| --- Log opened Mon Mar 28 00:00:14 2011 | ||
| --- Day changed Mon Mar 28 2011 | ||
| sonney2k | sergey_lisitsyn, it will be via git & github | 00:00 |
|---|---|---|
| sergey_lisitsyn | nice, I like github, thank you | 00:00 |
| sonney2k | yoh and me are just doing the final touches to the repository - so svn -> git conversion is still ongoing but expected to finish rather soon | 00:01 |
| sergey_lisitsyn | about open source experience - is it necessarily to take a part in shogun ideas? | 00:01 |
| yoh | sonney2k: ;) about deletes -- do not be greedy ;) | 00:01 |
| yoh | fetch all your students daily -- you will get it forever | 00:01 |
| sonney2k | sergey_lisitsyn, no but it is a plus / I similar experience can be as good though! | 00:02 |
| sonney2k | yoh, yes clones at shogun-toolbox.org of course :D | 00:02 |
| yoh | something funny... either majority of size comes from other heavy files or something... size of .git didn't drop after my 'filtering' :-/ | 00:03 |
| sergey_lisitsyn | sonney2k: I'm a bit uneasy with it, I had't take part in 'serious' open source projects | 00:03 |
| sergey_lisitsyn | yoh: приятно видеть здесь и говорящих на русском :) | 00:04 |
| yoh | sergey_lisitsyn: где тут по русски говорятьс? ;-) | 00:05 |
| sergey_lisitsyn | yoh: погуглил, если в человек МФТИ учился, то он явно по-русски говорит :) | 00:05 |
| sonney2k | yoh - did you check that now git log --name-only | grep .bz2 is empty | 00:05 |
| sergey_lisitsyn | в человек = человек | 00:05 |
| sergey_lisitsyn | тьуф блин, наоборот | 00:05 |
| sonney2k | I wonder how you guys type those letters.... | 00:06 |
| yoh | yeah yeah | 00:06 |
| yoh | but .git stays at 118M | 00:06 |
| sergey_lisitsyn | sonney2k: but on your homepage there is a notion about basic russian? :) | 00:06 |
| sonney2k | sergey_lisitsyn, IIRC I had russian for 10.5 years at school | 00:07 |
| sonney2k | but hey I forgot *everythin* | 00:07 |
| sonney2k | g | 00:07 |
| -!- skydiver [c255a037@gateway/web/freenode/ip.194.85.160.55] has joined #shogun | 00:07 | |
| sergey_lisitsyn | sonney2k: I had similar with german | 00:07 |
| sergey_lisitsyn | sonney2k: learnt for 3 years, don't remember anything | 00:08 |
| yoh | I guess due to all those: just run git log --stat | grep bytes | 00:08 |
| skydiver | can anybody help with error while running python examples: "ImportError: No module named sg" | 00:08 |
| yoh | sonney2k: I guess due to all those: just run git log --stat | grep bytes | 00:08 |
| sonney2k | err but wait, before it was 203M | 00:10 |
| yoh | what abour prunning compiled materials which you rmeoved later on ( */sg/inst/doc/{sg,manual}.pdf ) | 00:10 |
| sonney2k | so 118M is a big improvemnt | 00:10 |
| yoh | well -- I told ya that simply with repack you could get to 120M | 00:11 |
| sonney2k | there is also some .cpp.o crap | 00:11 |
| yoh | my aim was to bring it down to 20-30 ;) | 00:11 |
| sonney2k | which is strange | 00:11 |
| yoh | yeah | 00:11 |
| yoh | saw it | 00:11 |
| sergey_lisitsyn | sonney2k: in the description there are "Report success on the mailing list or IRC" after "Set up your platform to build a shogun snapshot ". what kind of success I should report? | 00:11 |
| sonney2k | sergey_lisitsyn, that you have managed to compile git head and run whatever toy example / minimal implementation of sth. that is the baseline for what you intend to do - we just want to make sure that it won't take you 1 month to get your environment to set up. | 00:13 |
| yoh | sonney2k: would you be interested in pruning all data/* at once as well? ;) | 00:14 |
| sergey_lisitsyn | sonney2k: I have already tried to develop a parzen window classifier :) | 00:14 |
| sonney2k | sergey_lisitsyn, that qualifies as a success story :D | 00:14 |
| sonney2k | (if done within shogun) | 00:14 |
| sergey_lisitsyn | sure in shogun :) | 00:15 |
| yoh | and for tarballs of dependencies... why do you keep them around (especially as tarballs)? | 00:15 |
| sonney2k | sergey_lisitsyn, that's a big plus then, just attach the patch | 00:15 |
| sergey_lisitsyn | sonney2k: tried but not complete! should I complete it? | 00:16 |
| sonney2k | yoh, not my code (I would never do this) | 00:16 |
| sergey_lisitsyn | sonney2k: just getting used with shogun interfaces, etc | 00:16 |
| sonney2k | sergey_lisitsyn, well would be cool... | 00:16 |
| sonney2k | yoh, I don't think this will help a lot | 00:17 |
| sergey_lisitsyn | sonney2k: so I could do it during this week, and then will send an application | 00:17 |
| yoh | solely tracking of changes in applications/ocr/data/ocr.svm.gz probably brought around a megabyte | 00:17 |
| sergey_lisitsyn | sonney2k: I already have in touch with C. Widmer about idea of implementing algos | 00:17 |
| sonney2k | yoh, seriously the .bz2 files are ~100MB in size. Removing them *must* save lots of space | 00:17 |
| yoh | heh | 00:18 |
| sergey_lisitsyn | sonney2k: have this problem with one dataset from UC Irvine - faces and non-faces, SVM files was a ~150 mb, but all the rest was 3 mb | 00:18 |
| sergey_lisitsyn | had* | 00:18 |
| skydiver | sergey, where r u from?) | 00:19 |
| sergey_lisitsyn | skydiver: Togliatty, Russia | 00:19 |
| skydiver | hello from SPb | 00:19 |
| yoh | Togliatty -- didn't know it is spelled like that ;) | 00:20 |
| sergey_lisitsyn | yoh: how else? | 00:20 |
| skydiver | maybe Toliatty? | 00:20 |
| sergey_lisitsyn | skydiver: 'italian' spelling | 00:20 |
| sergey_lisitsyn | like Spaghetti | 00:20 |
| sergey_lisitsyn | :) | 00:20 |
| skydiver | oh, i see | 00:20 |
| sergey_lisitsyn | I don't know, I thought it's a common spelling | 00:21 |
| sergey_lisitsyn | in wiki - Tolyatti | 00:21 |
| sonney2k | yoh, so what now? | 00:22 |
| yoh | patience ;) | 00:22 |
| sergey_lisitsyn | in deutsch wiki even Togliatti | 00:22 |
| sergey_lisitsyn | same thing with my name, I don't know how right it is :) | 00:22 |
| sergey_lisitsyn | skydiver: what idea do you like to 'work' for? | 00:23 |
| skydiver | i mostly like kernel feature analysis | 00:25 |
| sonney2k | yoh, could it be that some references are still lying around and you should do this here http://dound.com/2009/04/git-forever-remove-files-or-folders-from-history/ | 00:25 |
| skydiver | and you? | 00:25 |
| yoh | sonney2k: I will check but I already pruned reflog and other obvious victims | 00:25 |
| yoh | to no effect | 00:25 |
| sergey_lisitsyn | skydiver: implementing missing algorithms, mentor: Christian Widmer | 00:25 |
| sonney2k | yoh, *grmpf* | 00:26 |
| sergey_lisitsyn | skydiver: had you write to Conrad Rieck? I asked him for some details week ago | 00:26 |
| sergey_lisitsyn | sonney2k: btw, is Christian using IRC? | 00:27 |
| skydiver | has he replyed? | 00:27 |
| sergey_lisitsyn | skydiver: Rieck? Sure | 00:27 |
| * sonney2k hopes that we get enough slots from google | 00:27 | |
| yoh | wtf -- original is still there ;) | 00:27 |
| sonney2k | yoh, hah! | 00:27 |
| skydiver | sergey_lisitsyn: Konrad | 00:28 |
| sergey_lisitsyn | sonney2k: do you already have many candidates? | 00:28 |
| sonney2k | sergey_lisitsyn, so far not. I guess I will have to send an encouraging mail to the mentors to also idle here at least | 00:28 |
| sergey_lisitsyn | skydiver: yeap, he answered | 00:28 |
| sergey_lisitsyn | sonney2k: what kind of idle? | 00:28 |
| sergey_lisitsyn | skydiver: write me to 237338803, why we speak english there :) | 00:29 |
| sonney2k | sergey_lisitsyn, well I cannot know how many will apply but I am spending a lot of time answering emails... | 00:29 |
| yoh | I think I know | 00:29 |
| skydiver | sergey_lisitsyn: so why have you chosen another idea? | 00:29 |
| yoh | because those other branches are stand-alone and had no proper common root, they weren't rewritten/filtered at all | 00:30 |
| sergey_lisitsyn | sonney2k: my email to Chris was sent to you with answer, do you read all the 'appliers' mail? :) | 00:30 |
| sonney2k | yoh, I am listening... | 00:30 |
| skydiver | 237338803 - what is it? | 00:30 |
| yoh | and they would need to go through the same procedure manually | 00:30 |
| sergey_lisitsyn | skydiver: icq, what could it be else | 00:30 |
| sonney2k | yoh, you mean the structure etc branches? | 00:30 |
| sonney2k | So it is easy to fix right? | 00:30 |
| sonney2k | just some for loop around all of the branches? | 00:31 |
| yoh | those genefinder etc | 00:31 |
| sonney2k | yoh, yes | 00:31 |
| skydiver | sergey_lisitsyn: oh, i've forgotten when i had used icq... maybe jabber? | 00:31 |
| sergey_lisitsyn | skydiver: oh, that way :) minute | 00:31 |
| sonney2k | sergey_lisitsyn, so far yes. I sometimes don't respond when it is likely that the mentor will | 00:32 |
| sergey_lisitsyn | skydiver: sergey.lisitsyn@jabber.ru | 00:33 |
| sergey_lisitsyn | skydiver: I forgot it :D | 00:33 |
| sergey_lisitsyn | sonney2k: had I bored you? I have a few questions more | 00:34 |
| yoh | sonney2k: and because tags are dangling, i.e. not belonging to any branch -- not sure if they get rewritten properly... | 00:34 |
| sonney2k | yoh, but tags are just pointers to some checkout state? So there should not be a problem... | 00:35 |
| sonney2k | IMHO | 00:35 |
| yoh | sonney2k: yeah yea... but what was probably missing is " --tag-name-filter cat -- --all " | 00:35 |
| yoh | without it, filter-branch traversed only the space of commits reachable from master branch | 00:36 |
| sonney2k | yoh, o_O | 00:36 |
| yoh | since those release tags are off the master branch, they didn't get rewritten | 00:36 |
| sonney2k | makes sense | 00:36 |
| sonney2k | sergey_lisitsyn, don't ask to ask just ask... | 00:37 |
| sergey_lisitsyn | sonney2k: trying to be polite :) should I make a parzen classifier for shogun and then (after that) send an application and mention there that I already familiar with shogun interfaces, etc | 00:38 |
| sergey_lisitsyn | ? | 00:38 |
| yoh | yoh: initiated clean filter again also pruning those few .pdf and .o files | 00:39 |
| sonney2k | I don't expect the few .pdf's to really help... but lets see if the big humps are now gone | 00:40 |
| sonney2k | sergey_lisitsyn, if you can afford it yes - I mean there is probably no one that applies for GSoC who has contributed to shogun yet | 00:41 |
| yoh | ok ... dinner atm... will be back in 10-15 min | 00:41 |
| sonney2k | it is 1am by then - I gotta go sleeping. It is only 5hrs until I have to get up. | 00:42 |
| sergey_lisitsyn | sonney2k: okay, I'll try to make it, it seems simple, parzen is like the kNN | 00:42 |
| sergey_lisitsyn | sonney2k: oh, okay, good night | 00:43 |
| sonney2k | sergey_lisitsyn, I really hope that we get enough slots - would be bad having to make some unfair dice-roll based decisions. | 00:44 |
| sonney2k | yes good n8 | 00:44 |
| sonney2k | l8r | 00:44 |
| sergey_lisitsyn | sonney2k: I hope too :) | 00:44 |
| yoh | sonney2k: what??? going to sleep already??? come on -- it is monday already, time to work! ;) | 00:52 |
| skydiver | sonney2k: maybe you can advice algo i can implement now for shogun - i've got about week to dive into shogun code :) | 00:54 |
| skydiver | sonney2k: i think it is better way to learn toolbox using (or extending) it | 00:56 |
| -!- aifargonos [~aifargono@46.18.27.35] has quit [Ping timeout: 276 seconds] | 01:06 | |
| sergey_lisitsyn | yoh: а ты вроде не контрибьютор, просто помогаешь проекту? | 01:07 |
| sergey_lisitsyn | или я ошибаюсь? | 01:07 |
| yoh | sonney2k: since you left it on -- ufff! so, upon my last refiltering attempt -- did everything sweet, all tags/branches got rewritten BUT now it refuses to compress the archive into a single pack... so there is 20MB pack and then bulk of objects totalling to 153MB ... WTF? | 01:09 |
| yoh | sergey_lisitsyn: типа того | 01:10 |
| yoh | хотя я кажись когда то пару фиксов закомиттил | 01:10 |
| sergey_lisitsyn | yoh: понятно :) | 01:10 |
| yoh | мы его исплользуем (частично) в www.pymvpa.org | 01:10 |
| skydiver | кстати вопрос по поводу shogun | 01:10 |
| sergey_lisitsyn | всё, исчезла аглицкая речь) | 01:11 |
| skydiver | кем будут ревьювится коммиты? | 01:11 |
| yoh | sonney2k: the ultimate rewrite command was git filter-branch --tree-filter 'rm -rf `cat /home/yoh/proj/shogun/shogun-remastering/files2delete.txt `; find -iname *.bz2 -delete; find -iname *.o -delete; ' --tag-name-filter cat -- --all | 01:11 |
| skydiver | т.е. вообще насколько сложен процесс попадания патча в trunk | 01:11 |
| yoh | where that files2delete contained the actual files to be removed | 01:11 |
| skydiver | помнится в прошлом году некотоыре ребята в рассылку gsoc писали что бывает жесть в этих вопросах | 01:11 |
| sergey_lisitsyn | yoh: адский агрегат. вы в придачу к pymvpa тянете сёгун? | 01:12 |
| yoh | skydiver: ну это Вы Soerena спрашиваете | 01:12 |
| yoh | если вы его беднягу не завалите то думаю он выберется -- с GIT после того как разберется то должно пойти всё легче | 01:13 |
| yoh | особенно с прибамбасами что github даёт (pull requests, code review etc) | 01:13 |
| yoh | sergey_lisitsyn: а чего его тянуть? он сам идёт ;-) | 01:13 |
| yoh | я ещё по совместительсву Debian developer, so in my world installing shogun for pymvpa is just a matter of adding shogun-python-module into Recommends: of pymvpa package ;) | 01:14 |
| sergey_lisitsyn | yoh: ну ты вообще. :) | 01:15 |
| yoh | sonney2k: I will email you my COMPLETE protocol of struggle + minimalistic set of commands which should have theoretically done it | 01:15 |
| yoh | may be fresh minds of monday will be more git-cooperating ;) | 01:16 |
| yoh | sergey_lisitsyn: well, сонядветыщи тоже дебианоид -- все мы такия I guess ;) | 01:17 |
| yoh | sonney2k: meanwhile -- kids are already all around ;-) | 01:17 |
| yoh | пока ребяты -- удачи с GSOC! | 01:18 |
| sergey_lisitsyn | спасибо :) | 01:18 |
| sergey_lisitsyn | удачи | 01:18 |
| skydiver | thnx | 01:18 |
| yoh | kstati а вы то дебианоиды? | 01:27 |
| skydiver | а что значит сей термин?) | 01:29 |
| sergey_lisitsyn | yoh: я к сожалению этот самый, убунтунутый | 01:29 |
| -!- skydiver [c255a037@gateway/web/freenode/ip.194.85.160.55] has quit [Quit: Page closed] | 01:42 | |
| yoh | sergey_lisitsyn: nu znachit debianoid ;) pochemu zhe k sozhaleniyu? | 01:48 |
| sergey_lisitsyn | yoh: ну типа не pure дебиан :) | 01:48 |
| yoh | skydiver: user of www.debian.org... well you left already ;) | 01:48 |
| yoh | sergey_lisitsyn: nu u kazhdogo svoi grehi | 01:49 |
| yoh | i nedostatki | 01:49 |
| yoh | ;-) | 01:49 |
| sergey_lisitsyn | блин а за что такие мучения с этой рассылкой? | 01:50 |
| sergey_lisitsyn | почему не гугл группы или ещё что-нибудь | 01:50 |
| yoh | nu u menya s googlegroups kak raz mucheniya obychno ;) | 01:51 |
| yoh | ladno -- poshel dityov ukladivat' | 01:51 |
| sergey_lisitsyn | да я тут чуть крышей не поехал пока не нашёл последние сообщения | 01:52 |
| sergey_lisitsyn | yoh: from shogun.Features import RealFeatures, Labels ImportError: No module named shogun.Features | 01:57 |
| sergey_lisitsyn | yoh: не знаешь что с такой бедой сделать? :) ставлю первый раз | 01:57 |
| sergey_lisitsyn | отбой, ядибил | 02:01 |
| -!- blackburn [~qdrgsm@109.226.104.31] has joined #shogun | 02:07 | |
| -!- sergey_lisitsyn [6de2681f@gateway/web/freenode/ip.109.226.104.31] has quit [Quit: Page closed] | 02:07 | |
| blackburn | перекантовался :) | 02:07 |
| blackburn | yoh: у вас вечереет, наверное, поэтому доброй ночи :) | 02:14 |
| blackburn | good night everyone | 02:14 |
| -!- blackburn [~qdrgsm@109.226.104.31] has quit [Quit: Leaving.] | 02:14 | |
| yoh | sonney2k: ha -- not sure what went wrong last time, now .git is just 23MB ;) | 02:30 |
| -!- sonney2k [~sonney2k@f053046209.adsl.alicedsl.de] has quit [Ping timeout: 246 seconds] | 03:04 | |
| -!- sonney2k [~sonney2k@f053046246.adsl.alicedsl.de] has joined #shogun | 03:16 | |
| -!- dverev [b49531e3@gateway/web/freenode/ip.180.149.49.227] has joined #shogun | 03:20 | |
| -!- dverev [b49531e3@gateway/web/freenode/ip.180.149.49.227] has quit [Quit: Page closed] | 04:18 | |
| -!- Yin [cf2e5c11@gateway/web/freenode/ip.207.46.92.17] has joined #shogun | 04:37 | |
| -!- alesis-novik [~alesis@188.74.87.84] has joined #shogun | 05:04 | |
| alesis-novik | Anyone from Shogun around to talk about GSoC? | 05:07 |
| -!- alesis-novik [~alesis@188.74.87.84] has quit [Quit: Leaving] | 05:21 | |
| -!- epps [~epps@84.18.157.200] has joined #shogun | 06:01 | |
| -!- epps [~epps@84.18.157.200] has quit [Changing host] | 06:01 | |
| -!- epps [~epps@unaffiliated/epps] has joined #shogun | 06:01 | |
| -!- heatkernel [b89842eb@gateway/web/freenode/ip.184.152.66.235] has joined #shogun | 06:42 | |
| heatkernel | hello | 06:42 |
| heatkernel | I had an inquiry regarding the google "summer of code" projects | 06:43 |
| heatkernel | I am writing the contacts listed for invidual | 06:45 |
| heatkernel | project ideas | 06:45 |
| siddharth | Hello | 07:28 |
| siddharth | In Student Application for Shogun,its written "Set up your platform to build a shogun snapshot" | 07:29 |
| siddharth | So what do we have to provide in application? | 07:29 |
| -!- epps [~epps@unaffiliated/epps] has quit [Ping timeout: 246 seconds] | 07:51 | |
| -!- aifargonos [~aifargono@46.18.27.35] has joined #shogun | 08:04 | |
| -!- siddharth [~siddharth@117.211.88.150] has quit [Ping timeout: 264 seconds] | 08:12 | |
| sonney2k | yoh, then please provide it to me... | 08:14 |
| -!- aifargonos [~aifargono@46.18.27.35] has quit [Ping timeout: 276 seconds] | 08:27 | |
| -!- Tanmoy [75d35896@gateway/web/freenode/ip.117.211.88.150] has joined #shogun | 08:47 | |
| Tanmoy | @sploving has ur R bugs been fixed | 08:48 |
| -!- aifargonos [~aifargono@46.18.27.4] has joined #shogun | 08:48 | |
| sploving | hello, not yet | 08:49 |
| Tanmoy | hmmm lets try it once more | 08:49 |
| sploving | The bugs are because of SWIG | 08:50 |
| Tanmoy | did u copy all the .so files correctly | 08:50 |
| Tanmoy | actually i get it | 08:50 |
| Tanmoy | the swig seems to be working fine for me | 08:50 |
| Tanmoy | are ur test suites running fine | 08:50 |
| sploving | Swig does not support R very well | 08:51 |
| Tanmoy | yup | 08:51 |
| Tanmoy | tryin to fix tht | 08:51 |
| Tanmoy | surprisingly the static moduels r kinda easy | 08:51 |
| sploving | yep. | 08:51 |
| sploving | I have not dived into why r modular has such bug | 08:52 |
| Tanmoy | yup looking for it | 08:52 |
| sploving | I am not sure how long time it takes. I also need time to write the proposal. | 08:53 |
| Tanmoy | u mean the one on integrating Lua ? i did read tht | 08:53 |
| sploving | yep. | 08:54 |
| Tanmoy | wld be happy to c it after Gsoc | 08:54 |
| sonney2k | regarding R modular | 08:58 |
| sonney2k | sploving, Tanmoy: you should pass --disable-reference-counting to configure | 08:59 |
| sploving | I know that. | 09:00 |
| Tanmoy | hmm did try tht yday | 09:00 |
| sploving | succeed? | 09:00 |
| sonney2k | then it all works but you have memory leaks :-/ | 09:01 |
| Tanmoy | surprisingly in the configure it says its checked for python python modular r | 09:02 |
| Tanmoy | but not for r modular | 09:02 |
| sonney2k | there seems to be something wrong in swig's R module regarding reference counting, i.e. we call ref() to increment reference counts for objects and unref() to decrease /destroy them | 09:02 |
| Tanmoy | yup | 09:02 |
| sonney2k | Tanmoy, well it is disabled by default (for that reason...) | 09:02 |
| Tanmoy | i think in the mailing list thts what @sploving was trying to say | 09:02 |
| sonney2k | so just specify it: --interfaces=libshogun,r_modular | 09:03 |
| sploving | also with --disable-reference-counting | 09:04 |
| sonney2k | sploving, yes | 09:04 |
| sploving | sonney2k, have you concated with swig r developer? | 09:04 |
| sploving | about this bug | 09:04 |
| sploving | do they know? | 09:05 |
| sonney2k | sploving, yes but I think he needs a small test case to reproduce | 09:05 |
| sonney2k | shogun is too big for that... | 09:05 |
| sonney2k | hard to isolate what the problem really is then | 09:05 |
| sploving | you mean that in swig r testsuite it works well, but for shogun, it produce such problem | 09:06 |
| sonney2k | sploving, I don't know if swigs-R test suite contains such tests... | 09:07 |
| sonney2k | I would be surprised... | 09:07 |
| sonney2k | but there is another bug in swig that affects us | 09:07 |
| sonney2k | for octave_modular. | 09:07 |
| sploving | exit on crash? | 09:07 |
| sonney2k | it will crash on exit... | 09:07 |
| sonney2k | yes | 09:07 |
| Tanmoy | crash on exit | 09:08 |
| sonney2k | this could be worked around though by just creating a single shogun.so | 09:08 |
| sonney2k | err shogun.oct | 09:08 |
| sonney2k | will all the modules inside. | 09:08 |
| -!- blackburn [~qdrgsm@109.226.104.31] has joined #shogun | 09:08 | |
| sonney2k | and it was actually a bug in octave | 09:08 |
| sonney2k | but devels have fixed that in their development branch | 09:09 |
| blackburn | hi there | 09:09 |
| sonney2k | ...but the swig octave developer hasn't yet | 09:09 |
| sploving | you mean single modle is OK, more than one has bug | 09:09 |
| -!- aifargonos [~aifargono@46.18.27.4] has quit [Ping timeout: 252 seconds] | 09:10 | |
| sonney2k | hi... | 09:10 |
| sonney2k | sploving, single module yes. modules that depend on each other can be dangerours (when one unloads the other...) | 09:11 |
| blackburn | sonney2k: last night we discussed a little contribution of me for shogun, I'd like to get familiar with interfaces through writing a similar to kNN classifier. where should I send it? I mean I have no access to repo | 09:12 |
| sploving | sonney2k, i know that. | 09:13 |
| sonney2k | blackburn, send a patch to the mailinglist... | 09:14 |
| blackburn | sonney2k: oh, this way, thank you, I'll try | 09:14 |
| sonney2k | blackburn, note that shogun's repository will soon be on github | 09:16 |
| sonney2k | (well it is actually but I will reset it soon) | 09:16 |
| blackburn | sonney2k: may be I should send it to github? :) | 09:17 |
| sonney2k | blackburn, no the mailinglist ist the right place but yes you can fork it when I am ready (later today) | 09:17 |
| blackburn | sonney2k: I'm confused with all this mailinglists and SVN :) never used either | 09:18 |
| sonney2k | blackburn, then it is about time :D | 09:19 |
| sonney2k | blackburn, these things are the reasons open source works :) | 09:19 |
| blackburn | sonney2k: I mean I always worked with mercurial. there is no problem with it at all | 09:20 |
| blackburn | sonney2k: if I will have a questions with sending patch, I'll write to you there, okay? | 09:20 |
| sonney2k | blackburn, well I am way older I guess... used CVS before... | 09:20 |
| sonney2k | blackburn, just do svn diff and email it to the mailinglist | 09:21 |
| sonney2k | (as attachment) | 09:21 |
| blackburn | sonney2k: oh, I see | 09:21 |
| blackburn | sonney2k: thank you | 09:21 |
| blackburn | sonney2k: is there a difference with time of application sent? I mean should I sent it earlier to get | 09:22 |
| blackburn | 'slot' | 09:22 |
| sonney2k | blackburn, no not at all | 09:22 |
| blackburn | sonney2k: ok, thank you once more :) | 09:22 |
| sonney2k | blackburn, and we don't really decide how many slot we will get - google does ... | 09:23 |
| blackburn | sonney2k: may it be less than ideas count? | 09:23 |
| sonney2k | I guess 2-5 (though I would love to have 10 slots) | 09:24 |
| blackburn | oh | 09:24 |
| blackburn | it seriously less than ideas count | 09:25 |
| blackburn | many ideas could 'disappear', yeap? | 09:25 |
| sonney2k | blackburn, postponed then yes and maybe some tasks are easier than thought so ideas could merge | 09:25 |
| blackburn | sonney2k: if it isn't a secret, how many students applying now? | 09:26 |
| sonney2k | blackburn, I don't know | 09:26 |
| blackburn | even roughly? you said much of emails is sent to you, etc | 09:27 |
| sonney2k | there have been many questions though - but that doesn't mean nr_applicants++ | 09:27 |
| * blackburn calculating chances :) | 09:27 | |
| sonney2k | at GSoC students can apply for up to 20 projects | 09:27 |
| blackburn | I think I will try only for 'my' idea, discussed with mentor. or it isn't a right way? | 09:28 |
| sonney2k | so even if a student applies for shogun he might not want it because he might got accepted for some oter thing. | 09:28 |
| blackburn | sonney2k: I see, I hope to get into 'accepted' list :) | 09:30 |
| sonney2k | I think that is the right way. If we have clashes to sort out we will probably also talk to everyone involved individually | 09:30 |
| blackburn | sonney2k: leaving now, thank you for all of this information, good luck | 09:31 |
| blackburn | goodbye everyone | 09:32 |
| sonney2k | l8r | 09:32 |
| -!- blackburn [~qdrgsm@109.226.104.31] has quit [Quit: Leaving.] | 09:32 | |
| sploving | sonney2k, you said in the email, support a new language is more important in fixing these bugs? | 09:36 |
| sonney2k | sploving, no but it is a good way to get into shogun's swig / typemaps | 09:37 |
| sonney2k | and I cannot mentor fixing these bugs - they might be really really tough to fix | 09:37 |
| sploving | I think so. So if write a proposal supporting lua, I am not sure if I write the plans for fixing these bugs, as I am not sure of them | 09:38 |
| sonney2k | sploving, well just write that you intend to do java first and put it in good shape, then do ruby,lua or attempt bug fixes for swig-r / octave | 09:41 |
| sploving | so many~//1 | 09:43 |
| sploving | I just intended to interface lua | 09:44 |
| sploving | as lua is my most like languge | 09:44 |
| sonney2k | sploving, I see | 09:45 |
| sonney2k | but it is really easy | 09:46 |
| sonney2k | if you can do one language I think the others take you 1-3 days | 09:46 |
| sonney2k | it is really only about writing the typemaps | 09:46 |
| sploving | docs and exmples, testsuites these all take time | 09:47 |
| sonney2k | sploving, examples - right. I forgot about the examples | 09:48 |
| sonney2k | docs and testsuite are automagically done (from the examples) | 09:48 |
| sonney2k | sploving, we write examples as functions and dump what the functions return - this way we can re-use them for the test suite | 09:49 |
| sonney2k | also we write documentation for examples just once | 09:49 |
| sonney2k | some script merges then the example with its doc | 09:49 |
| sploving | oh that is good | 09:50 |
| sonney2k | also the website will automagically have the documented examples listed | 09:50 |
| sploving | then in the plan, i do not spent time on the doc and testsuite | 09:51 |
| sonney2k | well maybe 8hrs for docs per language - to just write how you should call configure and how to start shogun in that language | 09:51 |
| sonney2k | and a pointer to the minimal example | 09:51 |
| sonney2k | (it should take less but hey...conservative planning is better) | 09:52 |
| sonney2k | sploving, but really if you can please start with java - I think that language has the biggest demand and the best numerical libraries... | 09:53 |
| sploving | sonney2k, ok. I learned java before, and it is not a problem for me. | 09:53 |
| sploving | so i think java + lua, if time permit, i will attemp to fix bugs | 09:54 |
| sonney2k | sploving, I don't mind all your lua etc plans and will support you | 09:54 |
| sonney2k | sploving, regarding the bug fixes - you coud make a plan: small examples to test for the problem, connect with swig/octave/R maintainers and then bug hunt :) | 09:55 |
| -!- aifargonos [~aifargono@46.18.27.4] has joined #shogun | 09:56 | |
| sploving | sonney2k, ok~ when i am familiar with shogun, i will have a try | 09:57 |
| sonney2k | sploving, I think I will prepare shogun's configure to get you all the languages lua, ruby, java and you just do what you manage to do with java #1 prio and then it is up to you. | 09:57 |
| sonney2k | k | 09:57 |
| sploving | ok | 09:58 |
| sploving | sonney2k, Now I need prepare for the proposal, when i have question, i will ask you again. | 09:59 |
| sonney2k | sploving, k | 09:59 |
| sonney2k | thanks and good luck! | 09:59 |
| sploving | you are welcom.:) | 09:59 |
| Yin | hi Soeren | 10:32 |
| Yin | i am reading libsvm integration part now. | 10:32 |
| Yin | it seems that the main input is a kernel matrix. | 10:32 |
| Yin | so that you changed the svm_node data structure. | 10:32 |
| sonney2k | Yin, yes | 10:33 |
| Yin | orginally it contains index and value. now it only contains a index. | 10:33 |
| sonney2k | and then some minor things like I can enable bias now etc | 10:33 |
| sonney2k | err disable | 10:33 |
| Yin | i see. thanks. | 10:35 |
| Yin | another question. regarding the sgd and vw. the current framework seems does not support stream input. ( i am not sure about this, I am just exploring the classes now) | 10:36 |
| sonney2k | well it is something new to implement | 10:38 |
| Yin | is on-deman testing supported? suppose i have a big testing dataset, the testing will go one by one without the need to load all the test data into the memory. | 10:39 |
| -!- heatkernel [b89842eb@gateway/web/freenode/ip.184.152.66.235] has quit [Quit: Page closed] | 10:45 | |
| sonney2k | Yin, this only depends on the OnlineFeature object to be implemented | 10:46 |
| sonney2k | Yin: data is accessed example by example for most methods | 10:46 |
| sonney2k | so yes | 10:46 |
| -!- blackburn [~qdrgsm@109.226.104.31] has joined #shogun | 11:08 | |
| sonney2k | yoh, alrighty, the new git repository is both on github / shogun-toolbox.org. That will really save some bandwith :) | 11:10 |
| -!- siddharth [~siddharth@117.211.88.150] has joined #shogun | 11:14 | |
| -!- sonney2k [~sonney2k@f053046246.adsl.alicedsl.de] has quit [Ping timeout: 246 seconds] | 11:25 | |
| -!- gxr [55d846c4@gateway/web/freenode/ip.85.216.70.196] has joined #shogun | 11:26 | |
| -!- seviyor [c1e20418@gateway/web/freenode/ip.193.226.4.24] has joined #shogun | 11:31 | |
| -!- siddharth [~siddharth@117.211.88.150] has quit [Ping timeout: 260 seconds] | 11:33 | |
| -!- gxr [55d846c4@gateway/web/freenode/ip.85.216.70.196] has quit [Quit: Page closed] | 11:33 | |
| -!- alex__ [82955843@gateway/web/freenode/ip.130.149.88.67] has joined #shogun | 12:09 | |
| -!- sonney2k [~sonney2k@89.204.153.96] has joined #shogun | 12:11 | |
| -!- siddharth [~siddharth@117.211.88.150] has joined #shogun | 12:17 | |
| -!- skydiver [4deac315@gateway/web/freenode/ip.77.234.195.21] has joined #shogun | 12:21 | |
| -!- alex__ [82955843@gateway/web/freenode/ip.130.149.88.67] has left #shogun [] | 12:27 | |
| -!- aifargonos [~aifargono@46.18.27.4] has quit [Read error: Operation timed out] | 12:34 | |
| -!- sonney2k [~sonney2k@89.204.153.96] has quit [Ping timeout: 246 seconds] | 12:38 | |
| Tanmoy | the project on latent variable SVM's | 12:46 |
| Tanmoy | is the key to solve the bundle method as described in the ICML paper | 12:46 |
| Tanmoy | or is there some other solvers u have in mind | 12:47 |
| -!- jabbok [56220ef5@gateway/web/freenode/ip.86.34.14.245] has joined #shogun | 12:56 | |
| jabbok | #topic | 12:56 |
| -!- sonney2k [~sonney2k@no.ml.tu-berlin.de] has joined #shogun | 13:03 | |
| -!- dvevre [b49531e3@gateway/web/freenode/ip.180.149.49.227] has joined #shogun | 13:12 | |
| -!- aifargonos [~aifargono@46.18.27.4] has joined #shogun | 13:32 | |
| -!- Yin [cf2e5c11@gateway/web/freenode/ip.207.46.92.17] has quit [Quit: Page closed] | 13:37 | |
| aifargonos | Hi! | 13:46 |
| aifargonos | I have found some email address on the old GSoC page for shogun. But I did not find it on melange page ... | 13:47 |
| aifargonos | I wrote an email with question to it .. can I expect answer ?? | 13:48 |
| aifargonos | the address was .: sonne@debian.org | 13:49 |
| aifargonos | is anybody answering on this address ?? | 13:50 |
| blackburn | aifargonos: it shoud be Sören, sonney2k | 13:52 |
| aifargonos | I guessed this ... | 13:52 |
| aifargonos | but I did not receive the answer so far ... seems like writing questions here is faster .. | 13:53 |
| aifargonos | so the question was .: Do you have any preferences among the project ideas on your idea page ?? | 13:58 |
| -!- yin [~chatzilla@207.46.92.18] has joined #shogun | 13:58 | |
| aifargonos | I mean: Which projects are more important for shogun ?? | 13:59 |
| -!- dvevre [b49531e3@gateway/web/freenode/ip.180.149.49.227] has quit [Quit: Page closed] | 13:59 | |
| aifargonos | would some projects be more benefical for shogun than others ?? | 14:00 |
| -!- sonney2k [~sonney2k@no.ml.tu-berlin.de] has quit [Quit: Leaving] | 14:15 | |
| -!- jabbok [56220ef5@gateway/web/freenode/ip.86.34.14.245] has left #shogun [] | 14:24 | |
| -!- jabbok [56220ef5@gateway/web/freenode/ip.86.34.14.245] has joined #shogun | 14:25 | |
| -!- sonney2k [~sonney2k@no.ml.tu-berlin.de] has joined #shogun | 14:29 | |
| -!- cwidmer [~quassel@connect.tuebingen.mpg.de] has joined #shogun | 14:38 | |
| -!- rieck [~rieck@bane.ml.tu-berlin.de] has joined #shogun | 14:42 | |
| * rieck 's back | 14:42 | |
| sonney2k | ...in da house | 14:42 |
| rieck | wuzzup witg +o? | 14:42 |
| -!- rieck [~rieck@bane.ml.tu-berlin.de] has left #shogun [] | 14:43 | |
| -!- rieck [~rieck@bane.ml.tu-berlin.de] has joined #shogun | 14:43 | |
| yoh | sonney2k: so I guess the script worked just fine, right? ;) | 14:44 |
| sonney2k | yoh, yes :) | 14:46 |
| -!- mode/#shogun [+o sonney2k] by ChanServ | 14:46 | |
| aifargonos | sorry for spam, but I'll ask again .: | 14:47 |
| aifargonos | Do you have any preferences among the project ideas on your idea page ?? | 14:47 |
| aifargonos | would some projects be more benefical for shogun than others ?? | 14:47 |
| -!- rieck [~rieck@bane.ml.tu-berlin.de] has quit [Quit: need to re-active irc skillz.] | 14:48 | |
| @sonney2k | aifargonos, well it is best to choose a (set of) topics you like most | 14:49 |
| aifargonos | sonney2k: that is true, but I like (more or less) a lot of them, so I am looking for other criteria :-) .. | 14:51 |
| -!- knrrrd [~knrrrd@bane.ml.tu-berlin.de] has joined #shogun | 14:51 | |
| * knrrrd confused | 14:52 | |
| -!- mode/#shogun [+o knrrrd] by sonney2k | 14:54 | |
| -!- mode/#shogun [-o knrrrd] by ChanServ | 14:54 | |
| knrrrd | pwned | 14:54 |
| -!- mode/#shogun [+o knrrrd] by ChanServ | 14:55 | |
| blackburn | sonney2k: need your opinion: KLR vs. dimensionality reduction. what do you think is prior? | 14:59 |
| blackburn | i mean there is an idea of implementing algos | 14:59 |
| blackburn | exactly nobody can implement all the things, so I have to choose | 14:59 |
| -!- knrrrd [~knrrrd@bane.ml.tu-berlin.de] has quit [Quit: leaving] | 15:00 | |
| -!- dvevre [b49531e3@gateway/web/freenode/ip.180.149.49.227] has joined #shogun | 15:01 | |
| jabbok | hi all, I was thinking that some feature maps such as these http://berkeley.intel-research.net/arahimi/random-features/ could be useful together with the integration of liblinear you mentioned in the proposed ideas could make a good project for GSoc, what do you think? | 15:02 |
| @sonney2k | jabbok, makes a lot of sense! | 15:03 |
| @sonney2k | blackburn, up to you? cwidmer ... | 15:04 |
| blackburn | sonney2k: we talked about it and he adviced me to ask you about your opinion | 15:04 |
| cwidmer | I think dim reduction could nicely complement what we have in shogun | 15:05 |
| * sonney2k agrees | 15:05 | |
| @sonney2k | we have some PCACut / kPCA in there but no more... | 15:05 |
| blackburn | so, my application will contain dim reduction + KLR (optional) | 15:05 |
| jabbok | I didn't have enough time to scan all of shogun's features, are there any feature maps such as vedaldi's recent work or something like that implemented? http://www.vlfeat.org/~vedaldi/assets/pubs/vedaldi11efficient.pdf | 15:05 |
| blackburn | sonney2k: cwidmer: thank you | 15:06 |
| @sonney2k | jabbok, we have something very related | 15:07 |
| jabbok | sonney2k: could you recomend a paper? | 15:09 |
| @sonney2k | jabbok, it is COFFIN http://sonnenburgs.de/soeren/item/coffin/ | 15:10 |
| @sonney2k | jabbok, idea is very simple: | 15:10 |
| @sonney2k | you have a few basic operations that a linear SVM needs. We implement these very abstractly in http://www.shogun-toolbox.org/doc/classshogun_1_1CDotFeatures.html and thus we can use string kernels etc in primal solvers | 15:11 |
| -!- knrrrd [~knrrrd@twoface.ml.tu-berlin.de] has joined #shogun | 15:15 | |
| jabbok | sonney2k: ok, thanks a lot, I'll take a closer look at COFFIN's internals. Are there any chances for such a project random-feature-maps + liblinear + coffin to be accepted, should I pursue with the application and the rest? | 15:20 |
| @sonney2k | jabbok, sure ... in the worst case we have to push you into slightly other directions (like onlinefeatures or so) - but I am happy | 15:21 |
| -!- knrrrd [~knrrrd@twoface.ml.tu-berlin.de] has left #shogun [] | 15:21 | |
| jabbok | sonney2k: thanks, I'll continue with the standard application procedure and if I'll have any other questions I'll use both mail and IRC | 15:22 |
| -!- blackburn [~qdrgsm@109.226.104.31] has quit [Quit: Leaving.] | 15:38 | |
| -!- aifargonos [~aifargono@46.18.27.4] has quit [Ping timeout: 246 seconds] | 15:45 | |
| -!- knrrrd [~knrrrd@bane.ml.tu-berlin.de] has joined #shogun | 15:46 | |
| knrrrd | ops? | 15:47 |
| knrrrd | you are not running any autoop scripts? | 15:47 |
| -!- mode/#shogun [+o knrrrd] by ChanServ | 15:49 | |
| @knrrrd | Cool | 15:56 |
| -!- meanerelk [~meanerelk@cpe-75-187-51-178.columbus.res.rr.com] has joined #shogun | 15:56 | |
| -!- aifargonos [~aifargono@46.18.27.35] has joined #shogun | 15:58 | |
| -!- dvevre [b49531e3@gateway/web/freenode/ip.180.149.49.227] has quit [Quit: Page closed] | 15:59 | |
| -!- knrrrd changed the topic of #shogun to: Shogun Machine Learning Toolbox | We have been accepted for GSoC 2011 | Check out our ideas list http://bit.ly/gM05O0 | GSoC Timeline http://bit.ly/gy7Pdi | How to apply http://bit.ly/i0ThO8 | 16:01 | |
| -!- knrrrd [~knrrrd@bane.ml.tu-berlin.de] has left #shogun [] | 16:04 | |
| -!- sonney2k [~sonney2k@no.ml.tu-berlin.de] has quit [Ping timeout: 246 seconds] | 16:11 | |
| -!- yorkwo [8dd5adb2@gateway/web/freenode/ip.141.213.173.178] has joined #shogun | 16:14 | |
| -!- knrrrd [~knrrrd@bane.ml.tu-berlin.de] has joined #shogun | 16:19 | |
| -!- yorkwo [8dd5adb2@gateway/web/freenode/ip.141.213.173.178] has quit [Quit: Page closed] | 16:25 | |
| -!- jabbok [56220ef5@gateway/web/freenode/ip.86.34.14.245] has quit [Quit: Page closed] | 16:28 | |
| -!- knrrrd [~knrrrd@bane.ml.tu-berlin.de] has quit [Quit: ZNC - http://znc.sourceforge.net] | 16:46 | |
| -!- knrrrd [~knrrrd@bane.ml.tu-berlin.de] has joined #shogun | 16:49 | |
| -!- mode/#shogun [+o knrrrd] by ChanServ | 16:50 | |
| -!- knrrrd [~knrrrd@bane.ml.tu-berlin.de] has left #shogun [] | 16:55 | |
| -!- knrrrd [~knrrrd@bane.ml.tu-berlin.de] has joined #shogun | 16:56 | |
| -!- mode/#shogun [+o knrrrd] by ChanServ | 16:57 | |
| -!- vladimir [~Adium@188.19.57.96] has joined #shogun | 17:37 | |
| -!- vladimir3 [~Adium@188.19.57.96] has joined #shogun | 17:38 | |
| -!- vladimir [~Adium@188.19.57.96] has quit [Read error: Connection reset by peer] | 17:38 | |
| -!- vladimir [~Adium@188.19.57.96] has joined #shogun | 17:46 | |
| -!- vladimir3 [~Adium@188.19.57.96] has quit [Read error: Connection reset by peer] | 17:46 | |
| -!- vladimir [~Adium@188.19.57.96] has quit [Client Quit] | 17:46 | |
| -!- aifargonos [~aifargono@46.18.27.35] has quit [Ping timeout: 246 seconds] | 17:54 | |
| -!- l0uis_ [~l0uis@madmax.fitnr.com] has joined #shogun | 17:58 | |
| -!- aifargonos [~aifargono@193.206.186.107] has joined #shogun | 18:09 | |
| -!- sonney2k [~sonney2k@82.113.99.21] has joined #shogun | 18:46 | |
| sonney2k | in case you haven't noticed - we have switched to git - just now | 18:56 |
| sonney2k | see https://github.com/shogun-toolbox/shogun | 18:56 |
| sonney2k | thanks to yoh we have a tiny ~20MB repository with all the changes for the source code | 18:57 |
| -!- sonney2k [~sonney2k@82.113.99.21] has quit [Ping timeout: 246 seconds] | 19:01 | |
| serialhex | tanmoy, are you awake? | 19:40 |
| serialhex | ...ok, either you are awake and not paying attention, or asleep and not able to type and sleep at the same time :P either way i'm on and will be checking back periodically so we can get this problem with python support not compiling fixed | 19:42 |
| serialhex | ...and it's pretty sweet that you're on github now, i just forked the source. this should make it MUCH easier to code for the GSoC! thanks to whoever did that!! | 19:43 |
| -!- epps [~epps@84.18.157.200] has joined #shogun | 19:45 | |
| -!- epps [~epps@84.18.157.200] has quit [Changing host] | 19:45 | |
| -!- epps [~epps@unaffiliated/epps] has joined #shogun | 19:45 | |
| -!- aifargonos [~aifargono@193.206.186.107] has quit [Ping timeout: 240 seconds] | 19:56 | |
| Tanmoy | serialhex | 20:02 |
| Tanmoy | im here | 20:02 |
| Tanmoy | lets go on with the work | 20:02 |
| -!- Tanmoy [75d35896@gateway/web/freenode/ip.117.211.88.150] has quit [Ping timeout: 253 seconds] | 20:16 | |
| -!- Tanmoy [75d35896@gateway/web/freenode/ip.117.211.88.150] has joined #shogun | 20:16 | |
| serialhex | you back tanmoy? | 20:17 |
| Tanmoy | yup | 20:18 |
| serialhex | ok, it said your ping timed out... i just got your message a little while ago | 20:18 |
| Tanmoy | u seem to be away | 20:18 |
| serialhex | ...hmm, i'm still getting used to irc, but i'm here | 20:19 |
| serialhex | anyway, i just forked the github repo, and downloaded it & ran ./configure on it, but it still tells me that it's not finding the python/numpy developer files, yet yast tells me i have them | 20:20 |
| Tanmoy | ok | 20:20 |
| serialhex | everything else seems to work fine | 20:20 |
| Tanmoy | first u need python-dev | 20:20 |
| Tanmoy | so can u get tht | 20:21 |
| Tanmoy | frm synaptic package manager | 20:21 |
| Tanmoy | i have sent u a message | 20:21 |
| serialhex | ok | 20:21 |
| serialhex | ...yeah, yast says i have it :-/ going to refresh it, and numpy | 20:22 |
| Tanmoy | yup | 20:22 |
| Tanmoy | i thin ur using ubuntu 10.04 | 20:22 |
| -!- sonney2k [~sonney2k@f053046246.adsl.alicedsl.de] has joined #shogun | 20:22 | |
| Tanmoy | s/thin/think | 20:23 |
| serialhex | opensuse 11.4 | 20:23 |
| Tanmoy | k | 20:23 |
| serialhex | ok... thats finished... testing now... | 20:24 |
| serialhex | same: Checking for Python/Python Numpy Developer Files ... no | 20:24 |
| serialhex | maybe just get the source myself & install that way? | 20:25 |
| Tanmoy | yeah | 20:25 |
| -!- skydiver [4deac315@gateway/web/freenode/ip.77.234.195.21] has quit [Ping timeout: 252 seconds] | 20:26 | |
| Tanmoy | seems better | 20:26 |
| serialhex | ok... uninstall from yast first to prevent conflicts? | 20:26 |
| Tanmoy | yes | 20:27 |
| Tanmoy | which python version | 20:27 |
| Tanmoy | 2.7 right ? | 20:27 |
| serialhex | ok... 2.7 or 3? i'm not sure if numpy supports 3 yet... | 20:27 |
| serialhex | thats what i was going with | 20:27 |
| Tanmoy | yup 2.7 | 20:27 |
| serialhex | ok | 20:28 |
| -!- dvevre [b49531e3@gateway/web/freenode/ip.180.149.49.227] has joined #shogun | 20:29 | |
| serialhex | ...bunches of packages depend on python... :-/ | 20:31 |
| Tanmoy | ? like | 20:31 |
| serialhex | ...almost too many to list, i'm just going to remove them all and not worry too much about it | 20:32 |
| Tanmoy | nd u dnt have the dependencies reqd ? | 20:32 |
| Tanmoy | well yeah just get the package | 20:33 |
| Tanmoy | well slowly c what happens | 20:33 |
| serialhex | ok... np | 20:33 |
| serialhex | ...do i need to do anything special to get the dev files set up? or simply b/c i'm using the source it'll install them? | 20:40 |
| Tanmoy | nothng special is reqd | 20:41 |
| Tanmoy | i normally do it frm synaptic package manager | 20:41 |
| serialhex | yeah... it's easier, but you dont get to see the coolness that is ./configure; make; sudo make install that you cant see in windows :P | 20:42 |
| Tanmoy | :) | 20:43 |
| Tanmoy | lets get shogun running | 20:43 |
| Tanmoy | thn u can do it the cool way | 20:43 |
| serialhex | python is make-ing right now... i'll let you know when it's done (it may be a bit, i'm running a 3ghz p4) | 20:45 |
| Tanmoy | k | 20:47 |
| serialhex | python is done... numpy next | 20:48 |
| Tanmoy | get plain numpy | 20:49 |
| serialhex | ok | 20:50 |
| -!- dvevre_ [b49531e3@gateway/web/freenode/ip.180.149.49.227] has joined #shogun | 20:54 | |
| serialhex | ...shit, i think that broke chrome :-/ i'll see if it can be fixed later tho, using firefox (backups rock!) | 20:55 |
| -!- dvevre [b49531e3@gateway/web/freenode/ip.180.149.49.227] has quit [Ping timeout: 252 seconds] | 20:56 | |
| Tanmoy | is it done ? | 21:03 |
| serialhex | compiling now | 21:04 |
| serialhex | done! trying shogun ./configure | 21:07 |
| * sonney2k feels the tense in the air | 21:07 | |
| serialhex | ffs! | 21:07 |
| serialhex | Checking for Python numpy support ... 1.6.0b1 | 21:08 |
| serialhex | Checking for Python/Python Numpy Developer Files ... no | 21:08 |
| serialhex | yeah, i want to kick it, can i kick it? | 21:08 |
| -!- epps [~epps@unaffiliated/epps] has quit [Read error: Connection reset by peer] | 21:08 | |
| -!- eppz [~epps@unaffiliated/epps] has joined #shogun | 21:08 | |
| Tanmoy | just try make and make install and see if any errors are coming | 21:09 |
| sonney2k | serialhex, please run configure with the option --interfaces=libshogun,python_modular | 21:09 |
| sonney2k | then when it fails look at configure.log | 21:09 |
| sonney2k | well post the last lines | 21:09 |
| serialhex | ok... | 21:09 |
| Tanmoy | sonney2k his python-dev is not being found | 21:09 |
| sonney2k | Tanmoy, maybe paths are just wrong? | 21:09 |
| sonney2k | Or did you forget to add some enable-shared thingy? | 21:10 |
| sonney2k | (when compiling python) | 21:10 |
| Tanmoy | he downloaded frm the source if i am right | 21:10 |
| serialhex | yep, right from source... | 21:11 |
| serialhex | sry, late lunch | 21:11 |
| serialhex | ...how much of the log file do you want? though it seems to be looking in the right spot for the files now :-/ | 21:13 |
| sonney2k | the last 10-30 lines not sure | 21:13 |
| serialhex | PyMODINIT_FUNC initsg(void); | 21:13 |
| serialhex | cc -I. -I.. -fPIC -g -Wall -O9 -fexpensive-optimizations -frerun-cse-after-loop -fcse-follow-jumps -finline-functions -fschedule-insns2 -fthread-jumps -fforce-addr -fstrength-reduce -funroll-loops -march=native -mtune=native -pthread ./configure-5679-14378.c -o ./configure-4865-14378.o -Xlinker --no-undefined -fno-strict-aliasing -fPIC -I/usr/local/include/python2.7 -I/usr/local/lib/python2.7/site-packages/numpy/core/ | 21:13 |
| serialhex | include -lutil -ldl -shared -L/usr/local/lib/python2.7/config -lpython2.7 | 21:13 |
| serialhex | ./configure-5679-14378.c: In function ‘initsg’: | 21:13 |
| serialhex | ./configure-5679-14378.c:11:12: warning: unused variable ‘outArray’ | 21:13 |
| serialhex | ./configure-5679-14378.c: At top level: | 21:13 |
| serialhex | /usr/local/lib/python2.7/site-packages/numpy/core/include/numpy/__multiarray_api.h:1532:1: warning: ‘_import_array’ defined but not used | 21:13 |
| serialhex | /usr/local/lib/python2.7/config/libpython2.7.a(signalmodule.o): In function `timeval_from_double': | 21:13 |
| serialhex | /home/serialhex/Downloads/Python-2.7.1/./Modules/signalmodule.c:112: undefined reference to `floor' | 21:13 |
| serialhex | /home/serialhex/Downloads/Python-2.7.1/./Modules/signalmodule.c:112: undefined reference to `floor' | 21:13 |
| serialhex | /home/serialhex/Downloads/Python-2.7.1/./Modules/signalmodule.c:113: undefined reference to `fmod' | 21:13 |
| serialhex | /home/serialhex/Downloads/Python-2.7.1/./Modules/signalmodule.c:113: undefined reference to `fmod' | 21:13 |
| serialhex | /usr/local/lib/python2.7/config/libpython2.7.a(floatobject.o): In function `float_divmod': | 21:13 |
| serialhex | /home/serialhex/Downloads/Python-2.7.1/Objects/floatobject.c:740: undefined reference to `fmod' | 21:13 |
| serialhex | /usr/local/lib/python2.7/config/libpython2.7.a(floatobject.o): In function `float_rem': | 21:13 |
| serialhex | /home/serialhex/Downloads/Python-2.7.1/Objects/floatobject.c:718: undefined reference to `fmod' | 21:14 |
| serialhex | /usr/local/lib/python2.7/config/libpython2.7.a(floatobject.o): In function `float_pow': | 21:14 |
| serialhex | /home/serialhex/Downloads/Python-2.7.1/Objects/floatobject.c:902: undefined reference to `pow' | 21:14 |
| serialhex | /home/serialhex/Downloads/Python-2.7.1/Objects/floatobject.c:853: undefined reference to `fmod' | 21:14 |
| serialhex | /home/serialhex/Downloads/Python-2.7.1/Objects/floatobject.c:878: undefined reference to `fmod' | 21:14 |
| serialhex | /home/serialhex/Downloads/Python-2.7.1/Objects/floatobject.c:843: undefined reference to `fmod' | 21:14 |
| serialhex | /usr/local/lib/python2.7/config/libpython2.7.a(floatobject.o): In function `_Py_double_round': | 21:14 |
| serialhex | /home/serialhex/Downloads/Python-2.7.1/Objects/floatobject.c:1123: undefined reference to `fmod' | 21:14 |
| serialhex | /usr/local/lib/python2.7/config/libpython2.7.a(longobject.o): In function `PyLong_FromString': | 21:14 |
| serialhex | /home/serialhex/Downloads/Python-2.7.1/Objects/longobject.c:1851: undefined reference to `log' | 21:14 |
| serialhex | /home/serialhex/Downloads/Python-2.7.1/Objects/longobject.c:1851: undefined reference to `log' | 21:14 |
| serialhex | /usr/local/lib/python2.7/config/libpython2.7.a(complexobject.o): In function `_Py_c_pow': | 21:14 |
| serialhex | /home/serialhex/Downloads/Python-2.7.1/Objects/complexobject.c:139: undefined reference to `hypot' | 21:14 |
| serialhex | /home/serialhex/Downloads/Python-2.7.1/Objects/complexobject.c:140: undefined reference to `pow' | 21:14 |
| serialhex | /home/serialhex/Downloads/Python-2.7.1/Objects/complexobject.c:141: undefined reference to `atan2' | 21:14 |
| serialhex | /home/serialhex/Downloads/Python-2.7.1/Objects/complexobject.c:145: undefined reference to `sincos' | 21:14 |
| serialhex | /home/serialhex/Downloads/Python-2.7.1/Objects/complexobject.c:144: undefined reference to `exp' | 21:14 |
| serialhex | /home/serialhex/Downloads/Python-2.7.1/Objects/complexobject.c:145: undefined reference to `log' | 21:14 |
| serialhex | /usr/local/lib/python2.7/config/libpython2.7.a(complexobject.o): In function `_Py_c_abs': | 21:14 |
| serialhex | /home/serialhex/Downloads/Python-2.7.1/Objects/complexobject.c:210: undefined reference to `hypot' | 21:14 |
| serialhex | collect2: ld returned 1 exit status | 21:14 |
| serialhex | no object file created | 21:14 |
| serialhex | Result is: no | 21:14 |
| serialhex | ########################################## | 21:14 |
| serialhex | 38 :P | 21:14 |
| sonney2k | serialhex, I think it indeed is not configured with shared library support | 21:14 |
| sonney2k | see the libpython2.7.a? | 21:15 |
| sonney2k | it should be libshogun2.7.so | 21:15 |
| serialhex | ...ok cause i'm looking at the files it's saying it cant find :P | 21:15 |
| serialhex | err... undef ref :-/ | 21:16 |
| serialhex | how do i compile that?? i just compiled python from src | 21:16 |
| sonney2k | serialhex, go to the python dir and issue ./configure --help | 21:17 |
| sonney2k | is there some shared option in there? | 21:17 |
| serialhex | ok | 21:17 |
| serialhex | bunches of opts, one min... | 21:18 |
| sonney2k | serialhex, well ./configure --help 2>&1 | grep share | 21:19 |
| sonney2k | --enable-shared | 21:19 |
| sonney2k | that is what you need | 21:19 |
| serialhex | yep, thats it :D | 21:19 |
| serialhex | building now | 21:20 |
| sonney2k | I am pretty sure that this will solve your problem. | 21:20 |
| serialhex | then all i have to do is fix my broken crhome :-/ | 21:20 |
| -!- ChanServ changed the topic of #shogun to: Shogun Machine Learning Toolbox | We have been accepted for GSoC 2011 | Check out our ideas list http://bit.ly/gM05O0 | GSoC Timeline http://bit.ly/gy7Pdi | How to apply http://bit.ly/i0ThO8 | This channel is logged. | 21:21 | |
| sonney2k | serialhex, not sure what chrome has to do with that? | 21:22 |
| serialhex | ...it requires python, and i uninstalled the pyton from yast so i wouldnt have conflicts... apparently it somehow broke chrome (i'v still got firefox, but i prefer chrome) | 21:23 |
| serialhex | would i have to re-build numpy too? | 21:29 |
| -!- aifargonos [~aifargono@46.18.27.35] has joined #shogun | 21:30 | |
| serialhex | ...it dosnt matter, apparently that broke python :-/ | 21:33 |
| -!- alesis-novik [~alesis@188.74.87.84] has joined #shogun | 21:38 | |
| serialhex | ffs! | 21:39 |
| serialhex | how would i uninstall it & start from scratch? | 21:40 |
| sonney2k | make uninstall? | 21:46 |
| serialhex | ok... i figured its something like that... | 21:47 |
| Tanmoy | make clean | 21:50 |
| serialhex | ahh, ok :) | 21:52 |
| Tanmoy | sorrie missed some thngs | 21:52 |
| Tanmoy | where r we know | 21:52 |
| Tanmoy | @cwidmer u thr ? | 21:52 |
| serialhex | ...i tried to ./configure --enable-shared & make and it broke python... trying with a clean slate | 21:53 |
| serialhex | ffs... i get finished compiling python and i get this: | 22:04 |
| serialhex | serialhex@livecd:~/Downloads/Python-2.7.1> python | 22:04 |
| serialhex | Traceback (most recent call last): | 22:04 |
| serialhex | File "/usr/local/lib/python2.7/site.py", line 62, in <module> | 22:04 |
| serialhex | import os | 22:04 |
| serialhex | File "/usr/local/lib/python2.7/os.py", line 398, in <module> | 22:04 |
| serialhex | import UserDict | 22:04 |
| serialhex | File "/usr/local/lib/python2.7/UserDict.py", line 83, in <module> | 22:04 |
| serialhex | import _abcoll | 22:04 |
| serialhex | File "/usr/local/lib/python2.7/_abcoll.py", line 11, in <module> | 22:04 |
| serialhex | from abc import ABCMeta, abstractmethod | 22:04 |
| serialhex | File "/usr/local/lib/python2.7/abc.py", line 8, in <module> | 22:04 |
| serialhex | from _weakrefset import WeakSet | 22:04 |
| serialhex | File "/usr/local/lib/python2.7/_weakrefset.py", line 5, in <module> | 22:04 |
| serialhex | from _weakref import ref | 22:04 |
| serialhex | ImportError: No module named _weakref | 22:04 |
| Tanmoy | hmmm | 22:08 |
| Tanmoy | now it seems python does not work | 22:08 |
| serialhex | yeah... i'm just trying a vanilla ./configure & make to see if that fixes it :-/ | 22:09 |
| serialhex | & at the same time trying to find out how to uninstall if need be | 22:10 |
| sonney2k | serialhex, try looking for __weakref | 22:10 |
| Tanmoy | do chk the log with ./configure | 22:10 |
| sonney2k | it is no use to uninstall and retry | 22:10 |
| sonney2k | I suspect your PYTHONPATH / LD_LIBRARY_PATH are wrong | 22:10 |
| alesis-novik | Anyone here willing to talk about general expectations of students for GSoC? | 22:12 |
| serialhex | ffs! vanilla make is now broken :( | 22:13 |
| serialhex | i gtg to a class... i'll be back in ~3 hours to re visit this... :-/ | 22:13 |
| Tanmoy | k | 22:13 |
| Tanmoy | ill prolly be off to sleep | 22:13 |
| Tanmoy | but do keep me posted | 22:14 |
| serialhex | ok... i'll be on all day tomorrow too, so we can work on it more then | 22:14 |
| serialhex | will do! | 22:14 |
| Tanmoy | yup | 22:15 |
| sonney2k | alesis-novik, I am not sure what to you want to hear? | 22:16 |
| sonney2k | program mostly without a lot of guidance of course :) | 22:17 |
| alesis-novik | Well, I'm interested in the implementation of EM idea. While the idea, states the final goal, what is expected during the pre-summer period, for on | 22:18 |
| alesis-novik | s/on/one | 22:18 |
| sonney2k | alesis-novik, I think you have to ask vojtech about this - looks like he is not only so via email... | 22:18 |
| alesis-novik | I'll do that, thanks. Another thing: were the projects ideas on the web-site generally suggested with GSoC being a full-time job in mind (or is this mentor specific)? | 22:22 |
| sonney2k | alesis-novik, they are all intended to be full-time jobs yes. | 22:24 |
| sonney2k | though it can be that some ideas can be done faster there are always 'extensions' | 22:25 |
| alesis-novik | As I thought. | 22:28 |
| alesis-novik | Thanks for the answers. I wish I would have found this toolbox 3 months ago... | 22:29 |
| -!- dvevre_ is now known as dvevre | 22:30 | |
| sonney2k | alesis-novik, why so? | 22:34 |
| alesis-novik | Had a data mining course which I could have used this in | 22:36 |
| sonney2k | I see... | 22:36 |
| -!- dvevre [b49531e3@gateway/web/freenode/ip.180.149.49.227] has quit [Quit: Page closed] | 22:40 | |
| -!- Tanmoy [75d35896@gateway/web/freenode/ip.117.211.88.150] has quit [Ping timeout: 252 seconds] | 23:00 | |
| -!- eppz [~epps@unaffiliated/epps] has quit [Ping timeout: 246 seconds] | 23:54 | |
| --- Log closed Tue Mar 29 00:00:36 2011 | ||
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!