--- Log opened Mon Nov 04 00:00:10 2013 | ||
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has joined #shogun | 00:27 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 00:27 | |
shogun-notifier- | shogun: Fernando Iglesias :develop * 914e85a / doc/ipython-notebooks/ica/bss_audio.ipynb: https://github.com/shogun-toolbox/shogun/commit/914e85abf2d32fd18a17307168314acc9cbbf8a9 | 00:38 |
---|---|---|
shogun-notifier- | shogun: Separate fix markdown cell in bss_audio notebook into two cells | 00:38 |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Read error: Connection reset by peer] | 00:38 | |
shogun-notifier- | shogun: so that the small description shown in the webpage is not too long | 00:38 |
shogun-notifier- | shogun: Fernando Iglesias :develop * 303809f / doc/ipython-notebooks/ica/bss_audio.ipynb: https://github.com/shogun-toolbox/shogun/commit/303809f484f37c5c8335915722cc559676e0e701 | 00:38 |
shogun-notifier- | shogun: Merge pull request #1746 from iglesias/develop | 00:38 |
shogun-notifier- | shogun: | 00:38 |
shogun-notifier- | shogun: Separate fix markdown cell in bss_audio notebook into two cells | 00:38 |
-!- wiking [~wiking@info2k1.hu] has joined #shogun | 00:38 | |
@iglesiasg | wiking, sonney2k, hey guys, a question about the notebooks and the webpage | 00:39 |
@iglesiasg | so if I update one of the notebooks shown in http://shogun-toolbox.org/page/documentation/notebook | 00:39 |
@iglesiasg | as I have done with the abstract of one of the ICA notebooks | 00:39 |
-!- pickle271 [~kevin@24-212-221-132.cable.teksavvy.com] has joined #shogun | 00:39 | |
@iglesiasg | what needs to be done so that it gets updated in the website? | 00:40 |
@iglesiasg | pickle271, hey there! How are you doing? | 00:40 |
pickle271 | good man! | 00:40 |
pickle271 | just finished moving to Ottawa Canada | 00:40 |
pickle271 | got internet setup in my new apt this morning | 00:40 |
pickle271 | starting work tomorrow! | 00:40 |
@iglesiasg | pickle271, started the new job already? | 00:40 |
@iglesiasg | nice then! | 00:40 |
@iglesiasg | pickle271, btw, I did a tiny mod in one of your notebooks | 00:41 |
@iglesiasg | pickle271, see here http://shogun-toolbox.org/page/documentation/notebook | 00:41 |
@iglesiasg | the BSS audio one had a little bit long abstract to be shown in the website | 00:41 |
@iglesiasg | so I have basically divided that markdown cell into two cells | 00:41 |
@iglesiasg | pickle271, hope there is no problem for you with that! | 00:41 |
pickle271 | ahh cool | 00:41 |
pickle271 | looks good! | 00:42 |
@iglesiasg | all right | 00:42 |
@iglesiasg | no idea when it will get updated in the website though | 00:42 |
@iglesiasg | with any luck it will happen after the nightly build | 00:42 |
pickle271 | yeah the release looks good! | 00:42 |
pickle271 | I hope to have time to do some shogun stuff in a 2 weeks or so | 00:43 |
pickle271 | I have a fun example project in mind! | 00:43 |
@iglesiasg | cool | 00:43 |
@iglesiasg | what is it about? | 00:43 |
pickle271 | still a wip | 00:43 |
pickle271 | basically making a basic AI for mario kart using some vision and ml | 00:44 |
@iglesiasg | oh, it sounds challenging! | 00:44 |
@iglesiasg | so nice :) | 00:44 |
pickle271 | yeah i dunno if it will go anywhere I've just been playing around | 00:45 |
pickle271 | I have the framework setup | 00:45 |
pickle271 | now I need to do the actual machine learning part | 00:45 |
@iglesiasg | how is the ml part? | 00:46 |
@iglesiasg | I mean, what data to use and what is the objective of the ml module | 00:46 |
pickle271 | not totally sure yet, I was actually going to see what people thought was best | 00:46 |
pickle271 | I basically have a data set that is the frames of the game and what a user (aka me) was doing on the gamepad at that time | 00:47 |
pickle271 | and thats all I have done | 00:47 |
shogun-buildbot | build #273 of osx1 - libshogun is complete: Failure [failed compile] Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/273 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 00:47 |
@iglesiasg | I see | 00:47 |
shogun-buildbot | build #274 of osx1 - libshogun is complete: Failure [failed compile] Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/274 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 00:48 |
lisitsyn | oh guys never go to computer vision | 00:48 |
lisitsyn | :D | 00:48 |
pickle271 | I saw a really old video / paper on NNs that did something similar with a real car | 00:48 |
lisitsyn | stupid amount of time tuning parameters and stuff argh | 00:48 |
pickle271 | it only drove like 15 Km per hour though | 00:48 |
pickle271 | lisitsyn: haha I know, I kept trying to drift my MASc over to ML | 00:49 |
lisitsyn | pickle271: I am still working on that shit remember? | 00:49 |
pickle271 | lisitsyn: yeah I do, man that sucks | 00:49 |
lisitsyn | everything is overdued arght | 00:49 |
lisitsyn | however I resolved most of the issues | 00:49 |
pickle271 | lisitsyn: also in CV there is this nagging notion of - should I make this whole alg better or should I add a special case to deal with this particular problem | 00:50 |
@iglesiasg | pickle271, yeah, it also reminds me about sth Ng did with an helicopter that could fly upside down | 00:50 |
lisitsyn | pickle271: apparently that shitloader is moving with up to 20 degree angle | 00:50 |
pickle271 | iglesiasg: yeah, its deffs been done before | 00:50 |
lisitsyn | so I had to ouhhuh | 00:50 |
pickle271 | iglesiasg: but no one has opened up their framework | 00:50 |
lisitsyn | pickle271: for now I am using sobel response to maintain background model :D | 00:51 |
lisitsyn | apparently it is better | 00:51 |
pickle271 | iglesiasg: so exposing the problem in terms of input and output is my main goal, not performance | 00:51 |
@iglesiasg | pickle271, I see. In any case, it would be something interesting to work on, and really cool results very likely! | 00:51 |
pickle271 | lisitsyn: cool, I'm not familiar with that technique in that context | 00:51 |
lisitsyn | pickle271: well just MOG2 from opencv on top of sobel | 00:52 |
lisitsyn | pickle271: somehow it works :D | 00:52 |
pickle271 | lisitsyn: that makes some sense actually | 00:52 |
lisitsyn | pickle271: the problem is that it outputs a bunch of points | 00:52 |
lisitsyn | I mean no strong areas | 00:52 |
lisitsyn | so camshift will come mad | 00:52 |
lisitsyn | I just fit a rectangle | 00:52 |
lisitsyn | I am afraid it is not robust enough but trainset works.. | 00:53 |
shogun-notifier- | shogun: Fernando Iglesias :develop * 91a70c6 / doc/ipython-notebooks/ica/bss_image.ipynb: https://github.com/shogun-toolbox/shogun/commit/91a70c672873a7d24b367b7a6324b04b4e2e2acb | 00:53 |
shogun-notifier- | shogun: Fix two tiny typos in BSS image notebook | 00:53 |
shogun-notifier- | shogun: Fernando Iglesias :develop * 9361cab / doc/ipython-notebooks/ica/bss_image.ipynb: https://github.com/shogun-toolbox/shogun/commit/9361cab7db278fe40a2b96da023ddccc1ab5658c | 00:53 |
shogun-notifier- | shogun: Merge pull request #1747 from iglesias/develop | 00:53 |
shogun-notifier- | shogun: | 00:53 |
shogun-notifier- | shogun: Fix two tiny typos in BSS image notebook | 00:53 |
pickle271 | lisitsyn: thats all you can do | 00:54 |
pickle271 | lisitsyn: the trainset I mean | 00:54 |
lisitsyn | pickle271: yeah but trying to keep in mind other cases | 00:54 |
lisitsyn | when e.g. a man walks here | 00:54 |
lisitsyn | no idea how would it work | 00:54 |
pickle271 | lisitsyn: I was working on an industry project and they never gave us data, so we had to keep trying to make our own that fit their use case | 00:54 |
pickle271 | project crashed and burned | 00:54 |
lisitsyn | haha | 00:55 |
lisitsyn | pickle271: no these guys gave me a lot of data | 00:55 |
lisitsyn | 2.8 gb of videos | 00:55 |
lisitsyn | pickle271: I feel pretty guilty because they consult me when I ask and etc | 00:55 |
lisitsyn | but I still overdue that shit :D | 00:55 |
@iglesiasg | all right guys, time to sleep for me | 00:55 |
@iglesiasg | see you tomorrow! | 00:55 |
lisitsyn | iglesiasg: good night! | 00:55 |
pickle271 | iglesiasg: night | 00:56 |
@iglesiasg | pickle271, good luck with the new job! | 00:56 |
pickle271 | iglesiasg: thanks! | 00:56 |
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has quit [Quit: Ex-Chat] | 00:56 | |
pickle271 | oh beans | 00:56 |
pickle271 | meant to ask him about his google interview | 00:56 |
lisitsyn | pickle271: whom? | 00:56 |
pickle271 | iglesias | 00:56 |
lisitsyn | ah | 00:56 |
pickle271 | I'll catch him later | 00:57 |
-!- zxtx [~zv@ip-64-134-197-166.public.wayport.net] has joined #shogun | 01:01 | |
-!- zxtx [~zv@ip-64-134-197-166.public.wayport.net] has quit [Ping timeout: 248 seconds] | 01:07 | |
shogun-buildbot | build #265 of precise - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/265 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 01:14 |
-!- zxtx [~zv@ip-64-134-197-166.public.wayport.net] has joined #shogun | 01:22 | |
shogun-buildbot | build #266 of precise - libshogun is complete: Success [build successful] Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/266 | 01:32 |
-!- travis-ci [~travis-ci@ec2-23-20-252-49.compute-1.amazonaws.com] has joined #shogun | 01:38 | |
travis-ci | [travis-ci] it's Fernando Iglesias'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/13451462 | 01:38 |
-!- travis-ci [~travis-ci@ec2-23-20-252-49.compute-1.amazonaws.com] has left #shogun [] | 01:38 | |
shogun-buildbot | build #100 of clang34 - undefined behaviour analysis is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20undefined%20behaviour%20analysis/builds/100 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 01:45 |
-!- pickle271 [~kevin@24-212-221-132.cable.teksavvy.com] has left #shogun [] | 02:00 | |
shogun-buildbot | build #99 of clang34 - thread analysis is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/99 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 02:01 |
-!- zxtx [~zv@ip-64-134-197-166.public.wayport.net] has quit [Ping timeout: 245 seconds] | 02:02 | |
shogun-buildbot | build #275 of osx1 - libshogun is complete: Failure [failed compile] Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/275 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 02:05 |
-!- zxtx [~zv@ip-64-134-197-166.public.wayport.net] has joined #shogun | 02:19 | |
-!- zxtx [~zv@ip-64-134-197-166.public.wayport.net] has quit [Ping timeout: 260 seconds] | 02:36 | |
shogun-buildbot | build #118 of clang34 - static analysis is complete: Failure [failed analyse] Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20static%20analysis/builds/118 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 02:47 |
-!- zxtx [~zv@ip-64-134-197-166.public.wayport.net] has joined #shogun | 02:51 | |
-!- wiking_ [~wiking@info2k1.hu] has joined #shogun | 02:57 | |
shogun-buildbot | build #101 of clang34 - undefined behaviour analysis is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20undefined%20behaviour%20analysis/builds/101 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 02:59 |
-!- wiking [~wiking@info2k1.hu] has quit [Read error: Connection reset by peer] | 02:59 | |
shogun-buildbot | build #100 of clang34 - thread analysis is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/100 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 03:14 |
-!- zxtx [~zv@ip-64-134-197-166.public.wayport.net] has quit [Ping timeout: 246 seconds] | 03:23 | |
-!- zxtx [~zv@ip-64-134-197-166.public.wayport.net] has joined #shogun | 03:39 | |
-!- sonne|osx_ [~sonne@f053047066.adsl.alicedsl.de] has joined #shogun | 03:45 | |
-!- sonne|osx [~sonne@f053034105.adsl.alicedsl.de] has quit [Ping timeout: 245 seconds] | 03:47 | |
-!- sonne|osx_ is now known as sonne|osx | 03:47 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 03:53 | |
shogun-buildbot | build #119 of clang34 - static analysis is complete: Failure [failed analyse] Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20static%20analysis/builds/119 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 03:56 |
-!- zxtx [~zv@ip-64-134-197-166.public.wayport.net] has quit [Ping timeout: 245 seconds] | 04:02 | |
-!- Jagged [~NULL@vtluug/member/jagged] has joined #shogun | 04:02 | |
shogun-buildbot | build #102 of clang34 - undefined behaviour analysis is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20undefined%20behaviour%20analysis/builds/102 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 04:05 |
-!- zxtx [~zv@ip-64-134-197-166.public.wayport.net] has joined #shogun | 04:06 | |
shogun-buildbot | build #101 of clang34 - thread analysis is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20thread%20analysis/builds/101 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 04:20 |
shogun-buildbot | build #120 of clang34 - static analysis is complete: Failure [failed analyse] Build details are at http://buildbot.shogun-toolbox.org/builders/clang34%20-%20static%20analysis/builds/120 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 05:02 |
-!- zxtx [~zv@ip-64-134-197-166.public.wayport.net] has quit [Ping timeout: 272 seconds] | 05:04 | |
-!- zxtx [~zv@129-79-241-148.dhcp-bl.indiana.edu] has joined #shogun | 05:22 | |
-!- sonne|osx [~sonne@f053047066.adsl.alicedsl.de] has quit [Quit: sonne|osx] | 06:33 | |
-!- zxtx [~zv@129-79-241-148.dhcp-bl.indiana.edu] has quit [Read error: Operation timed out] | 08:20 | |
-!- Saurabh7 [~Saurabh7@117.200.191.204] has joined #shogun | 08:28 | |
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has joined #shogun | 08:51 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 08:51 | |
-!- zxtx [~zv@2001:558:6033:150:90ff:7a7c:7d5b:84d] has joined #shogun | 09:01 | |
-!- sonne|osx [~sonne@82.113.121.187] has joined #shogun | 09:12 | |
sonne|osx | uhh la la la | 09:12 |
lisitsyn | sonne|osx: la la la uhh! | 09:13 |
@iglesiasg | you in France guys? :) | 09:16 |
-!- sonne|osx [~sonne@82.113.121.187] has quit [Quit: sonne|osx] | 09:19 | |
-!- sonne|osx [~sonne@82.113.121.187] has joined #shogun | 09:25 | |
lisitsyn | iglesiasg: JA | 09:28 |
lisitsyn | :) | 09:29 |
@iglesiasg | oui then :D | 09:29 |
lisitsyn | iglesiasg: NEIN :D | 09:35 |
lisitsyn | iglesiasg: have you seen 'predator' tracker? | 09:36 |
@iglesiasg | lisitsyn, nope | 09:36 |
lisitsyn | iglesiasg: http://www.youtube.com/watch?v=1GhNXHCQGsM | 09:36 |
lisitsyn | iglesiasg: it learns the appearance of object to support further tracking | 09:37 |
@iglesiasg | lisitsyn, cool stuff | 09:39 |
lisitsyn | iglesiasg: don't know whether it is state of the art still though | 09:39 |
lisitsyn | iglesiasg: but the idea is here now | 09:39 |
@iglesiasg | I think the fact that it can change from tracking one thing to another so quickly is advanced stuff | 09:40 |
@iglesiasg | like he gave the bounding box and started tracking another thing | 09:40 |
lisitsyn | iglesiasg: yeah it is pretty much advanced | 09:40 |
lisitsyn | iglesiasg: but I like the idea actually | 09:41 |
lisitsyn | when you maintain a model | 09:41 |
@iglesiasg | it is open source | 09:45 |
lisitsyn | iglesiasg: yeah with crazy amount of matlab code :) | 09:51 |
sonne|osx | as usual | 09:51 |
-!- sonne|osx [~sonne@82.113.121.187] has quit [Quit: sonne|osx] | 09:55 | |
besser82 | lisitsyn: shouldn't it be easy to run MatLab-code with Octave?!? | 10:28 |
lisitsyn | besser82: yes why not | 10:28 |
lisitsyn | besser82: I meant matlab = unreadable :D | 10:28 |
besser82 | lisitsyn: Ahh ;) | 10:28 |
lisitsyn | and like matlab software ~ bad designed software | 10:28 |
besser82 | lisitsyn: I have a personal matlab to python compiler :-P | 10:29 |
besser82 | lisitsyn: btw. Marburg University starts dropping Matlab ;) | 10:30 |
besser82 | lisitsyn: replacement: Pyton + Scipy | 10:30 |
lisitsyn | besser82: that's the human way to do science | 10:30 |
besser82 | lisitsyn: :D | 10:31 |
besser82 | lisitsyn: They actually dropped it because of increasing license-fees :-P | 10:31 |
lisitsyn | besser82: that's ok too, no matter how they come from the dark side :D | 10:32 |
besser82 | lisitsyn: XD | 10:33 |
besser82 | lisitsyn: Or or like Uncle Wong says: Da FOSSSSSSSS iz da so mut beeeeedaahhh ;) | 10:34 |
lisitsyn | besser82: foss be good | 10:35 |
lisitsyn | :D | 10:35 |
besser82 | lisitsyn: btw. I just engaged an "Applied Physics BSc." for optimizing the SVM^bright algos ;) | 10:36 |
lisitsyn | besser82: how? | 10:37 |
besser82 | lisitsyn: He's a friend of mine from the Fedora community and I asked him to do the sucking work :-P | 10:37 |
lisitsyn | ahh | 10:37 |
lisitsyn | besser82: do you want to release it standalone too? | 10:38 |
besser82 | lisitsyn: Like "Hey Budyy! Ya know about SVM^sh*t? I wanna make a FOSS alternative which is much better, but need someone to look if there's room for optimazation in algos" | 10:38 |
besser82 | lisitsyn: Yes | 10:39 |
lisitsyn | besser82: as shared lib? | 10:39 |
besser82 | lisitsyn: I actually planned it to be a C++ shlib && cli-tools | 10:39 |
besser82 | lisitsyn: with the possibility to dlopen custom-kernels through a factory ;) | 10:39 |
lisitsyn | besser82: I see | 10:40 |
besser82 | lisitsyn: and possibly some Cython-style Python-module | 10:40 |
lisitsyn | besser82: I actually would like to see shogun using some libs | 10:41 |
lisitsyn | instead of being thaat gigantic | 10:41 |
besser82 | lisitsyn: me too ;) | 10:41 |
lisitsyn | may be liblinear should be 'fixed' that way too | 10:41 |
besser82 | lisitsyn: What's the beef with liblinear? | 10:41 |
besser82 | lisitsyn: I'm maintaining that for FC | 10:42 |
besser82 | lisitsyn: and have good connection to upstream ;) | 10:42 |
lisitsyn | besser82: well usual liblinear doesn't support that 'coffin' thing | 10:42 |
lisitsyn | i.e. callback dot and add operations | 10:42 |
lisitsyn | they act like it is dense | 10:42 |
besser82 | lisitsyn: I see; so it was customized by S?ren, rye? | 10:43 |
lisitsyn | yeah | 10:43 |
lisitsyn | besser82: multiclass was customized by me | 10:43 |
lisitsyn | iirc | 10:43 |
besser82 | lisitsyn: Lemme see, if can get a patch set upstream ;) | 10:43 |
lisitsyn | besser82: do you know its developers? | 10:43 |
besser82 | lisitsyn: Yeah, not in person. but yes | 10:44 |
lisitsyn | besser82: we can prepare a patch and propose | 10:44 |
lisitsyn | I have no idea whether sonne|work proposed it before | 10:44 |
lisitsyn | probably they know about that way | 10:44 |
besser82 | lisitsyn: And it shouldn't be much diffcult for me to get it in, with me wearing my "RedHat" :-P | 10:44 |
lisitsyn | besser82: if it was shared library I wouldn't mind if we used linking | 10:45 |
lisitsyn | no idea what others think | 10:45 |
lisitsyn | but I believe it is better | 10:45 |
besser82 | lisitsyn: liblinear Makefile builds a shlib by default ;) | 10:45 |
lisitsyn | besser82: well yeah I mean 'proper' shlib ;) | 10:46 |
sonne|work | besser82: it is not easy to do that with liblinear but possible - in the end you need to boil down all liblinear operations into dot, dense_dot, add_to_dense_vec such stuff | 10:46 |
lisitsyn | which can be used easier | 10:46 |
sonne|work | rather similar to what ocas does | 10:46 |
besser82 | lisitsyn: I see. So heavy API / ABI changes? | 10:46 |
lisitsyn | sonne|work: do you think it is better to have multiple shlibs around? | 10:46 |
lisitsyn | besser82: well yeah some changes | 10:46 |
sonne|work | in addition one would need to register print and error functions, malloc/free/calloc functions etc | 10:47 |
lisitsyn | ah yeah | 10:47 |
sonne|work | not easy but surely possible | 10:47 |
besser82 | sonne|work, lisitsyn: Who will dooooo it? | 10:47 |
lisitsyn | besser82: actually that's their drawback | 10:47 |
lisitsyn | I mean when you develop a library | 10:47 |
lisitsyn | you should think of providing that callbacks :D | 10:48 |
lisitsyn | besser82: I am still unavailable but I want to do that some day | 10:48 |
sonne|work | besser82: what is their incentive to provide a framework like this or even to accept patches like this? | 10:48 |
sonne|work | I don't see it happening | 10:48 |
lisitsyn | yeah I am afraid of that too | 10:48 |
lisitsyn | I see it is much easier to just maintain a fork of liblinear | 10:49 |
besser82 | sonne|work, lisitsyn: I can possibly get that stuff into SVM^bright as well ;) | 10:49 |
lisitsyn | with such changes | 10:49 |
besser82 | sonne|work, lisitsyn: If that would make sense? | 10:49 |
lisitsyn | besser82: no idea! | 10:49 |
sonne|work | besser82: it is work no idea who will do it | 10:49 |
lisitsyn | besser82: as for svm bright - please keep everything callbacked' :) | 10:50 |
besser82 | lisitsyn: Sure ;) | 10:50 |
lisitsyn | malloc/print/data access | 10:50 |
lisitsyn | errors too may be | 10:50 |
besser82 | lisitsyn: I actually planned to setup a nice class-based API/ABI ;) | 10:51 |
sonne|work | there is some update_linear component stuff in svm light which we have heavily changed | 10:51 |
sonne|work | all far from easy to migrate even w/ a clean API | 10:51 |
besser82 | lisitsyn: So full controll for your code ;) | 10:51 |
besser82 | lisitsyn: and some nice exception handling in case of problems ;) | 10:53 |
lisitsyn | besser82: yeah | 10:53 |
besser82 | lisitsyn: Everything in fully transparent ;) | 10:55 |
besser82 | lisitsyn: providing easy-all-in-one-calls && advanced its-all-up-to-you-headers ;) | 10:56 |
lisitsyn | besser82: haha too much marketing already ;) | 10:57 |
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has quit [Quit: Ex-Chat] | 11:06 | |
-!- thoralf [~thoralf@enki.zib.de] has joined #shogun | 11:53 | |
thoralf | Hey. | 11:53 |
besser82 | thoralf: Hall?chen :D | 12:05 |
thoralf | Guten Morgen :) | 12:06 |
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has joined #shogun | 13:44 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 13:44 | |
-!- benibadman [~benibadma@94.135.236.129] has joined #shogun | 13:54 | |
-!- wiking_ is now known as wiking | 14:05 | |
-!- wiking [~wiking@info2k1.hu] has quit [Changing host] | 14:05 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 14:05 | |
-!- mode/#shogun [+o wiking] by ChanServ | 14:05 | |
-!- Saurabh7 [~Saurabh7@117.200.191.204] has quit [Ping timeout: 245 seconds] | 14:22 | |
-!- Abhinav_ [~Abhinav@14.139.155.210] has joined #shogun | 14:40 | |
Abhinav_ | where can i find the shogun source | 14:40 |
Abhinav_ | if i want to work on it as a project | 14:40 |
Jagged | https://github.com/shogun-toolbox | 14:43 |
sonne|work | wiking: when I fiddle around in the shogun-cloud template - how could I install it later? | 14:43 |
@wiking | sonne|work: commit it and let me know... but i still need to sync | 14:52 |
@wiking | but never mind | 14:52 |
@wiking | just do it in the repo | 14:52 |
@wiking | and i'll fix it then | 14:52 |
sonne|work | wiking: ok | 14:52 |
-!- Abhinav_ [~Abhinav@14.139.155.210] has quit [Ping timeout: 246 seconds] | 14:56 | |
-!- pickle27 [~kevin@24-212-221-132.cable.teksavvy.com] has joined #shogun | 15:19 | |
-!- pickle27 [~kevin@24-212-221-132.cable.teksavvy.com] has left #shogun [] | 15:22 | |
-!- benibadm_ [~benibadma@94.135.236.129] has joined #shogun | 15:24 | |
-!- benibadman [~benibadma@94.135.236.129] has quit [Ping timeout: 272 seconds] | 15:27 | |
-!- lisitsyn [~lisitsyn@109-226-82-69.clients.tlt.100megabit.ru] has quit [Quit: Leaving.] | 16:27 | |
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has quit [Quit: Ex-Chat] | 16:29 | |
-!- sonne|osx [~sonne@82.113.106.57] has joined #shogun | 17:22 | |
sonne|osx | wiking: jinja2.exceptions.TemplateNotFound | 17:37 |
sonne|osx | TemplateNotFound: bootstrap_responsive.html | 17:37 |
-!- sonne|osx [~sonne@82.113.106.57] has quit [Quit: sonne|osx] | 17:40 | |
-!- benibadm_ [~benibadma@94.135.236.129] has quit [Read error: Operation timed out] | 17:52 | |
-!- thoralf [~thoralf@enki.zib.de] has quit [Quit: Konversation terminated!] | 17:52 | |
@wiking | sonne|work: wrong version of flask-bootstrap | 17:57 |
-!- zxtx [~zv@2001:558:6033:150:90ff:7a7c:7d5b:84d] has quit [Ping timeout: 245 seconds] | 18:14 | |
-!- zxtx [~zv@129-79-241-148.dhcp-bl.indiana.edu] has joined #shogun | 19:03 | |
-!- new_lido [~walid@41.218.173.235] has joined #shogun | 19:19 | |
-!- lisitsyn [~lisitsyn@80.252.20.67] has joined #shogun | 19:32 | |
-!- sonne|osx [~sonne@f053047066.adsl.alicedsl.de] has joined #shogun | 20:18 | |
-!- benibadman [~benibadma@port-92-206-121-116.dynamic.qsc.de] has joined #shogun | 20:20 | |
-!- benibadman [~benibadma@port-92-206-121-116.dynamic.qsc.de] has quit [Ping timeout: 240 seconds] | 20:24 | |
besser82 | wiking: Yo! | 20:34 |
besser82 | wiking: https://fedorahosted.org/fedora-infrastructure/ticket/4032#comment:6 | 20:34 |
besser82 | wiking: c-a has been fixed in meaning of storage :) | 20:34 |
@wiking | ok so now it's already there? | 20:35 |
besser82 | wiking: /dev/vdb is up and just needs formatting && fstab :D | 20:35 |
@wiking | besser82: shall i do it? | 20:35 |
besser82 | wiking: if you want :) | 20:35 |
@wiking | ok | 20:36 |
@wiking | let's see | 20:36 |
@wiking | fdisk :) | 20:36 |
besser82 | wiking: or sfdisk or parted :-P | 20:36 |
@wiking | hehehe done with fdisk :P | 20:36 |
@wiking | /dev/vdb1 63 52428799 26214368+ 83 Linux | 20:36 |
@wiking | should be good | 20:37 |
besser82 | sonney2k, sonne|osx, sonne|work: Yo! :D Do you have a minute? | 20:37 |
besser82 | wiking: +1 :D | 20:37 |
@wiking | besser82: where should i mount it? | 20:37 |
besser82 | wiking: I'd use /srv | 20:37 |
besser82 | wiking: because of selinux ;) | 20:37 |
@wiking | besser82: in /etc/fstab is it ok if i refer to it by /dev/vdb1 | 20:38 |
@wiking | or should it be UUID? | 20:38 |
besser82 | wiking: Like you want ;) | 20:38 |
besser82 | wiking: shouldn't change | 20:38 |
besser82 | wiking: but with autocheck enabled ;) like ... 1 1$ | 20:38 |
@wiking | /dev/vdb1 25G 44M 24G 1% /srv | 20:39 |
besser82 | wiking: and don't forget to `touch /.autorelabel && reboot` | 20:39 |
besser82 | wiking: to make selinux happy | 20:39 |
@wiking | eh? | 20:40 |
besser82 | wiking: this will make selinux restore all default context on new mounted fs ;) | 20:41 |
besser82 | wiking: and reboot | 20:41 |
besser82 | wiking: or shall I do just quick? | 20:42 |
@wiking | besser82: yeps | 20:43 |
besser82 | wiking: allrighty :D You will be logged out in a second ;) | 20:44 |
-!- sonne|osx [~sonne@f053047066.adsl.alicedsl.de] has quit [Quit: sonne|osx] | 20:44 | |
@wiking | i think i've logged out before | 20:49 |
@wiking | besser82: let me know when it's back | 20:49 |
@wiking | ah ok it's back | 20:51 |
@wiking | cool | 20:51 |
besser82 | wiking: Everything is fine? | 20:52 |
@wiking | seems like | 20:53 |
besser82 | wiking: for me too ;) | 20:53 |
@wiking | ok now we have to figure out how to share container instances between nodes :P | 20:58 |
@wiking | but since docker uses aufs | 20:58 |
@wiking | we can just use the diff of the container image | 20:58 |
@wiking | i.e. the changes since it got fired up | 20:59 |
@wiking | and that cannot be toooo much | 20:59 |
@wiking | i.e. couple of megabytes | 20:59 |
besser82 | wiking: I'm afraid AUFS and selinux don't play nice with oneanother :( | 20:59 |
@wiking | mmm well i guess we'll have to try :P | 21:00 |
-!- taylan [53a095ad@gateway/web/freenode/ip.83.160.149.173] has joined #shogun | 21:19 | |
taylan | Hi everyone..I got a couple of small questions about training a classifier with shogun.. | 21:21 |
taylan | - 1. Do I need to scale the features I provide? | 21:22 |
@wiking | taylan: 1) depends on the classifier you are trying to use. but as a general rule of thumb it's worth to try scaling | 21:23 |
taylan | 2. Is there a module which optimizes kernel parameters, or should I write a wrapper around the crossvalidation evaluator to optimize the choice of kernel parameters? | 21:23 |
@wiking | 2) modelselection should do this for you | 21:23 |
taylan | Thanks wiking! | 21:23 |
@wiking | nw | 21:23 |
taylan | I will check the model selection.. | 21:23 |
-!- taylan [53a095ad@gateway/web/freenode/ip.83.160.149.173] has quit [Quit: Page closed] | 21:30 | |
-!- iglesiasg [~iglesias@s83-179-44-135.cust.tele2.se] has joined #shogun | 21:32 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 21:32 | |
@wiking | awfully quite here in shogunland since release :P | 22:25 |
lisitsyn | wiking: see you in february :D | 22:48 |
-!- new_lido [~walid@41.218.173.235] has quit [Quit: Leaving] | 22:52 | |
-!- zxtx [~zv@129-79-241-148.dhcp-bl.indiana.edu] has quit [Ping timeout: 246 seconds] | 23:17 | |
-!- zxtx [~zv@129-79-241-148.dhcp-bl.indiana.edu] has joined #shogun | 23:18 | |
@wiking | :D | 23:20 |
--- Log closed Tue Nov 05 00:00:11 2013 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!