--- Log opened Fri Aug 19 00:00:00 2011 | ||
@sonney2k | yeah epic failure powered by w. gorbachew | 00:00 |
---|---|---|
blackburn | I hope we are only two subscribers of trac | 00:01 |
blackburn | I've generated a lot of mails | 00:01 |
blackburn | hey he is m. :) | 00:01 |
@sonney2k | but w - wodka | 00:02 |
@sonney2k | or just turn it upside down | 00:02 |
@sonney2k | http://shogun-toolbox.org:5347/builders/cmdline_static/builds/68/steps/test_1/logs/stdio | 00:02 |
@sonney2k | look at this everything fails but test still claims everything is good | 00:02 |
blackburn | wodka gorbachew nice | 00:02 |
shogun-buildbot | build #59 of octave_static is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org:5347/builders/octave_static/builds/59 blamelist: sonne@debian.org | 00:03 |
blackburn | Test successful: failure | 00:03 |
blackburn | :D | 00:03 |
@sonney2k | even better it works locally :D | 00:04 |
blackburn | sonney2k: wtf is: | 00:05 |
blackburn | https://github.com/shogun-toolbox/shogun/blob/master/examples/kernel_distance.py | 00:05 |
blackburn | why it is in examples | 00:06 |
@sonney2k | blackburn, what is the problem? | 00:07 |
@sonney2k | blackburn, when running tests - should it run all tests and then fail at the end if one of them failed? | 00:07 |
@sonney2k | or should it fail immediately? | 00:07 |
-!- blackburn [~blackburn@188.122.238.50] has quit [Read error: No route to host] | 00:09 | |
-!- blackburn [~blackburn@188.122.238.50] has joined #shogun | 00:10 | |
blackburn | make leaded to reboot heh | 00:10 |
blackburn | sonney2k: what is it, why it is root of examples? | 00:11 |
@sonney2k | blackburn, when running tests - should it run all tests and then fail at the end if one of them failed? | 00:11 |
@sonney2k | or should it fail immediately? | 00:11 |
@sonney2k | I tend to say at the end | 00:12 |
blackburn | at the end | 00:12 |
@sonney2k | doing | 00:12 |
blackburn | if we broke something in the middle | 00:12 |
blackburn | we will not see if something is broken too | 00:12 |
blackburn | -> harder to debuuug | 00:12 |
CIA-87 | shogun: Soeren Sonnenburg master * r5a0fd39 / (9 files in 9 dirs): exit 1 on error at the end of running check.sh - http://bit.ly/p5mauT | 00:14 |
@sonney2k | there we go | 00:15 |
@sonney2k | ahh root of examples | 00:16 |
@sonney2k | blackburn, bug :) | 00:16 |
@sonney2k | should be in python_static | 00:16 |
blackburn | yeah that's why I was tending to launch nuclear missile | 00:17 |
@sonney2k | blackburn, btw it would be nice to see the actual error | 00:17 |
@sonney2k | not just error | 00:17 |
@sonney2k | at least at the end | 00:17 |
@sonney2k | blackburn, this output http://shogun-toolbox.org:5347/builders/python_static/builds/59/steps/test_1/logs/stdio is pretty useless | 00:17 |
@sonney2k | ERROR - > great | 00:18 |
blackburn | hooray to the error! hah | 00:18 |
blackburn | no idea how to do that | 00:18 |
* sonney2k is a dedicated ERROR collector | 00:18 | |
@sonney2k | blackburn, it is easy - just store output in an array and print it at the end | 00:19 |
@sonney2k | I will do it but not now (need sleep...) | 00:19 |
blackburn | yeah | 00:19 |
@sonney2k | blackburn, I don't understand why tests are failing on the buildbot | 00:20 |
@sonney2k | they all work here locally... | 00:20 |
blackburn | btw is python static working now? | 00:20 |
blackburn | haven't checked for a while | 00:20 |
shogun-buildbot | build #69 of cmdline_static is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org:5347/builders/cmdline_static/builds/69 blamelist: sonne@debian.org | 00:27 |
@sonney2k | it should | 00:41 |
@sonney2k | blackburn, ahh it needs the data submodule to be checked out | 00:49 |
@sonney2k | that is why things fail | 00:49 |
blackburn | thought you're sleeping :D | 00:49 |
blackburn | sonney2k: I'm now having a zoo of 9 dim reduction preprocessors :D | 00:51 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting] | 00:53 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 00:53 | |
@sonney2k | blackburn, don't forget to feed all your animals | 00:54 |
blackburn | want to start implementing sammon but still have to fix pcas and get some promoter visualization | 00:54 |
CIA-87 | shogun: Soeren Sonnenburg master * r5ed8a9a / examples/undocumented/python_static/check.sh : don't run for loop in a sub shell - http://bit.ly/qpZfjx | 00:56 |
@sonney2k | so now really everything should fail :) | 00:57 |
@sonney2k | blackburn, I have to figure out how to use the github trigger | 00:57 |
@sonney2k | then directly after a commit the buildbot will start its job | 00:57 |
blackburn | github trigger? | 00:57 |
blackburn | ah | 00:57 |
@sonney2k | yes | 01:05 |
@sonney2k | hmmhh data is big | 01:06 |
@sonney2k | the git clone takes ages... | 01:06 |
shogun-buildbot | build #70 of cmdline_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org:5347/builders/cmdline_static/builds/70 | 01:11 |
shogun-buildbot | build #61 of python_static is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org:5347/builders/python_static/builds/61 blamelist: sonne@debian.org | 01:20 |
@sonney2k | blackburn, alright - now it in principle works http://shogun-toolbox.org:5347/builders/python_static/builds/61/steps/test_1/logs/stdio | 01:21 |
@sonney2k | there is one python test failing | 01:21 |
@sonney2k | but we indeed don't support that for python | 01:21 |
@sonney2k | so removing the example would fix it | 01:21 |
blackburn | sonney2k: you will be laughing but my mds is the slowest one ever | 01:22 |
blackburn | I didn't know there is fast tecnhique for double centering | 01:22 |
blackburn | :( | 01:22 |
@sonney2k | :) | 01:22 |
blackburn | I did it using blas | 01:22 |
blackburn | dgemm | 01:22 |
blackburn | fuck | 01:22 |
blackburn | so stupid | 01:23 |
@sonney2k | anyway I am dead tired now | 01:23 |
@sonney2k | cu tomorrow! | 01:23 |
@sonney2k | today even | 01:23 |
blackburn | yeah | 01:23 |
blackburn | see you | 01:23 |
blackburn | I will be fixing mds | 01:23 |
serialhex | having fun fixing mds blackburn?? :P | 01:25 |
blackburn | serialhex: yeah and wondering how stupid I am | 01:25 |
serialhex | hey,you knew there was something broken (or at least not working well) | 01:26 |
serialhex | and that's the first step to making it better! :D | 01:26 |
blackburn | I noticed it because pca makes it for 0.02s | 01:26 |
blackburn | and mds for 2s :D | 01:27 |
serialhex | wow! | 01:30 |
serialhex | that's crazy!!! | 01:31 |
serialhex | hey, the python rand() func, when you call it like rand(7) does it give you an int between 0 & 6 blackburn?? | 01:32 |
blackburn | ehh | 01:32 |
blackburn | is there any rand() function? | 01:33 |
serialhex | and what would "data=rand(dim, dim)" do??? | 01:33 |
serialhex | hmm... ok | 01:33 |
serialhex | idk | 01:33 |
blackburn | hmm may be it is numpy function | 01:34 |
blackburn | let me check | 01:34 |
blackburn | serialhex: yeah numpy func | 01:34 |
serialhex | ahh, it is numpy! | 01:34 |
serialhex | yes :P | 01:34 |
blackburn | returns dim*dim matrix of uniform distributed randoms | 01:34 |
serialhex | hmm, ok | 01:34 |
blackburn | randint for your task | 01:34 |
blackburn | not numpy | 01:34 |
blackburn | or so | 01:34 |
serialhex | well i'm converting classifier_custom_kernel_modular so i ned to know how those funcs work to translate them properly | 01:35 |
blackburn | ah | 01:35 |
blackburn | so it is matrix of dim*dim [0,1] uniform distr blabla | 01:35 |
serialhex | ok cool | 01:36 |
serialhex | i think i know how to solve this stuff! :) | 01:37 |
blackburn | really funny | 01:37 |
blackburn | I thought MDS is pretty slow | 01:37 |
blackburn | itself | 01:37 |
blackburn | I made distance calculation parallel | 01:38 |
blackburn | I made landmark approximation | 01:38 |
serialhex | ok | 01:38 |
blackburn | and it is fucking awful I did 2 matrix-matrix multiplications | 01:38 |
blackburn | when no need to do it | 01:38 |
serialhex | HA!!! lol | 01:39 |
blackburn | serialhex: another one fact about my stupidity | 01:41 |
blackburn | we have center_matrix function in math | 01:41 |
serialhex | ok | 01:41 |
blackburn | and I didn't use it oooooh | 01:42 |
serialhex | HAHA!!! :) | 01:42 |
serialhex | it's ok, you'll live! :P | 01:43 |
blackburn | what a shame | 01:43 |
serialhex | well i just had to look up the interface for the random number generator that *I* wrote oh-so-many months ago :P | 01:44 |
blackburn | hey isn't there any lib for that?! | 01:44 |
serialhex | probably, but we have one in shogun that i wrote at the beginning of summer, remember?? | 01:46 |
blackburn | not really | 01:47 |
blackburn | can't remember | 01:47 |
serialhex | really? you made the graph for me b/c i couldn't figure out how TF to do it! :P | 01:49 |
blackburn | graph?! | 01:49 |
blackburn | ah!!!!1 | 01:50 |
serialhex | yes... it was a normal random function... like 10k iterations to see if it was truely 'normal' | 01:50 |
blackburn | yeah | 01:50 |
blackburn | yes yes | 01:50 |
serialhex | (or normal enough) | 01:50 |
serialhex | yeah, seems you forgot too :P | 01:51 |
serialhex | geez! i've had like 3 cups of coffe and a cup of tea and i'm kinda tired!!! :( | 01:54 |
-!- in3xes [~in3xes@180.149.49.227] has quit [Remote host closed the connection] | 02:11 | |
blackburn | did it | 02:34 |
serialhex | sweet!! | 02:36 |
serialhex | hows it run now blackburn?? | 02:36 |
blackburn | 20-30 times faster | 02:36 |
CIA-87 | shogun: Sergey Lisitsyn master * r46db854 / src/shogun/preprocessor/MultidimensionalScaling.cpp : Efficient matrix centering instead of crappy dgemms for MDS - http://bit.ly/n2fIGh | 02:37 |
serialhex | DAMN!!!! | 02:37 |
serialhex | thats sweet! | 02:38 |
blackburn | serialhex: 15 times faster for 1000 | 02:39 |
blackburn | and 3x less memory usage.. | 02:39 |
blackburn | okay I have to sleep now | 02:40 |
blackburn | 4-40 here | 02:40 |
serialhex | DAMN!!!!! :P | 02:40 |
serialhex | ok man, g'nite and sleep well!!! | 02:40 |
blackburn | see you | 02:40 |
blackburn | good evening? :) | 02:40 |
-!- blackburn [~blackburn@188.122.238.50] has quit [Quit: Leaving.] | 02:40 | |
CIA-87 | shogun: Soeren Sonnenburg master * r90c20fe / examples/undocumented/python_static/kernel_linearbyte.py : remove unsupported example - http://bit.ly/ramReQ | 07:09 |
shogun-buildbot | build #63 of python_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org:5347/builders/python_static/builds/63 | 07:38 |
CIA-87 | shogun: Soeren Sonnenburg master * r9475428 / (9 files in 9 dirs): collect all errors in error.log and show them at the end in detail - http://bit.ly/nrNdZI | 09:55 |
shogun-buildbot | build #73 of cmdline_static is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org:5347/builders/cmdline_static/builds/73 blamelist: sonne@debian.org | 10:07 |
-!- blackburn [~blackburn@188.122.238.50] has joined #shogun | 10:12 | |
blackburn | sonney2k: have you read about my shame? :D | 10:14 |
shogun-buildbot | build #64 of python_static is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org:5347/builders/python_static/builds/64 blamelist: sonne@debian.org | 10:16 |
CIA-87 | shogun: Soeren Sonnenburg master * rdd04e8d / (9 files in 9 dirs): exit 0 upon examples running all OK - http://bit.ly/ocJzFP | 10:16 |
CIA-87 | shogun: Soeren Sonnenburg master * r675a2a7 / src/Makefile.template : fix r static doc install path - http://bit.ly/r3S9qp | 10:16 |
@sonney2k | blackburn, heroic work yes | 10:19 |
@sonney2k | blackburn, however I think I can beat you in all aspects of dumbness | 10:20 |
@sonney2k | but not today | 10:20 |
@sonney2k | today I am brilliant :D | 10:20 |
blackburn | heh | 10:20 |
blackburn | I should profile my algos asap | 10:22 |
blackburn | sonney2k: the most funny moment is that eigenproblem takes 0.7s for 10000 hah | 10:25 |
blackburn | so the bottleneck was NOT the eigenproblem | 10:25 |
shogun-buildbot | build #74 of cmdline_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org:5347/builders/cmdline_static/builds/74 | 10:28 |
blackburn | sonney2k: do you know how to do (I-W)(I-W)' without dgemm? | 10:28 |
shogun-buildbot | build #65 of python_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org:5347/builders/python_static/builds/65 | 10:28 |
blackburn | I'm now getting unsure there is no fast way to do such ops | 10:28 |
shogun-buildbot | build #65 of octave_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org:5347/builders/octave_static/builds/65 | 10:42 |
CIA-87 | shogun: Soeren Sonnenburg master * r201a958 / (12 files in 11 dirs): | 10:47 |
CIA-87 | shogun: add check.sh script to libshogun to follow the same testing principle | 10:47 |
CIA-87 | shogun: and really exit 1 on failure - http://bit.ly/oG3pbN | 10:47 |
shogun-buildbot | build #80 of libshogun is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org:5347/builders/libshogun/builds/80 blamelist: sonne@debian.org | 10:51 |
CIA-87 | shogun: Soeren Sonnenburg master * r316fcc9 / examples/undocumented/libshogun/check.sh : add escaped quotes to libshogun check - http://bit.ly/r6Zwp7 | 11:13 |
-!- heiko [~heiko@134.91.55.115] has joined #shogun | 11:31 | |
heiko | shogun-buildbot, status all | 11:45 |
shogun-buildbot | cmdline_static: idle, last build 56m49s ago: build successful | 11:45 |
shogun-buildbot | csharp_modular: idle, last build 13h13m49s ago: | 11:45 |
shogun-buildbot | java_modular: idle, last build 13h10m48s ago: | 11:45 |
shogun-buildbot | libshogun: idle, last build 29m26s ago: failed test_1 | 11:45 |
shogun-buildbot | lua_modular: idle, last build 13h13m49s ago: | 11:45 |
shogun-buildbot | nightly_all: idle, last build 13h53m51s ago: build successful | 11:45 |
shogun-buildbot | nightly_none: idle, last build 13h53m51s ago: build successful | 11:45 |
shogun-buildbot | octave_modular: idle, last build 12h55m48s ago: failed test test_1 | 11:45 |
shogun-buildbot | octave_static: idle, last build 1h03m04s ago: build successful | 11:45 |
shogun-buildbot | python_modular: idle, last build 13h13m49s ago: | 11:45 |
shogun-buildbot | python_static: idle, last build 57m27s ago: build successful | 11:45 |
shogun-buildbot | r_modular: idle, last build 13h13m49s ago: | 11:45 |
shogun-buildbot | r_static: idle, last build 1h11m15s ago: failed test | 11:45 |
shogun-buildbot | ruby_modular: idle, last build 13h13m49s ago: | 11:45 |
heiko | sonney2k, around? | 11:45 |
@sonney2k | heiko, yes | 11:46 |
heiko | sonney2k, did the patch work? | 11:46 |
@sonney2k | yes | 11:46 |
@sonney2k | perfect | 11:46 |
heiko | sonney2k, cool 8-) | 11:46 |
heiko | sonney2k, however there are still problems: | 11:46 |
@sonney2k | I am now trying hard to get the buildbot to work with all the make check-examples stuff | 11:46 |
@sonney2k | (for the last 20rhs already...) | 11:46 |
heiko | the hdf5 file leaks memory, even without serialization going on, json also, xml does read from uninit memory | 11:47 |
heiko | sonney2k, oh no, were you awake the whole night? | 11:47 |
@sonney2k | heiko, no | 11:48 |
heiko | puh :) | 11:48 |
@sonney2k | heiko, yeah I've seen your comment | 11:48 |
@sonney2k | we need migrations more urgently now | 11:48 |
heiko | ok | 11:48 |
heiko | I will go for it | 11:48 |
@sonney2k | so ignore for the time being | 11:48 |
CIA-87 | shogun: Soeren Sonnenburg master * r4d140a3 / (6 files): fix check.sh and add return 0 to examples missing the statement - http://bit.ly/o9hQhG | 12:07 |
blackburn | sonney2k: how can I set parallel for CDistance? | 12:13 |
@sonney2k | obj.parallel | 12:14 |
@sonney2k | (like for ever sgobject) | 12:14 |
blackburn | ?? | 12:14 |
blackburn | ah | 12:14 |
blackburn | it is public! | 12:14 |
blackburn | nice | 12:14 |
CIA-87 | shogun: Soeren Sonnenburg master * ra4e2ac1 / examples/undocumented/libshogun/check.sh : forgot escaping args - http://bit.ly/pRS9pk | 12:20 |
shogun-buildbot | build #83 of libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org:5347/builders/libshogun/builds/83 | 12:26 |
@sonney2k | finally | 12:27 |
heiko | finally :) | 12:33 |
shogun-buildbot | build #67 of octave_static is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org:5347/builders/octave_static/builds/67 blamelist: sonne@debian.org | 12:35 |
@sonney2k | hmm Inconsistency detected by ld.so: dl-close.c: 736: _dl_close: Assertion `map->l_init_called' failed! | 12:37 |
@sonney2k | ?!? | 12:37 |
heiko | ehm ?? never saw this | 12:41 |
@sonney2k | heiko, even better doesn't happen on my notebook but only on the buildbot | 12:44 |
heiko | sonney2k, I love these kind of errors | 12:44 |
heiko | different g++ version?= | 12:44 |
@sonney2k | heiko, ahh yeah there it is http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=633719 | 12:45 |
@sonney2k | valgrind doesn't say anything | 12:47 |
@sonney2k | great! | 12:47 |
* sonney2k is rebuilding octave3.2 | 12:50 | |
@sonney2k | yeah! | 13:02 |
@sonney2k | doesn't even build :D | 13:02 |
@sonney2k | o_O | 13:02 |
@sonney2k | and this issue is already filed as bug #637787 | 13:03 |
blackburn | huh LTSA weight calculation step is slower when parallel | 13:13 |
blackburn | hey have anybody real multiprocessor machine? heiko? sonney2k? | 13:28 |
heiko | blackburn, i have a centrino-duo | 13:29 |
heiko | i think there are two cores | 13:29 |
blackburn | I see | 13:29 |
heiko | but not two processors | 13:29 |
blackburn | same here | 13:29 |
CIA-87 | shogun: Soeren Sonnenburg master * r936bcdf / src/shogun/structure/DynProg.h : fix a few warnings in DynProg - http://bit.ly/mRXgmb | 13:30 |
blackburn | fuck yeah | 13:36 |
blackburn | spinlock rocks | 13:37 |
shogun-buildbot | build #72 of r_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org:5347/builders/r_static/builds/72 | 13:47 |
shogun-buildbot | build #68 of octave_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org:5347/builders/octave_static/builds/68 | 13:54 |
CIA-87 | shogun: Sergey Lisitsyn master * rc0f76ee / (2 files): Made neighbor search parallel in LocallyLinearEmbedding - http://bit.ly/o2bGh2 | 13:58 |
CIA-87 | shogun: Sergey Lisitsyn master * r92c21ec / (4 files): Spinlock instead of mutex - http://bit.ly/mQ4i4Y | 13:58 |
-!- blackburn [~blackburn@188.122.238.50] has quit [Quit: Leaving.] | 14:09 | |
shogun-buildbot | build #70 of octave_modular is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org:5347/builders/octave_modular/builds/70 | 14:20 |
-!- blackburn [~blackburn@188.122.238.50] has joined #shogun | 17:48 | |
blackburn | :D | 17:52 |
CIA-87 | shogun: Sergey Lisitsyn master * r77c1f2b / (3 files in 3 dirs): Made 100 warnings left - http://bit.ly/oqTXDv | 17:53 |
blackburn | oh it seems I got wrong | 17:53 |
blackburn | heiko: ich heisse sergey | 18:42 |
blackburn | :D | 18:42 |
heiko | blackburn, learning german for mariage? :D | 18:42 |
blackburn | genau | 18:42 |
heiko | hehe | 18:43 |
blackburn | in fact I really want to learn german | 18:43 |
blackburn | but english is prior of course haha | 18:43 |
heiko | what for? | 18:45 |
heiko | i mean german | 18:45 |
blackburn | idk may be possible country to live :) | 18:45 |
blackburn | or just for interest | 18:45 |
blackburn | it is pretty bad here you know :D | 18:45 |
-!- heiko [~heiko@134.91.55.115] has left #shogun [] | 19:28 | |
@sonney2k | blackburn, did you fix these already http://shogun-toolbox.org:5347/builders/libshogun/builds/86/steps/compile/logs/warnings%20(2)/text | 19:42 |
@sonney2k | ? | 19:42 |
blackburn | oh | 19:47 |
blackburn | these ones are new | 19:47 |
@sonney2k | blackburn, want to get back to 2500 warnings? | 19:49 |
@sonney2k | *eg* | 19:49 |
blackburn | don't understand what is you are talking about | 19:49 |
@sonney2k | blackburn, perestroika | 19:49 |
blackburn | NULL -> 0 is not a hard conversion | 19:49 |
CIA-87 | shogun: Sergey Lisitsyn master * r4824398 / (2 files): Fixed spinlock inits - http://bit.ly/omUI0l | 19:50 |
CIA-87 | shogun: Soeren Sonnenburg master * rdf7739a / (2 files): fix ppwm example and disable crashing parts in structure_* - http://bit.ly/nbYDH2 | 20:16 |
@sonney2k | blackburn, alright octave_modular and python_modular are working 100% w/ all examples | 20:17 |
@sonney2k | r_modular cannot work until swig is fixed | 20:18 |
@sonney2k | so what remains to do is proper installation for java_modular, lua_modular, csharp_modular and ruby_modular | 20:18 |
@sonney2k | and then running the examples we have / and their ports | 20:18 |
shogun-buildbot | build #65 of r_modular is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org:5347/builders/r_modular/builds/65 | 20:22 |
shogun-buildbot | build #62 of python_modular is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org:5347/builders/python_modular/builds/62 | 20:22 |
blackburn | I would like to remove all warns and finish pca/kpca | 20:28 |
blackburn | but need actual list of warnings | 20:29 |
blackburn | sonney2k: should we get applications back to work? | 20:44 |
shogun-buildbot | build #66 of r_modular is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org:5347/builders/r_modular/builds/66 blamelist: blackburn91@gmail.com | 20:58 |
@sonney2k | blackburn, yes | 20:59 |
@sonney2k | but first the examples | 20:59 |
blackburn | a lot of work still.. | 20:59 |
@sonney2k | blackburn, yes but I am positively surprised that octave/python_modular are good to go already | 21:00 |
@sonney2k | if we have migrations we could get the old tests for python back to work and then it is even clear that we have no regressions | 21:00 |
blackburn | I guess we have | 21:00 |
@sonney2k | I don't really expect any | 21:02 |
@sonney2k | 1-2 maybe | 21:02 |
@sonney2k | blackburn, can you write an analogous check.sh script like I wrote for python_modular for some other $LANG? | 21:09 |
@sonney2k | ruby / java / csharp / lua? | 21:09 |
blackburn | yeah but I have to finish kpca first :( | 21:09 |
blackburn | I don't understand what should I do with kpca .. | 21:12 |
blackburn | sonney2k: why preimage is always computed here in stprtool | 21:14 |
blackburn | I really lost the idea of kpca.. | 21:14 |
@sonney2k | blackburn, the idea is easy - just do pca in kernel feature space | 21:15 |
@sonney2k | the inverse mapping is nice to have but not essential | 21:16 |
@sonney2k | and I don't know how one could compute it if inputs are not just real valued numbers | 21:16 |
blackburn | so in vojtechs code kernel is used | 21:16 |
@sonney2k | (kpca can work with anything) | 21:16 |
blackburn | what is the lhs and rhs for tihs kernel? | 21:16 |
@sonney2k | like strings | 21:16 |
@sonney2k | lhs == train data, rhs == train or test data | 21:17 |
blackburn | so I have transformation matrix | 21:18 |
blackburn | but what is the kernel in apply? should it be centered? | 21:18 |
@sonney2k | lhs=train & rhs=test data | 21:21 |
@sonney2k | centering - i think so | 21:21 |
@sonney2k | look at vojtechs code he does some magic there | 21:21 |
shogun-buildbot | build #63 of python_modular is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org:5347/builders/python_modular/builds/63 blamelist: blackburn91@gmail.com | 21:21 |
blackburn | model.Alpha = (eye(num_data,num_data)-J)*A; | 21:22 |
blackburn | Jt=ones(num_data,1)/num_data; | 21:22 |
blackburn | model.b = A'*(J'*K*Jt-K*Jt); | 21:22 |
@sonney2k | ^ err that is an old build | 21:22 |
blackburn | this things are being computed in init | 21:22 |
blackburn | right? | 21:22 |
@sonney2k | yrah | 21:22 |
@sonney2k | (I cut my finger badly - so it is currently a bit hard to type) | 21:23 |
blackburn | oh | 21:23 |
blackburn | are you ok? | 21:23 |
@sonney2k | yes - no blood on the keyboard yet :) | 21:23 |
@sonney2k | blackburn, 100 warnings exactly now :) | 21:26 |
blackburn | yeah | 21:27 |
blackburn | I will try to grasp kpca 10-15 min more | 21:27 |
blackburn | and will remove all of them finally | 21:27 |
blackburn | no, no centering here | 21:29 |
CIA-87 | shogun: Soeren Sonnenburg master * rd0c6dda / (3 files): Fix documentation in structure. - http://bit.ly/qmmmxI | 21:33 |
CIA-87 | shogun: Soeren Sonnenburg master * r24610e5 / src/shogun/kernel/SNPStringKernel.h : fix documentation in SNPstringkernel - http://bit.ly/qKJ1NU | 21:36 |
CIA-87 | shogun: Soeren Sonnenburg master * rb6a1b12 / src/shogun/lib/DynInt.h : fix type definitions and properly document types up to 1024bit integers - http://bit.ly/pxQWlI | 21:47 |
CIA-87 | shogun: Soeren Sonnenburg master * r618179a / src/shogun/base/DynArray.h : remove extern SGIO statement - http://bit.ly/pbcuRU | 21:47 |
CIA-87 | shogun: Soeren Sonnenburg master * raa7f594 / src/shogun/distributions/HMM.h : document BaumWelchViterbiType - http://bit.ly/pfzNg6 | 21:47 |
@sonney2k | blackburn, but there must be ... in pca you also center the data before applying pca | 21:48 |
blackburn | no, kernel which eigenvectors are used is centered | 21:49 |
blackburn | in apply no centering | 21:49 |
shogun-buildbot | build #7 of nightly_none is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org:5347/builders/nightly_none/builds/7 | 21:53 |
@sonney2k | blackburn, how do we treat this? nightly_none fails (of course) because there are some things required to run the example which are just disabled | 21:55 |
@sonney2k | blackburn, should I disable running tests for nightly_* ? | 21:55 |
@sonney2k | and only check if things compile? | 21:55 |
blackburn | yeah disable it better | 21:55 |
shogun-buildbot | build #7 of nightly_all is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org:5347/builders/nightly_all/builds/7 | 21:56 |
blackburn | sonney2k: oh I accidentally did it | 21:58 |
@sonney2k | ? | 21:58 |
@sonney2k | did what | 21:58 |
blackburn | kpca apply | 21:58 |
@sonney2k | why does nightly_all still fail *grmpf* | 22:00 |
@sonney2k | I am again too tired to continue just now - I guess I will do java_modular check.sh in the morning | 22:01 |
blackburn | eh?? | 22:01 |
blackburn | there is check.sh already | 22:01 |
@sonney2k | a working check.sh | 22:01 |
@sonney2k | and it needs to integrate with make install too | 22:05 |
@sonney2k | I guess it is not that hard so when that works c# should be easy too | 22:06 |
blackburn | which check.sh should I do? | 22:06 |
@sonney2k | and then there is only lua and ruby left | 22:06 |
@sonney2k | blackburn, fix the warnings for now | 22:06 |
blackburn | okay | 22:06 |
@sonney2k | that's enough work | 22:06 |
blackburn | kpca is almost ready | 22:06 |
@sonney2k | then I guess we will have errors in java_modular tomorrow | 22:06 |
@sonney2k | and c# | 22:06 |
@sonney2k | and we need to complete examples, i.e. convert the ones from python_modular to octave/java/csharp/ruby/lua | 22:07 |
@sonney2k | at least a number of them | 22:07 |
@sonney2k | alright | 22:08 |
blackburn | sonney2k: should cleanup make preprocessor uninitialized? | 22:08 |
@sonney2k | cleanup should remove everything but essential base stuff | 22:10 |
@sonney2k | so yes | 22:10 |
@sonney2k | blackburn, http://jakemcgraw.com/imgs/langs.jpg | 22:10 |
blackburn | yeah truly thing | 22:11 |
CIA-87 | shogun: Sergey Lisitsyn master * r16c77fc / (9 files in 2 dirs): Made PCA preprocessor inherited from DimensionReductionPreprocessor, renamed KernelPCACut to KernelPCA - http://bit.ly/nISZUD | 22:55 |
CIA-87 | shogun: Sergey Lisitsyn master * rd88e533 / (2 files): Updated kPCA python modular example - http://bit.ly/qGqAt6 | 23:00 |
CIA-87 | shogun: Sergey Lisitsyn master * r92a2edb / (2 files): Updated kPCA ruby example - http://bit.ly/qxDHcU | 23:01 |
CIA-87 | shogun: Sergey Lisitsyn master * rae7dc3e / (2 files): Updated kPCA java modular example - http://bit.ly/qEkAJC | 23:04 |
CIA-87 | shogun: Sergey Lisitsyn master * rf192c44 / (17 files in 8 dirs): Warning fixes - http://bit.ly/rmc7rz | 23:16 |
-!- blackburn [~blackburn@188.122.238.50] has quit [Quit: Leaving.] | 23:40 | |
-!- blackburn [~blackburn@188.122.238.50] has joined #shogun | 23:40 | |
-!- blackburn [~blackburn@188.122.238.50] has quit [Client Quit] | 23:40 | |
--- Log closed Sat Aug 20 00:00:06 2011 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!