--- Log opened Sat Feb 16 00:00:14 2013 | ||
--- Day changed Sat Feb 16 2013 | ||
n4nd0 | sonney2k: where is that address from? | 00:00 |
---|---|---|
@sonney2k | n4nd0, you need to get the full trace | 00:08 |
n4nd0 | let me try again | 00:08 |
@sonney2k | Episode.IV / A.NEW.HOPE | 00:08 |
@sonney2k | etc | 00:09 |
n4nd0 | nothing here | 00:09 |
@sonney2k | star wars! | 00:09 |
n4nd0 | but I have seen something like through telnet some time | 00:09 |
@sonney2k | http://beaglenetworks.net/post/42707829171/star-wars-traceroute | 00:10 |
@sonney2k | the guy who did it | 00:10 |
n4nd0 | sonney2k: not around any more I guess? | 03:08 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 03:10 | |
-!- cavalier- [~cavalier-@70.118.110.73] has quit [] | 03:44 | |
-!- FSCV_ [~FSCV@173.254.212.46] has quit [Quit: Leaving] | 05:15 | |
-!- lambday [0e8b6149@gateway/web/freenode/ip.14.139.97.73] has joined #shogun | 07:54 | |
lambday | hello, this is soumyajit again.. I have a few queries regarding the "fast food" idea... I was unable to see the video but I went through the paper that talks about introducing Taylor features to approximate the kernel... | 08:08 |
lambday | firstly, they apply the method for Gaussian kernels but says "a similar approach could also work for other kernels which depend on distances or inner products between feature vectors, e.g. the sigmoid kernel."... does that mean all CDotKernels in shogun? | 08:09 |
lambday | secondly, how do you plan to implement this? say, for Gaussian kernel, extending it to another class that uses this method? and do it for all the kernels that the method applies to? | 08:10 |
lambday | thirdly, where can I look for some paper reference of the "fast food" talk? | 08:23 |
-!- sumit [73f91219@gateway/web/freenode/ip.115.249.18.25] has joined #shogun | 08:34 | |
blackburn | lambday: it is only a draft and more references will come once mentor of that idea describe it more | 09:25 |
lambday | blackburn: okay.. thanks :) | 09:26 |
-!- sumit [73f91219@gateway/web/freenode/ip.115.249.18.25] has quit [Ping timeout: 245 seconds] | 09:26 | |
-!- lambday [0e8b6149@gateway/web/freenode/ip.14.139.97.73] has left #shogun [] | 09:34 | |
-!- sumit [ca4eaca2@gateway/web/freenode/ip.202.78.172.162] has joined #shogun | 10:05 | |
-!- pari [6eeaec2a@gateway/web/freenode/ip.110.234.236.42] has joined #shogun | 11:21 | |
-!- flxb [~flx@g226040027.adsl.alicedsl.de] has joined #shogun | 11:41 | |
wiking | mmm | 11:55 |
wiking | i'm wondeirng about this fast food | 11:55 |
wiking | as there are other approaches as well | 11:55 |
wiking | like from vedaldi | 11:55 |
wiking | HKM ++ | 11:56 |
wiking | so to say ;) | 11:56 |
pari | hi all, I have just finished installing shogun. But i am not sure if everything is installed correctly. Is there any way to check it? | 11:57 |
wiking | pari well try to run an example? :) | 11:57 |
wiking | blackburn: ping | 11:57 |
pari | ok, thanks.. let me see | 11:59 |
pari | i couldn't get the example running i am getting the following error: basic_minimal.cpp:(.text+0x29): undefined reference to `shogun::init_shogun(void (*)(_IO_FILE*, char const*), void (*)(_IO_FILE*, char const*), void (*)(_IO_FILE*, char const*), void (*)(bool&, bool&))' basic_minimal.cpp:(.text+0x2e): undefined reference to `shogun::exit_shogun()' collect2: ld returned 1 exit status | 12:32 |
pari | what could be possibly wrong? | 12:32 |
wiking | mmm | 12:51 |
wiking | i reckon the path where the shogun library is installed (if u did make install) is not in the library path | 12:51 |
-!- pari [6eeaec2a@gateway/web/freenode/ip.110.234.236.42] has quit [Ping timeout: 245 seconds] | 12:54 | |
-!- flxb [~flx@g226040027.adsl.alicedsl.de] has quit [Ping timeout: 244 seconds] | 12:56 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 13:18 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 13:37 | |
-!- sumit [ca4eaca2@gateway/web/freenode/ip.202.78.172.162] has quit [Ping timeout: 245 seconds] | 13:37 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 15:28 | |
n4nd0 | sonney2k: around? | 15:40 |
n4nd0 | sonney2k: proof of concept for the website http://iglesiashogun.alwaysdata.net/ | 15:44 |
n4nd0 | I have isolated there the addition of the new bar | 15:44 |
n4nd0 | let me know if you were thinking of something like that | 15:44 |
blackburn | wiking: yeah fast food is not the only way | 16:16 |
-!- flxb [~flx@g226040027.adsl.alicedsl.de] has joined #shogun | 16:26 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 16:33 | |
-!- flxb [~flx@g226040027.adsl.alicedsl.de] has quit [Ping timeout: 252 seconds] | 16:45 | |
-!- heiko1 [~heiko@b019b9ec.bb.sky.com] has joined #shogun | 16:48 | |
@sonney2k | heiko1, hey there - I have a question about serialization again | 20:13 |
@sonney2k | when an array of SGVectors is de-serialized | 20:13 |
@sonney2k | then the serialization code will call SG_MALLOC(SGVector<type>, len) right? | 20:14 |
@sonney2k | wiking, I just did make valgrind in tests/unit | 20:14 |
@sonney2k | guess what - ocas examples have uninited memory reads | 20:14 |
@sonney2k | so no wonder we see weird stuff | 20:14 |
blackburn | sonney2k: is that mcocas? | 20:32 |
wiking | sonney2k: lol? | 20:33 |
@sonney2k | ocas/mcocas | 20:45 |
blackburn | sonney2k: so libocas? I tested mc one before IIRC | 20:59 |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 21:06 | |
shogun-notifier- | shogun: Sergey Lisitsyn :master * 749d966 / .gitignore: https://github.com/shogun-toolbox/shogun/commit/749d96694ba2f35b84f786bccb57614179415fec | 21:06 |
shogun-notifier- | shogun: Ignore gcda file extension | 21:06 |
shogun-notifier- | shogun: Sergey Lisitsyn :master * c4d485e / src/shogun/lib/tapkee/tapkee_shogun.cpp: https://github.com/shogun-toolbox/shogun/commit/c4d485eea0740e9357812b55a166e18120d5c6c4 | 21:06 |
shogun-notifier- | shogun: Fixed totally wrong tapkee integration feature access callback | 21:06 |
-!- in3xes [~in3xes@122.174.121.230] has joined #shogun | 21:21 | |
shogun-notifier- | shogun: Sergey Lisitsyn :master * 4091d89 / src/Makefile.template: https://github.com/shogun-toolbox/shogun/commit/4091d89436dcfda623a16ce7cfd59033db045fc5 | 21:30 |
shogun-notifier- | shogun: Added gcda for clear target | 21:30 |
shogun-notifier- | shogun: Sergey Lisitsyn :master * 9860c73 / src/shogun/converter/LaplacianEigenmaps.cpp: https://github.com/shogun-toolbox/shogun/commit/9860c736661ff4aba365c0986911d9883b176cfb | 21:30 |
shogun-notifier- | shogun: Removed unnecessary and memory-wise wrong distance initialization in laplacian eigenmaps | 21:30 |
@sonney2k | wiking, please run make valgrind in unit tests there are also errors in LatentModel | 21:48 |
shogun-notifier- | shogun: Soeren Sonnenburg :master * 99a8236 / src/shogun/.project: https://github.com/shogun-toolbox/shogun/commit/99a82365fa46dede16d93d0ef12be9027125b532 | 21:54 |
shogun-notifier- | shogun: add eclipse project file | 21:54 |
shogun-notifier- | shogun: Soeren Sonnenburg :master * 2b60dbe / / (9 files): https://github.com/shogun-toolbox/shogun/commit/2b60dbe596d28aaabed9122799b1aede31c484c5 | 21:54 |
shogun-notifier- | shogun: Fix serialization of Sparse features | 21:54 |
shogun-notifier- | shogun: | 21:54 |
shogun-notifier- | shogun: - remove inplace (de-)constructors for SGSparse | 21:54 |
shogun-notifier- | shogun: - add malloc support for SGSparseVector<T> | 21:54 |
shogun-notifier- | shogun: - enable unit test for sparse feature serialization | 21:54 |
heiko1 | sonney2k, eclipse project file? | 21:55 |
@sonney2k | heiko1, yeah | 21:55 |
@sonney2k | with own shogun formatter :D | 21:56 |
heiko1 | sonney2k, array of SGVectors, mmmh cannot really remember got to check the code | 21:56 |
@sonney2k | heiko1, I managed to fix this | 21:56 |
@sonney2k | heiko1, I did some magic to SG_MALLOC | 21:56 |
heiko1 | sonney2k, that is cool, so I can just always press ctrl-shift-f and formatting will be fine? | 21:56 |
heiko1 | sonney2k, cool! | 21:56 |
@sonney2k | whenever you do SG_MALLOC(SGVector<T>, len) then it will call new SGVector<T>[len](); | 21:57 |
@sonney2k | this will call the inplace constructor automagically | 21:57 |
@sonney2k | and SG_FREE() will know what to do too | 21:57 |
heiko1 | I see | 21:57 |
heiko1 | cool! | 21:57 |
@sonney2k | so this will e.g.make dynarray of SGVector work etc | 21:57 |
heiko1 | btw thanks for fixing this base-class problem | 21:58 |
@sonney2k | slightly stupid to have so many sg_malloc functions now but this fixes things at least in a nice way for the user | 21:58 |
@sonney2k | underneath it is hacky though | 21:58 |
heiko1 | yeah, agree, this is a good solution | 21:58 |
@sonney2k | but I don't see a way around this if we don't want to switch to new[] / new() | 21:59 |
@sonney2k | which is bad because we then can no longer interact with C malloc'd memory directly | 21:59 |
heiko1 | nono, agreed on this | 21:59 |
@sonney2k | (no more realloc / direclty using python matrix memory etc) | 21:59 |
@sonney2k | heiko1, so while we are at it - serialization I mean - what would it cost to support SGMatrixList / SGNDArray? | 22:00 |
blackburn | sonney2k: so you like working in eclipse now? | 22:00 |
heiko1 | sonney2k, you mean how much effort that would be? | 22:00 |
heiko1 | Can't really tell, these things tend to be a bit unpredictable | 22:01 |
heiko1 | For what do we need those? | 22:01 |
heiko1 | anything in particular? | 22:01 |
@sonney2k | e.g. serialization of QDA | 22:01 |
@sonney2k | but yes in general I would like to have serialization support for all SG* datatypes | 22:02 |
heiko1 | yes true | 22:02 |
heiko1 | dont know, at some point just has to take the pain :) | 22:02 |
heiko1 | sonney2k, btw I should this week finish my MMD stuff, so I can spend some time on doing things | 22:02 |
heiko1 | but whats top priority in your opinion? | 22:02 |
@sonney2k | heiko1, getting shogun 2.1 out | 22:03 |
heiko1 | there is also this svm C thing | 22:03 |
heiko1 | yeh regarding that | 22:03 |
@sonney2k | I don't know if we should add new features... | 22:03 |
@sonney2k | maybe better fix warnings / bugs etc | 22:03 |
heiko1 | ok | 22:03 |
heiko1 | I will focus on that then | 22:04 |
heiko1 | also a bit documentation | 22:04 |
heiko1 | there are also still some bugs in the streaming features | 22:04 |
heiko1 | and unit tests would also be good | 22:04 |
@sonney2k | heiko1, if you find bugs you should put them to github issues | 22:05 |
heiko1 | so far I did all these "tests" in the examples | 22:05 |
heiko1 | sonney2k, I know, the ones I mentioned are in there in fact | 22:05 |
@sonney2k | then we can write a unit test for that and fix the bug | 22:05 |
heiko1 | yes | 22:05 |
heiko1 | I am really happy about the unit tests | 22:05 |
@sonney2k | heiko1, so what would be needed for SGNDArray & serialization? | 22:05 |
@sonney2k | I mean I think we need a new type or? | 22:06 |
@sonney2k | heiko1, btw eclipse is dog slow... but it is very easy to jump around in tons of code / do refactoring | 22:06 |
heiko1 | sonney2k, I dont know currently, would have to look in the code and think about it, its some time since I last read code for this | 22:06 |
heiko1 | sonney2k, eclipse is very good in my opinion, for exactly those things, searching classes, refactoring, all this blabla that makes life easier | 22:07 |
heiko1 | I also like the debugger | 22:07 |
heiko1 | I was doing lots of emacs before | 22:07 |
heiko1 | but eclipse is just more comfortable | 22:07 |
heiko1 | and less RMI also :) | 22:07 |
@sonney2k | heiko1, well I am using it at work so why not give it a try for C++ /' shogun | 22:09 |
heiko1 | sonney2k, I dont use it for compiling though | 22:09 |
@sonney2k | but it is too slow :/ | 22:09 |
heiko1 | yes | 22:09 |
heiko1 | I mean | 22:09 |
heiko1 | its not that bad | 22:09 |
@sonney2k | I would be interested in trying shogun & java | 22:10 |
heiko1 | slower than emacs - of course | 22:10 |
@sonney2k | it is here | 22:10 |
@sonney2k | but my machine is aging... | 22:10 |
heiko1 | what takes long? | 22:10 |
heiko1 | ah ok, yes | 22:10 |
@sonney2k | click ... wait ... click ... wait | 22:10 |
heiko1 | I remember on my old machine, it was a bit annoying | 22:10 |
heiko1 | but you can disable a few things | 22:10 |
heiko1 | then it gets faster | 22:10 |
@sonney2k | heiko1, I think doing ndarray is actually easy | 22:10 |
@sonney2k | one needs to write one vector with the dimension sizes | 22:11 |
@sonney2k | then all that is needed is to write out the whole data | 22:11 |
@sonney2k | in the same way like any vector/matrix | 22:11 |
heiko1 | how is the data stored in the ndarray? | 22:11 |
@sonney2k | just a ptr to data | 22:11 |
@sonney2k | linear | 22:11 |
@sonney2k | in memory | 22:11 |
@sonney2k | like for sgmatrix | 22:11 |
heiko1 | so all in a row | 22:12 |
heiko1 | I was wondering whether one would jiust have multiple arrays at different positions or like this | 22:12 |
@sonney2k | no why should one | 22:15 |
@sonney2k | that is what sgmatrix list is for | 22:15 |
@sonney2k | it can handle e.g. images of different size | 22:15 |
@sonney2k | heiko1, I just had a look. I think all that is needed is to refactor these new_cont_begin stuff | 22:22 |
@sonney2k | to get a SGVector as argument with the dims | 22:22 |
heiko1 | sonney2k, with this structure it should be quite easy to do the nda, | 22:22 |
@sonney2k | if dims==1/2 | 22:22 |
heiko1 | yes | 22:22 |
heiko1 | exactly | 22:22 |
@sonney2k | then it does the same like before | 22:22 |
@sonney2k | that's it | 22:22 |
heiko1 | the migration stuff is a bit more tricky | 22:22 |
@sonney2k | but there are none | 22:23 |
heiko1 | cannot remember the methods names but there are some which also do the case distinctions over the types | 22:23 |
heiko1 | let me check | 22:23 |
heiko1 | TParameter::allocate_data_from_scratch | 22:24 |
heiko1 | void TParameter::copy_data | 22:24 |
heiko1 | well they should not be too hard to extend | 22:24 |
heiko1 | since already working | 22:24 |
blackburn | a few memory errors left! | 22:27 |
@sonney2k | heiko1, looks easy indeed | 22:28 |
@sonney2k | the biggest issue is to do the refactoring of new_cont_begin but that's it | 22:28 |
@sonney2k | blackburn, context? | 22:28 |
heiko1 | sonney2k, maybe you find some inconsistencies on the way :) | 22:29 |
blackburn | sonney2k: I mean in make check-valgirnd | 22:29 |
@sonney2k | blackburn, in libshogun examples you mean? | 22:29 |
blackburn | yes | 22:29 |
@sonney2k | blackburn, go go :) | 22:30 |
blackburn | I fixed a great bug with tapkee | 22:30 |
@sonney2k | not sure if there is anything like a 'great' bug :) | 22:30 |
@sonney2k | horrible maybe? | 22:30 |
blackburn | well all linear methods weren't working | 22:30 |
blackburn | that's just great | 22:31 |
@sonney2k | great == cool | 22:31 |
blackburn | really? | 22:31 |
blackburn | :) | 22:31 |
@sonney2k | yeah | 22:31 |
@sonney2k | doesn't mean big | 22:32 |
blackburn | *sarcasm* | 22:32 |
blackburn | :D | 22:32 |
@sonney2k | yeah sarcasm would work for great bug :D | 22:32 |
blackburn | thanks for pointing this out! | 22:33 |
blackburn | okay one more fix | 22:33 |
@sonney2k | heiko1, blackburn btw if n4nd0 pops up - I like the http://iglesiashogun.alwaysdata.net/# draft and want it for the main site! | 22:33 |
@sonney2k | I have to repair my bike now | 22:34 |
blackburn | I don't mind and don't understand how to put it on | 22:34 |
heiko1 | I can only see the menu | 22:34 |
blackburn | sonney2k: that's *great* time to do that | 22:34 |
blackburn | february night :D | 22:34 |
@sonney2k | heiko1, yeah it is a mockup | 22:34 |
shogun-notifier- | shogun: Sergey Lisitsyn :master * 538ee43 / src/shogun/lib/tapkee/tapkee_methods.hpp: https://github.com/shogun-toolbox/shogun/commit/538ee431c9d1c5a08526e4a95515187a69c6a3c7 | 22:35 |
shogun-notifier- | shogun: Commented out unused projecting functions to avoid unhandled memory | 22:35 |
blackburn | oh another great bug | 22:40 |
shogun-notifier- | shogun: Sergey Lisitsyn :master * 6cbe6de / src/shogun/converter/LinearLocalTangentSpaceAlignment.cpp: https://github.com/shogun-toolbox/shogun/commit/6cbe6de109f1fc2beca0906b036bb6231b5fd579 | 22:49 |
shogun-notifier- | shogun: Fixed lost nullspace shift in LLTSA | 22:49 |
* sonney2k watches a youtube howto do bicycle gearing | 22:49 | |
blackburn | sonney2k: gearing like? | 22:49 |
@sonney2k | like switching gears | 22:50 |
@sonney2k | from bevel to bevel | 22:50 |
blackburn | sonney2k: which part do you switch? | 22:50 |
blackburn | uh! that appears to be hard to speak about these mechanisms in english hah | 22:51 |
@sonney2k | alright time to get hands dirty | 22:51 |
-!- flxb [~flx@e179034251.adsl.alicedsl.de] has joined #shogun | 22:52 | |
shogun-buildbot_ | build #216 of rpm1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/216 blamelist: Soeren Sonnenburg <sonne@debian.org> | 22:52 |
@sonney2k | wish me luck | 22:54 |
-!- in3xes [~in3xes@122.174.121.230] has quit [Ping timeout: 255 seconds] | 22:58 | |
shogun-notifier- | shogun: Heiko Strathmann :master * afc7499 / tests/unit/features/ (2 files): https://github.com/shogun-toolbox/shogun/commit/afc74996818241677d2bd1f8153135b94890392e | 23:15 |
shogun-notifier- | shogun: renamed test | 23:15 |
shogun-notifier- | shogun: Heiko Strathmann :master * 4a6b977 / tests/unit/statistics/ (5 files): https://github.com/shogun-toolbox/shogun/commit/4a6b9772de47ea58e12500cb5f8c642998164714 | 23:15 |
shogun-notifier- | shogun: unit tests for kernel selection | 23:15 |
shogun-notifier- | shogun: Heiko Strathmann :master * 6e4af9c / tests/unit/ (7 files): https://github.com/shogun-toolbox/shogun/commit/6e4af9c75e5654671127062684be9e6988ce565c | 23:15 |
shogun-notifier- | shogun: Merge branch 'mmd_kernel_selection' | 23:15 |
shogun-notifier- | shogun: Heiko Strathmann :master * 0ab4a45 / tests/unit/ (7 files): https://github.com/shogun-toolbox/shogun/commit/0ab4a45d7e60a2bc16af3b8286b2c6af42bec37f | 23:15 |
shogun-notifier- | shogun: Merge pull request #879 from karlnapf/master | 23:15 |
shogun-notifier- | shogun: | 23:15 |
shogun-notifier- | shogun: unit tests for kernel selection | 23:15 |
@sonney2k | blackburn, I didn't know that my bike's gear can be switched that smoothly! | 23:43 |
* sonney2k is happy | 23:44 | |
blackburn | sonney2k: so you mean gears like say 1/2/3 and 1/2/3/4/5/6/7? | 23:44 |
@sonney2k | yes | 23:44 |
blackburn | so what did you do with it? | 23:44 |
@sonney2k | just fine tune it | 23:44 |
blackburn | ah yeah that's a good thing | 23:44 |
@sonney2k | I could only use 3/7 gears | 23:45 |
@sonney2k | 3 of 7 | 23:45 |
@sonney2k | and these were stuttering | 23:45 |
@sonney2k | I mean switching between them was horrible | 23:45 |
blackburn | huh that's not normal | 23:45 |
@sonney2k | now it is totally smooth | 23:45 |
blackburn | when I ride bike I switch them pretty often I think | 23:45 |
blackburn | sonney2k: do you ride bike already? | 23:47 |
blackburn | as I can see it is +2 in berlin | 23:47 |
@sonney2k | makes a lot of sense to keep the pace/energy requirements consistent and just switching gears to adjust | 23:47 |
@sonney2k | blackburn, every day I do! | 23:48 |
blackburn | sonney2k: well in general I prefer to keep them at something normal (3/3 for me) | 23:49 |
blackburn | but when road gets smooth (it happens 30% of time in russia :D).. | 23:50 |
blackburn | sonney2k: isn't that a little cold to ride a bike now? :) | 23:52 |
--- Log closed Sun Feb 17 00:00:39 2013 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!