-!- saket [0e8b5206@gateway/web/freenode/ip.14.139.82.6] has joined #shogun12:20
-!- saket [0e8b5206@gateway/web/freenode/ip.14.139.82.6] has quit [Client Quit]12:24
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun12:26
shogun-notifier-shogun: Monique :develop * d9ca1a1 / examples/undocumented/python_modular/inverse_covariance_estimation_demo.py: https://github.com/shogun-toolbox/shogun/commit/d9ca1a1d3e03394af836f09111d0e4e688ba393012:26
shogun-notifier-shogun: Graphical example for SparseInverseCovariance utility class that estimates sparse inverse covariance by given covariance matrix. Plots a heat map and the graphs of dependency between variables12:26
shogun-notifier-shogun: Monique :develop * 7952ae8 / examples/undocumented/python_modular/inverse_covariance_estimation_demo.py: https://github.com/shogun-toolbox/shogun/commit/7952ae8619d5a405e3341b1d76c99992f932d2b412:26
shogun-notifier-shogun: remove inverse_covariance_estimation_demo.py from python_modulat directory12:26
shogun-notifier-shogun: monicadragan :develop * e6df0b0 / examples/undocumented/python_modular/graphical/inverse_covariance_estimation_demo.py: https://github.com/shogun-toolbox/shogun/commit/e6df0b04ea00473e6b66dc9e9680d55436df466a12:26
shogun-notifier-shogun:  Graphical example for SparseInverseCovariance utility class that estimates sparse inverse covariance by given covariance matrix. A heat map and the graphs of dependency between variables are plotted.12:26
shogun-notifier-shogun: monicadragan :develop * f77d761 / examples/undocumented/python_modular/graphical/inverse_covariance_estimation_demo.py: https://github.com/shogun-toolbox/shogun/commit/f77d761e9d68f34585dd6c6d1ecf359e1675d72112:26
shogun-notifier-shogun: Refactoring: loop replace12:26
shogun-notifier-shogun: monicadragan :develop * f627b92 / examples/undocumented/python_modular/graphical/inverse_covariance_estimation_demo.py: https://github.com/shogun-toolbox/shogun/commit/f627b92e91381dcfce57c02d063950f7791b088f12:26
shogun-notifier-shogun: Refactoring: loop replace12:26
shogun-notifier-shogun: Heiko Strathmann :develop * fca8ced / examples/undocumented/python_modular/graphical/inverse_covariance_estimation_demo.py: https://github.com/shogun-toolbox/shogun/commit/fca8ced2318976fe1ee7db1f4525dcc6090188da12:26
shogun-notifier-shogun: Merge pull request #1035 from monicadragan/master12:26
shogun-notifier-shogun:12:26
shogun-notifier-shogun: Issue #863: Add sparse inverse covariance graphical example - dev branch12:26
shogun-notifier-shogun: Heiko Strathmann :develop * 8fed6a8 / / (6 files): https://github.com/shogun-toolbox/shogun/commit/8fed6a8b15b2e46aa3f08e19a0fa5711988ce6e512:27
shogun-notifier-shogun: -moved sigmoid fitting of BinaryLabels to CStatistics and make BinaryLabels use it12:27
shogun-notifier-shogun: -updated/added unit-tests12:27
shogun-notifier-shogun: Heiko Strathmann :develop * 98048ce / / (6 files): https://github.com/shogun-toolbox/shogun/commit/98048ce69e5d157e64b4117694064f30e88626fb12:27
shogun-notifier-shogun: Merge pull request #1038 from karlnapf/develop12:27
shogun-notifier-shogun:12:27
shogun-notifier-shogun: fit sigmoid refactoring12:27
shogun-buildbotbuild #891 of bsd1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/891  blamelist: monicadragan <monique@work.(none)>, Monique <monique@work.(none)>12:29
@HeikoSthis makes no sense buildbot12:30
shogun-buildbotbuild #892 of bsd1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/892  blamelist: monicadragan <monique@work.(none)>12:31
@HeikoSsonney2k: any idea what causes this?12:31
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun12:32
shogun-buildbotbuild #893 of bsd1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/89312:37
shogun-buildbotbuild #894 of bsd1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/894  blamelist: Monique <monique@work.(none)>12:38
@wikinguhuuu12:44
-!- iglesiasg [~nando@n177-p222.kthopen.kth.se] has joined #shogun12:45
-!- mode/#shogun [+o iglesiasg] by ChanServ12:45
@wikingHeikoS: there was a merge for master?:)12:46
@HeikoSwiking: no12:46
@wikingmmm i wonder then why bsd bot tries to run on master :(12:46
@HeikoSmaybe I did something woring, let me check12:46
@HeikoSnop12:47
@HeikoShttps://github.com/shogun-toolbox/shogun/commits/master12:47
@HeikoSlast commits from apr 0412:47
@HeikoShttps://github.com/shogun-toolbox/shogun/commits/develop12:47
@HeikoSand those I just made are here12:47
-!- foulwall_ [~foulwall@li379-21.members.linode.com] has quit [Remote host closed the connection]13:20
-!- Viksvampire19 [~AndChat49@223.185.83.159] has quit [Quit: Bye]13:31
@wikinglisitsyn: pong13:31
-!- abinash [75ef5e6e@gateway/web/freenode/ip.117.239.94.110] has quit [Ping timeout: 245 seconds]13:47
-!- poojits [75d35d4a@gateway/web/freenode/ip.117.211.93.74] has quit [Ping timeout: 245 seconds]13:50
-!- ozan [~ozan@ee-167-250.eee.metu.edu.tr] has quit [Quit: Leaving]14:04
-!- van51 [~van51@athedsl-401908.home.otenet.gr] has left #shogun ["QUIT :Leaving."]14:05
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds]14:21
-!- alibezz [~alibezz@177.182.164.8] has joined #shogun14:28
alibezzlisitsyn: ping14:29
alibezziglesiasg: ping14:29
@iglesiasgalibezz: pong14:30
alibezziglesiasg: i started the implementation of the Sammon mapping. things are going a little slow because of my other tasks, but they're indeed evolving!... hehe i'm sort of basing my implementation on the IsoMap one14:31
alibezzlike, seeing how the methods generally work in the lib etc14:31
alibezzdo you think it's a good idea?14:31
-!- sids_aquarius [~sidi_@li400-124.members.linode.com] has joined #shogun14:32
@iglesiasgalibezz: yeah it sounds good indeed14:33
alibezziglesiasg: an for step 7 of my issue https://github.com/shogun-toolbox/shogun/issues/102514:33
@iglesiasgalibezz: are you using sammong mapping's code from the matlab toolbox for DR?14:33
alibezzi'm writing the class under converter. is it the right thing to do?14:33
alibezzyes i am. i still have some doubts on how to test my implementation14:34
alibezzshould i write a little program or does shogun have some prototypes for isomap for example, and so i can easily modify it?14:34
alibezziglesiasg: sorry i didn't put your tag in the beginning of my messages. hehehe14:38
@iglesiasgalibezz: give me some minutes please :)14:39
alibezziglesiasg: no worries! ;) just wanted to make sure you saw the messages. it's all good!14:40
-!- harry27 [~ceo@202.3.77.220] has joined #shogun15:00
-!- harry27 [~ceo@202.3.77.220] has left #shogun []15:01
-!- cameron__ [~quassel@global-2-1.nat.csx.cam.ac.uk] has joined #shogun15:02
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]15:27
@iglesiasgalibezz: re15:38
@iglesiasgalibezz: sorry for the long dealy15:38
@iglesiasgdelay*15:38
@iglesiasgalibezz: so it can be done with a class in converter15:41
@iglesiasgalibezz: however, the DR in shogun is provided via tapkee15:41
@iglesiasgso implementing sammon mapping in tapkee directly would be a better choice15:41
@iglesiasgsomething similar to what lisitsyn described in the instructions in the github issue15:42
-!- wencan [~wencan@c-71-61-182-121.hsd1.pa.comcast.net] has joined #shogun15:48
-!- wencan [~wencan@c-71-61-182-121.hsd1.pa.comcast.net] has quit [Client Quit]15:53
alibezziglesias: yup. the class just calls the implementation from tapkee16:04
@iglesiasgalibezz: aaah ok sorry, all good then16:04
alibezziglesiasg: but it's gonna take me a while to have it all done. I have to prepare some lectures... hahahahha16:04
@iglesiasghehe I see. No worries16:05
alibezziglesiasg: the idea of sammon mapping is interesting though16:05
@iglesiasgalibezz: aham! I don't really know about it16:05
@iglesiasgalibezz: so it tries to minimize the error in the distance between input and projected space I read16:07
alibezziglesiasg: yup. I'm going to implement using gradient descent, just like in the matlab code.16:08
alibezziglesiasg: but maybe sgd would perform better. it's something to test in the future16:08
alibezziglesiasg: how can i test my code? should i write a little program or do they have something for isomap that i can reuse?16:08
@iglesiasgalibezz: maybe one of the examples under examples/undocumented?16:09
@iglesiasgalibezz: or even any of tapkee examples16:09
alibezziglesiasg: chill! didn't check these folders yet. ha16:09
@lisitsynkind of re16:10
@lisitsynalibezz: oh the idea of sammon mapping is as old as my parents probably :D16:11
@iglesiasglisitsyn: around 1969 according to wikipedia16:11
@iglesiasggood number btw16:11
@lisitsyna little younger then16:11
@lisitsynthe sammon mapping is I mean16:12
alibezzlisitsyn: iglesiasg: well my parents are def older. i'm 25. hehe they're from the 50s16:16
@lisitsynalibezz: 22 and 1961 here16:16
alibezzlisitsyn: you guys are rather young! when did you start coding?16:17
@lisitsynalibezz: as professionals or?16:18
@iglesiasgalibezz: I am 23 and started at university 5 years ago16:18
@iglesiasghaha but I am no professional :)16:18
@iglesiasg5 years ago with pascal omg16:18
@lisitsynI've been paid for programming for ehmm somewhere 1.5 to 2 years so far16:19
alibezzlisitsyn: weren't you coding before that?16:20
@lisitsynalibezz: of course, for some more years16:20
@lisitsynfrom school may be16:20
alibezzgot it16:20
alibezzcool16:20
alibezzi started coding when i was 1716:21
@lisitsyn>5 probably16:21
@iglesiasgooh you are old hardcore programmers :P16:21
alibezzbut mostly in research projects16:21
@lisitsynI remember I was coding some puzzles when I was 1116:21
@lisitsynwith pascal16:21
alibezzhahaha i don't have that much of experience in c++ tho16:21
@iglesiasglisitsyn: oh that's super cool16:21
@iglesiasglisitsyn: I regret none taught me anything about programming when I was younger16:22
alibezzyea same here16:22
@iglesiasgI didn't even know it existed before I started uni lol16:22
alibezzi wish i've learnt it earlier16:22
alibezzi spent most of my life studying french and english. ha16:22
@iglesiasgoh la al16:23
@iglesiasgoh la la hehe16:23
alibezzyea my life was kinda crazy. I like programming but I wish I knew C++ better. hhehhe well, it's all good16:24
alibezzliving is learning lolz16:24
votjakovrliving is machine learning :)16:26
alibezzvotjakovr: i see what you did there! hahahaha def sounds like a slogan16:27
-!- romi__ [~mizobe@187.74.6.200] has quit [Remote host closed the connection]16:28
-!- wencan [~wencan@c-71-61-182-121.hsd1.pa.comcast.net] has joined #shogun16:30
-!- iglesiasg [~nando@n177-p222.kthopen.kth.se] has quit [Quit: Lost terminal]16:31
votjakovralibezz: yep) i think so too16:31
-!- iglesiasg [~nando@n177-p222.kthopen.kth.se] has joined #shogun16:35
-!- mode/#shogun [+o iglesiasg] by ChanServ16:35
-!- wencan [~wencan@c-71-61-182-121.hsd1.pa.comcast.net] has quit [Quit: wencan]16:36
alibezziglesiasg: it's workers' day in Brazil. supposed to be a holiday for all workers. make a dance if you're Brazilian and are working! \o\ /o/ \o/ /o\16:38
alibezzhaha16:38
alibezzthe whole idea of holidays never make sense in grad school. haha16:38
@iglesiasgalibezz: also in the office here :(16:39
alibezzyea i have to give classes tomorrow. haha16:39
alibezziglesiasg:is it a holiday in spain too?16:40
@iglesiasgalibezz: you are the lecturer then? about?16:40
@iglesiasgalibezz: in Spain and in Sweden :)16:40
alibezziglesiasg: Information Retrieval. it's a lecture about language models16:40
alibezziglesiasg: we need to teach here in order to get our degree16:40
@iglesiasgalibezz: this is my office partner today http://robotnyheter.se/wp-content/uploads/2011/11/KTH-roboten_Dora.jpg :D16:41
-!- wencan [~wencan@c-71-61-182-121.hsd1.pa.comcast.net] has joined #shogun16:42
@lisitsyntotally back16:43
-!- gsomix_ [~gsomix@95.67.181.139] has joined #shogun16:47
-!- van51 [~van51@athedsl-401908.home.otenet.gr] has joined #shogun16:47
-!- sumit [73f91219@gateway/web/freenode/ip.115.249.18.25] has joined #shogun16:48
-!- gsomix [~gsomix@109.169.243.24] has quit [Ping timeout: 252 seconds]16:50
-!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has quit [Quit: Leaving]16:57
-!- sumit [73f91219@gateway/web/freenode/ip.115.249.18.25] has quit [Ping timeout: 245 seconds]17:05
alibezzhahahaha17:07
@iglesiasgalibezz?17:08
alibezzhey I'm laughing at your partner17:09
alibezzhehe a robot17:09
@iglesiasgaaah ok yeah hehe17:09
-!- gsomix_ is now known as gsomix17:13
gsomixhi, guys17:13
@iglesiasghey gsomix17:13
gsomixso, I have planted potatoes. time to coding. >:317:14
@iglesiasghaha cool17:15
foulwalljoin #shogun17:34
-!- foulwall [~foulwall@li379-21.members.linode.com] has joined #shogun17:34
cameron__hi17:47
@iglesiasghi cameron__17:50
cameron__I'm trying to understand all the python tests work17:51
cameron__I have found the directory where you keep all your test data in the tests/integration/tests17:51
cameron__I'm wondering how are those files produced17:52
@iglesiasgcameron__: they are in the data submodule17:55
@iglesiasganother repository17:55
cameron__i am reading the generate_testdata.py file in /tests/integration17:56
cameron__if I run this, it will regenrate all the test data files17:56
alibezziglesiasg: could you please tell me where the tapkee examples are?18:00
alibezziglesiasg: the undocumented ones don't really have much to do with isomap (or any other of the implemented mappings)18:01
@iglesiasgalibezz: really?18:01
@iglesiasglet me check18:01
alibezziglesiasg: seems so18:01
@iglesiasgalibezz: what about18:02
@iglesiasghttps://github.com/shogun-toolbox/shogun/blob/master/examples/undocumented/libshogun/converter_isomap.cpp18:02
alibezzoh in libshogun? duh18:02
alibezziglesiasg: i was looking for something in the python directories dunno why18:02
@iglesiasgalibezz: aham ok18:03
alibezziglesiasg: sorry for that. and thanks ;)18:03
@iglesiasgalibezz: no problem ;)18:03
@iglesiasgthere should be another example for isomap in python_modular or so in any case18:03
alibezznot really to be honest. or i'm crazy.18:03
alibezzbecause this is one of the dirs i checked out :O18:03
@iglesiasghttps://github.com/shogun-toolbox/shogun/blob/master/examples/undocumented/python_modular/converter_isomap_modular.py18:04
@iglesiasg?18:04
@iglesiasgalibezz: maybe you checked static interfaces18:04
alibezziglesiasg: yeah maybe. or maybe i' m really crazy. there's always a possibility lolz18:09
@iglesiasgheh18:10
@wikinglisitsyn: around?18:21
cameron__the generate_testdata.py complains about unknown category classifier.18:25
-!- ozan [~ozan@ee-167-250.eee.metu.edu.tr] has joined #shogun18:26
@iglesiasgcameron__: how are you running it?18:28
cameron__iglesiasg: just by ./generator_testdata.py18:28
@iglesiasgcameron__: so did you modify anyting to add a new classifier or sth?18:29
@iglesiasgcameron__: dummy question -- have you compiled and installed for python_modular?18:31
cameron__iglesiasg: yes, i've been playing around with shogun in python18:32
cameron__iglesiasg: nvm, i'll try and figure it out myself. should just be sth minor18:32
@iglesiasgcameron__: ok18:33
@iglesiasgcameron__: I cannot test right now in my machine if it works, but I'd bet it should18:33
@lisitsynwiking: yes18:37
-!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has joined #shogun18:39
-!- foulwall [~foulwall@li379-21.members.linode.com] has quit [Ping timeout: 264 seconds]18:42
cameron__votjakovr: Hi, did u fix the GP regression problem with -ve values?18:43
-!- van51 [~van51@athedsl-401908.home.otenet.gr] has left #shogun ["PING 1367426647"]18:44
votjakovrcameron__: Hi, yep. Now  variances is not negative18:46
cameron__votjakovr: great work. what was the problem actually?18:46
-!- foulwall_ [~foulwall@2001:da8:215:c252:a90a:af19:4758:3a27] has joined #shogun18:50
-!- iglesiasg [~nando@n177-p222.kthopen.kth.se] has quit [Quit: Lost terminal]18:54
votjakovrcameron__: i think somewhere in LAPACK calls and matrix manipulations, but i'm not sure.18:57
--- Log closed Wed May 01 19:10:49 2013
--- Log opened Wed May 01 19:10:56 2013
-!- shogun-toolbox [~shogun@7nn.de] has joined #shogun19:10
-!- Irssi: #shogun: Total of 21 nicks [4 ops, 0 halfops, 0 voices, 17 normal]19:10
-!- Irssi: Join to #shogun was synced in 6 secs19:10
-!- viksvampire19 [~chatzilla@115.184.88.5] has joined #shogun19:14
viksvampire19Hi everyone..19:21
alibezzlisitsyn: ping19:30
@lisitsynalibezz: pong19:31
alibezzlisitsyn: I'm trying to run the converter_isomap_modular.py example19:31
alibezzand it gives me this error: IOError: [Errno 2] No such file or directory: '../data/fm_train_real.dat'19:31
alibezzalthough i downloaded the data19:31
@lisitsynhow?19:31
alibezzand made the symbolic link point to it19:31
@lisitsynbahh why did you do that19:31
alibezzwell the data wasnt here hehe i just pointed it19:32
@lisitsynall you needed was 'git submodule init --update'19:32
alibezzoh nice19:32
alibezzlet me see if it works then19:32
@lisitsynjust restore your symlinks19:32
@lisitsynthey were here19:32
alibezzok no worries19:33
@sonney2klisitsyn, ok so 17 proposals by now19:36
@lisitsynsonney2k: + votjakovr + gsomix19:36
@sonney2klisitsyn, please mark the ones you intend to mentor19:36
@lisitsynsonney2k: with wish to mentor?19:36
@sonney2klisitsyn, http://shogun-toolbox.org/page/Events/gsoc201219:37
@sonney2klooking at this image19:37
@sonney2kwe got 25 applications in the last 3 days19:37
@lisitsynsonney2k: see # increased just before19:37
@lisitsynsonney2k: lets calm down - everybody gets less it seems19:37
@sonney2kmaybe it is the same - lets see19:38
@sonney2kwell we have maybe -10 compared to 201219:38
@lisitsynsonney2k: so we decided about 4?19:39
@sonney2klisitsyn, and please accept the one you would want to mentor most19:39
@sonney2klisitsyn, no why?19:39
@lisitsynsonney2k: well more or less19:39
@sonney2knothing is set in stone19:39
@lisitsynokay 519:39
@sonney2kyes more or less19:39
@lisitsynno I don't mean in stone19:39
@lisitsynHeikoS: where is your supertask guy (tm)?19:40
@HeikoSlisitsyn: having exams :) but wrote me an email, I told him to apply asap19:41
alibezzlisitsyn: until which day can I send the patch?19:44
alibezzlisitsyn: like, I'm gonna send it anyway, of course19:45
@lisitsynalibezz: good question19:45
alibezzlisitsyn: but to be able to participate of the GSoC selection19:45
@lisitsynlet me check the date19:46
@lisitsynalibezz: okay it is very unlikely to be that long but assignment due is May 2419:46
alibezzlisitsyn: okay. fortunately I'll send the sammon mapping until Friday ;) let's see how things work19:47
alibezzhopefully...19:47
@sonney2kalibezz, we decide before may 2019:48
@sonney2kso anything until then my make us reconsider19:48
@lisitsynalibezz: ^ so I will have some time to change my mind19:49
@lisitsynwe have to decide ideas ASAP19:49
@sonney2klisitsyn, ?19:49
alibezzlisitsyn: sure, makes sense19:49
@lisitsynsonney2k: I mean to ask for slots19:49
@lisitsynbut not real assignment19:49
@sonney2kyes we make our slot request based on the number of excellent proposals we get19:50
@sonney2kso we already *now* put names to tasks to get an idea of #slots19:50
alibezzyeah the things is that week is full of deadlines for me. haha i wish i had known of the applications earlier =(19:50
alibezzbut i totally understand it all. if it doesn't work, it's ok! i like helping anyways =)19:51
@sonney2kalibezz, well may 20 is the hard deadline19:51
@sonney2kon may 22 there will be a dedup meeting19:51
@sonney2kerr may 24 is the dedup meeting19:52
@sonney2kbut in any case google has the last word and does the official announcement on may2719:52
-!- Netsplit *.net <-> *.split quits: syst3mw0rm19:53
@lisitsynyeah so one more month of suffering19:53
-!- Netsplit over, joins: syst3mw0rm19:53
gsomixlisitsyn, everyday I'm suffering, life is pain, haha... :(19:54
alibezzlisitsyn: git submodule init --update didn't work actually19:54
@sonney2ksubmodule update --init19:55
@lisitsynoops :D19:55
@sonney2kalibezz, other way round!19:55
@lisitsynmy mistake sorry19:55
alibezzhuaauahua makes more sense19:55
alibezzit's because both init and update are commands19:55
alibezzwho knows... maybe --update was a path. duh19:55
-!- ozan [~ozan@ee-167-250.eee.metu.edu.tr] has quit [Quit: Leaving]19:56
alibezzlisitsyn: i'll need some style tips today, later. sammon mapping has a lot of parameters and i'd like to discuss what's the most elegant way of dealing with them in the shogun architecture. but I'll write/make a bunch of tests here before. and finish this crazy lecture i'm preparing. hahaah19:57
@lisitsynalright19:57
alibezzlisitsyn: well, the git update worked, but the error remains20:01
alibezzlisitsyn: IOError: [Errno 2] No such file or directory: '../data/fm_train_real.dat'20:01
alibezzany idea?20:01
@lisitsynalibezz: you said you changed symlinks20:02
@lisitsynare they listed in 'git status' output?20:02
alibezzactually i'm working in a brand new branch at this very moment20:02
alibezzso no symlinks were changed here20:02
@lisitsynis there a file ../data/fm_train_real.dat then?20:02
alibezzyup20:03
@lisitsynwhat is the dir you run example in20:03
@lisitsyn?20:03
alibezzif you could execute ~/shogun/examples/undocumented$ python python_modular/converter_isomap_modular.py20:03
alibezzi'd be glad ;)20:03
@lisitsynrun in python_modular folder20:04
@lisitsynthe path is not absolute that's why you get troubles20:04
alibezzyeah it's not absolute but i thought the execution would take the file where the python code is as a reference, not from where i'm running20:05
alibezzbut it worked. ;)20:05
-!- foulwall_ [~foulwall@2001:da8:215:c252:a90a:af19:4758:3a27] has quit [Ping timeout: 245 seconds]20:05
@lisitsynno, why? pwd is pwd20:05
alibezzlisitsyn: yes it is. it's from where i'm running not from where it's at. just got confused. he20:06
alibezzwell now i can start writing tests and going further with the sammon mapping20:07
@lisitsynnice20:07
-!- viksvampire19 [~chatzilla@115.184.88.5] has quit [Quit: ChatZilla 0.9.90 [Firefox 20.0/20130329030832]]20:22
-!- iglesiasg [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun20:33
-!- mode/#shogun [+o iglesiasg] by ChanServ20:33
-!- kakashi_ [~kakashi_@nltk/kakashi] has quit [Ping timeout: 240 seconds]20:34
-!- kakashi_ [~kakashi_@nltk/kakashi] has joined #shogun20:35
@sonney2klisitsyn, ok so I am convinced now that we will get 30 proposals until the end of may 3rd20:35
@lisitsynhaha20:35
@iglesiasgis it starting to grow exponentially?20:38
@lisitsyniglesiasg: 17 so far20:38
@iglesiasgyou know what to do if you get too many20:39
@iglesiasghttps://i.chzbgr.com/maxW500/7373469696/h5FA4741A/20:39
@lisitsynhahaha20:40
@iglesiasghaha20:40
@iglesiasgjust kidding20:40
gsomixiglesiasg, I want to eat because of this gif.20:49
-!- mikhailBelous [~quassel@37.98.242.99] has joined #shogun20:50
@iglesiasghaha20:51
mikhailBelousIn aplication you said that report when I succsesufuly download and compile project and get used to github lot time ago.I have done it20:52
@lisitsynmikhailBelous: alright - now we'd like to see contributions - that's a hard requirement for successful applicants20:55
mikhailBelousWhat the dead for contribution?21:00
mikhailBelous*deadline21:00
mikhailBelous?21:00
@lisitsynmikhailBelous: somewhat near 20 May I'd say21:00
mikhailBelousfine21:01
@lisitsynapplication deadline is 3 May21:01
mikhailBelouscontribution21:01
@lisitsyncontribution is 20 May21:01
@lisitsynapplication is 3 May21:02
-!- hushell [43bd6474@gateway/web/freenode/ip.67.189.100.116] has joined #shogun21:06
mikhailBelousIf is it problem that I can't be in the city during 20 july - 10 august, but worked it before?21:14
@lisitsynmikhailBelous: just mention that in application21:14
-!- KMcQuisten [d8338942@gateway/web/freenode/ip.216.51.137.66] has joined #shogun21:20
KMcQuistenHello all21:20
@lisitsynKMcQuisten: hey21:20
KMcQuistenHope all is going well on your side of the world21:21
@lisitsynhah yeah21:21
@lisitsynwhat about you?21:21
KMcQuistenNot too bad.  I put a question up on the mailing list a bit earlier today, and I partially answered my own question, but i'm still having some trouble.21:22
@lisitsynKMcQuisten: oh sorry hadn't enough time to check your question21:23
@lisitsynwill take a look a bit later21:23
KMcQuistenRighto.21:23
KMcQuistenI didn't know if it was something that i was just missing or if it was a deeper issue21:23
KMcQuistenSometimes, I must admit, I am an idiot who doesn't see what's in front of his own face *grin*21:24
-!- KMcQuisten [d8338942@gateway/web/freenode/ip.216.51.137.66] has quit [Quit: Page closed]21:45
-!- mikhailBelous [~quassel@37.98.242.99] has quit [Read error: Connection reset by peer]21:45
-!- zhh210 [80b493c5@gateway/web/freenode/ip.128.180.147.197] has joined #shogun21:49
-!- alibezz [~alibezz@177.182.164.8] has left #shogun []21:51
-!- zhh210 [80b493c5@gateway/web/freenode/ip.128.180.147.197] has left #shogun []21:52
-!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has quit [Quit: Leaving]22:02
-!- pickle27 [~kevin@rcv3-lab-pc.ee.queensu.ca] has joined #shogun22:08
-!- mikhailBelous [~quassel@37.98.242.99] has joined #shogun22:32
-!- Asix3 [~asix3@d47-69-144-148.nap.wideopenwest.com] has quit [Quit: Leaving]22:43
-!- mikhailBelous [~quassel@37.98.242.99] has quit [Remote host closed the connection]22:51
cameron__hi,22:56
cameron__just wondering why when you fetch from github from shogun22:56
cameron__the data folder is empty22:56
@iglesiasgcameron__: it is in another repo, a submodule22:57
@iglesiasggit submodule update or so should work22:57
cameron__thanks23:02
cameron__Trying to update the submodule , but failed the travis build.23:50
@iglesiasgcameron__: mmm I don't understand23:53
@iglesiasgyou cannot get the data repository because the travis build failed?23:53
cameron__i got it23:53
cameron__made a change23:53
cameron__tried to do commit and push23:53
cameron__i did it inside in the submoduel folder, then in the shogun folder23:54
cameron__changes appeared on github. but didn't seem right23:54
cameron__https://github.com/goldbug/shogun/commit/7790dd146bf5e315faba55ff926d9f7e080b540823:54
--- Log closed Thu May 02 00:00:29 2013