pickle27 @sonney2k @sonney2k --- Log opened Mon Jul 29 00:00:12 2013 --- Day changed Mon Jul 29 2013 sonney2k: right, I don't see anything about RealFeatures in here though 00:00 > library(modshogun) 00:00 Welcome! This is SHOGUN version 2.2.0 00:00 > x=GaussianKernel() 00:00 > x 00:00 An object of class "_p_shogun__CGaussianKernel" 00:00 Slot "ref": 00:00 Error in slot(object, what) : 00:00 no slot of name "ref" for this object of class "_p_shogun__CGaussianKernel" 00:00 just SGVector and SGMatrix 00:00 the question is what is this ref slot 00:00 sonney2k: I guess that's lispy name for method 00:01 they follow some scheme-like thing inside R 00:01 lisitsyn, for what? 00:01 sonney2k: for ref() 00:01 member function 00:02 whatever 00:02 what method do we need? 00:02 sonney2k: I guess it tries to call ref() 00:02 pickle27, ahh sure typemaps are just SGVector/SGMatrix stuff 00:03 pickle27, so whenever you have a function signature 00:03 foo(SGVector x) 00:03 you can write foo(c(1,2,3)) from R 00:03 similar for output 00:03 (in and out typemaps!) 00:03 lisitsyn, IDK I tried some minimal R example where I used ref/unref and they *were* called 00:04 no issues 00:04 sonney2k: no I was just guessing so everything can happen 00:04 sonney2k: okay - so then where do I need to be looking for figuring out the RealFeatures problem 00:05 lisitsyn, bah slotNames(x) 00:12 [1] "ref" 00:12 wtf? 00:12 pickle27, well lets start with some simple example first 00:12 x=RealLabels(c(1.0,2.2)) 00:13 when this is called 00:13 pickle27, go to line 36 in swig_typemaps.i 00:13 you will see a typecheck typemap 00:13 and a in typemap 00:14 for SGVector types 00:14 yeah Im there 00:14 the typecheck typemap well test if the type matches 00:14 if it does it returns the equiv of true (1) otherwise 0 00:14 so my first guess is that this always returns 0 00:15 maybe add some SG_SPRINT there to see the result 00:16 pickle27, if the typecheck typemap returns 1 00:16 the in typemap will be reached 00:16 and that one then just has to assign an appropriate SGVector to $1 00:17 currently it doesn't copy the memory and so the simple example (if it would pass the typecheck typemap) woudl just crash 00:17 since SGVector will at some point deallocate the memory when you exit R 00:18 and R likely too 00:18 -!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has left #shogun ["Fallen asleep!"] 00:19 sonney2k: so bascially print the result of line 49 00:21 pickle27, questions? 00:21 i mean 40 00:21 (TYPEOF($input) == r_type && Rf_ncols($input)==1 ) 00:21 pickle27, I am just trying this with all fields 00:21 -!- travis-ci [~travis-ci@ec2-54-243-23-104.compute-1.amazonaws.com] has joined #shogun 00:22 [travis-ci] it's Soeren Sonnenburg's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/9585748 00:22 -!- travis-ci [~travis-ci@ec2-54-243-23-104.compute-1.amazonaws.com] has left #shogun [] 00:22 pickle27, but then the big issue is to figure out why refcounting doesn't work 00:22 pickle27, I am not even getting towards the typecheck 00:23 so one now would have to debug the generated R code 00:24 pickle27, hmm 00:25 if I do 00:25 x=RegressionLabels() 00:25 > x$set_labels(c(1.0,2.3)) 00:25 > x$get_label(0) 00:26 [1] 1 00:26 > x$get_label(1) 00:26 [1] 2.3 00:26 it seems alright 00:26 maybe it just' doesn't work in the constructor style syntax 00:26 hmm 00:27 yeah seems like 00:28 with RealFeatures I get the (expected) crasher when I do 00:28 > x=RealFeatures() 00:28 > x$set_features(matrix(c(1,2,3,4,5,6),2,3)) 00:28 sonney2k: I get this when I run that 00:31 Error: evaluation nested too deeply: infinite recursion / options(expressions=)? 00:31 pickle27, there are certainly more issues but maybe it is a start 00:31 * sonney2k Zzzzz. 00:31 wierd this time I got a backtrace 00:31 ahg I hate when I accidently ctrl-l irc 00:32 -!- pickle27 [~Kevin@d67-193-243-174.home3.cgocable.net] has quit [Quit: Leaving] 00:35 ok guys, good night! 00:35 -!- pickle27 [~Kevin@d67-193-243-174.home3.cgocable.net] has joined #shogun 00:35 -!- iglesiasg [~Fernando@s83-179-44-135.cust.tele2.se] has quit [Quit: Leaving] 00:35 pickle27, memory corruption! 00:43 sonney2k: I ran it again and I couldn't get another backtrace 00:43 pickle27, you can fix it easily be copying the memory before gving it sgvector/amtrix 00:43 but it did happen once 00:43 build #1103 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1103 blamelist: Soeren Sonnenburg 00:48 sonney2k: fix it in the typemap func? 01:01 build #1420 of deb3 - modular_interfaces is complete: Failure [failed install test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1420 blamelist: Soeren Sonnenburg 01:11 okay thats enough for me today, see you all tomorrow 01:36 -!- pickle27 [~Kevin@d67-193-243-174.home3.cgocable.net] has quit [Quit: Leaving] 01:36 -!- nube [~rho@49.244.47.72] has joined #shogun 02:37 -!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] 02:55 -!- nube [~rho@49.244.47.72] has quit [Quit: Leaving.] 03:29 build #472 of nightly_default is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/472 04:12 -!- nube [~rho@116.90.239.3] has joined #shogun 05:25 -!- nube [~rho@116.90.239.3] has left #shogun [] 05:28 -!- lisitsyn [~lisitsyn@92-240-133-94.clients.tlt.100megabit.ru] has quit [Quit: Leaving.] 08:06 good morning 08:37 -!- van51 [~van51@ppp-94-66-56-238.home.otenet.gr] has joined #shogun 09:28 hello 09:30 -!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has joined #shogun 10:18 -!- iglesiasg [~Fernando@s83-179-44-135.cust.tele2.se] has joined #shogun 10:19 -!- mode/#shogun [+o iglesiasg] by ChanServ 10:19 good morning to all 10:20 iglesiasg: good morning :) 10:22 -!- van51 [~van51@ppp-94-66-56-238.home.otenet.gr] has quit [Quit: Leaving.] 10:41 -!- van51 [~van51@94.66.56.238] has joined #shogun 10:49 -!- vgorbati [c3ee5cb1@gateway/web/freenode/ip.195.238.92.177] has joined #shogun 10:59 -!- lisitsyn [~lisitsin@mxs.kg.ru] has joined #shogun 11:27 vgorbati: hey haven't seen you for a while? 11:28 how is it going? are you done with your thesis? 11:28 -!- gsomix_ [~gsomix@95.67.137.114] has joined #shogun 11:50 -!- gsomix [~gsomix@109.169.249.121] has quit [Ping timeout: 264 seconds] 11:53 lisitsyn: hello there) I basically rested all the time:) How is Berlin? 12:01 lisitsyn: do you mean master's thesis? 12:02 vgorbati: berlin is nice indeed :) 12:03 vgorbati: well some thesis you were working on 12:03 lisitsyn: I was working on monograph for my university supervisor (is it a correct word for it?), and yes, I believe it is done now:) 12:05 -!- foulwall [~user@116.114.141.87] has joined #shogun 13:15 -!- van51 [~van51@94.66.56.238] has quit [Quit: Leaving.] 13:37 -!- foulwall [~user@116.114.141.87] has quit [Ping timeout: 240 seconds] 13:48 -!- gsomix_ [~gsomix@95.67.137.114] has quit [Quit: Leaving] 14:24 -!- foulwall [~user@116.114.141.87] has joined #shogun 14:58 -!- iglesiasg [~Fernando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 256 seconds] 15:58 mmmm 16:33 -!- iglesiasg [~Fernando@s83-179-44-135.cust.tele2.se] has joined #shogun 16:43 -!- mode/#shogun [+o iglesiasg] by ChanServ 16:43 -!- van51 [~van51@ppp-94-66-56-238.home.otenet.gr] has joined #shogun 16:47 -!- foulwall [~user@116.114.141.87] has quit [Ping timeout: 240 seconds] 17:45 -!- foulwall [~user@116.114.133.251] has joined #shogun 17:46 -!- pickle27 [~Kevin@d67-193-243-174.home3.cgocable.net] has joined #shogun 17:54 -!- shogun-notifier- [~irker@7nn.de] has joined #shogun 18:04 shogun: Roman Votyakov :develop * c4b0bb6 / tests/unit/machine/gp/ (4 files): https://github.com/shogun-toolbox/shogun/commit/c4b0bb66f22d8895de7bf3a71554ad573c762ee2 18:04 shogun: add some unit-tests for likelihood models 18:04 shogun: Roman Votyakov :develop * f8f1b75 / src/shogun/classifier/GaussianProcessBinaryClassification.cpp: https://github.com/shogun-toolbox/shogun/commit/f8f1b751b8e046d9bd9afe3b787abb8bb78763d0 18:04 shogun: replace eigen3 exp() call with shogun exp() call 18:04 shogun: Heiko Strathmann :develop * b2727fc / / (5 files): https://github.com/shogun-toolbox/shogun/commit/b2727fc0b20e54d9b6ab6fb9ca556c759fa5b59c 18:04 shogun: Merge pull request #1325 from votjakovr/feature/gp_refactoring 18:04 shogun: 18:04 shogun: Add some unit-tests for likelihood models 18:04 -!- pickle27 [~Kevin@d67-193-243-174.home3.cgocable.net] has quit [Quit: Leaving] 18:07 build #1104 of cyg1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1104 blamelist: Heiko Strathmann , Roman Votyakov 18:09 foulwall, hey there - any news on the README? 18:25 Hey sonney2k, already finished, and now working on migration for tapkee demos. 18:27 -!- iglesiasg [~Fernando@s83-179-44-135.cust.tele2.se] has quit [Quit: Leaving] 18:27 foulwall, please send a PR then 18:27 ok sonney2k 18:28 heh I thought that the  after foulwall was dust on my screen :p 18:29 van51, heh! 18:30 van51, so how is it going now? 18:30 sonney2k: well first of all, the CHashedDocDotFeatures and the converter have identical results 18:31 van51, btw regarding benoits email - we already use bitmasks right? 18:31 sonney2k: in the doc features 18:31 van51, excellent! 18:31 van51, and speed? 18:31 sonney2k: in dense/sparse I had switched to just mod 18:31 sonney2k: for speed I'm not sure, it depends on the normalization 18:32 sonney2k: I mean unnormalized dot features take a lot time 18:32 van51, well of course with same normalization 18:32 sonney2k: but the sparsefeatures returned from the converter are not normalized 18:32 van51, oh you should add that to the converter too! 18:33 sonney2k: I tried but it would be the same normalziation. In the dot features we normalize the result, while there we create the entire vector so we would normalize each feature 18:33 wouldn't 18:34 dot product result* 18:34 van51, parse error 18:34 sonney2k: heh 18:34 sonney2k: in the dot features we compute the dense dot for instance 18:35 sonney2k: and we normalize it 18:35 van51, you jsut normalize each feature by dividing with 1/sqrt(count) 18:35 van51, sure but in dotfeatures add_to_dense_vec you do the same normalization per feature 18:36 sonney2k: I tried it with 1/sqrt(doc.size()) and it was giving poor results 18:37 sonney2k: maybe I missed something 18:37 sonney2k: I can try it again though 18:37 sonney2k: btw I wanted to ask you about the CCommUlongKernel with svmLight that you had told me 18:37 sonney2k: I tried to compare their speed first, but when I used the kernel it segfaulted in a radix_sort method 18:39 sonney2k: I'm guessing it ran out of memory 18:39 van51, maybe forgot that you need to use a double? 18:40 van51, it should give the exact same results 18:40 van51, hmmhh radix sort should work just fine (at least it did) 18:41 sonney2k: I'll reproduce it now 18:43 -!- pickle27 [~Kevin@d67-193-243-174.home3.cgocable.net] has joined #shogun 18:45 build #1421 of deb3 - modular_interfaces is complete: Failure [failed install test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1421 blamelist: Heiko Strathmann , Roman Votyakov 18:46 -!- travis-ci [~travis-ci@ec2-54-226-172-147.compute-1.amazonaws.com] has joined #shogun 18:53 [travis-ci] it's Heiko Strathmann'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/9610363 18:53 -!- travis-ci [~travis-ci@ec2-54-226-172-147.compute-1.amazonaws.com] has left #shogun [] 18:53 -!- foulwall [~user@116.114.133.251] has joined #shogun 18:53 pickle27, yes 18:55 pickle27, let me just add it 18:55 -!- foulwall [~user@116.114.133.251] has quit [Remote host closed the connection] 18:56 -!- foulwall [~user@116.114.133.251] has quit [Remote host closed the connection] 18:56 -!- foulwall [~user@116.114.133.251] has joined #shogun 18:56 sonney2k: this is my valgrind output 18:56 http://pastebin.com/wcT9rLpj 18:56 build #1105 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1105 blamelist: Roman Votyakov 18:58 van51, hmmhh looks like we broke it :/ 19:01 sonney2k: let me show you my code in case there is something wrong there 19:01 sonney2k: https://gist.github.com/van51/6105834 19:07 -!- lisitsyn [~lisitsin@mxs.kg.ru] has quit [Quit: Leaving.] 19:13 build #1422 of deb3 - modular_interfaces is complete: Failure [failed install test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1422 blamelist: Roman Votyakov 19:21 sonney2k: hi! i've found, that since doxygen 1.7.2 we can use MathJax instead of using prerendered bitmaps for rendering formulas. I think, that with MathJax formulas will be prettier. What do you think about using it? 19:40 votjakovr, sure! 19:56 votjakovr, that would speed up things too 19:56 votjakovr, only issue is that it won't detect errors in the file - I mean if any of us screws up a formula 19:58 sonney2k: hmm, yep, unfortunately 20:00 shogun: Soeren Sonnenburg :develop * eccafc8 / src/ (4 files): https://github.com/shogun-toolbox/shogun/commit/eccafc84132f31fe5bb2e03372200482710f87a9 20:11 shogun: add get_copy function to lib/memory to be used from typemaps for clones 20:11 pickle27, that fixes the crashers etc for me 20:12 pickle27, I could do 20:12 x=Regressionlabels() 20:12 x$set_labels(c(1,2,3)) 20:12 x$set_labels(c(1,2)) 20:12 ... 20:12 and also 20:12 x=RealFeatures() 20:13 x$set_feature_matrix(matrix(c(1,2,3,4,5,6),2,3)) 20:13 votjakovr, btw looks like doxygen also supports markdown syntax nowadays 20:18 sonney2k: oh, yeah, i see 20:20 sonney2k: okay i will merge and see if it works for me soon 20:21 -!- iglesiasg [~iglesias@2001:6b0:1:1da0:9ce1:9054:335:9d36] has joined #shogun 20:21 -!- mode/#shogun [+o iglesiasg] by ChanServ 20:21 hello! 20:22 iglesiasg, helllo! 20:22 -!- lisitsyn [~lisitsyn@213.87.133.71] has joined #shogun 20:23 kevin is a good boy today 20:24 pickle27, ;) 20:24 lisitsyn,  I will become a good boy soo, I promise :P 20:24 van51, votjakovr 20:24 haha 20:25 haha 20:25 foulwall, hushell 20:25 damn this phone typing is not so fast 20:25 van51, is the normalization working in the converter now? 20:26 iglesiasg, how is your marginal deals? 20:27 Hi lisitsyn 20:27 largely marginal 20:27 lisitsyn, LMNN? 20:27 foulwall, hey I just called to say 20:28 weekly report ;) 20:28 ok lisitsyn :) 20:28 iglesiasg: ja this one 20:28 iglesiasg, is it better performance wise now? 20:29 lisitsyn, so so. My implementation currently sucks a bit compared to the author's (and it is in Matlab, although with some parts in C) 20:29 lisitsyn, no, not yet 20:29 lisitsyn, I realized I have to change a sort of big thing now 20:29 lisitsyn, well, not that big actually 20:29 lisitsyn, you are with phone now, right? 20:29 iglesiasg: do you know the bottleneck? 20:29 otherwise I can show you 20:30 -!- foulwall [~user@116.114.133.251] has joined #shogun 20:30 iglesiasg: ueah, i can read things though 20:30 but not compile shogun 20:30 lisitsyn, haha indeed 20:30 let me show you then, on moment 20:30 -!- foulwall [~user@116.114.133.251] has quit [Client Quit] 20:30 one* 20:30 lisitsyn, https://github.com/iglesias/shogun/blob/feature/lmnn/src/shogun/metric/LMNNImpl.cpp#L135 20:31 lisitsyn, so I have that method right now that is computing the difference between every pair of feature vectors, and then computing the outer product of the diff vector with itself 20:32 foulwall, and where is the README PR ? 20:32 lisitsyn, right now that computes N^2 matrices of D^2 elements each one 20:33 lisitsyn, it is possible to make that smaller 20:33 sonney2k: sending 20:34 lisitsyn, since (x_i-x_j)*(x_i-x_j)' is the same as (x_j-x_i)*(x_j-x_i)' 20:34 lisitsyn, but I think that even if I take that into account, it is still something too big to keep in memory (it will still be the order of N^2 matrices) 20:35 iglesiasg: yeah quite big thing 20:36 lisitsyn, so I think that it is better to compute the outer products on demand; every time one is needed I compute it, even if that migh entail computing the same outer product more than once 20:36 lisitsyn, what do you think of doing it on demand? 20:37 lisitsyn, I think the author's implementation does it in this way 20:37 iglesiasg, would be slower I guess but takes less memory 20:37 if you dont have any N2 matrices more it totally makes sense 20:38 lisitsyn, then the bottleneck is this method https://github.com/iglesias/shogun/blob/feature/lmnn/src/shogun/metric/LMNNImpl.cpp#L316 20:38 are there any more N2 matrices? 20:39 iglesiasg, sounds like kernel caching 20:39 sonney2k, yeah, I thought that caching could help 20:39 so I think that for the moment I will modify the code to compute outer products on demand, and later we can see if it pays off to have caching 20:40 lisitsyn, what do you mean if there are more N2 matrices? 20:40 iglesiasg: I think caching makes sense here 20:40 iglesiasg: if thats the only NxN matrix here 20:41 lisitsyn, aah ok got it. Yes it is 20:41 lisitsyn, but note that it is not only a NxN matrix 20:41 then kill it yeah 20:41 lisitsyn, it is NxN matrices of DxD elements each (N=number of vectors, D=feature dimension) 20:42 oh 20:42 heh 20:42 because you compute the outer product of the difference of feature vectors 20:43 for each pair of feature vectors, you make its difference, and compute the outer product of the difference with itself 20:43 yeah thats not something to be in memory 20:43 -!- travis-ci [~travis-ci@ec2-72-44-42-216.compute-1.amazonaws.com] has joined #shogun 20:44 [travis-ci] it's Soeren Sonnenburg'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/9615083 20:44 -!- travis-ci [~travis-ci@ec2-72-44-42-216.compute-1.amazonaws.com] has left #shogun [] 20:44 by outer product you mean dger from blas 20:44 eight? 20:44 right? 20:44 lisitsyn, no idea :D 20:44 :D 20:44 lisitsyn, there are two ways you can multiply a vector with itself 20:45 yeah heard sth 20:45 lisitsyn, inner product -> you get an scalar (something like x'*x) 20:45 this one produces matrix 20:45 outer product -> you get a matrix (x*x') 20:45 exactly 20:45 dger :p 20:46 lisitsyn, I am checking dger 20:46 ddot and dger 20:46 lisitsyn, yeah it seems to be like that 20:46 lisitsyn, although they same in the doc that it is x*y' where x has m elements and n elements 20:46 iglesiasg, eigen has symmetric rank update 20:46 i used that thing in tapkees pca 20:47 lisitsyn, symmetric rank update? Do you think that could help me to make it faster? 20:47 lisitsyn, right now I get two eigen vectors, make their difference, and then the outer product 20:47 lisitsyn, https://github.com/iglesias/shogun/blob/feature/lmnn/src/shogun/metric/LMNNImpl.cpp#L154 20:48 that line and the next one 20:48 no probably wont help here 20:48 lisitsyn, aham. Why do you think so? 20:49 just trying to understand it better :) 20:49 iirc that rank update is better for multiple calls 20:49 lisitsyn, I guess it is a matter of testing it anyway, so I will keep in mind that there is another way of getting the outer product 20:50 iglesiasg: it employs symmetricity better somehow 20:51 -!- lisitsyn [~lisitsyn@213.87.133.71] has quit [Remote host closed the connection] 20:52 -!- lisitsyn [~lisitsyn@213.87.133.71] has joined #shogun 20:52 oh I get disconnected once turning my phone off 20:53 -!- lisitsyn [~lisitsyn@213.87.133.71] has left #shogun [] 20:57 build #1106 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1106  blamelist: Soeren Sonnenburg 21:03 ls 21:05 Documents Desktop Dropbox 21:06 foulwall, :P 21:06 haha iglesiasg 21:06 van51, minimalchange: https://gist.github.com/sonney2k/6106780 21:06 hahaha I do that too 21:06 except I have this thing where I like to clear my terminals using ctrl-l all the time 21:06 sonney2k: hey sorry I was helping my grandmother with something 21:06 and then I clear IRC and its annoying 21:06 hehe 21:07 sonney2k: what did you change? 21:07 the 8-1? 21:07 van51, yeah 21:07 like in the python example 21:07 sonney2k: what is the meaning of start then? 21:07 van51, where to start 21:08 sonney2k: :p 21:08 van51, well this is doing some bitheavy stuff 21:08 and basically it goes back 8 chars from the start so starting at index 7 is what you want to get 7...0 21:10 sonney2k: ah I see 21:11 sonney2k: well I'm going to get some food and work on some stuff later and overnight 21:11 sonney2k: btw I've also started the document example 21:11 sonney2k: and I was thinking of making a simple web ui, just to be able to write some lines and test it 21:12 van51, use the python_modular interface instead and some simple input or so and foulwall can later nicely integrate this 21:13 sonney2k: ok 21:14 sonney2k: so I'm off for about an hour 21:15 cu later 21:15 van51, btw it might be good to do that for your example too. prototyping with python really is much faster (and debugging too!) 21:17 sonney2k: I don't know python so I'm trying to avoid it 21:17 sonney2k: but I guess it's a good time to get to know as well :0 21:18 :)* 21:18 -!- van51 [~van51@ppp-94-66-56-238.home.otenet.gr] has quit [Quit: Leaving.] 21:21 sonney2k: I tried the new R modular and I ran those samples you posted and they work for me 21:23 sonney2k: it still seams like none of the methods are "binding" though 21:23 whenever I try and use a method I get a not found error 21:23 pickle27, that must be some swig issue - we have ton's of constructors 21:23 yeah 21:23 so figuring out which one to use is the task of swig 21:23 shogun: Soeren Sonnenburg :develop * 95def24 / src/shogun/machine/gp/InferenceMethod.h: https://github.com/shogun-toolbox/shogun/commit/95def246a179d7a0065c85cd1252e2f6cae136ef 21:26 shogun: fix typo in doxygen formula 21:26 shogun: Soeren Sonnenburg :develop * 3f113b7 / / (16 files): https://github.com/shogun-toolbox/shogun/commit/3f113b7a74d2a0ad54a9a484e93fca6ba9afde4f 21:26 shogun: add some get_strdup function to avoid compile errors (strdup is not ansi C!) 21:26 pickle27, back then I totally understood how that R stuff worked but I forgot 21:26 pickle27, the only thing I remember is that from the R code there are some tests run selecting the appropriate constructor 21:27 sonney2k: and I have no idea how it works 21:27 it sounds like its pretty tough to fix - maybe I'll just push an r_static example for now 21:28 -!- hushell [~hushell@c-24-21-169-136.hsd1.or.comcast.net] has quit [Ping timeout: 276 seconds] 21:29 -!- foulwall [~user@116.114.133.251] has joined #shogun 21:31 -!- foulwall [~user@116.114.133.251] has quit [Ping timeout: 246 seconds] 21:32 -!- sonney2k [~shogun@7nn.de] has quit [Ping timeout: 246 seconds] 21:33 -!- sonney2k_ [~shogun@7nn.de] has joined #shogun 21:33 -!- sonney2k_ is now known as sonney2k 21:34 -!- mode/#shogun [+o sonney2k] by ChanServ 21:35 pickle27, well you have to read swig code and the generated wrapper code modshogun.R etc 21:35 not fun 21:35 sonney2k: doesn't sound like it! 21:35 iglesiasg, any progress on the failing tests on the buildbot? 21:36 sonney2k, unfortunately not, sorry. I have focused on LMNN again since yesterday's eve 21:36 foulwall, ohh please don't put images in the repository 21:38 foulwall, please put them again under shogun-data/tapkee 21:39 ok sonney2k , I'll resend 21:39 foulwall, it might make sense to not have >1000 files too btw 21:40 foulwall, maybe better use a single .zip file with the images in there 21:40 like faces.zip 21:40 and mnist.zip 21:40 I see 21:41 foulwall, btw do you have a live demo already for that? 21:41 let me upload it on nn.7nn.de 21:41 sonney2k: check this, http://nn.7nn.de:8000/tapkee/entrance . 21:46 foulwall, nice! 21:47 build #1107 of cyg1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1107  blamelist: Soeren Sonnenburg 21:55 build #1108 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1108  blamelist: Soeren Sonnenburg 22:01 foulwall`, looking cool! 22:01 lisitsyn's gf will become worldwide famous eventually thanks to the faces embedding :) 22:02 iglesiasg, yeah the new lena :) 22:02 sonney2k, haha yeah 22:02 sonney2k, have you actually seen the original of the Lena's picture? 22:03 haha :) iglesiasg, I've seen. 22:03 it is, let's say, not precisely the one they show in class or in college textbooks :P 22:03 sonney2k: thanks:D 22:04 Lena is Swedish btw :) 22:04 poeple are pretty proud in here when they use the picture in class 22:04 -!- travis-ci [~travis-ci@ec2-72-44-42-216.compute-1.amazonaws.com] has joined #shogun 22:11 [travis-ci] it's Soeren Sonnenburg'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/9618211 22:11 -!- travis-ci [~travis-ci@ec2-72-44-42-216.compute-1.amazonaws.com] has left #shogun [] 22:11 -!- gsomix [~gsomix@95.67.137.114] has joined #shogun 22:11 good evening 22:11 report time? :) 22:11 gsomix, where is your PR? 22:12 sonney2k, with quoting and libsvm reader? not ready yet, I need little time for it. 22:14 gsomix, no the one that you submitted ~21hrs ago 22:15 where I wanted some tiny changes 22:15 shogun: Soeren Sonnenburg :develop * a6f59dd / src/configure,src/shogun/io/Parser.cpp: https://github.com/shogun-toolbox/shogun/commit/a6f59dd8e159a0945b75eb2aca45c1e6a11e24b3 22:16 shogun: use strtod if strtold is not available 22:16 sonney2k, ah, just a moment! 22:16 sonney2k, forgot it. =___= today was little holiday, sorry. 22:17 sonney2k, btw I haven't any warnings related to format on my machine. 22:17 gsomix, I need to merge yours before I can start to convert examples 22:17 gsomix, yeah sure 32bit vs 64bit 22:17 gsomix, so use the 'standard' ones and we will be fine 22:18 sonney2k, aha, indeed. 22:18 sonney2k, btw, can I help with examples& 22:18 *? 22:18 iglesiasg, wiking, votjakovr, pickle27 anyone here up to doing a doc sprint at google in october? 22:19 sonney2k, sure! 22:19 gsomix, well sure but we are quite a bit behind schedule :/ 22:19 probably no for me because I'm hoping to be doing some travelling in Oct 22:20 it is during oct14-18 http://www.booksprints.net/2013/07/2013-doc-camp-announced/ 22:20 sonney2k, what should we do to raise our odds of getting in? 22:21 iglesiasg, IDK 22:21 we certainly need 4 people that can do a stunt 22:22 sonney2k: looks very cool though, I'd be happy to help I just won 22:22 sonney2k, would you be in? 22:22 't be able to be there 22:22 iglesiasg, HeikoS, me + ?? 22:22 let's see if wiking and lisitsyn could make it 22:22 I am not sure I would go ipython-notebook for that 22:23 -!- hushell [~hushell@8-92.ptpg.oregonstate.edu] has joined #shogun 22:23 sonney2k, mmm I don't see why not 22:23 iglesiasg, they aim for a real book 22:24 sonney2k, people are doing real books these days based on ipython I think 22:24 copy & pasting parts from people's thesis would not be that easy then 22:25 sonney2k, lol 22:25 iglesiasg, I think that is our only chance to *finish* a book in 4 days 22:25 that is what they aim fo 22:25 are 22:25 iglesiasg, look at this http://www.booksprints.net/about/ 22:26 sonney2k, check thhis one: http://camdavidsonpilon.github.io/Probabilistic-Programming-and-Bayesian-Methods-for-Hackers/ 22:26 sonney2k, http://nbviewer.ipython.org/urls/raw.github.com/CamDavidsonPilon/Probabilistic-Programming-and-Bayesian-Methods-for-Hackers/master/Chapter1_Introduction/Chapter1_Introduction.ipynb 22:26 chapter 1 ^ 22:27 to me that looks like a nice real book 22:27 sonney2k, what do you think? 22:28 iglesiasg, looks *very* cool 22:29 indeed, we agree :) 22:29 iglesiasg, can this be converted to some epub /pdf etc? 22:30 or is it 'just' web? 22:30 sonney2k, no idea, honestly 22:30 -!- van51 [~van51@ppp-94-66-56-238.home.otenet.gr] has joined #shogun 22:30 let me have a quick look 22:30 van51, hey again 22:31 sonney2k: hi 22:31 they say so in the page, no idea how it looks like, let's see 22:31 iglesiasg, ok then it is pretty clear how to increase our chances - we should have some ideas what our target audience is, structure, running examples and methods we want to use 22:32 iglesiasg, but issue still remains that we are just a very small org (compared to e.g. KDE) 22:32 sonney2k, when is the deadline? 22:32 iglesiasg, in 3 days 22:32 sonney2k, I am thinking, if it is after GSoC ends 22:32 shit... 22:32 or 5? 22:32 iglesiasg, last year it was like this 22:32 I drop the idea then 22:33 before august 7! 22:33 if heiko is not around for discussion yeah then indeed 22:33 sonney2k, https://dl.dropboxusercontent.com/u/11020840/June82013.pdf 22:34 sonney2k, pdf version of the book 22:34 sonney2k, IMHO it looks nicer in the browser, but it is still cool I think 22:34 iglesiasg, I think we have a chance this year 22:35 iglesiasg, jsut because big data etc is such a big hype 22:35 sonney2k, compiling... 22:35 so we should try! 22:35 sonney2k, yep! 22:35 iglesiasg, very nice still!!! 22:36 so I vote sticking to ipython-notebook :) 22:36 build #1424 of deb3 - modular_interfaces is complete: Failure [failed compile octave_modular]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1424  blamelist: Soeren Sonnenburg 22:42 iglesiasg, sure then we just have to figure out how to integrate big chunks of latex 22:44 -!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 268 seconds] 22:45 sonney2k, yes 22:45 iglesiasg, no idea whether stuff like toc / index / references are possible 22:47 sonney2k, testing... 22:49 -!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has left #shogun ["Fallen asleep!"] 22:49 shogun-buildbot, force build 'cyg1 - libshogun' --branch=develop 22:51 build forced [ETA 17m23s] 22:51 I'll give a shout when the build finishes 22:51 shogun: Soeren Sonnenburg :develop * e6788a0 / src/interfaces/ (2 files): https://github.com/shogun-toolbox/shogun/commit/e6788a07fdb6845a38f92bc55489e80f1c33f7c1 22:53 shogun: drop get_copy from octave and perl 22:53 -!- travis-ci [~travis-ci@ec2-72-44-42-216.compute-1.amazonaws.com] has joined #shogun 22:54 [travis-ci] it's Soeren Sonnenburg'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/9620302 22:54 -!- travis-ci [~travis-ci@ec2-72-44-42-216.compute-1.amazonaws.com] has left #shogun [] 22:54 sonney2k, done. https://github.com/shogun-toolbox/shogun/pull/1323 22:59 compiling takes many time 22:59 shogun: Evgeniy Andreev :develop * 5798840 / / (5 files): https://github.com/shogun-toolbox/shogun/commit/5798840ae3254defe789403e3eac001cc665be31 23:01 shogun: fixed invalid read in Parser, CSVFile constructor is shortened 23:01 shogun: Soeren Sonnenburg :develop * f0ccda8 / / (5 files): https://github.com/shogun-toolbox/shogun/commit/f0ccda88694ea364c7e1b021c7dc2256648b001b 23:01 shogun: Merge pull request #1323 from gsomix/feature/parser 23:01 shogun: 23:01 shogun: Fixed invalid read in Parser, CSVFile constructor is shortened 23:01 sonney2k, so what to do with examples? 23:02 -!- pickle27 [~Kevin@d67-193-243-174.home3.cgocable.net] has quit [Quit: Leaving] 23:02 gsomix, maybe we should talk how to continue in general? 23:02 gsomix, it would be nice to have the other ascii formats working 23:03 sonney2k, yep. 23:03 gsomix, and the stringfeatures/sparsefeatures/densefeatures/matrixfeatures etc cleaned up / supported 23:03 sonney2k, so I'm working on libsvm file format now. 23:03 I mean this is the first and most important thing to finish 23:04 gsomix, sounds like I should rather convert the examples 23:04 while you continue your work 23:04 gsomix, afterwards I would like us to have some binary format based on protobuf and well there still is reading / writing matlab files 23:05 which likely is most important for users 23:05 build #1109 of cyg1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1109 23:06 sonney2k, aha 23:06 gsomix, but seriously I think most people will just use ascii and matlab format 23:07 sonney2k, other ascii formats. which? 23:07 matlab format please! 23:07 gsomix, when are we getting this? Really looking forward to it 23:08 sonney2k, matlab7 file format is last version, right? 23:09 gsomix, arff then van51 would say vw 23:09 gsomix, no 23:10 gsomix, matlab has a couple of formats 23:10 sonney2k, not compatible? 23:10 gsomix, as I wrote on the ideas page octave supports reading IIRC up to 7.2 23:10 afterwards they switched to sth hdf5 based 23:10 so you it would take some time 23:11 to do this 23:11 sonney2k, I should look into octave source? or there is specs on 7.2? 23:11 I found only 7.0.1 23:11 gsomix, http://www.mathworks.com/help/pdf_doc/matlab/matfile_format.pdf 23:12 sonney2k, huh! I never imagined that I could find it on mathworks site. 23:14 I mean MATLAB very closed and have not many docs about how it works inside. 23:15 *is very 23:15 gsomix, it seems not to be the latest format for sure 23:15 build #1110 of cyg1 - libshogun is complete: Failure [failed compile]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1110  blamelist: Soeren Sonnenburg 23:15 gsomix, maybe you can use octave's reader somehow 23:16 gsomix, only issue is we cannot test this 23:16 we have no matlab 23:16 so maybe we just get some .mat file from sb. 23:16 and also the ascii representation or sth 23:16 and then you can read/compare 23:16 sonney2k, ok, I can test it on my local machine. 23:17 and I can make .mat with ascii representation too 23:17 I have licensed matlab 2013a at lab. 23:18 sonney2k, I'm not full clear with protobuf now. 23:19 sonney2k, we plan to develop our file format based on protobuf? or so? 23:20 shogun: Soeren Sonnenburg :develop * 5733a65 / src/shogun/mathematics/Math.h: https://github.com/shogun-toolbox/shogun/commit/5733a65137c48909c95272fbbc6df57af5a1bb67 23:21 shogun: define _USE_MATH_DEFINES to get non-standard constant M_PI 23:21 gsomix, yes it is basically just a binary format / fast & size efficient to store values 23:23 have you guys started doing the mid-term evaluation? 23:23 * sonney2k gets hungry 23:24 * sonney2k off -> food 23:24 iglesiasg, nope. I have many work that I promised to finish before midterm. 23:24 gsomix, you will make on time! :) 23:25 build #1423 of deb3 - modular_interfaces is complete: Failure [failed compile octave_modular]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1423  blamelist: Soeren Sonnenburg 23:25 shogun-buildbot, force build 'cyg1 - libshogun' --branch=develop 23:28 build forced [ETA 16m24s] 23:28 I'll give a shout when the build finishes 23:28 build #1111 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1111 23:28 shogun-buildbot, force build 'cyg1 - libshogun' --branch=develop 23:29 build forced [ETA 16m24s] 23:29 I'll give a shout when the build finishes 23:29 build #1112 of cyg1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1112 23:30 shogun-buildbot, force build 'cyg1 - libshogun' --branch=develop 23:31 build forced [ETA 16m24s] 23:31 I'll give a shout when the build finishes 23:31 -!- travis-ci [~travis-ci@ec2-54-226-172-147.compute-1.amazonaws.com] has joined #shogun 23:36 [travis-ci] it's Soeren Sonnenburg'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/9621665 23:36 -!- travis-ci [~travis-ci@ec2-54-226-172-147.compute-1.amazonaws.com] has left #shogun [] 23:36 iglesiasg, I guess failing unit tests will be the only thing to show why matrixlist is unhappy 23:38 sonney2k, yeah, maybe. But valgrinds seems to complain about stuff going on when unref-ing 23:39 iglesiasg, yeah exactly 23:40 build #1425 of deb3 - modular_interfaces is complete: Failure [failed test python_modular]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1425  blamelist: Soeren Sonnenburg 23:41 sonney2k, that confuses me though 23:41 iglesiasg, why? 23:41 sonney2k, where can the bug be if the error arises in the destructor or when un-referencing? 23:41 some kind of double free 23:41 build #1113 of cyg1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/1113 23:47 evalution submitted 23:49 --- Log closed Tue Jul 30 00:00:53 2013