--- Log opened Fri Feb 08 00:00:41 2013 | ||
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 244 seconds] | 00:57 | |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun | 01:11 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Ping timeout: 246 seconds] | 03:32 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 03:33 | |
-!- Netsplit *.net <-> *.split quits: syst3mw0rm, blackburn, sonne|work, shogun-buildbot, cavalier-, @sonney2k, zxtx, wiking | 04:13 | |
-!- sonney2k [~shogun@7nn.de] has joined #shogun | 05:25 | |
-!- ServerMode/#shogun [+o sonney2k] by zelazny.freenode.net | 05:25 | |
-!- blackburn [~blackburn@85.114.170.181] has joined #shogun | 05:26 | |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun | 05:27 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 05:27 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 05:27 | |
-!- cavalier- [~cavalier-@70.118.110.73] has joined #shogun | 05:27 | |
-!- sonne|work [~sonnenbu@sams-office-nat.tomtomgroup.com] has joined #shogun | 05:27 | |
-!- syst3mw0rm [~quassel@aws.aamirkhan.co.in] has joined #shogun | 05:27 | |
-!- syst3mw0rm [~quassel@aws.aamirkhan.co.in] has quit [Ping timeout: 246 seconds] | 05:31 | |
-!- syst3mw0rm [~quassel@aws.aamirkhan.co.in] has joined #shogun | 05:33 | |
-!- FSCV_ [~FSCV@204.45.132.2] has quit [Ping timeout: 255 seconds] | 07:41 | |
-!- FSCV [~FSCV@173.254.212.46] has joined #shogun | 07:55 | |
-!- FSCV [~FSCV@173.254.212.46] has quit [Excess Flood] | 07:58 | |
-!- FSCV [~FSCV@173.254.212.46] has joined #shogun | 07:58 | |
-!- n4nd0 [~nando@n134-p10.kthopen.kth.se] has joined #shogun | 09:05 | |
-!- sumit [ca4eaca2@gateway/web/freenode/ip.202.78.172.162] has joined #shogun | 09:48 | |
-!- sumit [ca4eaca2@gateway/web/freenode/ip.202.78.172.162] has quit [Quit: Page closed] | 09:59 | |
@sonney2k | blackburn, so you don't want gsoc to happen this year? | 10:19 |
---|---|---|
blackburn | sonney2k: why so? | 10:19 |
@sonney2k | blackburn, I didn't see any emails... | 10:20 |
blackburn | sonney2k: I wanted to discuss with you | 10:20 |
blackburn | sonney2k: wrote in jabber actually :D | 10:20 |
@sonney2k | I am not at work... | 10:21 |
@sonney2k | so no jabber | 10:21 |
blackburn | sonney2k: I am convincing igor carron now | 10:22 |
blackburn | to mentor something | 10:22 |
blackburn | http://nuit-blanche.blogspot.ru/ his blog | 10:22 |
@sonney2k | blackburn, that comes later | 10:23 |
@sonney2k | first announce our interest | 10:23 |
blackburn | okay | 10:23 |
@sonney2k | look at what I did last year... | 10:23 |
@sonney2k | (email archives to the list) | 10:23 |
blackburn | sonney2k: should I ask to do everything in public? | 10:23 |
@sonney2k | blackburn, but yes cool | 10:24 |
@sonney2k | blackburn, ? | 10:24 |
@sonney2k | HEAD^^ igor | 10:24 |
blackburn | well for example should I ask possible mentors to announce their interest | 10:24 |
blackburn | in public | 10:24 |
blackburn | or it is ok to contact me | 10:24 |
blackburn | I don't mind any | 10:24 |
@sonney2k | blackburn, what I did: | 10:25 |
blackburn | yeah I am searching that mail | 10:25 |
@sonney2k | 1) tell everyone on the mailinglist we want to participate in gsoc | 10:25 |
@sonney2k | 2) give them a list of old ideas | 10:26 |
@sonney2k | 3) mention the new ideas url that we want to update | 10:26 |
@sonney2k | 4) talk to everyone (phone/chat but in private) who could potentially mentor and what exactly | 10:26 |
@sonney2k | 5) update the ideas list | 10:26 |
@sonney2k | 6) apply for gsoc (write proposal) | 10:27 |
@sonney2k | 7) announce that we are in or not | 10:27 |
blackburn | sounds like an algorithm, thanks :) | 10:27 |
blackburn | sonney2k: so it seems I should take it over completely in any case? | 10:27 |
blackburn | like I should say I am responsible for all the stuff about that? | 10:28 |
blackburn | okay I'll write some big mail tonight | 10:30 |
@sonney2k | get stuff done man | 10:30 |
@sonney2k | not like helping me to fix this bug | 10:31 |
blackburn | sonney2k: ? | 10:31 |
blackburn | sonney2k: well I asked that cause if I don't say anything - you will get mails | 10:32 |
blackburn | which you'd need to redirect anyway then | 10:32 |
@sonney2k | blackburn, recall the bug about failing GPs? | 10:35 |
@sonney2k | blackburn, just post it publicly on the mailinglist - people will know what to do | 10:35 |
@sonney2k | and draft the gsoc2013 ideas list website | 10:36 |
blackburn | sonney2k: ahh that was irony meaning I didn't help you to fix the bug ;D | 10:56 |
blackburn | sonney2k: quite acid! | 10:57 |
@sonney2k | blackburn, which it was... | 11:01 |
blackburn | I keep not getting what you mean | 11:02 |
@sonney2k | blackburn, recall that you said you will have a look at the bug? | 11:03 |
blackburn | whatever | 11:03 |
blackburn | I said I tried to fix that but didn't managed | 11:03 |
blackburn | sonney2k: I need admin rights in shogun-toolbox org at github | 11:04 |
@sonney2k | the compute incremental hash function was called on elements of an empty array (a check was missing...) | 11:05 |
@sonney2k | blackburn, what for? | 11:05 |
blackburn | sonney2k: at least to control service hooks | 11:06 |
@sonney2k | which hook is missing? | 11:06 |
blackburn | haha okay | 11:06 |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 11:23 | |
shogun-notifier- | shogun: Soeren Sonnenburg :master * 680939f / src/Makefile.template: https://github.com/shogun-toolbox/shogun/commit/680939fd00c27f5fdc4890542ad3194784984e28 | 11:23 |
shogun-notifier- | shogun: fix (harmless) error messages on make install | 11:23 |
shogun-notifier- | shogun: | 11:23 |
shogun-notifier- | shogun: better avoid them to not confuse us users or devlopers :) | 11:23 |
wiking | fl.mmmm | 11:29 |
blackburn | wiking: hello to you too | 11:29 |
wiking | sonney2k: here? | 11:29 |
wiking | hellobello | 11:29 |
wiking | :> | 11:29 |
@sonney2k | wiking, wassup god of the north? | 11:30 |
wiking | sonney2k: unittesting | 11:31 |
wiking | sonney2k: have u seen somewhere that stuff explained? | 11:31 |
* sonney2k wants - check | 11:31 | |
wiking | sonney2k: ok | 11:31 |
wiking | sonney2k: i'll try to see how we can do it | 11:31 |
wiking | but afaik it'll only work with ubuntu/debian systems | 11:31 |
wiking | as i have no clue how the others are handling those | 11:31 |
wiking | i mean the gtest+gmock package | 11:32 |
@sonney2k | wiking, perfectly ok with me - IMHO none of the devs uses anything els | 11:34 |
@sonney2k | e | 11:34 |
wiking | sonney2k: meeee | 11:34 |
wiking | sonney2k: i use os | 11:34 |
wiking | osx | 11:34 |
@sonney2k | yeah but you can take care of yourself :D | 11:34 |
wiking | good | 11:35 |
n4nd0 | I am using arch | 11:35 |
wiking | n4nd0: hehehe ok then you can help me | 11:35 |
n4nd0 | but I can try to care of me too :D | 11:35 |
wiking | n4nd0: can u check if you have google unit test package? | 11:35 |
wiking | n4nd0: and google mock package? | 11:36 |
n4nd0 | wiking: I installed gtest this morning actually, I needed it for ROS - uni work | 11:36 |
wiking | n4nd0: ehehhe u use ROS? cool | 11:36 |
n4nd0 | wiking: I have started today | 11:36 |
wiking | n4nd0: ok can you send me in pastebin or somewhere what is the package content for gtest? | 11:37 |
n4nd0 | I am going to use it for my thesis | 11:37 |
wiking | n4nd0: when r u finishing? | 11:37 |
n4nd0 | wiking: http://code.google.com/p/googletest/ enough? | 11:37 |
n4nd0 | wiking: I have started this week :D | 11:37 |
wiking | n4nd0: mmmm yeah that one | 11:37 |
n4nd0 | I plan to be finished by August - September | 11:37 |
wiking | but do you have that in package on arch? | 11:37 |
wiking | or you installed it from src/svn? | 11:38 |
n4nd0 | wiking: yep, it is in arch | 11:38 |
n4nd0 | it can be installed with pacman | 11:38 |
n4nd0 | I believe it was only available through AUR not so long ago | 11:38 |
wiking | ok | 11:39 |
wiking | can u give me the content of package? | 11:39 |
wiking | sonney2k: have u heard of this: google-glog | 11:39 |
n4nd0 | wiking: let me see | 11:39 |
wiking | sonney2k: but imho log4cxx is more mature | 11:39 |
wiking | n4nd0: thnx just put the whole file list into pastebin or something | 11:40 |
wiking | n4nd0: with debian you would do dpkg -L <package name> | 11:40 |
@sonney2k | wiking, if I would switch then to log4cxx but there is no urgent need to do it | 11:40 |
wiking | but i don't know for arch | 11:40 |
n4nd0 | I had to installed that one too before log4cxx, it seems that people from ROS the same tools | 11:40 |
wiking | sonney2k: k | 11:40 |
blackburn | wiking: what can you mentor? | 11:40 |
n4nd0 | wiking: I have a few files here under /usr/include/gtest, I am not sure which one you need | 11:41 |
n4nd0 | stupid question probably hehe ^ | 11:41 |
wiking | blackburn: mmm well i'd happily mentor either a full redesign of optimization. to finally have a good general interface for optimizers | 11:41 |
wiking | blackburn: or continue working with SO or latent stuff | 11:41 |
n4nd0 | I'd really like to continue with SO! | 11:41 |
blackburn | wiking: can you please try to describe these ideas next week | 11:42 |
blackburn | ? | 11:42 |
n4nd0 | atm I am reading the tutorial that Cheng posted on the mailing list regarding SO in CV | 11:42 |
@sonney2k | hmmh does anyone know what special stuff happened on february 1st? we had >600 new visitors that day | 11:42 |
n4nd0 | sonney2k: something workshop related maybe? | 11:42 |
wiking | sonney2k: i guess it's because they've realised from last year | 11:43 |
@sonney2k | n4nd0, your email? | 11:43 |
wiking | sonney2k: that they need to have a PR before doing anything with shogun in gsoc ;) | 11:43 |
n4nd0 | sonney2k: I just checked and I sent it February the 2nd, 01:12 am CET | 11:43 |
wiking | sonney2k: you thaught them well ;) | 11:43 |
wiking | blackburn: till next week? | 11:43 |
wiking | n4nd0: ok so nothing else just /usr/include/gtest ? | 11:43 |
wiking | n4nd0: do you maybe have something under /usr/src? | 11:44 |
wiking | n4nd0: or maybe /usr/lib | 11:44 |
wiking | ? | 11:44 |
blackburn | wiking: till end of next week I think | 11:44 |
@sonney2k | all these new visitors came from china! | 11:44 |
blackburn | oh sh | 11:44 |
wiking | blackburn: okey | 11:44 |
blackburn | n4nd0: we'd need a mentor for you | 11:44 |
wiking | n4nd0: wantme? :) | 11:44 |
n4nd0 | wiking: I have stuff in /usr/src/gtest | 11:45 |
wiking | i mean as a mentor :) | 11:45 |
n4nd0 | wiking: sure :) | 11:45 |
wiking | n4nd0: ooooh COOOL | 11:45 |
blackburn | coupling you with wiking would be not so good I think - we need fresh people | 11:45 |
wiking | blackburn: :D | 11:45 |
wiking | blackburn: fresh meat :) | 11:45 |
blackburn | yeah in means of mentoring too | 11:45 |
wiking | n4nd0: ok then it's the same as in debian | 11:45 |
wiking | n4nd0: can u check if there's google mock package for arch? | 11:45 |
blackburn | however if you want it so much we'd think about it | 11:45 |
n4nd0 | wiking: these files are under /usr/src/gtest/src | 11:46 |
n4nd0 | wiking: gtest-all.cc gtest.h gtest_main.cc | 11:46 |
wiking | n4nd0: yep that's cool | 11:46 |
n4nd0 | wiking: I have not found it in pacman repos | 11:46 |
wiking | n4nd0: aha ok! | 11:46 |
n4nd0 | there should be sth in AUR I guess | 11:46 |
n4nd0 | let me see | 11:46 |
wiking | thnx | 11:46 |
@sonney2k | wiking, n4nd0, blackburn once we have some more tests active we should really release 2.1 | 11:47 |
n4nd0 | wiking: here you go https://aur.archlinux.org/packages/gmock/ | 11:47 |
@sonney2k | not like last year | 11:47 |
-!- alexlovesdata [~binder@2001:638:806:e001:eda1:c6f9:301a:5329] has joined #shogun | 11:47 | |
n4nd0 | I was thinking these days | 11:47 |
@sonney2k | last year we failed to release before n4nd0 made us drown in patches | 11:47 |
n4nd0 | what about a project just focused on examples?? | 11:47 |
n4nd0 | :( | 11:47 |
blackburn | alexlovesdata: hey there - good time to visit us | 11:47 |
@sonney2k | and so we had no more time to release | 11:47 |
blackburn | alexlovesdata: want student this year? | 11:47 |
wiking | sonney2k: yep | 11:48 |
n4nd0 | I think Georg Zeller did great last year helping me, we could ask him too | 11:48 |
wiking | n4nd0: thnx | 11:48 |
n4nd0 | wiking: do you need me to install it? | 11:49 |
@sonney2k | n4nd0, then please ask him if he has anything | 11:49 |
wiking | n4nd0: well i'm just checking now whether i can find out what's the content of this package... i guess from the tarball i can get that info | 11:49 |
n4nd0 | wiking: ok | 11:49 |
n4nd0 | sonney2k: ok, I am mailing him right now | 11:49 |
@sonney2k | wiking, well make it work for debian / ubuntu first then one step at a time... | 11:51 |
n4nd0 | blackburn: what is your @hsogun-toolbox.org address, I'll CC you | 11:51 |
wiking | sonney2k: yeah i'm on it now | 11:51 |
blackburn | n4nd0: sergey.lisitsyn@ | 11:51 |
n4nd0 | blackburn: ty | 11:52 |
@sonney2k | does anyone know what http://www.weibo.com/ is? | 11:53 |
blackburn | sonney2k: ?????????????? | 11:53 |
blackburn | how can you don't understand? | 11:53 |
blackburn | :D | 11:53 |
blackburn | if I press translate in my chrome it says it is microblogging | 11:54 |
@sonney2k | so someone blogged about shogun there I guess | 11:54 |
@sonney2k | alexlovesdata, did you ask about rooms at TU for the after workshop party? | 11:55 |
blackburn | http://qing.weibo.com/1653362857/628c50a9330015y2.html | 11:55 |
blackburn | may be that | 11:55 |
n4nd0 | published 2012-04-24 I think | 11:56 |
@sonney2k | blackburn, rather not yeah exactly^ | 11:56 |
blackburn | hah yeah | 11:56 |
blackburn | wrong | 11:56 |
@sonney2k | blackburn, btw if you want to know how shogun is used - check this http://scholar.google.de/scholar?hl=en&q=shogun+toolbox&btnG=&as_sdt=1%2C5&as_sdtp= | 12:01 |
@sonney2k | A food image recognition system with multiple kernel learning :D | 12:01 |
@sonney2k | How far can you get with a modern face recognition test set using only simple features? | 12:01 |
alexlovesdata | student for what, blackburn? | 12:01 |
blackburn | alexlovesdata: gsoc | 12:01 |
@sonney2k | Geotagged photo recognition using corresponding aerial photos with multiple kernel learning | 12:02 |
blackburn | sonney2k: old.. - just nothing from 2012 | 12:02 |
@sonney2k | A surf-based spatio-temporal feature for feature-fusion-based action recognition | 12:02 |
alexlovesdata | @sonney2k ... not yet ... I want to check that mean assert but now | 12:02 |
alexlovesdata | that mean assert bug now | 12:02 |
@sonney2k | look there are you :) ????????????? ???????? ?????? ? ??????? ?????? ??????? ???????? ? ?????????? ??????????????? ?????????? | 12:03 |
blackburn | sonney2k: lol | 12:03 |
blackburn | you scary people with these words | 12:04 |
blackburn | sonney2k: can you translate it? | 12:04 |
blackburn | with your mind I mean | 12:04 |
@sonney2k | almost | 12:05 |
@sonney2k | some words are missing | 12:05 |
@sonney2k | alexlovesdata, please put this next on the list then | 12:05 |
blackburn | sonney2k: which? | 12:05 |
@sonney2k | alexlovesdata, about the assertion failure | 12:05 |
@sonney2k | it assumes that 0.9 accuracy on that data set is possible | 12:05 |
@sonney2k | did anyone test if this is reasonable? | 12:06 |
@sonney2k | I mean what does KNN get on this dataset? | 12:06 |
alexlovesdata | it has to do with traffic sign detection | 12:06 |
alexlovesdata | @sonney2K: i finmd this non reasonable | 12:06 |
blackburn | sonney2k: alexlovesdata: may be some student polishing mkl? | 12:06 |
@sonney2k | blackburn, ????????????? ???????? | 12:06 |
alexlovesdata | because we use some data and some kernels and non-sparse MKL | 12:06 |
blackburn | sonney2k: recognition, road | 12:06 |
alexlovesdata | Hogs for traffic sign detection | 12:07 |
blackburn | alexlovesdata: bingo | 12:07 |
alexlovesdata | but i do not know opornyi vektor | 12:07 |
blackburn | alexlovesdata: stupid to write about that, isn't it? :D | 12:07 |
@sonney2k | alexlovesdata, support vector | 12:07 |
blackburn | alexlovesdata: support vector | 12:07 |
alexlovesdata | ohh, thx blackburn | 12:07 |
alexlovesdata | @sonney2k: I want to check the performance of that example | 12:08 |
blackburn | original vapnik's name :D | 12:08 |
alexlovesdata | and then if it is reasonably high with good kernels remove that assertion | 12:08 |
blackburn | not the name of vapnik I mean - but the method's name | 12:08 |
alexlovesdata | hehehe, I got it blackburn ... opornyi vektor ;) | 12:08 |
@sonney2k | blackburn, why did you say there is nothing from 2012 in the scholar list: Predicting the outcome of renal transplantation, Forensic identification with environmental samples ... | 12:08 |
@sonney2k | lots of | 12:08 |
blackburn | sonney2k: hm well there are some | 12:09 |
blackburn | top matches are from other libraries though | 12:09 |
alexlovesdata | @sonney2k: I will take care of the room question, ... for how many people and what days? | 12:09 |
alexlovesdata | aboutl iterature and things using shogun: my Plos ONE paper, my CVIU paper | 12:10 |
@sonney2k | alexlovesdata, I guess <20 people for the weekend sat/sunday | 12:10 |
alexlovesdata | @ sonney2k thanks! I willask andrea about our room first ... why not? | 12:10 |
@sonney2k | blackburn, let me guess weka and libsvm/liblinear? | 12:10 |
alexlovesdata | andrea = Klaus secretary | 12:10 |
blackburn | sonney2k: I mean other libraries that publish mention us | 12:10 |
blackburn | like 'toolbox for doing all the shit' | 12:11 |
blackburn | and they say there is other toolbox, shogun | 12:11 |
@sonney2k | blackburn, sure they have to compare against | 12:12 |
alexlovesdata | @blackburn: about Gsoc 2013 ... not sure yet ... too many things now above my head | 12:12 |
-!- blackburn [~blackburn@85.114.170.181] has quit [Read error: Connection reset by peer] | 12:12 | |
alexlovesdata | I will know better in march because then some things have been decided | 12:13 |
-!- blackburn [~blackburn@85.114.170.181] has joined #shogun | 12:13 | |
@sonney2k | alexlovesdata, when is your defense btw? | 12:16 |
alexlovesdata | @sonney2k 27.2. 15.30 Marchstr 23 room 4.0something | 12:18 |
alexlovesdata | when I know a party location for the evening i will announce it officially | 12:18 |
@sonney2k | alexlovesdata, quite some time still but good luck nevertheless | 12:19 |
alexlovesdata | room 4.033 | 12:21 |
alexlovesdata | thx | 12:21 |
-!- n4nd0 [~nando@n134-p10.kthopen.kth.se] has quit [Quit: leaving] | 12:32 | |
-!- n4nd0 [~nando@n134-p10.kthopen.kth.se] has joined #shogun | 12:34 | |
alexlovesdata | question: where is data in the sgonu git? | 12:35 |
alexlovesdata | in the shogun git? | 12:35 |
n4nd0 | alexlovesdata: it is in a git submodule | 12:36 |
n4nd0 | alexlovesdata: do you want to get it locally? | 12:36 |
alexlovesdata | yes, exactly nando | 12:36 |
n4nd0 | alexlovesdata: go to the root of your git repo and run | 12:37 |
n4nd0 | git submodule update --init | 12:37 |
alexlovesdata | ahh ok, ... it is cloning data | 12:38 |
n4nd0 | alexlovesdata: :) | 12:38 |
-!- FSCV_ [~FSCV@173.254.212.46] has joined #shogun | 12:39 | |
alexlovesdata | @sonney2k: the example works fine with me with errors > 0.93 | 12:40 |
alexlovesdata | but the results are non-deterministic | 12:41 |
alexlovesdata | so it can happen that sometimes it drops < 0.9 | 12:41 |
alexlovesdata | I suggestion to lower the assertion to 0.75 | 12:41 |
alexlovesdata | as a fix | 12:41 |
@sonney2k | alexlovesdata, ohh I see | 12:41 |
@sonney2k | alexlovesdata, maybe we better initialize the random number generator | 12:42 |
alexlovesdata | I will try to push a commit ... i must learn that anyway :) | 12:42 |
alexlovesdata | @sonney2k ... good idea | 12:42 |
-!- FSCV [~FSCV@173.254.212.46] has quit [Ping timeout: 260 seconds] | 12:42 | |
@sonney2k | alexlovesdata, I guess it is due to the way the x-validation split is chosen? | 12:42 |
alexlovesdata | I will try to init the random number generator and check whether this is deterministic then | 12:43 |
@sonney2k | kthx | 12:43 |
alexlovesdata | good! : CMath::init_random(1); yields a reproducible result | 12:51 |
alexlovesdata | above 0.9 | 12:51 |
alexlovesdata | I will push a whatever fix | 12:51 |
@sonney2k | alexlovesdata, that will be your first pull request right? did it just take 2 years to convince you github is not your enemy ;) | 12:58 |
alexlovesdata | hahaha, sonne, no ... I just needed some hours of time without haste ... nothing can be really appreciated when running too fast | 13:22 |
alexlovesdata | I will chekc how to init a pull request from my remote to shogun ... | 13:22 |
-!- n4nd0 [~nando@n134-p10.kthopen.kth.se] has quit [Ping timeout: 252 seconds] | 13:24 | |
alexlovesdata | I needed to understand how to update my remote/master from shogun's upstream master first | 13:24 |
blackburn | alexlovesdata: you need fork | 13:28 |
blackburn | do you have one? | 13:28 |
alexlovesdata | not yet I think | 13:36 |
alexlovesdata | I am reading http://git-scm.com/book/ch5-2.html for that | 13:37 |
-!- n4nd0 [~nando@n134-p10.kthopen.kth.se] has joined #shogun | 13:39 | |
blackburn | alexlovesdata: ask me in case you need any assist | 13:41 |
@sonney2k | blackburn can you merge it? | 14:04 |
blackburn | yeah sure | 14:05 |
blackburn | I was talking to alexlovesdata about github stuff | 14:05 |
shogun-notifier- | shogun: AlexBinder :master * dfe64a8 / examples/undocumented/libshogun/evaluation_cross_validation_multiclass_mkl.cpp: https://github.com/shogun-toolbox/shogun/commit/dfe64a849684b3801ce1d5999e954d4e1e07f0b6 | 14:05 |
shogun-notifier- | shogun: initialized random number generator for reproducible outcomes in the light of an assertion ASSERT(result->mean>0.9); | 14:05 |
shogun-notifier- | shogun: Sergey Lisitsyn :master * 253b5bc / examples/undocumented/libshogun/evaluation_cross_validation_multiclass_mkl.cpp: https://github.com/shogun-toolbox/shogun/commit/253b5bc7d3930dd70f9b0d434e68d8da47f7d495 | 14:05 |
shogun-notifier- | shogun: Merge pull request #859 from AlexBinder/master | 14:05 |
shogun-notifier- | shogun: | 14:05 |
shogun-notifier- | shogun: init random number generator in evaluation_cross_validation_multiclass_mkl.cpp | 14:05 |
@sonney2k | pending valgrind runs this should bring us back to normal. | 14:07 |
@sonney2k | Zzzz... | 14:07 |
alexlovesdata | do we have Kernel fisher discriminant in shogun? | 14:11 |
blackburn | no | 14:11 |
alexlovesdata | should be easy to implement ... ten lines in matlab when we use spectral regression | 14:11 |
blackburn | if you give matlab code to me I can implement that using eigen3 | 14:12 |
blackburn | or you can :) | 14:12 |
n4nd0 | or maybe it is a good patch for someone who wants to get involved in gsoc ;) | 14:12 |
blackburn | n4nd0: you? | 14:13 |
blackburn | :D | 14:13 |
n4nd0 | blackburn: haha why not | 14:13 |
n4nd0 | but I was thinking of new-comers | 14:13 |
blackburn | n4nd0: if you want to do some monotone work - feel free to check for some code in .h | 14:13 |
blackburn | you know what to do with that code ;) | 14:14 |
n4nd0 | aham | 14:14 |
n4nd0 | are we purifying the code from those things? | 14:14 |
n4nd0 | online inline methods to survive? | 14:14 |
n4nd0 | s/online/only | 14:14 |
blackburn | I don't see much help from inline methods | 14:15 |
alexlovesdata | what is .h ? | 14:15 |
n4nd0 | I did in my head only inline -> online | 14:15 |
n4nd0 | lol | 14:15 |
blackburn | alexlovesdata: extension, header :) | 14:15 |
alexlovesdata | ye ... but what has that t odo with monotone work? | 14:15 |
blackburn | the more code we have in .h the slower compilation is | 14:15 |
alexlovesdata | I know .h is a C header | 14:15 |
n4nd0 | exactly | 14:15 |
blackburn | alexlovesdata: we need to get all the code to .cpp | 14:15 |
blackburn | alexlovesdata: yes apparently .hpp is a proper extension | 14:16 |
alexlovesdata | I never code routines in .h | 14:16 |
alexlovesdata | not even inits ... | 14:16 |
blackburn | alexlovesdata: that's gut :) | 14:16 |
blackburn | I am the guilty who did that I think | 14:16 |
alexlovesdata | I could also try to fight with eigen3 myself ... but thanks you for the offer | 14:17 |
blackburn | alexlovesdata: feel free to ask me about eigen3 as well - I spent a lot of nights having luv time with eigen3 | 14:17 |
blackburn | there are some issues when you don't understand how to do something - it takes time to google | 14:18 |
alexlovesdata | ye ... true | 14:22 |
alexlovesdata | http://www.shogun-toolbox.org/doc/en/current/methods.html preprocessors lacks http://www.shogun-toolbox.org/doc/en/current/classshogun_1_1CRandomFourierGaussPreproc.html | 14:23 |
wiking | n4nd0: here? | 14:37 |
wiking | n4nd0: what is the make/configure env for ROS? cmake? | 14:37 |
n4nd0 | wiking: they use cmake yes | 14:38 |
wiking | n4nd0: ok | 14:39 |
wiking | n4nd0: thnx | 14:39 |
n4nd0 | np | 14:40 |
n4nd0 | wiking: what is it you are working on? | 14:40 |
n4nd0 | just curious | 14:40 |
wiking | n4nd0: trying to integrate into shogun to compile the gtest with the same flags as shogun is build with | 14:40 |
wiking | as that's the proper way to use gtest | 14:40 |
n4nd0 | aham I see | 14:40 |
blackburn | alexlovesdata: hah - >50% of methods are missing there | 14:45 |
-!- heiko [~heiko@nat-175-169.internal.eduroam.ucl.ac.uk] has joined #shogun | 14:45 | |
blackburn | I am trying to replace one pointer field in kernel with shared_ptr | 14:46 |
heiko | sonney2k, sonne|work, blackburn around? | 14:46 |
blackburn | it goes ok actually - a few changes | 14:46 |
blackburn | heiko: ja | 14:46 |
heiko | hi :) | 14:46 |
heiko | I just saw a cool talk | 14:46 |
blackburn | about love? | 14:46 |
heiko | on a hybrid cholesky algorithm which works on multiple core cpus and GPU | 14:46 |
blackburn | nerd | 14:46 |
blackburn | :D | 14:46 |
heiko | :) | 14:46 |
heiko | I talked the author into integrating that into shogun | 14:46 |
n4nd0 | hi heiko | 14:47 |
n4nd0 | is it CUDA? | 14:47 |
heiko | n4nd0, hi :) | 14:47 |
heiko | yes | 14:47 |
n4nd0 | nice | 14:47 |
wiking | mmm | 14:47 |
blackburn | heiko: may be he should rather integrate it to eigen | 14:47 |
heiko | quite interesting thing he did | 14:47 |
blackburn | heiko: cholesky is needed in gps, right? | 14:47 |
heiko | cholesky is needed everywhere where there is the word "Gaussian" | 14:48 |
blackburn | eigen has pretty nice cholesky anyway | 14:48 |
blackburn | are there such big matrices? | 14:48 |
heiko | blackburn, the GPU is used | 14:48 |
heiko | this changes things | 14:49 |
heiko | with eigen, you are right, but eigen has no cuda dependecies :) | 14:49 |
blackburn | heiko: how does he code gpu? it is a great matter | 14:49 |
heiko | cuda c | 14:49 |
blackburn | heiko: eigen runs on cuda | 14:49 |
heiko | really? | 14:49 |
heiko | mmh | 14:49 |
heiko | maybe then this would be a better idea you are right | 14:49 |
heiko | but you know what? | 14:49 |
blackburn | yeah gael works on some port | 14:49 |
heiko | it will be exclusive to shogun otherwise :D | 14:49 |
n4nd0 | haha | 14:50 |
heiko | *and* he could try to make it mature in a safe environment | 14:50 |
n4nd0 | SHOGUN - Machine learning and linear algebra toolbox :D | 14:50 |
heiko | and then once its working fine, put it into eigen | 14:50 |
heiko | n4nd0, well :) | 14:50 |
blackburn | cuda is nvidia property I am not the biggest fan of it | 14:50 |
n4nd0 | blackburn: why not? | 14:50 |
heiko | blackburn, agreed, but everyone has a GPU | 14:50 |
heiko | so one should use them | 14:50 |
blackburn | I don't | 14:50 |
blackburn | :D | 14:51 |
heiko | its like 20 times faster on 10000 matrices | 14:51 |
blackburn | HD4000 is not a gpu :D | 14:51 |
heiko | blackburn, doesnt it have opencl bindings? | 14:51 |
heiko | it is a gpu for sure | 14:51 |
blackburn | heiko: kind of useless | 14:52 |
blackburn | heiko: but we need gpu desperately - can't say anything against it | 14:52 |
n4nd0 | isn't any graphics card a GPU? | 14:52 |
heiko | blackburn, so what about the argument, that all the people were very interested in it | 14:52 |
heiko | n4nd0 I think yes | 14:52 |
blackburn | n4nd0: well sure but compare some nvidia board in big computer | 14:52 |
heiko | but cuda does only work on nvidia | 14:52 |
blackburn | and hd4000 in my notebook :D | 14:52 |
n4nd0 | blackburn: sure | 14:53 |
-!- SCV__ [~FSCV@187.210.54.165] has joined #shogun | 14:53 | |
-!- FSCV_ [~FSCV@173.254.212.46] has quit [Ping timeout: 255 seconds] | 14:53 | |
blackburn | heiko: I am not against that | 14:54 |
heiko | http://www.csml.ucl.ac.uk/events/76 | 14:54 |
heiko | have a look, it also depends whether the guy is willing to port it to shogun (requires at least some work) | 14:55 |
blackburn | heiko: we all need new technologies to have more freedom | 14:55 |
heiko | blackburn, indeed | 14:56 |
n4nd0 | blackburn: you made me think when you said the nvidia thing | 14:56 |
n4nd0 | I assumed that CUDA must be used if one wants to use GPU | 14:56 |
blackburn | n4nd0: no there is opencl | 14:57 |
blackburn | which is powerful as well | 14:57 |
n4nd0 | I see | 14:57 |
blackburn | and *open* :) | 14:57 |
n4nd0 | hehe | 14:58 |
blackburn | I don't like to be stallman but still | 14:58 |
n4nd0 | from wikipedia | 14:58 |
n4nd0 | "if the OpenCL implementation is correctly tweaked to suit the target architecture, it performs no worse than CUDA" | 14:58 |
n4nd0 | it sounds like CUDA tends to perform better :) | 14:58 |
blackburn | n4nd0: yes that's natural | 14:59 |
n4nd0 | but limited to nvidia | 14:59 |
n4nd0 | yes, I continued reading and it's true | 14:59 |
blackburn | you pay for generality | 14:59 |
heiko | cuda is completely closed | 14:59 |
heiko | the used to give away source code | 14:59 |
heiko | but not anymore | 14:59 |
heiko | the guy just told me | 14:59 |
heiko | so this is a bad thing in general, I agree with blackburn | 14:59 |
blackburn | I can't run cuda on my machine | 14:59 |
heiko | blackburn, the guy also talked about opencl, seems to be not too complicated to port it from cude to opencl | 15:00 |
n4nd0 | how different is to program in OpenCL and CUDA? | 15:00 |
heiko | maybe we can convince him | 15:00 |
n4nd0 | aham | 15:00 |
blackburn | quite different I think - it depends on the complexity | 15:00 |
heiko | n4nd0, I think its very similar, though opencl may not have all the features | 15:00 |
blackburn | hah | 15:00 |
heiko | blackburn, lol :) | 15:00 |
blackburn | well | 15:00 |
blackburn | if it is matrix algorithm | 15:00 |
blackburn | shoudl be similar I believe | 15:01 |
heiko | yes, this is what I meant | 15:01 |
blackburn | in case of something with kernels - it could get worse | 15:01 |
heiko | kernels on gpus? | 15:01 |
heiko | this is crazy :) | 15:01 |
blackburn | heiko: gpu kernels | 15:01 |
heiko | haha :) | 15:01 |
heiko | I see | 15:01 |
blackburn | you can tailor some code to run on gpu | 15:01 |
blackburn | using C | 15:01 |
n4nd0 | blackburn: both cuda and opencl use the concept of kernels right? | 15:01 |
blackburn | I think so n4nd0 | 15:01 |
heiko | Metropolis-Hasting-Kernel, Gaussian Kernel, Covariance Kernek, GPU-kernel, Linux-kernel, Kernel of a linear map | 15:02 |
heiko | aaaaaah | 15:02 |
n4nd0 | haha people seem to like that word | 15:02 |
blackburn | heiko: epanechnikov kernel | 15:02 |
blackburn | heiko: in density estimation | 15:02 |
heiko | that sound too Russian for me :) | 15:02 |
heiko | bb in 5 mins | 15:02 |
blackburn | heiko: it is | 15:02 |
n4nd0 | blackburn: AFAIK you need kernels to use the GPU at least in CUDA | 15:03 |
blackburn | n4nd0: yes but you don't have to write complex stuff if you do something with matrices I think | 15:03 |
n4nd0 | aham | 15:04 |
blackburn | n4nd0: however as I see you'd need to write kernel anyway yes | 15:04 |
n4nd0 | yeah | 15:04 |
blackburn | there is an idea that keep me nervous | 15:05 |
n4nd0 | which one? | 15:05 |
blackburn | I am curious if it is possible to write algorithms using some 'workers' | 15:05 |
n4nd0 | blackburn: tell me more :) | 15:06 |
blackburn | so you want to train svm which means you run a few workers that do that | 15:06 |
blackburn | they should be kind of actors or os | 15:06 |
n4nd0 | mmm | 15:06 |
blackburn | that idea is so simple but I can't understand how to do that in general | 15:06 |
n4nd0 | workers are to be run in parallel I understand | 15:06 |
blackburn | not only parallel but also distributed | 15:06 |
blackburn | proablby mahout works somehow like that | 15:07 |
n4nd0 | aham | 15:07 |
n4nd0 | what is it that you don't understand how to make it general? | 15:07 |
blackburn | would be so cool to have something like general principle of doing that | 15:08 |
heiko | re | 15:09 |
blackburn | probably I am talking about map-reduce :D | 15:09 |
n4nd0 | blackburn: hehe maybe | 15:09 |
blackburn | I wouldn't like to write map-reduce code - it is mahout stuff | 15:09 |
n4nd0 | blackburn: you can use frameworks like hadoop for the map/rduce part | 15:10 |
heiko | btw I advertised the workshop a bit here | 15:10 |
blackburn | n4nd0: that's what mahout does | 15:10 |
heiko | few people said they might come | 15:10 |
n4nd0 | heiko: that's cool :) | 15:10 |
n4nd0 | blackburn: true | 15:11 |
heiko | all the kernel people are here in London, so we might already have two of them :) I will also harass the others | 15:11 |
n4nd0 | heiko: what kernel are we talking about now? :D | 15:11 |
heiko | haha :D | 15:11 |
heiko | linux-kernel | 15:11 |
blackburn | so, heiko, all mercer cult fellows? | 15:12 |
heiko | blackburn, yes, more the RKHS guys with characteristic kernels, distribution embeddings, kernels on MDPs | 15:12 |
heiko | etc | 15:12 |
blackburn | hare krishna hare RKHS | 15:12 |
heiko | I will also invite John-Shawe-Taylor :) | 15:12 |
heiko | but he is ever busy | 15:12 |
blackburn | no idea who it is | 15:13 |
heiko | made them popular for ML | 15:13 |
blackburn | nice | 15:13 |
heiko | blackburn, I also have some other ideas on things to integrate into shogun | 15:14 |
heiko | one being the computation of determinants of very large matrices | 15:14 |
blackburn | heiko: what about this year gsoc? | 15:14 |
heiko | blackburn, what do you mean? | 15:14 |
blackburn | heiko: student/mentor? :) | 15:14 |
heiko | blackburn, I can do a mentor this year | 15:14 |
blackburn | you probably have two choices only | 15:14 |
heiko | since I have a job finally :) | 15:15 |
blackburn | yeah same here | 15:15 |
heiko | nice | 15:15 |
heiko | Could also implement stuff myself, but maybe its less stressful to tell others what to do | 15:15 |
blackburn | gsoc have the same monetary value for me now | 15:15 |
blackburn | so I wouldn't like to left my job because of gsoc :) | 15:16 |
heiko | yeah, I mean are you even a student now? | 15:16 |
blackburn | heiko: I am | 15:16 |
heiko | I am not, would have to become one in order to participate | 15:16 |
heiko | oh, ok | 15:16 |
heiko | hopefully we can manage the gsoc this year | 15:16 |
blackburn | and will be for an year more after that | 15:16 |
blackburn | :D | 15:16 |
heiko | but I will have a bit more time from spring on | 15:16 |
blackburn | I could participate twice more | 15:16 |
n4nd0 | heiko: but you are phd student now right? | 15:16 |
heiko | n4nd0, not yet | 15:17 |
heiko | hopefully in sept | 15:17 |
n4nd0 | heiko: aaaah ok | 15:17 |
heiko | just had my interview last week | 15:17 |
n4nd0 | oh, I wish you good luck with the selection then! | 15:17 |
heiko | thanks :) | 15:17 |
n4nd0 | aiming at UCL? | 15:17 |
heiko | yes, http://www.gatsby.ucl.ac.uk/ | 15:17 |
blackburn | heiko: I will announce gsoc initiative tonight | 15:18 |
heiko | but its quite competetive, the alternative is here: http://www.ucl.ac.uk/statistics/people/heiko-strathmann | 15:18 |
heiko | blackburn, initiative? like ideas etc? | 15:18 |
heiko | when do we have to apply? | 15:18 |
blackburn | heiko: gsoc is not yet announced :) | 15:18 |
heiko | n4nd0, how are your studies going? | 15:18 |
heiko | blackburn, so what will you announce? | 15:19 |
blackburn | heiko: that we are going to apply | 15:19 |
heiko | isn't it better to wait until it is announced? | 15:19 |
n4nd0 | heiko: fine, I started this week my master thesis at KTH finally | 15:19 |
blackburn | heiko: sonney2k doesn't think so | 15:19 |
heiko | n4nd0, nice master in? CS? | 15:19 |
heiko | blackburn, well ok, actually I would agree | 15:19 |
n4nd0 | heiko: the plan is to be done with the few things I have left for Spain by May-June and finish the master thesis by August-September | 15:20 |
heiko | since we are currently slow with everything | 15:20 |
blackburn | heiko: but I am checking for mentorship already | 15:20 |
n4nd0 | heiko: half way between CS and EE, it is a robotics master | 15:20 |
heiko | n4nd0, very cool! | 15:20 |
n4nd0 | heiko: but I do CS stuff | 15:20 |
heiko | you have to show some cool robot videos | 15:20 |
n4nd0 | heiko: I have something from a course I did a couple of years ago | 15:20 |
heiko | n4nd0, are you using your special svms somewhere in that? | 15:21 |
n4nd0 | heiko: no hehe | 15:21 |
blackburn | heiko: I wanted to talk about GPs :D | 15:21 |
blackburn | heiko: just in a few words - why? | 15:21 |
blackburn | what makes it special? | 15:21 |
heiko | n4nd0, show me, I also have one cool video of an AR-drone | 15:21 |
heiko | blackburn, what? | 15:21 |
blackburn | heiko: why should I choose GP and when? | 15:21 |
n4nd0 | heiko: http://www.youtube.com/watch?v=oqfwsNfYS8A | 15:22 |
n4nd0 | heiko: this is very crappy compared to a drone hehe | 15:22 |
n4nd0 | but we built it from scratch | 15:22 |
blackburn | n4nd0: you never showed that to me - it was a mistake I married you! | 15:22 |
heiko | n4nd0, cool! :) | 15:22 |
n4nd0 | blackburn: haha | 15:22 |
n4nd0 | I think I had posted that video around here before! | 15:22 |
n4nd0 | heiko: this is the lab where I am at KTH http://www.kth.se/en/csc/forskning/cvap | 15:23 |
heiko | cool stuff | 15:23 |
n4nd0 | I want to become a phd here too | 15:23 |
heiko | blackburn, well you can use them to model complex functions | 15:24 |
heiko | within a probabilistic framework | 15:24 |
heiko | so quite flexible, they are used everywhere | 15:24 |
heiko | and fully Bayesian, which is sometimes very useful | 15:24 |
blackburn | heiko: my gf is working with hyperspectral images (d ~80) - should I recommend to try GPs? | 15:25 |
blackburn | task is classification mostly | 15:25 |
heiko | what is the goal? | 15:25 |
blackburn | heiko: top accuracy I think | 15:25 |
heiko | to obtain high-accuracy predictions? | 15:25 |
heiko | then better use an SVM | 15:25 |
heiko | although they are similar | 15:25 |
blackburn | but training samples are very noised | 15:25 |
heiko | but if you are interested in predictive distributions, Bayesian kernel selecton or similar, use GPs | 15:25 |
blackburn | heiko: so it is still kind of theoretical? | 15:26 |
heiko | blackburn, no, just depends on what you want | 15:26 |
heiko | predictive distributions can be very useful | 15:27 |
blackburn | heiko: how can one use it? | 15:27 |
heiko | since you for example have probabilities for multiclass problems | 15:27 |
heiko | very useful | 15:27 |
heiko | but not so much for binary decisions | 15:27 |
heiko | and kernel selection also, you can have a posterior distribution over Gaussian kernel sizes in every dimension for example | 15:27 |
blackburn | heiko: predictions are multiclass in that case (water, asphalt, ...) | 15:28 |
heiko | blackburn, use svms, much easier if you just want to predict :) | 15:29 |
blackburn | heiko: well her supervisor wants to promote some method so they use it yet | 15:29 |
heiko | blackburn, probabilitites can be useful | 15:29 |
heiko | since you can add a decision "dont know" | 15:29 |
blackburn | yes that's clear | 15:29 |
n4nd0 | heiko: wow the drone is following his face? | 15:29 |
heiko | n4nd0, yes | 15:30 |
n4nd0 | heiko: that's pretty cool | 15:30 |
heiko | using opencv | 15:30 |
heiko | this ar drone is really cool | 15:30 |
blackburn | heiko: is it you on the video with long hair? :) | 15:30 |
heiko | yeh with the red trousers, thats me :) | 15:30 |
heiko | about 1,5 yrs ago | 15:31 |
blackburn | looks pink :D | 15:31 |
heiko | now I look more exhausted ;) | 15:31 |
blackburn | heiko: even more? | 15:31 |
blackburn | :) | 15:31 |
heiko | hah | 15:31 |
blackburn | on the 0006 it detected your groin heiko | 15:32 |
heiko | that was a cool thing, this printer company, oce, bought us a few of those drones and sponsored us playing with them for one week - in holland | 15:32 |
n4nd0 | oh that's nice! | 15:32 |
blackburn | heiko: did you implement that? | 15:33 |
blackburn | it looks like you were kinda presenting it | 15:33 |
heiko | blackburn, yes | 15:34 |
heiko | used opencv to track the face | 15:34 |
heiko | then generated movement commands from that | 15:34 |
blackburn | heiko: typical VJ stuff? | 15:34 |
heiko | where the difficulty was to estimate the distance to the face with one camera | 15:34 |
heiko | since face size is not linear to distance | 15:34 |
blackburn | heiko: yes d^2 | 15:34 |
heiko | did a power-regression | 15:34 |
heiko | blackburn, no interestingly not | 15:35 |
heiko | another function | 15:35 |
blackburn | hmm yes sure | 15:35 |
blackburn | I am wrong | 15:35 |
heiko | and with that we got a +-10cm estimate | 15:35 |
blackburn | it looks like something else | 15:35 |
heiko | which can be used to move back and forth | 15:35 |
blackburn | I can imagine a plot of that function | 15:35 |
alexlovesdata | hey n4ndo ... you are also a computer vision guy ... | 15:35 |
heiko | very simple, but looks cool | 15:35 |
blackburn | well when it comes to say 5 meters | 15:35 |
blackburn | 5.5 and 4.5 - no matter | 15:35 |
blackburn | size won't change at all :D | 15:36 |
heiko | blackburn, yes also the camera resolution is horrible as you can see | 15:36 |
blackburn | alexlovesdata: what to do if not CV | 15:36 |
blackburn | :D | 15:36 |
heiko | but our "data" (measured distance by hand) was explained with almost no error | 15:36 |
heiko | guys, I gotta do some work now, see you later! | 15:36 |
blackburn | cu | 15:36 |
n4nd0 | bye! | 15:37 |
n4nd0 | alexlovesdata: I am more into robotics localization, mapping, SLAM and that stuff | 15:37 |
n4nd0 | but yeah there is lot of CV used in that too | 15:37 |
n4nd0 | kinect sensors, kinect sensors eveywhere :) | 15:37 |
@sonney2k | alexlovesdata, I just git pull'd and guess what I still have the issue plus a couple more due to blackburn's arpack removal | 15:51 |
@sonney2k | and fibonaccyheap and convertree | 15:52 |
blackburn | sonney2k: ? | 15:52 |
blackburn | shit | 15:52 |
n4nd0 | what happens with them? | 15:52 |
blackburn | sonney2k: 2 mins | 15:52 |
alexlovesdata | you still have that assertion error, sonney2k ?? | 15:54 |
alexlovesdata | what accuracy is printed before the assert fails? | 15:54 |
alexlovesdata | I get a 0.94 ... | 15:54 |
shogun-notifier- | shogun: Sergey Lisitsyn :master * 7d0a59b / examples/undocumented/libshogun/ (3 files): https://github.com/shogun-toolbox/shogun/commit/7d0a59b5373bccd869b44bbdc370e62319349569 | 15:55 |
shogun-notifier- | shogun: Removed fibheap and arpack examples | 15:55 |
@sonney2k | alexlovesdata, 0.76 | 15:57 |
@sonney2k | blackburn, btw why did you remove arpack / fibheap? | 15:57 |
blackburn | sonney2k: why are not needed anymore | 15:58 |
blackburn | duplicated | 15:58 |
blackburn | sonney2k: I strongly don't expect we use it somewhere out of tapkee | 15:58 |
@sonney2k | I see ok | 15:59 |
blackburn | so the only reason they appeared in shogun was dimred | 15:59 |
blackburn | fibonacci heap stayed in tapkee | 15:59 |
blackburn | arpack wrapper was replaced with better one | 15:59 |
wiking | ok here comes a semi-working google unit tester | 15:59 |
@sonney2k | alexlovesdata, did you git pull and have > 0.9 accuracy? | 15:59 |
@sonney2k | alexlovesdata, look the buildbot is also not happy http://shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/655/steps/test%20libshogun/logs/stdio | 16:00 |
wiking | sonney2k: do we have a $() variable for ar, i.e. static linking | 16:01 |
n4nd0 | cu later people | 16:02 |
-!- n4nd0 [~nando@n134-p10.kthopen.kth.se] has quit [Quit: leaving] | 16:03 | |
@sonney2k | wiking, hmmhh don't think so | 16:04 |
blackburn | wtf | 16:05 |
blackburn | covertree | 16:05 |
blackburn | :E | 16:05 |
blackburn | not used anymore as well - sonney2k | 16:05 |
shogun-notifier- | shogun: Sergey Lisitsyn :master * 9c17a1d / examples/undocumented/libshogun/ (2 files): https://github.com/shogun-toolbox/shogun/commit/9c17a1d8f0fb20da361ba58046a4e32874850a54 | 16:06 |
shogun-notifier- | shogun: Removed covertree example | 16:06 |
@sonney2k | wiking, ar is hardcoded - just found it line 576 in Makefile.template | 16:07 |
@sonney2k | wiking, feel free to introduce a variable for it | 16:07 |
blackburn | sonney2k: would you like if we put shared_ptr from boost to our lib? | 16:07 |
@sonney2k | blackburn, haha have you hear yandex is more popular than bing :D | 16:08 |
blackburn | sonney2k: actually it looks like changing internal (class field) pointers to shared_ptrs is not that difficult | 16:08 |
@sonney2k | blackburn, this is something we cannot do inside gsoc | 16:08 |
blackburn | sonney2k: no, really? bing sucks I knew that | 16:09 |
blackburn | but I don't use yandex at all | 16:09 |
@sonney2k | I am afraid this might be much more work | 16:09 |
blackburn | sonney2k: it could introduce bugs - that's I am more afraid of - but I changed one field in CKernel to shared_ptr and just had to change a few casts in siblings | 16:10 |
blackburn | and getter/setter | 16:10 |
-!- SCV__ [~FSCV@187.210.54.165] has quit [Quit: Leaving] | 16:10 | |
@sonney2k | and I would not want to risk to get the same chaos we had last gsoc when we had all these transitions not yet done | 16:10 |
@sonney2k | if we want to do such stuff then directly after gsoc | 16:10 |
blackburn | okay | 16:10 |
@sonney2k | blackburn, hope that is OK... I recall that you became quite desperate last summer when nothing worked... | 16:11 |
blackburn | sonney2k: I don't remember that exactly | 16:11 |
@sonney2k | blackburn, did we have chatlogs back then? I certainly remember... | 16:12 |
blackburn | sonney2k: well I am desperate 90% of time ;D | 16:12 |
-!- FSCV [~FSCV@187.210.54.165] has joined #shogun | 16:12 | |
blackburn | but yes I agree it is not the time to do that | 16:13 |
@sonney2k | gtg | 16:13 |
-!- travis-ci [~travis-ci@ec2-23-22-8-110.compute-1.amazonaws.com] has joined #shogun | 16:14 | |
travis-ci | [travis-ci] it's Sergey Lisitsyn's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/4667693 | 16:14 |
-!- travis-ci [~travis-ci@ec2-23-22-8-110.compute-1.amazonaws.com] has left #shogun [] | 16:14 | |
-!- alexlovesdata [~binder@2001:638:806:e001:eda1:c6f9:301a:5329] has left #shogun [] | 16:16 | |
wiking | mmm this is a shit | 16:36 |
wiking | gtest is as a source in ubuntu | 16:36 |
wiking | but not gmock :S | 16:36 |
blackburn | wiking: yes but it was known before? | 16:43 |
wiking | blackburn: well yes and no.... :) | 16:59 |
-!- sumit [ca4eaca2@gateway/web/freenode/ip.202.78.172.162] has joined #shogun | 17:36 | |
wiking | ok i think this is more or less done | 17:37 |
wiking | one will have to have google mock source as well at hand to use it | 17:38 |
wiking | the package in ubuntu is faulty | 17:38 |
shogun-notifier- | shogun: vigsterkr :master * 0c0581e / src/configure,tests/unit/Makefile: https://github.com/shogun-toolbox/shogun/commit/0c0581eedaeab0beb973241535788e350a083914 | 17:43 |
shogun-notifier- | shogun: Fix google mocking framework detection | 17:43 |
shogun-notifier- | shogun: Google unit testing and mocking framework has to be compiled | 17:43 |
shogun-notifier- | shogun: with the same compiler flags as the code you are testing. | 17:43 |
shogun-notifier- | shogun: Viktor Gal :master * 24a7529 / src/configure,tests/unit/Makefile: https://github.com/shogun-toolbox/shogun/commit/24a7529c8af20750f31d316fd098b9c98b3518fb | 17:43 |
shogun-notifier- | shogun: Merge pull request #860 from vigsterkr/utest | 17:43 |
shogun-notifier- | shogun: | 17:43 |
shogun-notifier- | shogun: Fix google mocking framework detection | 17:43 |
wiking | someone shold test is | 17:44 |
wiking | shoudl test this as well | 17:44 |
wiking | works for me on ubuntu LTS 12.x | 17:44 |
-!- travis-ci [~travis-ci@ec2-23-22-8-110.compute-1.amazonaws.com] has joined #shogun | 17:56 | |
travis-ci | [travis-ci] it's Viktor Gal's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/4669658 | 17:56 |
-!- travis-ci [~travis-ci@ec2-23-22-8-110.compute-1.amazonaws.com] has left #shogun [] | 17:56 | |
-!- heiko [~heiko@nat-175-169.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.] | 18:05 | |
-!- heiko [~heiko@nat-175-169.internal.eduroam.ucl.ac.uk] has joined #shogun | 18:05 | |
wiking | heiko: there you go | 18:34 |
wiking | heiko: unit testing with gmock | 18:34 |
heiko | wiking, the unit test thing? | 18:34 |
heiko | wiking, you are awesome! | 18:34 |
heiko | I will use that for the stuff I am currently doing, after figuring out how | 18:34 |
wiking | heiko: ok | 18:34 |
heiko | have some time on the weekend, nice work | 18:34 |
wiking | heiko: note that you'lll need google mock source... the package in ubuntu is not good | 18:35 |
heiko | ok | 18:35 |
heiko | just the source? | 18:35 |
wiking | yeps | 18:35 |
wiking | and of course gtest src as well | 18:35 |
wiking | but for that the ubuntu package is good | 18:35 |
heiko | ok | 18:36 |
heiko | cool | 18:36 |
heiko | will try out soon | 18:36 |
heiko | this will help us a lot | 18:36 |
heiko | once it works fine, we can only accept commits if they include some tests | 18:36 |
heiko | so we wont have to rely on those big picture tests in the c++ examples | 18:36 |
blackburn | wiking: I got may be stupid question | 18:37 |
wiking | blackburn: go ahead | 18:38 |
blackburn | okay we have latent svm | 18:38 |
blackburn | what is the difference between it and sliding window detector? | 18:38 |
blackburn | do I understand right it is the same at some point? | 18:38 |
blackburn | it is rather a matter of training | 18:39 |
blackburn | right? | 18:39 |
wiking | mm | 18:46 |
wiking | well the detection is quite similar | 18:47 |
wiking | it's just rahter yeah the training | 18:47 |
blackburn | wiking: have you seen that paper? http://rogerioferis.com/VisualRecognitionAndSearch/material/Class8Pedestrian3.pdf | 18:48 |
blackburn | wiking: pedestrian detection at 45 fps | 18:48 |
shogun-buildbot | build #777 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/777 blamelist: Viktor Gal <vigsterkr@gmail.com> | 18:55 |
-!- heiko [~heiko@nat-175-169.internal.eduroam.ucl.ac.uk] has left #shogun [] | 19:09 | |
-!- sumit [ca4eaca2@gateway/web/freenode/ip.202.78.172.162] has quit [Ping timeout: 245 seconds] | 19:18 | |
shogun-notifier- | shogun: Soeren Sonnenburg :master * 159ddb0 / / (4 files): https://github.com/shogun-toolbox/shogun/commit/159ddb02b278342f1b37097f4bd2cb3d12de061d | 19:40 |
shogun-notifier- | shogun: fix miscelleaneous erros/warnings | 19:40 |
shogun-notifier- | shogun: | 19:40 |
shogun-notifier- | shogun: - fix example compiler error | 19:40 |
shogun-notifier- | shogun: - fix doxygen warnings | 19:40 |
shogun-notifier- | shogun: - fix compiler flag | 19:40 |
@sonney2k | wiking, which packages does one need to install for gmock/gtest? | 19:43 |
@sonney2k | libgtest-dev and google-mock? | 19:43 |
wiking | sonney2k: apt-get install libgtest-dev | 19:45 |
wiking | and download gmock src | 19:45 |
wiking | as the one in ubuntu is broken (does not provide sources under /usr/src) | 19:46 |
wiking | sonney2k: http://googlemock.googlecode.com/files/gmock-1.6.0.zip | 19:46 |
wiking | and then just | 19:46 |
@sonney2k | it is there /usr/src/gmock/src | 19:46 |
wiking | is it? | 19:46 |
@sonney2k | debian here though | 19:46 |
wiking | ah ok | 19:46 |
wiking | ubuntu does not give have that | 19:46 |
wiking | but yeah | 19:46 |
shogun-buildbot | build #778 of deb3 - modular_interfaces is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/778 | 19:47 |
wiking | by default it treis to find it under /usr/src/gmock | 19:47 |
wiking | so then it should be fine for you | 19:47 |
@sonney2k | g++ -I/usr/src/gmock -c /usr/src/gmock/src/gmock-all.cc | 19:47 |
wiking | nono | 19:47 |
wiking | don't do that | 19:47 |
wiking | just cd tests/unit | 19:47 |
wiking | make | 19:47 |
@sonney2k | g++ -I/usr/src/gtest -c /usr/src/gtest/src/gtest-all.cc | 19:47 |
wiking | after ./configure | 19:47 |
@sonney2k | wiking, no I mean it fails to find things on configure | 19:47 |
wiking | mmm | 19:48 |
wiking | what does it say | 19:48 |
wiking | i mean configure | 19:48 |
@sonney2k | brb | 19:48 |
wiking | Checking for Google C++ Mocking Framework ... yes | 19:49 |
wiking | if it's not yes then plz copy-paste here the error line | 19:49 |
blackburn | check da announcement | 19:51 |
wiking | which? | 19:52 |
blackburn | at mailing list | 19:54 |
@sonney2k | wiking, it just says | 20:00 |
@sonney2k | ============ Checking for Google C++ Mocking Framework ============ | 20:01 |
@sonney2k | Result is: no | 20:01 |
@sonney2k | you could have made it a little more verbose :-] | 20:01 |
@sonney2k | wiking, can I split it up in separate gmock/ gtest tests? | 20:04 |
@sonney2k | wiking, hmmhh afk? then let me just do it | 20:05 |
blackburn | http://www.google-melange.com/gsoc/homepage/google/gsoc2013 hah google can't suggest another theme this time | 20:05 |
@sonney2k | blackburn, so they are running out of money? | 20:06 |
blackburn | sonney2k: that's a great implication out of my mind | 20:07 |
blackburn | too complex for me :) | 20:07 |
blackburn | may be it feels ok for them to keep old graphics | 20:07 |
* sonney2k continous the gmock/gtest endeavor | 20:07 | |
blackburn | sonney2k: we should name shogun versions by bacterias | 20:08 |
blackburn | sonney2k: escherichia coli | 20:08 |
@sonney2k | blackburn, or virii | 20:08 |
@sonney2k | but only the very bad bad ones | 20:08 |
blackburn | sonney2k: yes, may be 3.0 should be called HIV | 20:09 |
@sonney2k | blackburn, when we introduce self modifying codes and template meta programming then it surely will be! | 20:09 |
blackburn | sonney2k: we should take some bacteria branch with increasing complexity | 20:10 |
blackburn | and then put next name each time we release :) | 20:10 |
wiking | sonney2k: i've tried that | 20:16 |
wiking | sonney2k: i mean splitting | 20:16 |
@sonney2k | wiking, give a a few more minutes... | 20:17 |
wiking | sonney2k: but one without the ohter is useless | 20:17 |
wiking | sonney2k: but not in configure.log | 20:17 |
wiking | sonney2k: the simple output of configure | 20:17 |
wiking | sonney2k: as there's more info where it actually died | 20:17 |
wiking | sonney2k: ah ok i see it now | 20:19 |
wiking | i forgot to print an error | 20:19 |
shogun-notifier- | shogun: Soeren Sonnenburg :master * 77862dd / src/configure: https://github.com/shogun-toolbox/shogun/commit/77862dda510a8c39790cd36d7de46a14d035e84c | 20:25 |
shogun-notifier- | shogun: hopefully fix gmock/gtest detection | 20:25 |
@sonney2k | wiking, please try if it works ^ | 20:25 |
@sonney2k | wiking, I think we might want to make this more robust and actually run a cxx_check compiling something like g++ -I/usr/src/gtest -c /usr/src/gtest/src/gtest-all.cc | 20:27 |
@sonney2k | blackburn, nice announcement :) | 20:29 |
blackburn | thanks | 20:29 |
wiking | sonney2k: well | 20:30 |
@sonney2k | blackburn, now we only need the ideas url | 20:30 |
@sonney2k | wiking, ? | 20:30 |
blackburn | sonney2k: I'll add one once we get some | 20:30 |
@sonney2k | blackburn, just copy it from last year for now | 20:30 |
wiking | sonney2k: works | 20:30 |
@sonney2k | recall everyone can edit now so it is a good start | 20:31 |
@sonney2k | in particular the optimization thing is still valid | 20:31 |
@sonney2k | and I want someone to code examples and write d3js / django interactive code | 20:31 |
@sonney2k | for some show down | 20:31 |
@sonney2k | let me ask Mr. ICA | 20:33 |
blackburn | sonney2k: please cc me if that's ok | 20:33 |
@sonney2k | if he would be interested in getting ICA in | 20:33 |
blackburn | sonney2k: ICA is pretty small | 20:34 |
blackburn | even fastICA is not that much code | 20:34 |
@sonney2k | there are tons of algorithms for that | 20:34 |
@sonney2k | a. ziehe - tdsep | 20:35 |
blackburn | sonney2k: haha I didn't expect you write in that lang | 20:40 |
@sonney2k | blackburn, I can hardly talk in $OTHERLANG to a friend :D | 20:41 |
blackburn | sonney2k: google translate helped me a little | 20:41 |
@sonney2k | blackburn, lets hope AZ is in. he did some cool stuff on that matter and has some nice examples too! | 20:45 |
blackburn | sonney2k: nice | 20:46 |
@sonney2k | blackburn, if I wasn't afraid of multiboost I would ask them | 20:46 |
blackburn | sonney2k: why you are afraid? | 20:46 |
@sonney2k | mission impossible | 20:46 |
@sonney2k | we also need some small ideas to be implemented for potential candidates... | 20:47 |
@sonney2k | wiking, ok nice | 20:47 |
@sonney2k | wiking, how do I run a test now? | 20:48 |
blackburn | sonney2k: what is impossible in multiboost? | 20:49 |
wiking | sonney2k: in ./tests/unit run make | 20:50 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 20:50 | |
blackburn | sonney2k: oh gosh that's a list! | 20:54 |
@sonney2k | blackburn, https://www.ohloh.net/p/shogun | 20:54 |
blackburn | by gunnar | 20:54 |
blackburn | that's THE LIST | 20:54 |
@sonney2k | notice the peak? | 20:54 |
@sonney2k | at ~2010 | 20:54 |
blackburn | of code? | 20:55 |
@sonney2k | that is when I merged and removed their code | 20:55 |
blackburn | of commits? | 20:55 |
@sonney2k | yes | 20:55 |
@sonney2k | code | 20:55 |
blackburn | ah | 20:55 |
blackburn | heh | 20:55 |
@sonney2k | LoC | 20:55 |
@sonney2k | blackburn, which list? | 20:55 |
@sonney2k | boosting list? | 20:55 |
blackburn | sonney2k: gunnar forwarded my mail | 20:55 |
@sonney2k | heh :) | 20:56 |
wiking | sonney2k: works? | 20:58 |
wiking | sonney2k: should get some errors | 21:00 |
blackburn | sonney2k: 'das Programm announced noch' is that german or you switched lang accidentally? | 21:03 |
blackburn | I am just curious | 21:03 |
@sonney2k | blackburn, my german is ruined :) | 21:05 |
blackburn | sonney2k: I feel inspired when trying to read deutsch | 21:06 |
blackburn | looks so fancy :) | 21:07 |
blackburn | noch wissen! wir ob wir! aus berline! | 21:07 |
blackburn | berlin* | 21:07 |
blackburn | machinnes lernen | 21:07 |
blackburn | :D | 21:07 |
blackburn | Un?berwachtes Lernen! | 21:09 |
blackburn | oh correct is maschinelles lernen | 21:10 |
n4nd0 | wiking: do you think gtest/gmock should work fine here? | 21:17 |
n4nd0 | I can test it if so | 21:17 |
shogun-notifier- | shogun: Sergey Lisitsyn :master * 55a234d / src/shogun/kernel/string/LocalAlignmentStringKernel.cpp: https://github.com/shogun-toolbox/shogun/commit/55a234d58576e3bffaa12a98fee48cda77177e26 | 21:23 |
shogun-notifier- | shogun: Removed shameful not stable mark from Local Alignment string kernel | 21:23 |
@sonney2k | n4nd0, please try | 21:25 |
n4nd0 | sonney2k: ok | 21:26 |
n4nd0 | just need to configure for it and build I guess | 21:27 |
@sonney2k | blackburn, n4nd0, wiking - what does ./evaluation_cross_validation_multiclass_mkl give you when you run it? | 21:27 |
blackburn | oh well let me compile latest | 21:27 |
@sonney2k | *lol* http://en.rocketnews24.com/2013/02/04/chinas-photoshop-experts-will-fix-your-bad-photos-but-be-careful-what-you-wish-for/ | 21:27 |
blackburn | sonney2k: was http://images.webpark.ru/uploads54/120913/Photoshop_11.jpg -> http://images.webpark.ru/uploads54/120913/Photoshop_12.jpg | 21:30 |
blackburn | and http://images.webpark.ru/uploads54/120913/Photoshop_13.jpg | 21:31 |
@sonney2k | wiking, can you please fix the 2 failing gtests ? I want to enable them in the builds! | 21:31 |
@sonney2k | buildbots taht is | 21:31 |
@sonney2k | otherwise we will never have them | 21:31 |
@sonney2k | blackburn, *lol* | 21:32 |
blackburn | sonney2k: http://ru.fishki.net/picsw/092010/01/post/fotoshop/foto_20.jpg -> http://ru.fishki.net/picsw/092010/01/post/fotoshop/foto_21.jpg | 21:33 |
blackburn | http://ru.fishki.net/picsw/092010/01/post/fotoshop/foto_68.jpg | 21:34 |
blackburn | well anyway the best is that: | 21:39 |
blackburn | http://www.1trash.net/wp-content/uploads/2012/11/1.jpg | 21:39 |
blackburn | they asked - photoshop me please to not look drunk | 21:39 |
blackburn | result: | 21:39 |
blackburn | http://www.1trash.net/wp-content/uploads/2012/11/2.jpg http://www.1trash.net/wp-content/uploads/2012/11/3.jpg | 21:40 |
blackburn | http://www.1trash.net/wp-content/uploads/2012/11/5.jpg http://www.1trash.net/wp-content/uploads/2012/11/6.jpg | 21:40 |
blackburn | http://www.1trash.net/wp-content/uploads/2012/11/7.jpg http://www.1trash.net/wp-content/uploads/2012/11/8.jpg | 21:40 |
@sonney2k | *megarotflbtc* | 21:43 |
wiking | n4nd0: try | 21:47 |
wiking | sonney2k: btw make unit-tests in ./shogun/src should work as well | 21:48 |
wiking | hopefully | 21:48 |
blackburn | sonney2k: I can't compile examples here | 21:49 |
blackburn | hah | 21:49 |
blackburn | yeah your thing with ; | 21:50 |
blackburn | sonney2k: why did you do that? it is better to have ; | 21:50 |
@sonney2k | wiking, there is no make unit-tests? | 21:51 |
@sonney2k | blackburn, was not possible | 21:52 |
blackburn | sonney2k: mean assert failed here | 21:52 |
blackburn | 0.76 | 21:52 |
blackburn | hah well setting seed won't work | 21:52 |
@sonney2k | blackburn, issue is that when you have a more complex macro that includes if (cond) { sth; } | 21:52 |
blackburn | it is not machine independent | 21:52 |
n4nd0 | wiking: moment please | 21:53 |
blackburn | sonney2k: ah you added unlikely | 21:54 |
@sonney2k | blackburn, well it fails everywhere I tried | 21:54 |
blackburn | good | 21:54 |
@sonney2k | so I wonder if someone here gets a 0.9 | 21:54 |
@sonney2k | n4nd0, you still have a 32bit system? | 21:54 |
n4nd0 | sonney2k: I am using 64 OS | 21:55 |
@sonney2k | blackburn, yeah and I need these extra { } brackets | 21:55 |
blackburn | sonney2k: you could put it to { } block | 21:55 |
@sonney2k | so only fix would be to change coding style to require | 21:55 |
@sonney2k | if () | 21:55 |
blackburn | then ; would work | 21:55 |
@sonney2k | { | 21:55 |
@sonney2k | } | 21:55 |
@sonney2k | else | 21:55 |
@sonney2k | { | 21:55 |
@sonney2k | } | 21:55 |
@sonney2k | alwyas | 21:55 |
blackburn | ah it is already in | 21:55 |
blackburn | sh | 21:56 |
@sonney2k | blackburn, sure long tiem | 21:56 |
@sonney2k | time | 21:56 |
blackburn | so { ... }; is not possible? | 21:56 |
@sonney2k | it is but not | 21:56 |
@sonney2k | if () {}; else {} | 21:56 |
@sonney2k | the else is then out of context | 21:56 |
blackburn | no I mean | 21:56 |
blackburn | ah | 21:57 |
@sonney2k | but if we had if () { {}; } else {} all good | 21:57 |
blackburn | okay got it | 21:57 |
@sonney2k | so that is why maybe we should change it to using {} always | 21:57 |
blackburn | macro w/o ; in the end is bad | 21:57 |
@sonney2k | blackburn, log4cxx does that too... | 21:57 |
blackburn | sonney2k: if we get igor we could get some factorizations/sparse stuff | 21:59 |
blackburn | I don't give a fuck what they are doing but sounds magical | 21:59 |
@sonney2k | blackburn, watch your language... | 22:01 |
blackburn | pooo | 22:01 |
blackburn | sonney2k: are you getting too old for such language? :) | 22:02 |
blackburn | anyway sonney2k not only ICA but some NMF or so would be nice | 22:03 |
blackburn | it is really cool as we are now learning features not only classifiers | 22:03 |
blackburn | desperately need group lasso specialist | 22:09 |
@sonney2k | wiking, haha you unit tests found an error in SGVector::max - the function you introduced utilizing atlas to find the max :) | 22:15 |
@sonney2k | scary! | 22:18 |
@sonney2k | all hail unit tests! | 22:19 |
blackburn | sonney2k: does shogun pass unit tests? | 22:19 |
@sonney2k | you mean the 8 we have :D | 22:20 |
blackburn | yes still | 22:20 |
@sonney2k | 6/8 currently | 22:20 |
@sonney2k | but I think I fixed the max issue now | 22:20 |
blackburn | sonney2k: we need code coverage computation | 22:21 |
blackburn | to understand how it is covered | 22:21 |
@sonney2k | 0.000001 % :D | 22:21 |
blackburn | it is so nice for python but I don't know about C++ | 22:21 |
blackburn | yes | 22:21 |
blackburn | but if we see now it is 0.1% we are happy | 22:21 |
@sonney2k | heh | 22:21 |
blackburn | sonney2k: lol, gcov | 22:23 |
blackburn | it is already here | 22:23 |
@sonney2k | so what is it? | 22:23 |
blackburn | something gnu to compute coverage | 22:23 |
wiking | sonney2k: :D | 22:25 |
@sonney2k | looks like this is more like grof | 22:25 |
@sonney2k | alright 1 test fails now :) | 22:25 |
blackburn | yeah I don't understand how to use it | 22:26 |
@sonney2k | wiking, good boy, bad boy! | 22:26 |
wiking | sonney2k: well then good that we have unit test finally :P | 22:26 |
blackburn | wiking: any idea? | 22:26 |
wiking | sonney2k: heheh i wonder what was happening there with the max thing :) | 22:26 |
blackburn | about coverage | 22:26 |
wiking | blackburn: about.... sorry i was showering till now | 22:26 |
@sonney2k | wiking, yeah we need them with programmers like you ;-) | 22:26 |
* wiking hides | 22:26 | |
wiking | sonney2k: but actually what happened in that max calculation? | 22:26 |
blackburn | we need to know what is tested and what is not | 22:26 |
wiking | blackburn: ah that's easy | 22:27 |
blackburn | wiking: how? | 22:27 |
wiking | blackburn: almost nothing is tested :D | 22:27 |
blackburn | wiking: I will find you | 22:27 |
wiking | blackburn: and kill me? :) | 22:27 |
blackburn | ja! | 22:27 |
wiking | blackburn: ok you want it per object or per object per function? | 22:27 |
blackburn | I want it somehow | 22:27 |
wiking | because the per obj is quite easy | 22:28 |
wiking | as you can see for each .cc you need a _unittest.cc | 22:28 |
wiking | so that's quite easy | 22:28 |
blackburn | no that's not interesting | 22:28 |
wiking | but i reckon that's not so nice | 22:28 |
wiking | i think we need an obj->func coverage | 22:29 |
blackburn | I found that gconv can be used with google test | 22:29 |
blackburn | http://meekrosoft.wordpress.com/2010/06/02/continuous-code-coverage-with-gcc-googletest-and-hudson/ | 22:29 |
@sonney2k | wiking, the cblas_idamax compute argmax_i(abs(x[i])) | 22:29 |
wiking | blackburn: mmm | 22:30 |
@sonney2k | wiking, but we used it to compute max | 22:30 |
blackburn | blas is pure evil | 22:30 |
@sonney2k | dangerous for negative numbers... | 22:30 |
wiking | sonney2k: i was totaly convinced from some doc that it's a simple max for double | 22:30 |
@sonney2k | wiking, yeah but it is not | 22:31 |
blackburn | http://www.netlib.org/blas/idamax.f | 22:31 |
blackburn | IDAMAX finds the index of element having max. absolute value. | 22:31 |
blackburn | wiking: never trust anything but netlib | 22:31 |
wiking | blackburn: yep | 22:31 |
wiking | i can see that now | 22:31 |
wiking | ok well it's good for abs_max :) | 22:31 |
wiking | if we'd have that | 22:31 |
wiking | btw: blackburn well i guess we could use that | 22:31 |
blackburn | wiking: yes, I'd like to try | 22:32 |
wiking | blackburn: but i guess some simple ctags hack could do it as well no? | 22:32 |
@sonney2k | interestingly we use it in arg_max too | 22:32 |
@sonney2k | but we have no test for that that could fail | 22:32 |
* sonney2k runs away screaming | 22:32 | |
blackburn | lol | 22:32 |
blackburn | sonney2k: well at least we can be sure iranian missile will fail | 22:32 |
@sonney2k | blackburn, sure of that? anything south of the equator will be mapped to north of it | 22:33 |
@sonney2k | same east of the null meridian -> west | 22:34 |
@sonney2k | west <-> east | 22:34 |
@sonney2k | scary no? | 22:34 |
blackburn | sonney2k: hmm | 22:34 |
blackburn | interesting what it will reach then | 22:34 |
blackburn | have you all seen new iranian aircraft? | 22:35 |
@sonney2k | indeed some drones attempting to land on berlins new airport ;^) | 22:35 |
blackburn | http://c.img22.rian.ru/images/92138/00/921380064.jpg | 22:36 |
wiking | blackburn: the plastic mockup? | 22:36 |
blackburn | wiking: ! | 22:36 |
blackburn | exactly | 22:36 |
wiking | :D | 22:36 |
wiking | no i haven't seen it :D | 22:36 |
wiking | blackburn: well let's try that gconv thingy then | 22:37 |
blackburn | wiking: I am trying it now | 22:37 |
blackburn | wiking: sonney2k: this way we can accept any patches that increase coverage | 22:42 |
@sonney2k | blackburn, so you mean tasks for gsoc applicants? | 22:42 |
blackburn | yes | 22:42 |
@sonney2k | how demotivating :/ | 22:43 |
blackburn | well may be | 22:43 |
@sonney2k | porting examples over from python_modualr to other $LANG would also be valuable | 22:44 |
wiking | :>>>> | 22:44 |
blackburn | sonney2k: isn't that demotivating? | 22:44 |
@sonney2k | blackburn, at least you see that stuff runs some real thing | 22:45 |
@sonney2k | and maybe even with fance gfx | 22:45 |
@sonney2k | fancy | 22:45 |
@sonney2k | ohh boy | 23:11 |
@sonney2k | look svm ocas is working again... | 23:11 |
@sonney2k | all due to that isamax thing | 23:11 |
wiking | sonney2k: \o/ | 23:13 |
wiking | sonney2k: megafail by me | 23:13 |
wiking | :( | 23:13 |
n4nd0 | wiking, sonney2k : sorry, I had an issue with my python, but now it is fine :) | 23:13 |
n4nd0 | wiking: Checking for Google C++ Testing Framework ... unit testing framework not detected, for unit testing please install it | 23:13 |
wiking | but seems not many people use shogun as nobody complained | 23:13 |
n4nd0 | Checking for Google C++ Mocking Framework ... unit testing framework not detected, for unit testing please install it | 23:13 |
n4nd0 | let me check configure too | 23:13 |
wiking | n4nd0: ok i guess it cannot find the src dir for either of those libraries | 23:14 |
wiking | it tries to find it under /usr/src/{gtest,gmock} by default | 23:14 |
wiking | if the src is somewhere else you have to specify it for the ./configure by --gmock=PATH --gtest=PATH | 23:15 |
shogun-notifier- | shogun: Soeren Sonnenburg :master * 0017902 / / (5 files): https://github.com/shogun-toolbox/shogun/commit/001790288e12ada4bd2d4956007b4f87834d84c7 | 23:15 |
shogun-notifier- | shogun: fix argmax/max in sgvector | 23:15 |
shogun-notifier- | shogun: | 23:15 |
shogun-notifier- | shogun: for float/double: isamax computes the argmax of abs(vector) | 23:15 |
shogun-notifier- | shogun: not the max! introduce new max_abs / arg_max_abs functions for that and | 23:15 |
shogun-notifier- | shogun: add tests for these too. | 23:15 |
n4nd0 | wiking: /usr/src/gtest is there | 23:16 |
wiking | n4nd0: where's your gmock src? | 23:16 |
wiking | if you dont have that you'll have to download it | 23:16 |
wiking | n4nd0: but still i wonder why it cannot find gtest src | 23:17 |
n4nd0 | wiking: yeah that was weird | 23:17 |
n4nd0 | gmock is in... | 23:17 |
n4nd0 | /usr/include/gmock | 23:18 |
wiking | n4nd0: but the src? | 23:18 |
wiking | because those are only the headers | 23:18 |
wiking | n4nd0: but still need to figure out why gtest fails for u | 23:19 |
wiking | n4nd0: anything informative in the configure.log? | 23:20 |
n4nd0 | wiking: does gmock dir look fine? http://pastebin.com/Vs0Da84e | 23:20 |
n4nd0 | checking configure.log now | 23:20 |
wiking | n4nd0: yeah headers looks great | 23:21 |
n4nd0 | wiking: nothing apart from what I pasted before, in the configure.log | 23:21 |
wiking | but you would need the src files as well: gmock-all.cc gmock-cardinalities.cc gmock.cc gmock-internal-utils.cc gmock_main.cc gmock-matchers.cc gmock-spec-builders.cc | 23:21 |
wiking | if u dont have those | 23:21 |
wiking | download gmock src somewhere | 23:21 |
wiking | and give the path for the configure script | 23:22 |
@sonney2k | wiking, please install gmock on your buildslaves... | 23:22 |
wiking | sonney2k: mmm letmecheck what freebsd has | 23:22 |
wiking | sonney2k: i have gtest installed on ubuntu | 23:22 |
@sonney2k | I hope I find the time to do it on all the rest and then we enable this thing | 23:22 |
@sonney2k | but for now | 23:22 |
@sonney2k | ZZZZZzzzz... | 23:22 |
wiking | but as said earlier on ubuntu | 23:22 |
wiking | gmock package is fucked | 23:22 |
wiking | so i cannot enable it :S | 23:23 |
blackburn | watch your language! :D | 23:23 |
@sonney2k | wiking, then put the files manually somewhere | 23:23 |
blackburn | ah soeren is sleeping | 23:23 |
blackburn | now we can say bad words | 23:23 |
n4nd0 | haha | 23:23 |
blackburn | jk | 23:23 |
wiking | sonney2k: /home/wiking/gmock-1.6.0 | 23:23 |
wiking | sonney2k: for the ubuntu buildbot | 23:23 |
wiking | sonney2k: but afaik you should not run make unit-tests there | 23:23 |
n4nd0 | wiking: I've got some python files in /usr/shar/gmock too | 23:23 |
wiking | sonney2k: as its doing static analyzer there | 23:24 |
blackburn | http://www.youtube.com/watch?feature=player_detailpage&v=zfc8x0NWtIE songs of sonney2k's youth | 23:24 |
n4nd0 | no clue of cc files though | 23:24 |
n4nd0 | wiking: ok gtest detected now | 23:25 |
wiking | n4nd0: cool now i think u should download gmock for yourself | 23:27 |
wiking | n4nd0: if the package doesn't have the src installed somewhere | 23:27 |
wiking | i mean the c++ src | 23:27 |
n4nd0 | wiking: I have found the sources in a zip in AUR | 23:27 |
wiking | of gmock that is | 23:27 |
n4nd0 | wiking: take a look at it | 23:29 |
n4nd0 | http://googlemock.googlecode.com/files/gmock-1.6.0.zip | 23:29 |
wiking | yep | 23:31 |
wiking | that's what u need | 23:31 |
wiking | :) | 23:31 |
n4nd0 | wiking: complete install or just point in the configure to the .src files? | 23:32 |
wiking | n4nd0: the former one... just point to the root dir of what you've got extracted from that zip | 23:33 |
wiking | i mean latter | 23:33 |
wiking | :P | 23:33 |
wiking | the compilation will be done by shogun | 23:33 |
n4nd0 | ok | 23:33 |
n4nd0 | wiking: let's try it out | 23:34 |
blackburn | what do you guys think if we make shogun a browser? | 23:34 |
wiking | if the ./configure finds both | 23:34 |
wiking | blackburn: most def :) | 23:34 |
wiking | n4nd0: just run make unit-tests | 23:34 |
wiking | that should both compile the shogun lib and try to execute the unit tests | 23:34 |
blackburn | we will get a lot of users by a matter of seconds | 23:34 |
blackburn | people would like to see how ML library changed to browser | 23:35 |
wiking | blackburn: i guess we'll compile it for all the mobile platforms as well :P | 23:35 |
n4nd0 | wiking: gmock no detected still | 23:36 |
wiking | n4nd0: ok how do you run ./configure? | 23:36 |
n4nd0 | wiking: a sec I think I know what it might be | 23:36 |
n4nd0 | mmm nothing it didn't work out | 23:37 |
n4nd0 | so I have | 23:37 |
n4nd0 | oh shit I am idiot | 23:37 |
n4nd0 | I still had the /usr/include/gmock from before... | 23:37 |
wiking | ok just paste here your ./configure line :) | 23:37 |
n4nd0 | got it detected :) | 23:38 |
n4nd0 | ./configure --disable-optimization --interfaces=python_modular --gtest=/usr/src/gtest/ --python=python2 --gmock=/usr/src/gmock | 23:38 |
wiking | mm and you extracted your gmock zip to /usr/src/gmock ? | 23:38 |
n4nd0 | yeah | 23:39 |
n4nd0 | but it is working now | 23:39 |
n4nd0 | Checking for Google C++ Testing Framework ... yes | 23:39 |
n4nd0 | Checking for Google C++ Mocking Framework ... yes | 23:39 |
wiking | heheh then you can just remove both --gtest=/usr/src/gtest/ and --gmock=/usr/src/gmock | 23:39 |
wiking | as it's by default looking for it at that location | 23:39 |
n4nd0 | hehe I like to type :P | 23:39 |
wiking | ok well now type | 23:39 |
wiking | make unit-tests | 23:39 |
wiking | and it should do the whole thing | 23:39 |
wiking | and on the end see the unit test running | 23:40 |
n4nd0 | make unit-tests is a gmock thing? | 23:40 |
n4nd0 | make unit-tests | 23:40 |
n4nd0 | make: Nothing to be done for `unit-tests'. | 23:40 |
wiking | mmm | 23:40 |
wiking | i wonder | 23:40 |
wiking | ok then just do this | 23:40 |
wiking | cd ../tests/unit && make | 23:40 |
n4nd0 | but I didn't make yet | 23:40 |
n4nd0 | maybe it needs to be done | 23:40 |
wiking | dont worry about that | 23:41 |
wiking | it should invoke automatically the compilation of shogun library | 23:41 |
n4nd0 | making in tests/unit ... | 23:41 |
wiking | yep | 23:41 |
wiking | so that'll make everything that u need for unit testing | 23:41 |
wiking | and then run the unit tests | 23:41 |
n4nd0 | ok | 23:43 |
n4nd0 | wiking: successful make | 23:48 |
--- Log closed Sat Feb 09 00:00:43 2013 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!