--- Log opened Mon May 13 00:00:42 2013 | ||
shogun-buildbot | build #1104 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/1104 blamelist: Soeren Sonnenburg <sonne@debian.org> | 00:07 |
---|---|---|
@sonney2k | lisitsyn, all integration test data needs to be regenerated | 00:13 |
@lisitsyn | sonney2k: again? | 00:13 |
@sonney2k | if you are still awake please do it ... run generator.py and commit | 00:14 |
@sonney2k | yeah | 00:14 |
@sonney2k | preprocessor is now a Dyn*array | 00:14 |
@sonney2k | not longer a ** | 00:14 |
@sonney2k | so reading serialized files failes now | 00:15 |
@lisitsyn | sonney2k: but why that affected data | 00:15 |
@lisitsyn | ahh | 00:15 |
@sonney2k | for *all* features | 00:15 |
@lisitsyn | got it | 00:15 |
@lisitsyn | yeah will do right | 00:15 |
@sonney2k | intrusive change | 00:15 |
@sonney2k | thanks | 00:15 |
@sonney2k | and cu | 00:15 |
-!- dvalcarce [5b757556@gateway/web/freenode/ip.91.117.117.86] has joined #shogun | 00:41 | |
shogun-notifier- | shogun: Sergey Lisitsyn :develop * 611ba7f / data: https://github.com/shogun-toolbox/shogun/commit/611ba7f1a8b397efc2fd3b96c41d71177db1c178 | 00:59 |
shogun-notifier- | shogun: Updated data reference | 00:59 |
-!- dvalcarce [5b757556@gateway/web/freenode/ip.91.117.117.86] has quit [Quit: Page closed] | 01:17 | |
-!- travis-ci [~travis-ci@ec2-107-20-75-71.compute-1.amazonaws.com] has joined #shogun | 01:24 | |
travis-ci | [travis-ci] it's Sergey Lisitsyn'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/7109116 | 01:24 |
-!- travis-ci [~travis-ci@ec2-107-20-75-71.compute-1.amazonaws.com] has left #shogun [] | 01:24 | |
shogun-buildbot | build #810 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/810 | 01:28 |
pickle27 | hey is anyone on here right now? | 01:29 |
van51 | pickle27: I am :p | 01:40 |
pickle27 | van51: cool, I wanted to ask somebody a question | 01:40 |
van51 | pickle27: shoot then and I hope I can help | 01:41 |
pickle27 | van51: so I've finished my eigen3 version of MCLDA and I want to get some feedback, is a pull request the preferred way to do this? Its what I did last time but thought I'd ask this time | 01:42 |
pickle27 | van51: basically I am planning to do more work on it so I don't want it merge yet | 01:43 |
pickle27 | van51: thoughts? | 01:44 |
van51 | pickle27: I am not sure if it is the preferred way, but I 've seen that if you reference the issue you are working on in your commit message and push in your repository, like #1061, then it makes a post under that issue with a link to your push | 01:45 |
van51 | and it shows the commit with the diff | 01:45 |
pickle27 | van51: interesting | 01:45 |
van51 | so maybe they can comment there instead on a PR | 01:45 |
van51 | of* | 01:45 |
van51 | but that is just an alternative that I can think of.. not sure if it is the best way to go :p | 01:47 |
shogun-buildbot | build #1105 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/1105 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 01:51 |
@lisitsyn | pickle27: yes just do a PR | 01:52 |
pickle27 | lisitsyn: okay will do | 01:53 |
@lisitsyn | argh I can't get asleep | 01:53 |
pickle27 | might help if you weren't on IRC :p | 01:53 |
@lisitsyn | no I was lying and trying to fall asleep | 01:53 |
pickle27 | lisitsyn: yeah it must be late there | 01:55 |
@lisitsyn | pickle27: 3:55 | 01:55 |
pickle27 | lisitsyn: oh wow thats real late | 01:55 |
@lisitsyn | pickle27: kind of, yes | 01:56 |
van51 | do you have somewhere to be in the morning? | 01:56 |
@lisitsyn | van51: of course, job :) | 01:56 |
van51 | damn | 01:56 |
@lisitsyn | actually 100 km bus trip first | 01:56 |
van51 | maybe you can get some sleep there :p | 01:56 |
@lisitsyn | van51: yeah that's what I usually do :D | 01:57 |
van51 | haha | 01:57 |
@lisitsyn | van51: good I am just a programmer not doing some cars or airplanes :D | 01:58 |
van51 | lisitsyn: yea | 01:59 |
van51 | or a surgeon even worse | 01:59 |
@lisitsyn | van51: http://www.surgeonsimulator2013.com/ | 01:59 |
van51 | hehe | 02:00 |
van51 | have you tried it? | 02:00 |
@lisitsyn | yeah failed badly | 02:00 |
@lisitsyn | van51: it is actually easy to break all the bones | 02:00 |
@lisitsyn | van51: there is a demo | 02:00 |
@lisitsyn | when you have to change a heart | 02:00 |
@lisitsyn | :D | 02:00 |
van51 | and you broke the heart as well | 02:01 |
van51 | :p | 02:01 |
van51 | I want to try it myself, I've seen a couple of videos and seems funny | 02:01 |
-!- van51 [~van51@athedsl-320452.home.otenet.gr] has left #shogun ["QUIT :Leaving."] | 03:15 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 04:00 | |
shogun-buildbot | build #393 of nightly_default is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/393 | 04:15 |
-!- foulwall [~foulwall@2001:da8:215:503:3d0f:f6ae:e6ff:aaef] has joined #shogun | 04:17 | |
foulwall | hi @sonney2k? | 04:18 |
-!- foulwall [~foulwall@2001:da8:215:503:3d0f:f6ae:e6ff:aaef] has quit [Ping timeout: 245 seconds] | 04:29 | |
-!- sijin [~smuxi@144.214.222.109] has quit [Excess Flood] | 04:48 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has joined #shogun | 04:59 | |
-!- sijin [~smuxi@144.214.222.109] has joined #shogun | 05:02 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has quit [Remote host closed the connection] | 05:20 | |
-!- abinash [75c8e44b@gateway/web/freenode/ip.117.200.228.75] has joined #shogun | 06:00 | |
-!- abinash [75c8e44b@gateway/web/freenode/ip.117.200.228.75] has quit [Ping timeout: 245 seconds] | 06:13 | |
-!- nube [~rho@116.90.239.3] has joined #shogun | 06:23 | |
-!- lisitsyn [~blackburn@188-122-243-31.clients.tlt.100megabit.ru] has quit [Quit: Leaving.] | 06:33 | |
-!- kakashi_ [~kakashi_@nltk/kakashi] has quit [Read error: Operation timed out] | 07:14 | |
-!- nube [~rho@116.90.239.3] has quit [Remote host closed the connection] | 07:18 | |
-!- kakashi_ [~kakashi_@nltk/kakashi] has joined #shogun | 07:20 | |
-!- nube [~rho@116.90.239.3] has joined #shogun | 07:47 | |
-!- nube [~rho@116.90.239.3] has quit [Ping timeout: 252 seconds] | 08:13 | |
-!- lambday [3ba21749@gateway/web/freenode/ip.59.162.23.73] has joined #shogun | 08:19 | |
@sonney2k | lambday, good morning! | 08:20 |
-!- nube [~rho@116.90.239.13] has joined #shogun | 08:22 | |
lambday | sonney2k: good morning :) | 08:22 |
@sonney2k | lambday, so you are back in the game? | 08:27 |
lambday | sonney2k: yes.. studying the KRYLSTAT code... | 08:28 |
lambday | Heiko suggested a few things about efficient implementation... | 08:28 |
lambday | trying to figure out which parts we can directly use, and which parts need rewriting | 08:29 |
@sonney2k | lambday, very good - keep it on :) | 08:29 |
lambday | sonney2k: thank you.. :) | 08:29 |
-!- abinash [75c8e14e@gateway/web/freenode/ip.117.200.225.78] has joined #shogun | 08:30 | |
@sonney2k | pickle27, you have to ask lisitsyn about eigen3 vectorization - he is the expert with that | 08:30 |
abinash | sonney2k: as per the issue #1080 "We have a fully functional maximum likelihood II framework for Gaussian process regression.". I am unable to get it where it is implemented? | 08:34 |
@sonney2k | sry gtg | 08:34 |
abinash | ok, but anybody else can help me out on this | 08:35 |
-!- nube [~rho@116.90.239.13] has quit [Read error: Connection reset by peer] | 08:41 | |
-!- sonne|work [~sonnenbu@sams-office-nat.tomtomgroup.com] has joined #shogun | 08:52 | |
-!- nube [~rho@36.253.14.206] has joined #shogun | 09:00 | |
-!- nube [~rho@36.253.14.206] has quit [Client Quit] | 09:04 | |
-!- iglesiasg [c1934d18@gateway/web/freenode/ip.193.147.77.24] has joined #shogun | 09:15 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 09:16 | |
-!- nube [~rho@116.90.239.3] has joined #shogun | 09:30 | |
-!- abinash [75c8e14e@gateway/web/freenode/ip.117.200.225.78] has quit [Ping timeout: 245 seconds] | 09:35 | |
-!- nube [~rho@116.90.239.3] has quit [Quit: Leaving.] | 09:52 | |
-!- nube [~rho@116.90.239.3] has joined #shogun | 09:57 | |
-!- nube [~rho@116.90.239.3] has quit [Ping timeout: 246 seconds] | 10:04 | |
-!- nube [~rho@116.90.239.3] has joined #shogun | 10:09 | |
-!- iglesiasg_ [c1934d18@gateway/web/freenode/ip.193.147.77.24] has joined #shogun | 10:14 | |
-!- HeikoS [~heiko@nat-173-117.internal.eduroam.ucl.ac.uk] has joined #shogun | 10:14 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 10:14 | |
sonne|work | HeikoS: moin! seen abinash's question | 10:14 |
-!- iglesiasg [c1934d18@gateway/web/freenode/ip.193.147.77.24] has quit [Ping timeout: 245 seconds] | 10:15 | |
@HeikoS | sonne|work: hi! could you copy paste it? | 10:15 |
-!- iglesiasg_ is now known as iglesiasg | 10:15 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 10:15 | |
@HeikoS | ah I see the mail | 10:15 |
sonne|work | yeah | 10:15 |
sonne|work | iglesiasg, HeikoS: could you please run ./tester evaluation_multiclassovrevaluation_modular.py on your machine (with up-to-date shogun!) | 10:23 |
sonne|work | I dont' understand the build error | 10:23 |
sonne|work | test error | 10:23 |
@iglesiasg | sonne|work: ok | 10:24 |
@HeikoS | compiling ... | 10:25 |
sonne|work | iglesiasg, HeikoS thanks! | 10:26 |
-!- Netsplit *.net <-> *.split quits: pickle27 | 10:27 | |
@HeikoS | sonne|work: http://pastebin.com/cATb3Vv9 | 10:31 |
sonne|work | HeikoS: interesting - so you have the same deviation as the buildbot | 10:32 |
-!- lisitsyn [~lisitsin@mxs.kg.ru] has joined #shogun | 10:32 | |
-!- mode/#shogun [+o lisitsyn] by ChanServ | 10:32 | |
@HeikoS | will this be solved by equals? | 10:32 |
sonne|work | HeikoS: the deviation is to large | 10:32 |
sonne|work | -0.927693761815 | 10:32 |
sonne|work | +0.928008821676 | 10:32 |
sonne|work | too | 10:32 |
@lisitsyn | hey there | 10:33 |
sonne|work | lisitsyn: hey | 10:36 |
sonne|work | lisitsyn: http://shogun-toolbox.org/buildbot/builders/nightly_default/builds/393/steps/test/logs/stdio | 10:37 |
sonne|work | have you seen this build failure? | 10:37 |
@lisitsyn | sonne|work: not yet | 10:37 |
sonne|work | it seems HeikoS can reproduce it on his machine http://pastebin.com/cATb3Vv9 | 10:37 |
@lisitsyn | okay some deviation here | 10:37 |
sonne|work | lisitsyn: but I couldn't | 10:37 |
sonne|work | big deviation yes | 10:37 |
@lisitsyn | sonne|work: I am unable to check that on my machine now unfortunately | 10:37 |
sonne|work | lisitsyn: but you ran generator.py last night right? | 10:38 |
@lisitsyn | sonne|work: yes | 10:38 |
sonne|work | so you don't have it then | 10:38 |
@lisitsyn | if it is not floating random | 10:38 |
@lisitsyn | oh! go wings go | 10:40 |
sonne|work | 1e-3 is not | 10:40 |
@lisitsyn | ducks were beaten hurray | 10:40 |
@lisitsyn | sonne|work: why not | 10:40 |
sonne|work | how? | 10:40 |
sonne|work | this test did work before | 10:40 |
sonne|work | after wiking rand change we have this now | 10:41 |
@lisitsyn | sonne|work: no idea just guess | 10:41 |
@lisitsyn | everything is so complex I don't exclude any reason | 10:42 |
sonne|work | HeikoS: could you please add predicted_labels and svm to be returned in ../examples/undocumented/python_modular/evaluation_multiclassovrevaluation_modular.py ? | 10:43 |
sonne|work | then run generator.py and push upload everything | 10:43 |
sonne|work | lisitsyn: well it could be if numerics is different then we get a different solution up to svm eps precision | 10:44 |
@lisitsyn | sonne|work: but that's some accuracy or so, right? | 10:45 |
@lisitsyn | I think they differ by one incorrect answer or so | 10:45 |
@iglesiasg | I am getting this warning everywhere after updating to latest Shogun | 10:47 |
@iglesiasg | ../shogun/lib/external/dSFMT/dSFMT.h:47:4: warning: #warning "DSFMT_MEXP is not defined. I assume DSFMT_MEXP is 19937." [-Wcpp] #warning "DSFMT_MEXP is not defined. I assume DSFMT_MEXP is 19937." | 10:47 |
@lisitsyn | iglesiasg: reconfigure | 10:48 |
@iglesiasg | lisitsyn: all right | 10:48 |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 10:52 | |
shogun-notifier- | shogun: Heiko Strathmann :develop * b6e89ed / examples/undocumented/python_modular/evaluation_multiclassovrevaluation_modular.py: https://github.com/shogun-toolbox/shogun/commit/b6e89edf56db03455dda64f0de35c0aa160916a0 | 10:52 |
shogun-notifier- | shogun: added svm and predicted labels to output | 10:52 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 52fac1b / data: https://github.com/shogun-toolbox/shogun/commit/52fac1bc58154e02ace571769eab72335db58ccf | 10:52 |
shogun-notifier- | shogun: updated data | 10:52 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 2e33ec3 / / (2 files): https://github.com/shogun-toolbox/shogun/commit/2e33ec3b3b180aa5582b8248f316a9394f2174cb | 10:52 |
shogun-notifier- | shogun: Merge pull request #1088 from karlnapf/develop | 10:52 |
shogun-notifier- | shogun: | 10:52 |
shogun-notifier- | shogun: update multiclass ovr example | 10:52 |
sonne|work | HeikoS: thx | 10:57 |
-!- van51 [~van51@195.251.235.94] has joined #shogun | 11:05 | |
-!- iglesiasg_ [c1934d18@gateway/web/freenode/ip.193.147.77.24] has joined #shogun | 11:16 | |
-!- iglesiasg [c1934d18@gateway/web/freenode/ip.193.147.77.24] has quit [Ping timeout: 245 seconds] | 11:17 | |
iglesiasg_ | sonne|work: same precision error with the tester here http://pastebin.com/q5pPEUMx | 11:18 |
-!- iglesiasg_ is now known as iglesiasg | 11:18 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 11:18 | |
-!- nube [~rho@116.90.239.3] has quit [Ping timeout: 246 seconds] | 11:21 | |
sonne|work | iglesiasg: ok then I am clueless and I guess HeikoS push will fix it | 11:31 |
-!- nube [~rho@116.90.239.3] has joined #shogun | 11:31 | |
shogun-buildbot | build #1106 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/1106 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 11:46 |
-!- lambday [3ba21749@gateway/web/freenode/ip.59.162.23.73] has quit [Ping timeout: 245 seconds] | 11:47 | |
shogun-buildbot | build #1107 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/1107 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 11:59 |
@HeikoS | ah I commited data seperately sorry | 11:59 |
-!- lambday [3ba21749@gateway/web/freenode/ip.59.162.23.73] has joined #shogun | 12:01 | |
sonne|work | HeikoS: when did you commit data? | 12:02 |
@HeikoS | sonne|work: two commits for example change and data in main repo | 12:03 |
@HeikoS | the data repo I pushed before | 12:03 |
sonne|work | HeikoS: I mean did you push the data repository first | 12:03 |
-!- goldbug [~quassel@global-2-11.nat.csx.cam.ac.uk] has joined #shogun | 12:03 | |
@HeikoS | yes, but the other thing was wring | 12:03 |
sonne|work | and then git add data / commit / push to shogun develop? | 12:03 |
@HeikoS | wrong | 12:03 |
sonne|work | HeikoS: so we have 2 submits instead of 1 now | 12:03 |
sonne|work | ok | 12:03 |
@HeikoS | yep | 12:03 |
sonne|work | hmmhh buildbot seems unhappy | 12:05 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 643467d / src/NEWS: https://github.com/shogun-toolbox/shogun/commit/643467d0a53da83fd700ecad836870d1af8af4e0 | 12:06 |
shogun-notifier- | shogun: update news | 12:06 |
-!- nube [~rho@116.90.239.3] has quit [Quit: Leaving.] | 12:09 | |
-!- van51 [~van51@195.251.235.94] has quit [Ping timeout: 252 seconds] | 12:17 | |
shogun-buildbot | build #813 of cyg1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/813 blamelist: Soeren Sonnenburg <sonne@debian.org> | 12:17 |
-!- iglesiasg [c1934d18@gateway/web/freenode/ip.193.147.77.24] has quit [Ping timeout: 245 seconds] | 12:18 | |
-!- iglesiasg [c1934d18@gateway/web/freenode/ip.193.147.77.24] has joined #shogun | 12:20 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 12:21 | |
-!- van51 [~van51@195.251.235.71] has joined #shogun | 12:34 | |
-!- goldbug [~quassel@global-2-11.nat.csx.cam.ac.uk] has quit [Remote host closed the connection] | 12:35 | |
shogun-buildbot | build #1108 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/1108 blamelist: Soeren Sonnenburg <sonne@debian.org> | 12:44 |
@lisitsyn | sonne|work: I think we just too weak to handle that bug | 13:05 |
@lisitsyn | :D | 13:05 |
@lisitsyn | ws are different.. | 13:06 |
sonne|work | lisitsyn: yeah but just w's | 13:08 |
@lisitsyn | sonne|work: isn't that enough? | 13:08 |
sonne|work | to fail anything is | 13:09 |
sonne|work | it could be different input etc | 13:09 |
sonne|work | lisitsyn: oh enums in python 3.4 http://eli.thegreenplace.net/2013/05/10/python-will-have-enums-in-3-4/ | 13:10 |
-!- lambday [3ba21749@gateway/web/freenode/ip.59.162.23.73] has quit [Ping timeout: 245 seconds] | 13:12 | |
@lisitsyn | sonne|work: Y | 13:12 |
@lisitsyn | sonne|work: looks totally redundant | 13:13 |
@lisitsyn | sonne|work: but that's trendy to put every known feature to PL | 13:14 |
sonne|work | lisitsyn: yeah certainly not earth shaking to have that | 13:16 |
@lisitsyn | sonne|work: can be easily emulated, makes no sense for me at all | 13:16 |
sonne|work | maybe they are immutable | 13:17 |
@lisitsyn | sonne|work: still not a great deal.. | 13:18 |
-!- iglesiasg [c1934d18@gateway/web/freenode/ip.193.147.77.24] has quit [Ping timeout: 245 seconds] | 13:21 | |
-!- Netsplit *.net <-> *.split quits: kakashi_, van51, @HeikoS | 13:21 | |
-!- Netsplit over, joins: van51, @HeikoS, kakashi_ | 13:25 | |
-!- iglesiasg [c1934d18@gateway/web/freenode/ip.193.147.77.24] has joined #shogun | 13:25 | |
-!- iglesiasg [c1934d18@gateway/web/freenode/ip.193.147.77.24] has quit [Client Quit] | 13:25 | |
-!- iglesiasg [c1934d18@gateway/web/freenode/ip.193.147.77.24] has joined #shogun | 13:38 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 13:41 | |
-!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has joined #shogun | 13:47 | |
-!- iglesiasg [c1934d18@gateway/web/freenode/ip.193.147.77.24] has quit [Ping timeout: 245 seconds] | 13:52 | |
sonne|work | van51: hey! how is you dyn*array patch going? | 13:57 |
sonne|work | complete massacre by now? | 13:57 |
van51 | sonne|work: hey! I have some tests failing and saying that "Number of features/kernels does not match - bailing out" and I am trying to figure out why | 14:14 |
van51 | maybe I should submit a PR and you can review it a bit? | 14:15 |
sonne|work | van51: keep trying, but if you are stuck for too long | 14:15 |
sonne|work | yes exactly | 14:15 |
sonne|work | we don't want you to spend your rest of your life with such 'fun' stuff | 14:15 |
-!- sijin_ [~smuxi@144.214.222.109] has joined #shogun | 14:15 | |
van51 | sonne|work: yea I am going to keep looking it, it's just that I didn't have that much time last week to work on it and I kind of postponed it | 14:16 |
van51 | hehe | 14:16 |
-!- sijin [~smuxi@144.214.222.109] has quit [Ping timeout: 264 seconds] | 14:16 | |
sonne|work | van51: mark the PR somehow as WIP | 14:16 |
van51 | sonne|work: ok! I'll probably do it tonight or first thing tomorrow morning if I don't manage to do it | 14:18 |
sonne|work | van51: ok cool | 14:18 |
-!- travis-ci [~travis-ci@ec2-184-72-169-82.compute-1.amazonaws.com] has joined #shogun | 14:29 | |
travis-ci | [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/7116775 | 14:29 |
-!- travis-ci [~travis-ci@ec2-184-72-169-82.compute-1.amazonaws.com] has left #shogun [] | 14:29 | |
-!- sijin_ [~smuxi@144.214.222.109] has quit [] | 14:30 | |
-!- sijin [~smuxi@144.214.222.109] has joined #shogun | 14:31 | |
-!- nube [~rho@49.244.108.149] has joined #shogun | 14:49 | |
-!- iglesiasg [c1934d1a@gateway/web/freenode/ip.193.147.77.26] has joined #shogun | 14:52 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 14:52 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 15:06 | |
@iglesiasg | wiking: I think make under tests/unit crashes | 15:13 |
@iglesiasg | no shogun::CMath::rand_state there | 15:14 |
@lisitsyn | iglesiasg: clean should help there | 15:14 |
@wiking | iglesiasg: hey | 15:16 |
@wiking | lisitsyn: yeah try a make clean && make | 15:17 |
@wiking | iglesiasg: ^ | 15:17 |
@wiking | ;) | 15:17 |
@wiking | iglesiasg: tell me if it still crashes | 15:17 |
@iglesiasg | wiking: running atm, thanks! | 15:17 |
@iglesiasg | lisitsyn: thank you too :) | 15:17 |
-!- travis-ci [~travis-ci@ec2-184-72-169-82.compute-1.amazonaws.com] has joined #shogun | 15:17 | |
travis-ci | [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/7118142 | 15:17 |
-!- travis-ci [~travis-ci@ec2-184-72-169-82.compute-1.amazonaws.com] has left #shogun [] | 15:17 | |
@wiking | mmm travis is still dying :( | 15:17 |
@wiking | mmm interesting.... only 1 of that is crashing on buildbot | 15:19 |
van51 | I thought that SG_UNREF(k) would set k to null after it's done | 15:23 |
van51 | and with the list way after the get_next_element() on the last element, k was set to null, and so after that there were tests of the form k==null | 15:25 |
@lisitsyn | van51: in general no | 15:25 |
van51 | but with the new array way, k was never set to null after the last element | 15:25 |
van51 | and so the next conditions failed | 15:25 |
van51 | I think I can fix it now :) | 15:25 |
-!- Netsplit *.net <-> *.split quits: votjakovr | 15:28 | |
van51 | another question, when I am issuing make tests on the src folder I am getting errors of the form "Error opening file '../data/fm_train_real.dat'" | 15:32 |
van51 | is that expected or not? :p | 15:33 |
@iglesiasg | van51: try fetching the data files | 15:33 |
@iglesiasg | git submodule init | 15:33 |
@iglesiasg | git submodule update | 15:33 |
@iglesiasg | in the root of your repo | 15:33 |
van51 | iglesiasg: it worked! thanks :) | 15:35 |
@iglesiasg | van51: nice! you are welcome :) | 15:35 |
-!- Netsplit *.net <-> *.split quits: kakashi_, sijin, @iglesiasg, @HeikoS, van51 | 15:51 | |
-!- lisitsyn [~lisitsin@mxs.kg.ru] has quit [Quit: Leaving.] | 15:51 | |
-!- Netsplit over, joins: @iglesiasg, sijin, van51, @HeikoS, kakashi_ | 15:55 | |
-!- kakashi_ [~kakashi_@nltk/kakashi] has quit [Remote host closed the connection] | 15:56 | |
-!- kakashi_ [~kakashi_@nltk/kakashi] has joined #shogun | 15:56 | |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 16:01 | |
shogun-notifier- | shogun: Heiko Strathmann :develop * 587ccb0 / src/shogun/lib/DataType.cpp,src/shogun/lib/DataType.h: https://github.com/shogun-toolbox/shogun/commit/587ccb031745f880ebdfc8a189a51713e8adf861 | 16:01 |
shogun-notifier- | shogun: added a proper equals method | 16:01 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 1302f34 / src/shogun/base/Parameter.cpp,src/shogun/base/Parameter.h: https://github.com/shogun-toolbox/shogun/commit/1302f3499f9469ebb2baa498ecfcc7c8a8de2d77 | 16:01 |
shogun-notifier- | shogun: added some first methods for equal methods on CSGObjects | 16:01 |
shogun-notifier- | shogun: Heiko Strathmann :develop * ff24461 / src/shogun/base/SGObject.cpp,src/shogun/base/SGObject.h: https://github.com/shogun-toolbox/shogun/commit/ff244613212122d54b3de609c7c4d95a45dd026e | 16:01 |
shogun-notifier- | shogun: first implementation of equals method (not completely working yet) | 16:01 |
shogun-notifier- | shogun: Heiko Strathmann :develop * c70d92d / tests/unit/base/Parameter_unittest.cc,tests/unit/base/SGObject_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/c70d92d8c746be567069a1a315c27c2ef59dc8fe | 16:01 |
shogun-notifier- | shogun: first set of unit tests for equals framework | 16:01 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 939cbb4 / src/NEWS: https://github.com/shogun-toolbox/shogun/commit/939cbb4104dd58d379cb4d1c36af9e8dbfe894dc | 16:01 |
shogun-notifier- | shogun: mention equals framework in new features | 16:01 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 44158f4 / src/shogun/base/Parameter.cpp,src/shogun/base/Parameter.h: https://github.com/shogun-toolbox/shogun/commit/44158f434f93db7bb095da008b3b04bb8ac42632 | 16:01 |
shogun-notifier- | shogun: added proper handling of string based types | 16:01 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 8a24dd0 / tests/unit/base/Parameter_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/8a24dd0be0424204accfd27eb8b27afbe59002db | 16:01 |
shogun-notifier- | shogun: new unit test for equal and string based types | 16:01 |
-!- sonne|work [~sonnenbu@sams-office-nat.tomtomgroup.com] has quit [Excess Flood] | 16:05 | |
-!- sonne|work [~sonnenbu@sams-office-nat.tomtomgroup.com] has joined #shogun | 16:06 | |
-!- sijin [~smuxi@144.214.222.109] has quit [Excess Flood] | 16:09 | |
shogun-buildbot | build #814 of cyg1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/814 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 16:09 |
shogun-buildbot | build #524 of rpm1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/524 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 16:11 |
-!- sijin [~smuxi@144.214.222.109] has joined #shogun | 16:11 | |
-!- sijin [~smuxi@144.214.222.109] has quit [Remote host closed the connection] | 16:13 | |
shogun-buildbot | build #978 of bsd1 - libshogun is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/978 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 16:14 |
shogun-buildbot | build #525 of rpm1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/525 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 16:14 |
shogun-buildbot | build #526 of rpm1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/526 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 16:16 |
shogun-buildbot | build #979 of bsd1 - libshogun is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/979 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 16:19 |
-!- van51 [~van51@195.251.235.71] has quit [Ping timeout: 255 seconds] | 16:19 | |
shogun-buildbot | build #527 of rpm1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/rpm1%20-%20libshogun/builds/527 | 16:21 |
shogun-buildbot | build #970 of deb2 - static_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/970 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 16:22 |
shogun-buildbot | build #980 of bsd1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/980 | 16:26 |
shogun-buildbot | build #815 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/815 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 16:27 |
shogun-buildbot | build #971 of deb2 - static_interfaces is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/971 | 16:34 |
shogun-buildbot | build #816 of cyg1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/816 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 16:43 |
shogun-buildbot | build #1109 of deb3 - modular_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1109 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 16:54 |
-!- nube [~rho@49.244.108.149] has quit [Ping timeout: 252 seconds] | 17:03 | |
-!- travis-ci [~travis-ci@ec2-54-234-90-246.compute-1.amazonaws.com] has joined #shogun | 17:08 | |
travis-ci | [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/7123641 | 17:08 |
-!- travis-ci [~travis-ci@ec2-54-234-90-246.compute-1.amazonaws.com] has left #shogun [] | 17:08 | |
-!- HeikoS [~heiko@nat-173-117.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.] | 17:11 | |
shogun-buildbot | build #1110 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/1110 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 17:13 |
-!- iglesiasg [c1934d1a@gateway/web/freenode/ip.193.147.77.26] has quit [Quit: Page closed] | 17:17 | |
-!- pickle27 [~kevin@rcv3-lab-pc.ee.queensu.ca] has joined #shogun | 17:18 | |
pickle27 | hello everyone | 17:19 |
@sonney2k | pickle27, good morning :) | 17:21 |
@sonney2k | HeikoS - needless to say your test works just fine here | 17:23 |
pickle27 | sonney2k: saw your comments on the PR I'll still be doing some work on the eigen3 version | 17:28 |
pickle27 | sonney2k: I'm kind of attached to the lapack one now lol | 17:28 |
-!- lisitsyn [~blackburn@37.61.179.179] has joined #shogun | 17:29 | |
-!- mode/#shogun [+o lisitsyn] by ChanServ | 17:29 | |
shogun-notifier- | shogun: iglesias :develop * 01205a7 / src/shogun/lib/tapkee/ (44 files): https://github.com/shogun-toolbox/shogun/commit/01205a7527236ea5f5e95cccf167650c038b80a7 | 17:33 |
shogun-notifier- | shogun: Tapkee update | 17:33 |
shogun-notifier- | shogun: Sergey Lisitsyn :develop * e8d0891 / src/shogun/lib/tapkee/ (44 files): https://github.com/shogun-toolbox/shogun/commit/e8d0891b31aebe3a8d08499ee34f7763f1c60262 | 17:33 |
shogun-notifier- | shogun: Merge pull request #1090 from iglesias/feature/tapkee_update | 17:33 |
shogun-notifier- | shogun: | 17:33 |
shogun-notifier- | shogun: Tapkee update | 17:33 |
pickle27 | lisitsyn: I'm about to read your paper - whats the best way to mark up a pdf on linux? anyone have any ideas? | 17:33 |
@lisitsyn | pickle27: evince? | 17:33 |
pickle27 | lisitsyn: yeah thats probably what I was going to use | 17:34 |
pickle27 | lisitsyn: thought I'd see if anyone had any suggestions | 17:34 |
@lisitsyn | pickle27: from the left | 17:34 |
@lisitsyn | there is 'thumbnails' thing | 17:34 |
@lisitsyn | select annotations here | 17:34 |
@lisitsyn | then you can put some annotations all around the pdf | 17:35 |
pickle27 | lisitsyn: okay thanks | 17:36 |
@sonney2k | pickle27, yeah that will be interesting to see how much eigen3 is slower | 17:39 |
@lisitsyn | eigen3 is slower? | 17:40 |
@lisitsyn | where on what? | 17:40 |
shogun-buildbot | build #817 of cyg1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/817 blamelist: iglesias <fernando.iglesiasg@gmail.com> | 17:40 |
pickle27 | lisitsyn: multiclass LDA | 17:41 |
@lisitsyn | pickle27: comparing to blas? | 17:41 |
pickle27 | lisitsyn: we haven't tested yet but our belief is that it will be slower | 17:41 |
pickle27 | lisitsyn: yeah | 17:41 |
@lisitsyn | belief?? | 17:41 |
pickle27 | lisitsyn: the svd in particular | 17:41 |
pickle27 | lisitsyn: theory | 17:41 |
@lisitsyn | pickle27: what theory? different methods? | 17:43 |
pickle27 | lisitsyn: the theory is that blas will be faster than eigen3 | 17:43 |
@lisitsyn | pickle27: funny :) | 17:43 |
@lisitsyn | any reason? | 17:43 |
pickle27 | lisitsyn: we'll see it'll be interesting | 17:43 |
@lisitsyn | pickle27: it depends what blas impl you use | 17:44 |
pickle27 | lisitsyn: yeah | 17:45 |
@lisitsyn | pickle27: http://eigen.tuxfamily.org/index.php?title=Benchmark and there are benchmarks | 17:45 |
pickle27 | lisitsyn: interesting | 17:47 |
pickle27 | lisitsyn: I'm still planning to test both my mclda versions for kicks | 17:47 |
@lisitsyn | pickle27: yeah I just mean the benchmark might not be really fair | 17:48 |
pickle27 | lisitsyn: very true | 17:48 |
@lisitsyn | pickle27: you are heavily using eigen map | 17:49 |
@lisitsyn | it may cause some limitations | 17:49 |
pickle27 | lisitsyn: yeah - also thats my only real option right? | 17:49 |
@lisitsyn | I am not very sure but alignment matters | 17:49 |
@lisitsyn | pickle27: I didn't check very thorougly | 17:50 |
pickle27 | lisitsyn: okay | 17:50 |
@lisitsyn | but it may cause something bad | 17:50 |
@lisitsyn | performance wise | 17:50 |
@lisitsyn | pickle27: but better check first :) | 17:51 |
pickle27 | lisitsyn: but if I want to return SGMatrix then eigen map is my only choice right | 17:51 |
@lisitsyn | yes sure | 17:51 |
@lisitsyn | okay be back a bit later | 17:52 |
-!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has joined #shogun | 17:58 | |
-!- travis-ci [~travis-ci@ec2-184-72-169-82.compute-1.amazonaws.com] has joined #shogun | 17:58 | |
travis-ci | [travis-ci] it's Sergey Lisitsyn'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/7126593 | 17:58 |
-!- travis-ci [~travis-ci@ec2-184-72-169-82.compute-1.amazonaws.com] has left #shogun [] | 17:58 | |
shogun-buildbot | build #818 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/818 | 17:59 |
shogun-buildbot | build #1111 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/1111 blamelist: iglesias <fernando.iglesiasg@gmail.com> | 18:27 |
shogun-buildbot | build #1112 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/1112 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 18:38 |
-!- HeikoS [~heiko@nat-173-117.internal.eduroam.ucl.ac.uk] has joined #shogun | 19:12 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 19:12 | |
-!- Netsplit *.net <-> *.split quits: @lisitsyn, sonne|work | 19:18 | |
-!- iglesiasg [d58f323c@gateway/web/freenode/ip.213.143.50.60] has joined #shogun | 19:19 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 19:19 | |
-!- Netsplit over, joins: sonne|work | 19:23 | |
-!- lisitsyn [~blackburn@37.61.179.179] has joined #shogun | 19:27 | |
-!- kevin_ [~kevin@rcv3-lab-pc.ee.queensu.ca] has joined #shogun | 19:33 | |
-!- sonne|work1 [~sonnenbu@sams-office-nat.tomtomgroup.com] has joined #shogun | 19:38 | |
-!- Netsplit *.net <-> *.split quits: pickle27, votjakovr, @iglesiasg, sonne|work | 19:43 | |
-!- Netsplit over, joins: votjakovr | 19:47 | |
-!- kevin_ is now known as pickle27 | 19:57 | |
@sonney2k | HeikoS, ping ping? | 19:57 |
@HeikoS | sonney2k: hi | 19:58 |
@sonney2k | HeikoS, I was trying to convert the tests to use your equals but all fail | 20:01 |
@HeikoS | okay | 20:01 |
@HeikoS | so do you have en example? | 20:02 |
@HeikoS | probably still contains some bugs that I need to take care of | 20:02 |
@HeikoS | but I tested it on a few more complex examples and always worked | 20:02 |
@HeikoS | sonney2k: so would be good to reproduce the example in c++ | 20:04 |
@sonney2k | ./tester.py kernel_gaussian_modular.py | 20:07 |
@sonney2k | [DEBUG] entering DynamicObjectArray::equals() | 20:08 |
@sonney2k | [DEBUG] comparing parameter 0 | 20:08 |
@sonney2k | [DEBUG] comparing parameter "array" to other's "array" | 20:08 |
@sonney2k | [DEBUG] comparing parameter 1 | 20:08 |
@sonney2k | [DEBUG] comparing parameter "num_elements" to other's "num_elements" | 20:08 |
@sonney2k | [DEBUG] leaving DynamicObjectArray::equals(): parameters at position 1 with name "num_elements" differs from other object parameter with name "num_elements" | 20:08 |
@HeikoS | sonney2k | 20:09 |
@HeikoS | so first problem: | 20:09 |
@sonney2k | HeikoS, can one somehow print the values? | 20:09 |
@HeikoS | there are many more debug messages | 20:09 |
@HeikoS | which print the values | 20:09 |
@HeikoS | however, they are not printed for some reason | 20:09 |
@HeikoS | I dont know why | 20:09 |
@sonney2k | how to enable them? | 20:09 |
@HeikoS | sonney2k: its a bug | 20:10 |
@sonney2k | HeikoS, well give me an example | 20:10 |
@HeikoS | remember I told you a couple of times that some debug messages dont appear | 20:10 |
@sonney2k | never heard of any issues | 20:10 |
@sonney2k | no | 20:10 |
@sonney2k | I fixed that last time | 20:10 |
@HeikoS | so this particular example of yours enters the TParameter::compare_ptype method | 20:11 |
@HeikoS | and also the TParmaeter::compare_stype | 20:11 |
-!- van511 [~van51@athedsl-320452.home.otenet.gr] has joined #shogun | 20:11 | |
@HeikoS | both print debug messages | 20:11 |
@HeikoS | but I could only see them when I made them PRINTS | 20:11 |
@HeikoS | sonney2k: oh just realised a bug | 20:13 |
@HeikoS | let me fix and push | 20:13 |
@sonney2k | HeikoS, just found the issue with SG_SDEBUG not being printed | 20:14 |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * f560645 / src/shogun/io/SGIO.h: https://github.com/shogun-toolbox/shogun/commit/f56064580242e029c108dc2ee3455b2162e1eddd | 20:16 |
shogun-notifier- | shogun: fix typo causing SG_SDEBUG not being printed with loglevel MSG_DEBUG | 20:16 |
@HeikoS | sonney2k: what caused it? | 20:18 |
@HeikoS | ah saw it | 20:18 |
@HeikoS | sonney2k: whats the new output? | 20:21 |
@sonney2k | HeikoS, compiling | 20:22 |
@HeikoS | sonney2k: I just realised I forgot to check some NULL cases | 20:22 |
@HeikoS | so need to add some parts | 20:22 |
@sonney2k | that might cause it | 20:22 |
@sonney2k | both preprocessor chains should be NULL | 20:23 |
shogun-buildbot | build #819 of cyg1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/819 blamelist: Soeren Sonnenburg <sonne@debian.org> | 20:24 |
@sonney2k | HeikoS, http://dpaste.com/1152486/ | 20:26 |
@sonney2k | now with output | 20:27 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 8dcb486 / src/shogun/base/Parameter.cpp: https://github.com/shogun-toolbox/shogun/commit/8dcb486cf14f1ae40cabe871208689ff9860e4c7 | 20:28 |
shogun-notifier- | shogun: added some NULL checks | 20:28 |
shogun-notifier- | shogun: Heiko Strathmann :develop * c94db34 / src/shogun/base/Parameter.cpp: https://github.com/shogun-toolbox/shogun/commit/c94db34d6906757b667398b2636cd417c29a183d | 20:28 |
shogun-notifier- | shogun: fixed copy paste errors | 20:28 |
shogun-notifier- | shogun: Heiko Strathmann :develop * ae08525 / src/shogun/base/Parameter.cpp: https://github.com/shogun-toolbox/shogun/commit/ae0852561c12bfaeef01a3440fd8be9bbe039fb8 | 20:28 |
shogun-notifier- | shogun: added some NULL checks | 20:28 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 43c9d2c / src/shogun/base/Parameter.cpp: https://github.com/shogun-toolbox/shogun/commit/43c9d2cf30bb071c3e67cddd497920c6f0b642b3 | 20:28 |
shogun-notifier- | shogun: added some NULL checks | 20:28 |
shogun-notifier- | shogun: Heiko Strathmann :develop * a4a7d8a / src/shogun/base/Parameter.cpp: https://github.com/shogun-toolbox/shogun/commit/a4a7d8aea0319fa0b4309bc4a94bfabbf640eed9 | 20:28 |
shogun-notifier- | shogun: fixed copy paste errors | 20:28 |
shogun-notifier- | shogun: Heiko Strathmann :develop * bf0ba16 / src/shogun/base/Parameter.cpp: https://github.com/shogun-toolbox/shogun/commit/bf0ba165e2fac2e1664d3ac2f2f19084e07ea519 | 20:28 |
shogun-notifier- | shogun: Merge branch 'sgobject_equals' into develop | 20:28 |
shogun-notifier- | shogun: Heiko Strathmann :develop * d5030fb / src/shogun/base/Parameter.cpp: https://github.com/shogun-toolbox/shogun/commit/d5030fb9729b8fb249d7d3a011252e7967365554 | 20:28 |
shogun-notifier- | shogun: Merge pull request #1092 from karlnapf/develop | 20:28 |
shogun-notifier- | shogun: | 20:28 |
shogun-notifier- | shogun: some NULL checks for equals method | 20:28 |
@HeikoS | sonney2k: ok, this seems to be something else | 20:29 |
@sonney2k | HeikoS, ohh maybe this could even be different... | 20:29 |
@HeikoS | I just fixed the NULL things | 20:29 |
@sonney2k | subtle case! | 20:29 |
@HeikoS | this seems to have to do with DynamicObject array | 20:29 |
@HeikoS | but i had this error before actually | 20:29 |
@HeikoS | forgot what caused it | 20:29 |
@sonney2k | I have save_serializable_pre | 20:30 |
@sonney2k | etc functions in there | 20:30 |
@sonney2k | to not store more than necessary | 20:30 |
@sonney2k | The dynamic object array is growing with chunks of size 128 IIRC | 20:31 |
@HeikoS | yes | 20:31 |
@HeikoS | this might not be saved | 20:31 |
@sonney2k | so I am trying to avoid saving 128 NULL objects | 20:31 |
@HeikoS | I see | 20:31 |
@sonney2k | if the array in fact has 0 size | 20:31 |
@sonney2k | so what now? | 20:32 |
@HeikoS | ignore-list? | 20:32 |
@sonney2k | call save_serializable_pre? | 20:32 |
@sonney2k | IDK | 20:33 |
@HeikoS | dont know | 20:33 |
@HeikoS | I gotta go now anyway, I will think about it | 20:33 |
@HeikoS | these are marginal cases though | 20:33 |
@HeikoS | the general equals works | 20:33 |
@sonney2k | I think that is the only case... | 20:33 |
@HeikoS | very nice to have the parameter framework | 20:33 |
@HeikoS | btw | 20:33 |
@HeikoS | makes things easy | 20:33 |
@HeikoS | the offset in the sparse vectors costed me some nerves though, we should document that properly | 20:34 |
@HeikoS | anyway, gotta go, have a nice evening! | 20:34 |
@sonney2k | yeah only dyn* arrays have this | 20:34 |
@sonney2k | HeikoS, cu! | 20:35 |
van511 | sonney2k: I 'm almost done, what remains(hopefully!) is to fix the usage of the class in an example | 20:38 |
van511 | so I 've made a PR so you can start reviewing it if you have some time | 20:38 |
shogun-buildbot | build #820 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/820 | 20:47 |
-!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has quit [Quit: Leaving] | 20:49 | |
@sonney2k | van511, sry - which example do you mean? | 20:59 |
@sonney2k | van511, a C++ one or which $LANG? | 20:59 |
van511 | a C++ one | 21:00 |
van511 | tbh I didn't check in the other languages | 21:00 |
-!- travis-ci [~travis-ci@ec2-50-16-64-171.compute-1.amazonaws.com] has joined #shogun | 21:01 | |
travis-ci | [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/7131757 | 21:01 |
-!- travis-ci [~travis-ci@ec2-50-16-64-171.compute-1.amazonaws.com] has left #shogun [] | 21:01 | |
-!- van511 [~van51@athedsl-320452.home.otenet.gr] has quit [Quit: Leaving.] | 21:01 | |
-!- van51 [~van51@athedsl-320452.home.otenet.gr] has joined #shogun | 21:02 | |
@sonney2k | van51, it should be pretty transparent from the other languages but you should see it in your PR | 21:03 |
@sonney2k | van51, travis has no mercy... | 21:03 |
shogun-buildbot | build #1113 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/1113 blamelist: Soeren Sonnenburg <sonne@debian.org> | 21:03 |
@sonney2k | van51, https://travis-ci.org/shogun-toolbox/shogun/builds/7132002 | 21:03 |
van51 | sonney2k: haha | 21:04 |
van51 | yea I am monitoring it | 21:04 |
van51 | sonney2k: btw, you also wanted me to split the combine_kernels() method into smaller pieces, right? | 21:05 |
@sonney2k | van51, yeah | 21:06 |
van51 | sonney2k: will do that now then | 21:06 |
@sonney2k | van51, but you may only introduce one bug at a time :D | 21:06 |
van51 | sonney2k: please, it's going to be bug-free :p | 21:07 |
@sonney2k | van51, I too write only bugfree code | 21:07 |
@sonney2k | but you know the wire used to transport my bugfree code is surrounded by swarms of bugs | 21:08 |
van51 | aargh, I have the same problem! | 21:09 |
@sonney2k | van51, if you know a good debugger let me know... I pay good prize! | 21:10 |
-!- HeikoS1 [~heiko@nat-173-117.internal.eduroam.ucl.ac.uk] has joined #shogun | 21:11 | |
van51 | sonney2k: I'll be on the lookout then ;) | 21:11 |
@sonney2k | HeikoS, hmmh I just ignored this variable but now I have other failures | 21:14 |
-!- HeikoS [~heiko@nat-173-117.internal.eduroam.ucl.ac.uk] has quit [Ping timeout: 268 seconds] | 21:15 | |
@sonney2k | lisitsyn, have you ever thought about dropping swig altogether? | 21:22 |
shogun-buildbot | build #1114 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/1114 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 21:23 |
@sonney2k | van51, this is a monster patch | 21:24 |
@sonney2k | van51, it is basically impossible to test if everything is OK | 21:25 |
van51 | sonney2k: what can we do then? | 21:25 |
@sonney2k | van51, all we can do is unit tests | 21:25 |
@sonney2k | van51, it will also unfortunately break all our combinedkernel/features related integration tests | 21:27 |
@sonney2k | really stupid | 21:27 |
@sonney2k | I don't know a way around this | 21:27 |
@sonney2k | let me check if our old tests covered combined* | 21:27 |
van51 | sonney2k: well, what needs to be done | 21:27 |
@sonney2k | yes we have one test for combined | 21:28 |
@sonney2k | van51, is that a Q? | 21:29 |
van51 | sonney2k: no I meant that what needs to be done will be done, even if it's tedious | 21:29 |
van51 | but now that you mention it, it is also a question :p | 21:30 |
van51 | because I'm going to need some guidance for sure | 21:31 |
@sonney2k | van51, the big problem I see is that this patch might introduce a bug in MKL & svmlight etc | 21:31 |
@sonney2k | but I have no idea how to easily test this | 21:31 |
@sonney2k | ohh wait maybe I have | 21:31 |
van51 | :) | 21:31 |
@sonney2k | lets take one very complex example | 21:31 |
@sonney2k | then save the output it produces | 21:31 |
@sonney2k | and compare again | 21:31 |
van51 | aah | 21:32 |
van51 | sounds good | 21:32 |
@sonney2k | if that matches all good | 21:32 |
@sonney2k | then run this with valgrind again | 21:32 |
@sonney2k | no memory leaks / uninited memory -> ready to merge | 21:32 |
@sonney2k | let me see what would work | 21:32 |
@sonney2k | van51, which $LANG can you compile on your machine? | 21:33 |
@sonney2k | van51, if you could install octave | 21:33 |
van51 | sonney2k: I can do octave, I have it already | 21:34 |
@sonney2k | there are 2 examples mkl_classify_christmas_star.m mkl_regression_3sine_waves.m | 21:34 |
@sonney2k | we should modify them a bit to store the actuall classifier outputs | 21:34 |
@sonney2k | but then if they still work fine we are good for sure | 21:34 |
@sonney2k | they heavily use the combinedfeatures/kernels | 21:34 |
@sonney2k | van51, actually just run mkl_classify_christmas_star | 21:35 |
@sonney2k | it will give you a varilable 'result' | 21:35 |
@sonney2k | result.trainout and result.testout | 21:36 |
@sonney2k | these are matrices with all outputs for various experiments | 21:36 |
@sonney2k | if these match with your changes we are fine | 21:36 |
van51 | sonney2k: ok | 21:36 |
van51 | sonney2k: first I have to compile the octave interface right? | 21:37 |
@sonney2k | with a one-liner we can add the same to the regression example | 21:37 |
@sonney2k | van51, just install liboctave-dev | 21:37 |
@sonney2k | then do ./configure --interfaces=octave_static --disable-optimization | 21:37 |
@sonney2k | (the latter only to save some compile time) | 21:38 |
van51 | sonney2k: Ok, I am compiling | 21:39 |
van51 | sonney2k: but on my edited code.. | 21:39 |
@sonney2k | van51, well you can just save the output in octave with | 21:41 |
@sonney2k | save somefile results | 21:42 |
@sonney2k | but if it even works it is good | 21:42 |
@sonney2k | I will give some minor feedback in the meantime | 21:43 |
van51 | sonney2k: cool | 21:43 |
van51 | I'll let you know | 21:43 |
@sonney2k | van51, what strikes me is that we have no more thread safety issues traversing the array... all just index lookups | 21:48 |
@sonney2k | very useful | 21:48 |
van51 | sonney2k: yea, that was my intention all along :p | 21:50 |
van51 | sonney2k: btw does that example take a long time or could it be stuck or something? | 21:52 |
@sonney2k | van51, if you did compile with --disable-optimization it might take a while | 21:54 |
@sonney2k | but it should be rather quick | 21:55 |
@sonney2k | let me compile the interface | 21:55 |
van51 | ok it finished | 21:56 |
van51 | sonney2k: it showed two warnings | 21:57 |
@sonney2k | it took maybe 10 seconds here | 21:59 |
@sonney2k | van51, please do | 22:00 |
@sonney2k | pcolor(w) | 22:00 |
@sonney2k | colorbar | 22:01 |
van51 | it says can't plot with an empty cb range | 22:01 |
van51 | and can't plot with an empty y range | 22:02 |
@sonney2k | van51, then please | 22:02 |
@sonney2k | type w | 22:02 |
@sonney2k | w | 22:02 |
@sonney2k | and copy the output somewhere | 22:02 |
van51 | everything is 0.2 | 22:02 |
@sonney2k | dpaste.com or so | 22:02 |
@sonney2k | errm | 22:02 |
@sonney2k | not good | 22:02 |
van51 | thought so | 22:02 |
@sonney2k | van51, ok let me compile your version | 22:12 |
@sonney2k | maybe you have some other libs installed or so | 22:12 |
@sonney2k | van51, if that doesn't help - you should write a unit test that checks the combined kernel | 22:13 |
van51 | sonney2k: Ok, I will | 22:14 |
lisitsyn | sonney2k: dropping swig? | 22:14 |
van51 | sonney2k: I'm sorry but I have to go for a while now | 22:14 |
@sonney2k | van51, wrt your patch - it is syntax wise all fine now | 22:14 |
van51 | wrt? | 22:14 |
@sonney2k | there is no obvious error in that and since it is 1000 loc or so I certainly won't spot the 1 potential line with a mistake | 22:15 |
@sonney2k | w.r.t. | 22:15 |
van51 | w.r.t.? :p | 22:15 |
lisitsyn | with respect to | 22:15 |
@sonney2k | lisitsyn, well think of we could generate interfaces | 22:15 |
van51 | aah, thanks | 22:15 |
@sonney2k | and use the paramter framework to pass around stuff | 22:15 |
@sonney2k | I mean *lean* interfaces | 22:16 |
lisitsyn | sonney2k: I don't really get it | 22:16 |
lisitsyn | sonney2k: something more stupid | 22:16 |
lisitsyn | for me | 22:16 |
lisitsyn | :) | 22:16 |
lisitsyn | sonney2k: so do not use swig at all? | 22:17 |
@sonney2k | van51, ok it takes ages | 22:18 |
@sonney2k | so sth is wrong | 22:18 |
van51 | sonney2k: ok | 22:18 |
van51 | i'll be back in 2 or 3 hours | 22:18 |
van51 | I'll start looking at it again then | 22:18 |
@sonney2k | then I am hopefully asleep | 22:18 |
@sonney2k | drug time | 22:19 |
van51 | ok, i'll bug you again tomorrow then :p | 22:20 |
@sonney2k | van51, but please try to get closer to the root cause | 22:20 |
van51 | sonney2k: I will :) | 22:21 |
van51 | sorry but I g2g for now | 22:21 |
@sonney2k | van51, cu | 22:21 |
van51 | bye | 22:21 |
@sonney2k | and thanks! | 22:21 |
-!- van51 [~van51@athedsl-320452.home.otenet.gr] has left #shogun ["PING 1368476562"] | 22:22 | |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * 2d30a5b / src/shogun/ (5 files): https://github.com/shogun-toolbox/shogun/commit/2d30a5b1ed6f6fa939cc7dbbde717f7a10e42cc1 | 22:23 |
shogun-notifier- | shogun: CSGObject::equals method | 22:23 |
shogun-notifier- | shogun: | 22:23 |
shogun-notifier- | shogun: -hardcoded ignore of 'num_elements' | 22:23 |
shogun-notifier- | shogun: -use ``only'' double in precision (floatmax_t is not supported by swig) | 22:23 |
shogun-buildbot | build #1174 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1174 blamelist: Soeren Sonnenburg <sonne@debian.org> | 22:33 |
-!- travis-ci [~travis-ci@ec2-184-72-169-82.compute-1.amazonaws.com] has joined #shogun | 22:47 | |
travis-ci | [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/7135409 | 22:47 |
-!- travis-ci [~travis-ci@ec2-184-72-169-82.compute-1.amazonaws.com] has left #shogun [] | 22:47 | |
-!- nube [~rho@49.244.108.149] has joined #shogun | 22:59 | |
-!- lisitsyn [~blackburn@37.61.179.179] has quit [Ping timeout: 245 seconds] | 23:29 | |
--- Log closed Tue May 14 00:00:44 2013 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!