--- Log opened Wed Dec 16 00:00:31 2015 | ||
-!- HeikoS [~heiko@98.143.210.167] has quit [Quit: Leaving.] | 00:29 | |
-!- shaochuan [~shaochuan@c-73-158-54-228.hsd1.ca.comcast.net] has joined #shogun | 00:44 | |
-!- HeikoS [~heiko@98.143.210.167] has joined #shogun | 01:00 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 01:00 | |
-!- HeikoS [~heiko@98.143.210.167] has quit [Client Quit] | 01:00 | |
rcurtin | lisitsyn: I think the idea of labeling source via git blame and ordering authors by lines of code committed is really cool. if you do this, maybe you can point me at the code you used so maybe I can do the same to our code in mlpack? :) | 01:33 |
---|---|---|
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 01:36 | |
shogun-buildbot | build #1054 of nightly_default is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/1054 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Sergey Lisitsyn <lisitsyn.s.o@gmail.com>, Wu Lin <yorker.lin@gmail.com>, yorkerlin <yorker.lin@gmail.com> | 04:49 |
shogun-buildbot | build #932 of nightly_all is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_all/builds/932 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Sergey Lisitsyn <lisitsyn.s.o@gmail.com>, Wu Lin <yorker.lin@gmail.com>, yorkerlin <yorker.lin@gmail.com> | 05:17 |
-!- HeikoS [~heiko@98.143.210.167] has joined #shogun | 05:37 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 05:37 | |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 05:43 | |
shogun-notifier- | shogun: Wu Lin :develop * 1881d79 / src/shogun/optimization/ (5 files): https://github.com/shogun-toolbox/shogun/commit/1881d79ee110a12a87b9f61895a56ca99d2a1e5f | 05:43 |
shogun-notifier- | shogun: Fix python_modular warnings by adding API documentation | 05:43 |
shogun-notifier- | shogun: Wu Lin :develop * 9646e59 / src/shogun/optimization/ (5 files): https://github.com/shogun-toolbox/shogun/commit/9646e5917190f946bfc2a465b532d411d25b1fe9 | 05:43 |
shogun-notifier- | shogun: Merge pull request #2948 from yorkerlin/develop | 05:43 |
shogun-notifier- | shogun: | 05:43 |
shogun-notifier- | shogun: Fix python_modular warnings by adding API documentation | 05:43 |
shogun-buildbot | build #392 of trusty - libshogun - viennacl is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/trusty%20-%20libshogun%20-%20viennacl/builds/392 blamelist: Wu Lin <yorker.lin@gmail.com> | 05:49 |
shogun-buildbot | build #3521 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/3521 blamelist: Wu Lin <yorker.lin@gmail.com> | 05:50 |
-!- HeikoS [~heiko@98.143.210.167] has quit [Quit: Leaving.] | 06:01 | |
-!- travis-ci [~travis-ci@ec2-54-145-108-159.compute-1.amazonaws.com] has joined #shogun | 06:23 | |
travis-ci | it's Wu Lin's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: https://travis-ci.org/shogun-toolbox/shogun/builds/97143711 | 06:23 |
-!- travis-ci [~travis-ci@ec2-54-145-108-159.compute-1.amazonaws.com] has left #shogun [] | 06:23 | |
-!- sanuj [~sanuj@117.204.249.17] has joined #shogun | 06:28 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 08:43 | |
-!- sanuj [~sanuj@117.204.249.17] has quit [Read error: Connection reset by peer] | 12:26 | |
-!- sanuj [~sanuj@117.204.249.17] has joined #shogun | 16:23 | |
-!- HeikoS [~heiko@98.143.210.167] has joined #shogun | 16:25 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 16:25 | |
-!- lupinix [~quassel@fedora/lupinix] has quit [Remote host closed the connection] | 16:46 | |
-!- lupinix [~quassel@v22014041761818086.yourvserver.net] has joined #shogun | 16:51 | |
-!- lupinix [~quassel@v22014041761818086.yourvserver.net] has quit [Changing host] | 16:51 | |
-!- lupinix [~quassel@fedora/lupinix] has joined #shogun | 16:51 | |
-!- HeikoS [~heiko@98.143.210.167] has quit [Quit: Leaving.] | 17:39 | |
-!- sanuj [~sanuj@117.204.249.17] has quit [Quit: Leaving] | 18:13 | |
-!- lordon [5f5ad198@gateway/web/freenode/ip.95.90.209.152] has joined #shogun | 18:27 | |
lordon | Hey guys, maybe one of you guys can help me, I want to do some machine learning stuff with shoguns Ruby modular interface. Unfortunately, when I do exactly the same with Python and Ruby I get different results (The Ruby one is rather bad...) - here you can find two Jupyter notebooks with PCA as example: Python: http://bit.ly/1I9hJhP Ruby: http://bit.ly/1NTw3M3 Any idea? | 18:31 |
-!- HeikoS [~heiko@98.143.210.167] has joined #shogun | 18:45 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 18:45 | |
-!- HeikoS [~heiko@98.143.210.167] has quit [Quit: Leaving.] | 19:08 | |
-!- HeikoS [~heiko@98.143.210.167] has joined #shogun | 19:10 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 19:10 | |
@besser82 | lordon, that might be because of the ruy-interpreter... | 20:02 |
@besser82 | ruby | 20:02 |
@besser82 | lordon, which OS are you using? | 20:03 |
-!- HeikoS [~heiko@98.143.210.167] has quit [Quit: Leaving.] | 20:17 | |
-!- HeikoS [~heiko@98.143.210.167] has joined #shogun | 20:17 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 20:17 | |
-!- HeikoS [~heiko@98.143.210.167] has quit [Quit: Leaving.] | 20:43 | |
-!- HeikoS [~heiko@98.143.210.167] has joined #shogun | 20:44 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 20:44 | |
-!- HeikoS [~heiko@98.143.210.167] has quit [Client Quit] | 20:45 | |
-!- HeikoS [~heiko@98.143.210.167] has joined #shogun | 20:47 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 20:47 | |
-!- HeikoS [~heiko@98.143.210.167] has quit [Ping timeout: 265 seconds] | 20:53 | |
lisitsyn | uhm this is quite bad | 21:42 |
lordon | besser82, I'm using Ubuntu versions 14.10 and 15.04 LTS, ruby 2.2.2 MRI (using rbenv) | 22:35 |
@besser82 | lisitsyn, ? | 22:39 |
lisitsyn | from the example I see typemaps seem to be ok | 22:40 |
lisitsyn | but somehow results differ | 22:40 |
@besser82 | lordon, so your Ruby interpreter is compiled from scratch? | 22:40 |
@besser82 | lisitsyn, mhh... so testsuite-data is faulty, too? :/ | 22:41 |
lisitsyn | why? | 22:41 |
@besser82 | lisitsyn, for me ruby-tests are passing... so the precomputed results are differing, i asume? | 22:42 |
@besser82 | like we compare with pickles in python-testsuite | 22:43 |
lisitsyn | not sure if it is covered | 22:43 |
@besser82 | lisitsyn, just checked... It isn't covered | 22:47 |
@besser82 | :( | 22:47 |
lisitsyn | I am not sure where error could happen | 22:47 |
lisitsyn | I mean under the hood it is the same C++ code | 22:47 |
lisitsyn | so typemaps? | 22:47 |
@besser82 | So we just check if functions in the interface are working | 22:47 |
lisitsyn | but if they glitch they like explode | 22:48 |
@besser82 | lisitsyn, maybe some difference in float handling with ruby? | 22:48 |
@besser82 | like the interp handled float-precision differently | 22:48 |
lisitsyn | uh should not change it that much | 22:48 |
@besser82 | like with python between ix86 and x86_64 :/ | 22:49 |
@besser82 | possibly it's some differnce in numpy vs. narray | 22:49 |
@besser82 | I'll check that tomorrow | 22:50 |
lisitsyn | besser82: oh would be nice if you get some time | 22:50 |
@besser82 | doing some random (but similar values) float arithmetics with numpy vs. narray and see the outcome | 22:51 |
@besser82 | lisitsyn, I'll do my best... :P | 22:51 |
@besser82 | lordon, ^ see backlog; might be because of narray... Needs some more investigation | 22:52 |
@besser82 | It's bed-time for me, now... Godd N8, folks! =) | 22:52 |
lordon | besser82 yes, rbenv install compiles ruby from scratch.. thank you guys! I was wondering because I thought the same C++ things were called, too... | 22:54 |
-!- lordon [5f5ad198@gateway/web/freenode/ip.95.90.209.152] has quit [Ping timeout: 252 seconds] | 23:00 | |
--- Log closed Thu Dec 17 00:00:32 2015 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!