--- Log opened Sun Mar 09 00:00:30 2014 | ||
-!- sonne|osx [~sonne@e179065162.adsl.alicedsl.de] has quit [Quit: sonne|osx] | 00:26 | |
-!- HeikoS [~heiko@p57972433.dip0.t-ipconnect.de] has joined #shogun | 00:37 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 00:37 | |
-!- funcode [~funcode@117.55.241.55] has quit [Quit: Leaving] | 00:54 | |
-!- gif [~gif@130.195.253.57] has joined #shogun | 01:19 | |
gif | sonney2k: what version did you refer to? the compiler? the mac os? | 01:21 |
---|---|---|
-!- HeikoS [~heiko@p57972433.dip0.t-ipconnect.de] has quit [Ping timeout: 240 seconds] | 01:33 | |
-!- sunil1337 [0e8bd057@gateway/web/freenode/ip.14.139.208.87] has joined #shogun | 02:09 | |
-!- sunil1337 [0e8bd057@gateway/web/freenode/ip.14.139.208.87] has quit [Ping timeout: 245 seconds] | 02:22 | |
-!- kislay [~kislay@93.114.45.13] has joined #shogun | 03:38 | |
shogun-buildbot_ | build #737 of nightly_default is complete: Failure [failed doc] Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/737 | 03:41 |
-!- KJainGSoC [783bc6ca@gateway/web/freenode/ip.120.59.198.202] has joined #shogun | 03:50 | |
-!- newlido [~quassel@41.218.181.148] has joined #shogun | 03:55 | |
-!- zxtx [~zv@c-98-223-196-32.hsd1.in.comcast.net] has quit [Ping timeout: 244 seconds] | 04:18 | |
-!- zxtx [~zv@c-98-223-196-32.hsd1.in.comcast.net] has joined #shogun | 04:33 | |
-!- newlido [~quassel@41.218.181.148] has quit [Ping timeout: 240 seconds] | 04:39 | |
-!- newlido [~quassel@41.218.173.221] has joined #shogun | 04:41 | |
-!- newlido [~quassel@41.218.173.221] has quit [Remote host closed the connection] | 04:54 | |
-!- newlido [~quassel@41.218.173.221] has joined #shogun | 04:55 | |
-!- kislay [~kislay@93.114.45.13] has quit [Ping timeout: 265 seconds] | 05:08 | |
-!- kislay [~kislay@93.114.45.13] has joined #shogun | 05:31 | |
-!- naywhayare [~ryan@spoon.lugatgt.org] has quit [Write error: Broken pipe] | 05:34 | |
-!- naywhaya1e [~ryan@spoon.lugatgt.org] has joined #shogun | 05:35 | |
-!- naywhaya1e is now known as naywhayare | 05:36 | |
-!- abinash_panda [0e8b2997@gateway/web/freenode/ip.14.139.41.151] has joined #shogun | 06:03 | |
-!- newlido [~quassel@41.218.173.221] has quit [Remote host closed the connection] | 06:22 | |
-!- newlido [~quassel@41.218.173.221] has joined #shogun | 06:22 | |
-!- kislay [~kislay@93.114.45.13] has quit [Read error: Connection reset by peer] | 06:22 | |
-!- dhruv13_ [~dhruv13@84.64.13.59] has joined #shogun | 06:23 | |
-!- dhruv13_ [~dhruv13@84.64.13.59] has quit [Client Quit] | 06:23 | |
-!- dhruv13_ [~dhruv13@84.64.13.59] has joined #shogun | 06:24 | |
-!- dhruv13 [~dhruv13@84.64.13.59] has quit [Excess Flood] | 06:25 | |
-!- gif [~gif@130.195.253.57] has quit [Remote host closed the connection] | 06:26 | |
-!- kislay_ [~kislay@93.114.45.13] has joined #shogun | 06:26 | |
-!- sonne|osx [~sonne@e179064197.adsl.alicedsl.de] has joined #shogun | 06:55 | |
-!- newlido_ [~quassel@41.218.172.4] has joined #shogun | 07:10 | |
-!- newlido [~quassel@41.218.173.221] has quit [Ping timeout: 244 seconds] | 07:11 | |
-!- cameo54321 [2729cfbf@gateway/web/freenode/ip.39.41.207.191] has joined #shogun | 07:20 | |
shogun-buildbot_ | build #71 of osx2 - python is complete: Failure [failed test python modular] Build details are at http://buildbot.shogun-toolbox.org/builders/osx2%20-%20python/builds/71 blamelist: Parijat Mazumdar <mazumdarparijat@gmail.com>, Orion Poplawski <orion@nwra.com>, Soeren Sonnenburg <sonne@debian.org>, iglesias <fernando.iglesiasg@gmail.com>, Bj?rn Esser | 08:01 |
shogun-buildbot_ | <bjoern.esser@gmail.com>, Thoralf Klein <thoralf.klein@zib.de>, lambday <heavensdevil6909@gmail.com>, Viktor Gal <vigsterkr@gmail.com>, Heiko Strathmann <heiko.strathmann@gmail.com>, tklein23 <tklein23@users.noreply.github.com>, Viktor Gal <viktor.gal@maeth.com>, Thoralf Klein <thoralf@fischlustig.de>, Sergey Lisitsyn <lisitsyn.s.o@gmail.com>, Fernando Iglesias | 08:01 |
shogun-buildbot_ | <fernando.iglesiasg@gmail.com>, Roman Votyakov <votjakovr@gmail.com> | 08:01 |
-!- dhruv13_ [~dhruv13@84.64.13.59] has quit [Ping timeout: 265 seconds] | 08:04 | |
-!- cameo54321 [2729cfbf@gateway/web/freenode/ip.39.41.207.191] has quit [Ping timeout: 245 seconds] | 08:06 | |
-!- kislay_ [~kislay@93.114.45.13] has quit [Ping timeout: 240 seconds] | 08:25 | |
shogun-buildbot_ | build #72 of osx2 - python is complete: Failure [failed test python modular] Build details are at http://buildbot.shogun-toolbox.org/builders/osx2%20-%20python/builds/72 blamelist: Soeren Sonnenburg <sonne@debian.org>, Thoralf Klein <thoralf.klein@zib.de> | 08:30 |
-!- kislay [~kislay@93.114.45.13] has joined #shogun | 08:40 | |
shogun-buildbot_ | build #405 of osx1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/405 blamelist: lambday <heavensdevil6909@gmail.com>, Soeren Sonnenburg <sonne@debian.org>, Viktor Gal <vigsterkr@gmail.com>, pl8787 <pl8787@gmail.com>, Viktor Gal <viktor.gal@maeth.com>, Heiko Strathmann | 08:48 |
shogun-buildbot_ | <heiko.strathmann@gmail.com> | 08:48 |
-!- abinash_panda [0e8b2997@gateway/web/freenode/ip.14.139.41.151] has quit [Ping timeout: 245 seconds] | 09:13 | |
-!- akpraharaj [67157d4c@gateway/web/freenode/ip.103.21.125.76] has joined #shogun | 09:22 | |
-!- kislay [~kislay@93.114.45.13] has quit [Read error: Connection reset by peer] | 09:22 | |
-!- akpraharaj [67157d4c@gateway/web/freenode/ip.103.21.125.76] has quit [Ping timeout: 245 seconds] | 09:28 | |
-!- newlido [~quassel@41.218.179.76] has joined #shogun | 09:30 | |
-!- newlido_ [~quassel@41.218.172.4] has quit [Ping timeout: 240 seconds] | 09:30 | |
-!- kislay [~kislay@93.114.45.13] has joined #shogun | 09:37 | |
-!- kislay [~kislay@93.114.45.13] has quit [Ping timeout: 264 seconds] | 10:14 | |
-!- newlido [~quassel@41.218.179.76] has quit [Remote host closed the connection] | 11:02 | |
-!- kislay [~kislay@93.114.45.13] has joined #shogun | 11:04 | |
-!- kislay [~kislay@93.114.45.13] has quit [Read error: Connection reset by peer] | 11:37 | |
-!- PirosB3 [~pirosb3@162.219.6.191] has joined #shogun | 11:48 | |
-!- dhruv13 [~dhruv13@84.64.14.9] has joined #shogun | 11:56 | |
PirosB3 | hi all | 12:00 |
PirosB3 | anyone on? | 12:00 |
dhruv13 | PirosB3: hello! | 12:02 |
PirosB3 | hey dhruv13 hows it going? | 12:02 |
dhruv13 | PirosB3: not bad! | 12:02 |
PirosB3 | coin anything interesting? | 12:02 |
PirosB3 | doin | 12:02 |
dhruv13 | PirosB3: kind of... | 12:03 |
dhruv13 | PirosB3: I'm new to a lot of stuff, but figuring it out seems fun! | 12:03 |
dhruv13 | PirosB3: (now) | 12:03 |
PirosB3 | what are you looking at? | 12:03 |
dhruv13 | PirosB3: K nearest neighbours implementation through kd trees | 12:04 |
PirosB3 | did you have expirence with ML before? | 12:04 |
dhruv13 | PirosB3: looking at a library which can be possibly bundled along with shogun, called nanoflann... | 12:04 |
PirosB3 | cool | 12:04 |
dhruv13 | PirosB3: not much with implementing ML algorithms, though i have used them for various programming contests | 12:05 |
dhruv13 | PirosB3: like kaggle | 12:05 |
PirosB3 | oh cool! | 12:05 |
dhruv13 | PirosB3: :) | 12:05 |
dhruv13 | PirosB3: how about you? | 12:06 |
PirosB3 | http://psi.cecs.anu.edu.au/spec/#schema | 12:06 |
PirosB3 | looking up stuff for my proposal | 12:06 |
PirosB3 | have you done that yet? | 12:06 |
dhruv13 | PirosB3: not really... I'mstill getting a hang of how shogun works... | 12:06 |
PirosB3 | yeah me too | 12:07 |
dhruv13 | PirosB3: maybe once i'm done with this task, i'll hunt around for stuff ;-) | 12:07 |
PirosB3 | ahahah :) | 12:08 |
PirosB3 | anyways thx for your help last time | 12:08 |
dhruv13 | PirosB3: anytime!! | 12:08 |
PirosB3 | wiking: are you on by any chance? | 12:13 |
-!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has joined #shogun | 12:23 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 12:23 | |
-!- khalednasr [2945ccf3@gateway/web/freenode/ip.41.69.204.243] has joined #shogun | 12:36 | |
-!- khalednasr [2945ccf3@gateway/web/freenode/ip.41.69.204.243] has quit [Client Quit] | 12:37 | |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 12:44 | |
shogun-notifier- | shogun: Parijat Mazumdar :develop * 8d4eeef / / (4 files): https://github.com/shogun-toolbox/shogun/commit/8d4eeef22a4eb22999837eb17a88dd3ad0f38a83 | 12:44 |
shogun-notifier- | shogun: added tree structures for id3 | 12:44 |
shogun-notifier- | shogun: Parijat Mazumdar :develop * eb4cb6e / / (14 files): https://github.com/shogun-toolbox/shogun/commit/eb4cb6e77e373f7ed8893183cceb47d09490dfef | 12:44 |
shogun-notifier- | shogun: BinaryTreeMachineNode created | 12:44 |
shogun-notifier- | shogun: Parijat Mazumdar :develop * b1d9a3d / / (7 files): https://github.com/shogun-toolbox/shogun/commit/b1d9a3ddd95cdde3dac969660e463012ff422001 | 12:44 |
shogun-notifier- | shogun: unit-tests added for valgrind check | 12:44 |
shogun-notifier- | shogun: Parijat Mazumdar :develop * 56a6ff3 / tests/unit/multiclass/tree/ (3 files): https://github.com/shogun-toolbox/shogun/commit/56a6ff36b53018b81b6e56a39b50a0b668f71892 | 12:44 |
shogun-notifier- | shogun: unittest names changed | 12:44 |
shogun-notifier- | shogun: Parijat Mazumdar :develop * fba3a47 / src/shogun/multiclass/tree/ConditionalProbabilityTree.cpp: https://github.com/shogun-toolbox/shogun/commit/fba3a47e19d41d52537601bbeb40f64112ea8312 | 12:44 |
shogun-notifier- | shogun: REQUIRE condition removed to counter travis failure | 12:44 |
shogun-notifier- | shogun: Fernando Iglesias :develop * 02a1773 / / (17 files): https://github.com/shogun-toolbox/shogun/commit/02a1773de106310dd86bacec24f86cd716e31866 | 12:44 |
shogun-notifier- | shogun: Merge pull request #1950 from mazumdarparijat/pca | 12:44 |
shogun-notifier- | shogun: | 12:44 |
shogun-notifier- | shogun: added data structures for decision tree implementation | 12:44 |
PirosB3 | hey dhruv13 do you know anything about multinomial naive bayes? | 12:45 |
@wiking | PirosB3: there's a lot of ML tutorials out there on the internet: http://lmgtfy.com/?q=multinomial+naive+bayes | 12:54 |
PirosB3 | lol, thx | 12:55 |
PirosB3 | wiking: do you have a sec? I wanted to ask you some questions regarding the SaaS. I have a better vision of it now and of it's resources. But I still have some doubts | 12:56 |
@wiking | PirosB3: email plz | 12:57 |
PirosB3 | ok | 12:57 |
@wiking | preferably mailing list | 12:57 |
shogun-buildbot_ | build #2754 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/2754 blamelist: Parijat Mazumdar <mazumdarparijat@gmail.com> | 13:01 |
shogun-buildbot_ | build #2755 of deb1 - libshogun is complete: Success [build successful] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/2755 | 13:09 |
@iglesiasg | all right, false alarm the fail in deb1 | 13:10 |
@iglesiasg | :) | 13:10 |
shogun-buildbot_ | build #406 of osx1 - libshogun is complete: Success [build successful] Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/406 | 13:19 |
-!- kislay [~kislay@93.114.45.13] has joined #shogun | 13:20 | |
dhruv13 | iglesiasg: hey! can i bother you with something? | 13:27 |
@iglesiasg | dhruv13, tell me | 13:28 |
dhruv13 | iglesiasg: well, is there a significant difference between approximate knn and what needs to be implemented for the kd tree issue? | 13:28 |
dhruv13 | iglesiasg: nanoflann implements ann | 13:29 |
@iglesiasg | dhruv13, so when I wrote the issue I was not thinking about ann actually | 13:29 |
dhruv13 | iglesiasg: please check https://code.google.com/p/nanoflann/ | 13:29 |
dhruv13 | iglesiasg: section 1.6 | 13:29 |
@iglesiasg | dhruv13, just a kd tree to do conventional exact nn search | 13:29 |
dhruv13 | iglesiasg: "what nanoflann can't do" | 13:30 |
@iglesiasg | dhruv13, I was thinking about going on to ann once this was done | 13:30 |
@iglesiasg | dhruv13, just because I think it is a more natural progression for the person working on that | 13:30 |
dhruv13 | iglesiasg: okay, then i think i will have to go with my own implementation of a kdtree | 13:30 |
@iglesiasg | learn kd-tree, use it, see limitations on real data and the need of ann | 13:30 |
dhruv13 | iglesiasg: okay! | 13:30 |
@iglesiasg | dhruv13, having said that | 13:30 |
@iglesiasg | dhruv13, that was just my idea, it is not something set on stone | 13:30 |
-!- akp_ [67157e4c@gateway/web/freenode/ip.103.21.126.76] has joined #shogun | 13:31 | |
@iglesiasg | dhruv13, ann is very useful. So the nanoflann integration could be very nice as well | 13:31 |
@iglesiasg | dhruv13, I don't see anything about exact/approximate nn search in section 1.6 btw | 13:32 |
dhruv13 | iglesiasg: oh sorry, that part is in about | 13:33 |
dhruv13 | iglesiasg: "Queries for neighbors around any arbitrary location in space can then be solved quickly and efficiently using Approximate Nearest Neighbor (ANN) algorithms." | 13:33 |
@iglesiasg | dhruv13, all right! | 13:34 |
dhruv13 | iglesiasg: the 1.6, point 3 confused me, so i referred to the paper | 13:34 |
dhruv13 | iglesiasg: the part about automatic configuration | 13:34 |
dhruv13 | iglesiasg: that seems to be in flann, but not nanoflann | 13:35 |
@iglesiasg | dhruv13, all right | 13:35 |
dhruv13 | iglesiasg: so, for now, i think i will work on nanoflann integration | 13:36 |
@iglesiasg | dhruv13, it sounds great! | 13:36 |
dhruv13 | iglesiasg: and then try out my own kdtree implementation | 13:36 |
dhruv13 | iglesiasg: it will take me a while, since i'm still working around the shogun feature class.... | 13:37 |
@iglesiasg | dhruv13, sure, no problem at all | 13:37 |
dhruv13 | iglesiasg: thanks! | 13:37 |
-!- PirosB3 [~pirosb3@162.219.6.191] has quit [Quit: PirosB3] | 13:38 | |
akp_ | are we looking for the basic knn algorithm using flann? | 13:38 |
shogun-notifier- | shogun: Viktor Gal :develop * 5675b9d / / (22 files): https://github.com/shogun-toolbox/shogun/commit/5675b9d0d54142118abb0b90dab5cd6701626dcc | 13:39 |
shogun-notifier- | shogun: CMake-ize documentum generation | 13:39 |
shogun-notifier- | shogun: Doxygen documentation was waaay out of date and it was still using the | 13:39 |
shogun-notifier- | shogun: old make system. updated: | 13:39 |
shogun-notifier- | shogun: * documentation is generated via cmake i.e. `make doc` | 13:39 |
shogun-notifier- | shogun: * outdated doxygen descriptions are removed | 13:39 |
-!- travis-ci [~travis-ci@ec2-23-22-234-209.compute-1.amazonaws.com] has joined #shogun | 13:39 | |
travis-ci | [travis-ci] it's Fernando Iglesias'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/20390929 | 13:39 |
-!- travis-ci [~travis-ci@ec2-23-22-234-209.compute-1.amazonaws.com] has left #shogun [] | 13:39 | |
-!- shogun-buildbot_ [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting] | 13:41 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 13:41 | |
@wiking | shogun-buildbot: force build --branch=develop 'nightly_default' | 13:41 |
shogun-buildbot | The build has been queued, I'll give a shout when it starts | 13:42 |
akp_ | I wanted to discuss about the OpenCV interface to shogun toolkit. | 13:45 |
akp_ | Do we need a complete implementation here or simply training?> | 13:46 |
@wiking | akp_: depends what u wanna do... | 13:47 |
akp_ | I am currently developing an interface for the dlib machine learning library. | 13:47 |
@wiking | ? | 13:47 |
@wiking | i guess you are then on the wrong channel mate | 13:48 |
@wiking | this is the shogun machine learning library channel | 13:48 |
akp_ | No. Sorry to mention the name of the library. But I think I can contribute to this library too. | 13:49 |
@wiking | no worries | 13:49 |
@wiking | that's great | 13:49 |
akp_ | Before that | 13:49 |
-!- akp__ [67157f54@gateway/web/freenode/ip.103.21.127.84] has joined #shogun | 13:51 | |
akp__ | I just wanted toknow if shogun provides a basic interface for implementing SVMs | 13:52 |
-!- akp___ [67157e4c@gateway/web/freenode/ip.103.21.126.76] has joined #shogun | 13:53 | |
-!- akp_ [67157e4c@gateway/web/freenode/ip.103.21.126.76] has quit [Ping timeout: 245 seconds] | 13:54 | |
akp___ | I wanted to know if shogun provides interface for nonlinear svms | 13:54 |
akp___ | can you shed some light on that? | 13:55 |
@wiking | http://shogun-toolbox.org/doc/en/current/ | 13:55 |
@wiking | it has everything there what u need to know about the interfaces | 13:55 |
-!- akp__ [67157f54@gateway/web/freenode/ip.103.21.127.84] has quit [Ping timeout: 245 seconds] | 13:56 | |
@wiking | lisitsyn: btw ^ | 13:56 |
akp___ | ignore my ignorance. What I think is a simple training interface using these non-linear svms | 13:56 |
@wiking | akp___: as said... the documentation is quite extensive + examples | 13:57 |
@wiking | i suggest check those | 13:57 |
akp___ | Well I think you misunderstood what I am saying. Apart from using the current features of shogun, I think we can integrate parts of OpenCV including feature extractio etc into shogun | 13:58 |
@wiking | akp___: we do not really want to integrate parts of openc into shogun | 13:59 |
@wiking | as shogun is a machine learning library | 13:59 |
@wiking | whilst opencv is rather an image processing library | 13:59 |
@wiking | but we'd like to have some example applications of opencv and shogun | 14:00 |
-!- cameo54321 [2729b2d3@gateway/web/freenode/ip.39.41.178.211] has joined #shogun | 14:00 | |
akp___ | yeah. That is what I am trying to say. | 14:00 |
@wiking | mmm | 14:01 |
@wiking | sorry i did not get that... | 14:01 |
akp___ | ok. So apart from training, which can be used in examples such as ocr and image classification, | 14:02 |
akp___ | Many other examples can be demonstrated. | 14:03 |
@wiking | yes indeed | 14:03 |
akp___ | I am working in the computer vision for past 6 months. | 14:04 |
akp___ | As i have said earlier, my notable implementations include a multi-object classifier and a text segmentation & recognition tool. | 14:05 |
akp___ | Having said that, can you say what specific applications of computer vision using your library are you looking for? Just some examples, or a specific example, like say a biometric application using | 14:07 |
akp___ | opencv and shogun | 14:07 |
@wiking | akp___: well it's really up to you | 14:08 |
@wiking | we have not fixed any specifics ... some people already did eigenfaces examples | 14:08 |
@wiking | but it can be anything you like | 14:08 |
@wiking | or you are interested in | 14:08 |
akp___ | ok, great! | 14:09 |
@wiking | but since there are like 2-3 students who started working on eigenfaces so that should be excluded from the possible options | 14:09 |
@wiking | apart from that | 14:09 |
@wiking | it's really up to u | 14:09 |
@wiking | we don't have limitations... but please make sure that if u start working on it that we expect you to create an ipython notebook out of that application | 14:10 |
akp___ | yeah. sure. | 14:11 |
akp___ | THe ipython should contain the appropriate prototype code at the time of proposal? | 14:12 |
akp___ | And apart from that,' | 14:13 |
akp___ | do we need to submit a demo application using OpenCV and Shogun at the time of proposal? | 14:14 |
akp___ | Can I discuss my idea with you in private? | 14:16 |
akp___ | chat I mean | 14:16 |
-!- akp_ [67157f4c@gateway/web/freenode/ip.103.21.127.76] has joined #shogun | 14:18 | |
@wiking | well for proposal you can use the mailing list | 14:18 |
@wiking | and discuss there the idea | 14:18 |
@wiking | we don't expect you to create already a notebook for the idea... but you should start working on entrance issues of shogun to get yourself familiarised with shogun | 14:19 |
akp_ | ok. | 14:20 |
shogun-buildbot | build #738 forced | 14:20 |
shogun-buildbot | I'll give a shout when the build finishes | 14:20 |
-!- akp___ [67157e4c@gateway/web/freenode/ip.103.21.126.76] has quit [Ping timeout: 245 seconds] | 14:20 | |
akp_ | let me wrap up my idea. I'll get back to you after a while. Any problem? | 14:21 |
-!- cameo54321 [2729b2d3@gateway/web/freenode/ip.39.41.178.211] has quit [Ping timeout: 245 seconds] | 14:21 | |
@wiking | akp_: use mailing list | 14:22 |
@wiking | that's read by everybody from the project so you'll get more feedback | 14:23 |
-!- akp__ [67157d4c@gateway/web/freenode/ip.103.21.125.76] has joined #shogun | 14:23 | |
-!- akp__ [67157d4c@gateway/web/freenode/ip.103.21.125.76] has quit [Client Quit] | 14:24 | |
-!- akp_ [67157f4c@gateway/web/freenode/ip.103.21.127.76] has quit [Ping timeout: 245 seconds] | 14:25 | |
-!- PirosB3 [~pirosb3@162.219.6.191] has joined #shogun | 14:26 | |
shogun-buildbot | Hey! build nightly_default #738 is complete: Exception [exception upload] | 14:28 |
shogun-buildbot | Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/738 | 14:28 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting] | 14:29 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 14:30 | |
@wiking | shogun-buildbot: force build --branch=develop 'nightly_default' | 14:30 |
shogun-buildbot | The build has been queued, I'll give a shout when it starts | 14:30 |
shogun-buildbot | build #739 forced | 14:33 |
shogun-buildbot | I'll give a shout when the build finishes | 14:33 |
shogun-buildbot | build #1865 of cyg1 - libshogun is complete: Failure [failed compile test] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1865 blamelist: Viktor Gal <viktor.gal@maeth.com> | 14:36 |
-!- dhruv13 [~dhruv13@84.64.14.9] has quit [Ping timeout: 264 seconds] | 14:42 | |
-!- travis-ci [~travis-ci@ec2-54-81-118-199.compute-1.amazonaws.com] has joined #shogun | 15:06 | |
travis-ci | [travis-ci] it's Viktor Gal'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/20392657 | 15:06 |
-!- travis-ci [~travis-ci@ec2-54-81-118-199.compute-1.amazonaws.com] has left #shogun [] | 15:06 | |
shogun-buildbot | Hey! build nightly_default #739 is complete: Success [build successful] | 15:12 |
shogun-buildbot | Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/739 | 15:12 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting] | 15:19 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 15:19 | |
shogun-notifier- | shogun: lambday :develop * b6a9867 / doc/ipython-notebooks/logdet/logdet.ipynb: https://github.com/shogun-toolbox/shogun/commit/b6a986772998b86cf7ec5bb7b7c73d990af9c551 | 15:27 |
shogun-notifier- | shogun: fix failing input error in logdet notebook | 15:27 |
shogun-notifier- | shogun: Viktor Gal :develop * 26aa304 / doc/ipython-notebooks/logdet/logdet.ipynb: https://github.com/shogun-toolbox/shogun/commit/26aa304ecd99ac4cee28c362c3ba11ba9a572b07 | 15:27 |
shogun-notifier- | shogun: Merge pull request #1968 from lambday/develop | 15:27 |
shogun-notifier- | shogun: | 15:27 |
shogun-notifier- | shogun: fix failing input error in logdet notebook | 15:27 |
-!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has quit [Quit: Leaving] | 15:30 | |
-!- PirosB3 [~pirosb3@162.219.6.191] has quit [Quit: PirosB3] | 15:36 | |
-!- HeikoS [~heiko@p4FE97CA6.dip0.t-ipconnect.de] has joined #shogun | 15:37 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 15:37 | |
@wiking | HeikoS: i'm just fixing documentation generation atm and i've just remembered last year you and i talking at TU-berlin to have ipython notebooks... it's fucking crazy how much of those we already have now | 15:41 |
@wiking | as well as how crazy outdated the doxygen was... | 15:41 |
@HeikoS | wiking: hey! | 15:42 |
@HeikoS | wiking: yeah, the last year brought great improvements | 15:43 |
@HeikoS | also unit tests ... | 15:43 |
@HeikoS | wiking: we should keep on pushing things :) | 15:43 |
@wiking | HeikoS: tell me about lars | 15:44 |
@wiking | HeikoS: what do u want there to be eigenized | 15:44 |
@HeikoS | wiking: mainly I want to get rid of the un-readable lapack calls | 15:45 |
@HeikoS | for multiplications | 15:45 |
@HeikoS | like matrix-vector | 15:45 |
@HeikoS | and matrix matrix | 15:45 |
@wiking | HeikoS: so the agenda is that we get rid of lapack? | 15:45 |
@wiking | why | 15:45 |
@HeikoS | wiking: I agree, we should do the factorisations via our own interface to exchange backends | 15:46 |
@HeikoS | wiking: but for these lower level things, I think maybe lets just use eigen | 15:46 |
@HeikoS | wiking: there have been many man ybugs caused by the confusing lapack api. | 15:46 |
@HeikoS | wiking: with eigen, code is much more clear and therefore easier to verify | 15:46 |
@HeikoS | wiking: lapack has all these parameters in the matrix-vector methods, people confuse them sometimes | 15:47 |
@HeikoS | wiking: so its mainly readability | 15:47 |
@HeikoS | wiking: also lapack is a pain to link against, eigen is much easier | 15:47 |
@wiking | still... there's magma which is exactly like lapack api ;) | 15:47 |
@HeikoS | wiking: and I dont know whether it makes sense to overload the SGMatrix multiplication api | 15:47 |
@wiking | so what r u going to do then? | 15:47 |
@HeikoS | wiking: thats all hidden in the backend | 15:47 |
@wiking | well that's why i would do the same | 15:47 |
@HeikoS | wiking: I want to have the algorithm code clean and easy | 15:48 |
@wiking | lars should be engine independent | 15:48 |
@HeikoS | wiking: do you also want that for low-level calls? such as matrix-.vector product | 15:48 |
@wiking | if u have lapack u have lapack based lars | 15:48 |
@wiking | if u have eigen u have that | 15:48 |
@wiking | HeikoS: well on the long run... | 15:48 |
@HeikoS | wiking: its very hard to have a unified interface for that | 15:48 |
@wiking | HeikoS: think about doint dot product of a large matrix *vector in gpu | 15:49 |
@HeikoS | wiking: makes things inefficient sometimes | 15:49 |
@HeikoS | think of transposes etc | 15:49 |
@HeikoS | wiking: matrix-vector doesnt make sense on gpu | 15:49 |
@HeikoS | wiking: maybe matrix matrix does | 15:49 |
@HeikoS | wiking: but matrix matrix is alsready openmp'ed in eigen | 15:49 |
@HeikoS | wiking: so I would go for that | 15:49 |
@wiking | HeikoS: well say matrix is n x n | 15:49 |
@wiking | and n is 10000000 | 15:49 |
@HeikoS | wiking: you cannot store that anyways | 15:49 |
@wiking | HeikoS: well maybe shogun cannot | 15:50 |
@HeikoS | wiking: the thing about the eigen3-ization also is that people wirte tests and review the code | 15:50 |
@HeikoS | wiking: thats a good side effect | 15:50 |
@wiking | but come one a normal tesla can store amazing amount of data | 15:50 |
@HeikoS | wiking: so I would vote for high level calls via our own backend | 15:50 |
@HeikoS | wiking: and multiplications in eigen | 15:50 |
@HeikoS | wiking: and maybe change that at some point once we have soe experience how the factorizations work | 15:50 |
@HeikoS | wiking: moving from lapack to eigen already has many advantages (intro tasks, tests, bugfixes, much cleaner code) | 15:51 |
@wiking | mmm well i just think that throwing out lapack is irresponsible | 15:51 |
-!- cameo54321 [2729ef55@gateway/web/freenode/ip.39.41.239.85] has joined #shogun | 15:51 | |
@wiking | HeikoS: btw stan... dude that's really a mess | 15:52 |
@wiking | HeikoS: nobody would like to have dependency on that ... not even u | 15:53 |
@wiking | HeikoS: it has no good build system and everything is bundled | 15:53 |
-!- kprah [67157d4c@gateway/web/freenode/ip.103.21.125.76] has joined #shogun | 15:54 | |
@HeikoS | wiking: wiking ok what about having some easy api for matrix-matrix and matrix-vector and do everything against that? | 15:54 |
@wiking | HeikoS: SGMatrix has that | 15:54 |
@HeikoS | wiking: Should not be part of SGMatrix but seperate SGMatrix is mainly a memory wrapper | 15:55 |
@wiking | if not then it's a shame and should be implemented :D | 15:55 |
@HeikoS | wiking: should be in linalg or so | 15:55 |
@wiking | so like a static function | 15:55 |
@wiking | mm(matrix,matrix) | 15:55 |
@wiking | ? | 15:55 |
@wiking | :D | 15:55 |
@HeikoS | wiking: although overloaded operators such as in eigen are much nicer | 15:55 |
@HeikoS | wiking: what about transposes and stuff | 15:55 |
@wiking | why not SGMatrix::operator*(sgmatrix) | 15:55 |
@HeikoS | wiking: eigen does all that in place | 15:55 |
kprah | Hello. | 15:56 |
@wiking | shogun-buildbot: force build --branch=develop 'nightly_default' | 15:56 |
@HeikoS | wiking: mmh, I mean it would be *great* not to write against libs but against a general iunterface | 15:56 |
shogun-buildbot | The build has been queued, I'll give a shout when it starts | 15:56 |
@wiking | HeikoS: i get that part | 15:56 |
@wiking | and i'm up for that | 15:56 |
@HeikoS | wiking: but I am also fine with only having high-level calls generalised | 15:57 |
@wiking | (see the eigensolver story) | 15:57 |
@HeikoS | wiking: since eigen is so good for low-level things | 15:57 |
-!- kprah_ [67157e4c@gateway/web/freenode/ip.103.21.126.76] has joined #shogun | 15:57 | |
@HeikoS | wiking: and its alspo open-mp'ed for matrix-matrix | 15:57 |
@HeikoS | wiking: we just have to activate that | 15:57 |
@HeikoS | wiking: so maybe lets do that for now? Could do another project at some point doing the multiplications | 15:57 |
@wiking | HeikoS: well if openmp is available then it should be already active.. or? | 15:57 |
@HeikoS | wiking: one has to set number of threads in eigen /openmp or something | 15:58 |
@HeikoS | they have a howto page | 15:58 |
kprah_ | Hello guys. Anyone free? | 15:58 |
@HeikoS | kprah, whats up? | 15:58 |
kprah_ | Hi. | 15:58 |
@HeikoS | wiking: eigen3 also has views and all that, which is really helpful | 15:58 |
@HeikoS | wiking: and lots of in-place magic | 15:58 |
@HeikoS | thats very hard to put under a unified interface | 15:58 |
@HeikoS | wiking: and most gain really is in high level calls | 15:58 |
kprah_ | The idea "LP/QP optimization framework caught my eye and I wanted to talk about that. | 15:59 |
@wiking | kprah_: shoot | 15:59 |
@HeikoS | wiking: vector products are so easy, that its not really a big deal with having different backends, and eigen3 works on parallel stuff | 15:59 |
kprah_ | I am a chemical engineering sophomore in my last semester., | 15:59 |
@HeikoS | kprah, for that you should try to catch one of the mentors, try writing to the mailing list and cc them | 15:59 |
@wiking | HeikoS: well that's the thing that it does but not always the best solution... | 16:00 |
@wiking | HeikoS: /me is the mentor ;) | 16:00 |
@wiking | :D | 16:00 |
kprah_ | Previous semester in my winter internship, my work invloved | 16:00 |
@HeikoS | wiking: haha ok then you deal with that :) | 16:00 |
cameo54321 | wiking: HeikoS thanks for bringing up the topic, | 16:00 |
@HeikoS | wiking: I am always for step-wise improvements, so lets to the factorizations, that would already be amazing | 16:00 |
-!- kprah [67157d4c@gateway/web/freenode/ip.103.21.125.76] has quit [Ping timeout: 245 seconds] | 16:00 | |
@wiking | HeikoS: we'll see how goes with the eigensolver unification | 16:01 |
kprah_ | usage of various solvers like bonmin and cplex to generate optimal prodUCTS FROM the given set of reactants. | 16:01 |
@HeikoS | wiking: but I think those tasks to eigen3-ize things are good since they produce tests and get people involved :) ... We should however work against the new stuff for factorazations, yeah | 16:01 |
@HeikoS | cameo54321: what? | 16:01 |
@wiking | kprah_: that sounds great | 16:01 |
kprah_ | I ended up using CPLEX and bonmin. | 16:02 |
kprah_ | My problem there was essentially a mixed integer quadratic programming problem. | 16:02 |
@wiking | kprah_: in that task you would need to work a lot designing an interface that can be used regardless of the available solver backend | 16:02 |
@wiking | so we could concentrate of course on a given problem as well | 16:03 |
kprah_ | Yeah, | 16:03 |
@wiking | but the main idea is to do some unification of dealing with solvers | 16:03 |
@wiking | as there's like kazzilion of them out there | 16:03 |
cameo54321 | I talked with wiking yesterday on bundling the stan issue.. Today I shifted more to exploring shogun's code, GP ipython notebook by roman and your kameleon mcmc implementation | 16:03 |
kprah_ | That depends a lot on the given constraints i.e. which solver you are using. | 16:03 |
kprah_ | One optimization solver that was one of the most principal one is OPTI toolbox | 16:04 |
kprah_ | it is open source and you can see it here. : http://www.i2c2.aut.ac.nz/Wiki/OPTI/ | 16:05 |
cameo54321 | @Heiko what if i start writing an ipython notebook for the kameleon mcmc by you? | 16:05 |
kprah_ | Basically it is a matlab toolbox. But it is very intelligent in itself and chooses the most appropriate solver depending on the condition that you are in | 16:06 |
kprah_ | I would like to port it to shogun , from matlab to c++. | 16:06 |
lisitsyn | dear HeikoS I would like to take an opportunity and participate in the project mentored by you, could you kindly tell me how to proceed next? | 16:07 |
-!- HeikoS [~heiko@p4FE97CA6.dip0.t-ipconnect.de] has quit [Read error: Operation timed out] | 16:07 | |
-!- kprah [67157d4c@gateway/web/freenode/ip.103.21.125.76] has joined #shogun | 16:07 | |
lisitsyn | kaboom I killed him! | 16:07 |
lisitsyn | :D | 16:07 |
kprah | wiking: what do you say? | 16:07 |
kprah | Also do we need mixed integer optimization here> | 16:08 |
kprah | ? | 16:08 |
@wiking | kprah: sounds good but porting from matlab to c++ is always a bad idea | 16:09 |
-!- cameo54321 [2729ef55@gateway/web/freenode/ip.39.41.239.85] has quit [Ping timeout: 245 seconds] | 16:09 | |
kprah | No actually the code that the author wrote is actually just a wrapper. | 16:09 |
@wiking | kprah: wrapper of? | 16:10 |
kprah | a matlab wrapper for accessing all those solvers. | 16:10 |
-!- kprah_ [67157e4c@gateway/web/freenode/ip.103.21.126.76] has quit [Ping timeout: 245 seconds] | 16:10 | |
@wiking | kprah: yeah that's what i'm saying as well | 16:11 |
@wiking | it's a good start to have a look at it | 16:11 |
@wiking | but to use that in any way as a reference is a bad idea | 16:11 |
@wiking | because you'll limit yourself and not use all the good stuff that c++ offers | 16:11 |
@wiking | anyhow of course it's good to have an idea from it | 16:11 |
kprah | Well actually there is a paper for it by the same author. | 16:12 |
-!- kprah_ [67157e4c@gateway/web/freenode/ip.103.21.126.76] has joined #shogun | 16:13 | |
-!- cameo54321 [2729ef55@gateway/web/freenode/ip.39.41.239.85] has joined #shogun | 16:13 | |
kprah_ | Well we can construct a similar wrapper here. | 16:13 |
kprah_ | Do you want to see the references? | 16:13 |
-!- cameo54321 [2729ef55@gateway/web/freenode/ip.39.41.239.85] has quit [Client Quit] | 16:14 | |
@wiking | well that's what i'm saying apart from having similar functionality it shouldn't be similar ;) | 16:14 |
@wiking | kprah: since actually what we would like to achive is not only wrapper for all solvers | 16:14 |
-!- cameo54321 [2729ef55@gateway/web/freenode/ip.39.41.239.85] has joined #shogun | 16:14 | |
@wiking | but as well do create building blocks for LP/QP solvers | 16:14 |
@wiking | like KKT solving etc | 16:14 |
kprah_ | Yeah. It is there I think. We can construct it. | 16:15 |
@wiking | since there are several functionalities in a qp solver | 16:15 |
@wiking | that is the same | 16:15 |
@wiking | regardless of the given QP solver... | 16:15 |
kprah_ | ok. | 16:15 |
kprah_ | So what do you suggest finally then? | 16:16 |
@wiking | and the idea was that one would create a reference cone LP/QP solver based on this framework.. | 16:16 |
@wiking | kprah: well in order to be able to do discuss something | 16:16 |
-!- kprah [67157d4c@gateway/web/freenode/ip.103.21.125.76] has quit [Ping timeout: 245 seconds] | 16:16 | |
@wiking | plans should be made so we can iterate them | 16:16 |
@wiking | best of course to do these type of things on the mailing list | 16:16 |
@wiking | since then a lot of people can participate | 16:17 |
@wiking | but in the meanwhile you should as well get familiarised with shogun's code | 16:17 |
@wiking | and start working on entrance tasks | 16:17 |
@wiking | so that you get a grip of how development goes in shogun | 16:17 |
-!- kprah [67157d4c@gateway/web/freenode/ip.103.21.125.76] has joined #shogun | 16:18 | |
kprah | Ok. | 16:18 |
kprah | I'll put up this on the mailing list then. | 16:19 |
kprah | BTW you said cone programming | 16:19 |
-!- kprah_ [67157e4c@gateway/web/freenode/ip.103.21.126.76] has quit [Ping timeout: 245 seconds] | 16:20 | |
kprah | a cone program is nothing a linear program with non linear constraints right? | 16:20 |
-!- kislay [~kislay@93.114.45.13] has quit [Read error: Connection reset by peer] | 16:20 | |
@wiking | kprah: i mean one good example for this would to see how one could unify (reduce code redundancy) among libbmrm.cpp libp3bm.cpp libppbm.cpp in shogun | 16:20 |
@wiking | kprah: http://en.wikipedia.org/wiki/Conic_optimization | 16:21 |
@wiking | http://cvxopt.org/ | 16:21 |
cameo54321 | lisitsyn: can you please guide me on what happened recently? | 16:22 |
lisitsyn | cameo54321: ?? | 16:22 |
cameo54321 | I mistakenly wrote @ with Heiko's name | 16:23 |
lisitsyn | ah okay | 16:23 |
kprah | So you need a general conic solver? | 16:23 |
cameo54321 | at the same time a disconnection happened at my end | 16:23 |
kprah | @wiking | 16:23 |
cameo54321 | IRC logs say you killed somebody, was it me :-) | 16:25 |
kprah | wiking : with any constraints and objectives? | 16:25 |
kprah | wiking : I mean with unknown nature of constraints as well as the objective function? | 16:26 |
@wiking | http://abel.ee.ucla.edu/cvxopt/userguide/coneprog.html | 16:26 |
lisitsyn | cameo54321: no, that was heiko | 16:26 |
-!- kprah_ [67157d4c@gateway/web/freenode/ip.103.21.125.76] has joined #shogun | 16:27 | |
cameo54321 | lisitsyn Oh | 16:27 |
lisitsyn | cameo54321: just kidding nevermind ;) HeikoS will get back I guess | 16:27 |
shogun-buildbot | build forced [ETA 38m51s] | 16:27 |
shogun-buildbot | I'll give a shout when the build finishes | 16:27 |
lisitsyn | as he is sometimes having some bad connection | 16:28 |
-!- HeikoS [~heiko@p4FE97CA6.dip0.t-ipconnect.de] has joined #shogun | 16:29 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 16:29 | |
shogun-buildbot | build #1866 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://buildbot.shogun-toolbox.org/builders/cyg1%20-%20libshogun/builds/1866 | 16:29 |
kprah_ | wiking : so basically a cone program is a quadratic optimization problem with any given linear or non linear constraint, right? | 16:30 |
@wiking | kprah_: are we talking about coneqp or conelp? | 16:30 |
cameo54321 | HeikoS Sorry I mistakenly wrote @ with your name | 16:31 |
-!- kprah [67157d4c@gateway/web/freenode/ip.103.21.125.76] has quit [Ping timeout: 245 seconds] | 16:31 | |
kprah_ | No both. | 16:31 |
kprah_ | It includes both | 16:31 |
kprah_ | a coneqp degenerates into a conelp if you set H = 0 | 16:32 |
kprah_ | that is a null matrix. | 16:32 |
kprah_ | It will accordingly change the algorithm used. | 16:33 |
-!- kprah [67157d4c@gateway/web/freenode/ip.103.21.125.76] has joined #shogun | 16:33 | |
@wiking | HeikoS: u realise that eigenizing everything will be funny to use the shogun wrappers | 16:34 |
@wiking | HeikoS: because then u will have to map back the matricies | 16:34 |
@HeikoS | wiking: what do yo u mean? | 16:34 |
@wiking | as obviously we dont want to support ::compute(Eigen::Matrix | 16:35 |
@HeikoS | ? | 16:35 |
-!- kprah__ [67157f4c@gateway/web/freenode/ip.103.21.127.76] has joined #shogun | 16:36 | |
kprah__ | wiking : sorry, my connection with irc is a little unstable. | 16:36 |
@wiking | kprah__: maybe mail? | 16:36 |
kprah__ | Sorry. Did you say anything before this? | 16:36 |
kprah__ | Yeah. We can try mail. | 16:37 |
-!- kprah_ [67157d4c@gateway/web/freenode/ip.103.21.125.76] has quit [Ping timeout: 245 seconds] | 16:37 | |
-!- kprah [67157d4c@gateway/web/freenode/ip.103.21.125.76] has quit [Ping timeout: 245 seconds] | 16:37 | |
@wiking | HeikoS: so say u eigenize LARS. which means you map every matrix to an eigen::matrix irhgt? | 16:38 |
@wiking | *right? | 16:38 |
@HeikoS | wiking: hey | 16:38 |
@HeikoS | wiking: yes | 16:38 |
@wiking | HeikoS: and then u say wanna call a QR factorization | 16:38 |
@HeikoS | wiking: no QR should be done against our new interface | 16:38 |
@wiking | yeah yeah | 16:38 |
@HeikoS | just for the algebra | 16:38 |
@wiking | that's what i'm talking about | 16:38 |
@wiking | so when u did all the algebra | 16:38 |
@HeikoS | wiking: eigen3 maps are O(1) to create | 16:39 |
@wiking | and then u want to do QR | 16:39 |
@wiking | then u will have to map back | 16:39 |
@HeikoS | and they directy modify the SGMatrix data | 16:39 |
@wiking | ok i'm just saying that it's a bit funky... | 16:39 |
kprah__ | wiking : so do you have anything to suggest? I am introducing myself on the mailing list. | 16:39 |
@HeikoS | wiking: nono | 16:40 |
-!- travis-ci [~travis-ci@ec2-54-81-118-199.compute-1.amazonaws.com] has joined #shogun | 16:40 | |
travis-ci | [travis-ci] it's Viktor Gal'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/20396379 | 16:40 |
-!- travis-ci [~travis-ci@ec2-54-81-118-199.compute-1.amazonaws.com] has left #shogun [] | 16:40 | |
@wiking | kprah__: mmm just start the discussion about your idea | 16:40 |
@HeikoS | wiking: you dont map back | 16:40 |
kprah__ | ok. Thanks. | 16:41 |
@HeikoS | wiking: you have an SGMatrix which you use to allocate memory, then you pass this to eigen which flls it | 16:41 |
@HeikoS | wiking: or you have a matrix which you give to eigen and it modifies it | 16:41 |
@HeikoS | wiking: so all good, thats why I like the lib so much | 16:41 |
shogun-notifier- | shogun: = :develop * d6df009 / src/shogun/mathematics/linalg/ratapprox/logdet/ (2 files): https://github.com/shogun-toolbox/shogun/commit/d6df009ca735ccfd4cd5cbf51e2f9c6885c622c5 | 16:43 |
shogun-notifier- | shogun: added constructors for LogDetEstimator | 16:43 |
shogun-notifier- | shogun: Sunil Mahendrakar :develop * af28014 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/af28014985670d69c4845aa41c9bf1faaab1b202 | 16:43 |
shogun-notifier- | shogun: added constructors LogDetEstimator,unit tests for those and added get methods for private variables of LogDetEstimator | 16:43 |
shogun-notifier- | shogun: Sunil Mahendrakar :develop * a4db061 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/a4db0613d18e73ff11c2d3a3f3f6467bbbbfd3e9 | 16:43 |
shogun-notifier- | shogun: fixed typos in #1940 | 16:43 |
shogun-notifier- | shogun: Sunil Mahendrakar :develop * eb4ac93 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/eb4ac93fbe2abe6e05c3c045ad10928ffc79dd85 | 16:43 |
shogun-notifier- | shogun: added docs to #1940, unneccesary data from unit tests | 16:43 |
shogun-notifier- | shogun: Sunil Mahendrakar :develop * 0ca5322 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/0ca532224f1af2549cfb4db0763d1e21778e7abf | 16:43 |
shogun-notifier- | shogun: removed constructor for dense matrix | 16:43 |
shogun-notifier- | shogun: Heiko Strathmann :develop * ce8102e / / (3 files): https://github.com/shogun-toolbox/shogun/commit/ce8102e3b444ab77f75fa6997ca67f19c4da512a | 16:43 |
shogun-notifier- | shogun: Merge pull request #1940 from sunil1337/develop | 16:43 |
shogun-notifier- | shogun: | 16:43 |
shogun-notifier- | shogun: added constructors for LogDetEstimator | 16:43 |
cameo54321 | HeikoS would you like to discuss the kameleon mcmc code? | 16:44 |
-!- kprah__ [67157f4c@gateway/web/freenode/ip.103.21.127.76] has quit [Ping timeout: 245 seconds] | 16:45 | |
cameo54321 | I was thinking if it will be a good idea to write an ipython notebook for it/ | 16:45 |
cameo54321 | ? | 16:45 |
@HeikoS | cameo54321: I have to go now, but happy to discuss soon. Maybe already send some questions for the mailinglist? | 16:45 |
@HeikoS | cameo54321: yeah maybe, but keep in mind its not based on Shogun, so that it a point for after the MCMC framework is there | 16:46 |
@HeikoS | cameo54321: maybe think how the class structure can be translated to shogun | 16:46 |
cameo54321 | Ok Sure.. | 16:46 |
cameo54321 | HeikoS: Last point, which class structure? | 16:48 |
@HeikoS | cameo54321: of the mcmc framework in kameleon mcmc and shogun-mcmc-prototype | 16:48 |
@HeikoS | see my repository | 16:48 |
cameo54321 | Ah ok. | 16:48 |
@HeikoS | cameo54321: running off, see you later | 16:48 |
cameo54321 | See you.. Thanks :-) | 16:49 |
-!- HeikoS [~heiko@p4FE97CA6.dip0.t-ipconnect.de] has quit [Read error: Operation timed out] | 16:50 | |
cameo54321 | lisitsyn: I still cannot believe it. The @ killed Heiko? | 16:51 |
cameo54321 | llistitsyn: It seems I'll have to learn how to use IRC.. | 16:52 |
cameo54321 | ! | 16:53 |
cameo54321 | Ok now I wrote the name wrong.. | 16:53 |
-!- PirosB3 [~pirosb3@162.219.6.191] has joined #shogun | 17:00 | |
lisitsyn | cameo54321: no | 17:01 |
shogun-buildbot | Hey! build nightly_default #740 is complete: Success [build successful] | 17:04 |
shogun-buildbot | Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/740 | 17:04 |
PirosB3 | hey guys, does anyone have a good example of init_shogun and exit_shogun written in python? | 17:08 |
@wiking | lisitsyn: oh nice have u seen this: https://github.com/shogun-toolbox/shogun/graphs/traffic | 17:09 |
lisitsyn | wiking: cool | 17:09 |
-!- kislay [~kislay@93.114.45.13] has joined #shogun | 17:09 | |
@wiking | lisitsyn: got that tweet ? :D | 17:23 |
@wiking | lisitsyn: seems somebody hacked his twitter acc | 17:24 |
lisitsyn | wiking: hmm yeah I guess this twitter is pwned | 17:24 |
-!- thoralf [~thoralf@p4FEC89F4.dip0.t-ipconnect.de] has joined #shogun | 17:25 | |
thoralf | Hey :) | 17:25 |
-!- Jey [~chatzilla@2601:4:700:363:c507:abec:6e4:77d8] has joined #shogun | 17:26 | |
-!- Jey [~chatzilla@2601:4:700:363:c507:abec:6e4:77d8] has left #shogun [] | 17:27 | |
-!- HeikoS [~heiko@p4FE97CA6.dip0.t-ipconnect.de] has joined #shogun | 17:48 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 17:48 | |
thoralf | Hey Heiko | 17:49 |
@HeikoS | thoralf: hey! | 17:49 |
thoralf | HeikoS: Do you have a few minutes for discussing the Multilabel stuff? | 17:50 |
@HeikoS | thoralf: yeah, good point | 17:50 |
@HeikoS | just going through the 50 emails today ;) | 17:50 |
@HeikoS | but almost done by noe | 17:51 |
thoralf | lol | 17:51 |
@HeikoS | thoralf: ok 2 mins, I might have to run off at some point | 17:51 |
@HeikoS | thoralf: so about those arrays with dynamic length | 17:51 |
@HeikoS | thoralf: this is a problem we often encountered, so it would be good to solve it somehow | 17:51 |
@HeikoS | thoralf: because then things are serialisable etc | 17:52 |
@HeikoS | thoralf: I think in fact there might already be a solution | 17:52 |
@HeikoS | thoralf: this one also doesnt work for the modular interfaces | 17:52 |
thoralf | Okay, first point: this array of vectors. Can you give me a pointer? | 17:53 |
thoralf | Second point: About the modular interfaces, are you sure it doesn't work? | 17:53 |
thoralf | Or what would prevent it from working in modular interfaces? | 17:54 |
@HeikoS | thoralf: there is no datatype for those arrays | 17:59 |
@HeikoS | thoralf: we can only map data that is possible to represent as shogun types | 18:00 |
@HeikoS | thoralf: which means everything that you can register using SG_ADD | 18:00 |
@wiking | HeikoS: "The '''compute(SGMatrix)''' method should just be a convenience wrapper for linear operators" why? | 18:01 |
@wiking | HeikoS: since actually internally it's the other way round | 18:01 |
@wiking | HeikoS: at least in case of DirectEigenSolver... | 18:02 |
@HeikoS | wiking: ah for direct solvers it can be like that | 18:02 |
@HeikoS | wiking: but then the direct solver has another interface than the sparse one | 18:02 |
@HeikoS | wiking: mmh | 18:03 |
@HeikoS | maybe that should be like that even | 18:03 |
@HeikoS | dense and sparse are two quite different cases | 18:03 |
@HeikoS | wiking: for direct solvers, we assume that we can store the matrix | 18:03 |
@wiking | Heyeah | 18:03 |
@HeikoS | wiking: but still, a linear operator is a more general object than a matrix | 18:03 |
@wiking | yeah i got that | 18:03 |
@HeikoS | wiking: and even direct eigensolvers can be applied to operators which are not matrices | 18:03 |
@HeikoS | matrices <= linear operators | 18:03 |
@wiking | i mean i know that in case of an iterative solver it'll be better to have a linop | 18:03 |
@HeikoS | wiking: but thats all not really needed | 18:04 |
@wiking | HeikoS: yeah i've finished basic cs course back in the days of my bsc ;) | 18:04 |
@HeikoS | wiking: haha ;) | 18:04 |
@HeikoS | wiking: but maybe we should stick to conventions there | 18:04 |
@wiking | HeikoS: well if we stick to linops | 18:04 |
@HeikoS | wiking: which is: direct solvers take matrices, and sparse ones take sparse matrices or operators | 18:04 |
@HeikoS | wiking: thats how matlab and numpy do it | 18:05 |
@wiking | then only the users will have to do | 18:05 |
@wiking | linop(matrix) | 18:05 |
@wiking | which is of course a bit of a shit | 18:05 |
@wiking | that's why i've added those | 18:05 |
@wiking | HeikoS: i mean what i'm trying to say is that obviously if the user is an idiot | 18:05 |
@wiking | that we cannot prevent :) | 18:05 |
@HeikoS | wiking: yeah, its annoying if one actually *wants* to use a matrix, there should be methods for that, that hide all the linear operator crap | 18:06 |
@HeikoS | haha ;) | 18:06 |
@wiking | HeikoS: eyeah so i would go with | 18:06 |
@wiking | EigenSolver has all sorts of pure virtual functions | 18:06 |
@wiking | and then some of them are implemented and some not | 18:06 |
@wiking | in the given implementation | 18:06 |
@wiking | but of course we can follow more strict rules | 18:07 |
@wiking | of c++ | 18:07 |
@HeikoS | wiking: yeah | 18:07 |
@wiking | and then only introduce compute(linop) in EigenSolver | 18:07 |
@HeikoS | wiking: I mean we can have both right? | 18:07 |
@wiking | and only in classs where it actually makes sense to do compute(SGMatrix) we introduce that | 18:07 |
@HeikoS | the direct eigensolver just can have another method | 18:07 |
@wiking | HeikoS: yea hyeah sure | 18:08 |
@wiking | HeikoS: just that then sometimes you'll have do do dynamic_cast | 18:08 |
@HeikoS | wiking: yeah exactly, so if you know you are using a direct solver, you can use the compute (SGMatrix), but for things as lambday 's framework, we dont actually know, and need a common interface | 18:08 |
@wiking | when u actually want to call that method that is only supported in a given child of EigenSolver | 18:08 |
@HeikoS | wiking: yeah, but then you need to be sure what it is, which is fine | 18:08 |
-!- PirosB3 [~pirosb3@162.219.6.191] has quit [Quit: PirosB3] | 18:08 | |
@wiking | i feel the former solution is more 'standard' | 18:09 |
@HeikoS | wiking: prefer that a lot to having all methods in the base class and then get errors at runtime | 18:09 |
@wiking | do you? | 18:09 |
@wiking | i'm just trying to say that i dont :D | 18:09 |
@wiking | it's kind of awkward when there's an API but then you get a runtime error :S | 18:09 |
@wiking | at least for me | 18:10 |
@wiking | lisitsyn: ^ | 18:10 |
lisitsyn | damn | 18:10 |
@wiking | btw i think i cleared up some doxygen shit from the mid 2000s ;) | 18:10 |
@wiking | lisitsyn: verdict? | 18:10 |
lisitsyn | wiking: I don't really get what's you are discussing | 18:11 |
@wiking | lisitsyn: base abstract class has all the methods and then depending on the actual implementation some of those methods are not implemented vs abstract base class has only a few methods and then the inherited classes extends the api | 18:13 |
@HeikoS | wiking: I think we should capture stuff at compile time | 18:13 |
@wiking | in short | 18:13 |
lisitsyn | oh thanks | 18:13 |
lisitsyn | that's really better ;) | 18:13 |
@HeikoS | wiking: but its not the biggest deal on earth, more important to actually do this unification, however its done | 18:13 |
@HeikoS | wiking: but run-time errors scare people, in contrast to compile errors which scare ourselves ;) | 18:13 |
lisitsyn | why do we need to have these non implemented methods? | 18:13 |
@wiking | HeikoS: then the former one is the right way | 18:13 |
@wiking | i mean latter | 18:14 |
@wiking | fuck | 18:14 |
@wiking | :D | 18:14 |
@wiking | lisitsyn: dont have to .. this is a design decision | 18:14 |
@wiking | lisitsyn: you know one requires dynamic cast the other will just fucking fail runtime :) | 18:14 |
lisitsyn | yeah | 18:14 |
lisitsyn | wiking: what's the class we are talking about? | 18:15 |
@wiking | lisitsyn: EigenSolver | 18:15 |
@wiking | lisitsyn: MagicMachine2.0 | 18:15 |
@HeikoS | wiking: I dont have hard feelings about either way, just a slight preference. | 18:16 |
@HeikoS | wiking, lisitsyn how to deal with those datastructors of thoralf? | 18:16 |
@HeikoS | arrays of different sized vectors | 18:16 |
@wiking | HeikoS: std::vector? :) | 18:16 |
@HeikoS | lisitsyn: you had a similar problem once right? | 18:16 |
@wiking | or what? | 18:17 |
@HeikoS | wiking: and interfaces? | 18:17 |
lisitsyn | HeikoS: yeah iirc | 18:17 |
@HeikoS | wiking: how to pass this around? | 18:17 |
lisitsyn | HeikoS: actually fernando had it | 18:17 |
lisitsyn | and well I did some multilabel stuff as well | 18:17 |
@wiking | HeikoS: if it's std::vector then it's easy T* &vector[0] :) | 18:17 |
@wiking | HeikoS: gimmelinecode | 18:18 |
@HeikoS | wiking: https://github.com/shogun-toolbox/shogun/pull/1958 | 18:18 |
@wiking | can u tell me a file/line in the diff | 18:19 |
@HeikoS | yeah | 18:19 |
@wiking | HeikoS: https://github.com/shogun-toolbox/shogun/pull/1958/files#diff-45553468fba83f5a1e68a228853c922dR136 | 18:20 |
@HeikoS | wiking: line 75 of MultilabelLabels.cpp | 18:20 |
@wiking | ? | 18:20 |
@HeikoS | wiking: yeah all that stuff | 18:20 |
@HeikoS | wiking: I dont like creating c -arrays . we moved to SG* for a reason | 18:20 |
@wiking | oh yeah :D | 18:20 |
lisitsyn | argh we continue to do shoguncode | 18:21 |
@HeikoS | lisitsyn, wiking it would be great if you joined the discussion in this PR | 18:21 |
@wiking | HeikoS: DynArray? | 18:21 |
@HeikoS | since I dont really know how we should do all that stuff | 18:21 |
@wiking | :) | 18:21 |
@HeikoS | wiking: yeah that would be another solution | 18:21 |
lisitsyn | HeikoS: I am trying to follow this up | 18:21 |
@HeikoS | and having different dimensions right? | 18:21 |
@wiking | can actually shogun/base/DynArray.h hold SGVector? :D | 18:22 |
lisitsyn | dynarray sgvector | 18:22 |
lisitsyn | sorry it is fuck o'clock | 18:23 |
lisitsyn | :D | 18:23 |
@wiking | heheheh | 18:23 |
@wiking | mmmm | 18:23 |
@wiking | now i recall a moment | 18:23 |
@wiking | when i had this shit as well :DDD | 18:23 |
@wiking | but cannot recall where was it actually | 18:23 |
@HeikoS | wiking, lisitsyn gotta run off now, see you guys later | 18:25 |
@wiking | laterz | 18:26 |
-!- kprah [67157f4c@gateway/web/freenode/ip.103.21.127.76] has joined #shogun | 18:27 | |
-!- kprah_ [67157f4c@gateway/web/freenode/ip.103.21.127.76] has joined #shogun | 18:30 | |
-!- HeikoS [~heiko@p4FE97CA6.dip0.t-ipconnect.de] has quit [Ping timeout: 264 seconds] | 18:30 | |
kprah_ | wiking: where can I start? | 18:31 |
@wiking | kprah_: i reckon u did not see my message there | 18:31 |
@wiking | kprah_: r u familiar with bundle solvers? | 18:31 |
-!- kprah [67157f4c@gateway/web/freenode/ip.103.21.127.76] has quit [Ping timeout: 245 seconds] | 18:31 | |
-!- kprah__ [67157f4c@gateway/web/freenode/ip.103.21.127.76] has joined #shogun | 18:32 | |
@wiking | kprah_: https://github.com/shogun-toolbox/shogun/tree/develop/src/shogun/structure here are 3 different bundle solvers: libppbm, libbmrm and libp3bm they should share quite amount of code but they dont... so this could be a good start how to unify a bit the stuff there... | 18:33 |
-!- kprah___ [67157d4c@gateway/web/freenode/ip.103.21.125.76] has joined #shogun | 18:33 | |
kprah___ | i suppose yu are referring to solvers | 18:33 |
kprah___ | capable of solving multiple natured problems? | 18:34 |
-!- witness___ [uid10044@gateway/web/irccloud.com/x-qjlutmnzduulfvgk] has quit [Quit: Connection closed for inactivity] | 18:34 | |
kprah___ | ignore the connection problems please. | 18:34 |
@wiking | kprah_: i'm talking now about 3 given implementation in shogun | 18:34 |
@wiking | that should be worked on | 18:34 |
@wiking | check the code of those solvers | 18:35 |
@wiking | and see how you could unify them reducing code redundancy | 18:35 |
-!- kprah____ [67157f4c@gateway/web/freenode/ip.103.21.127.76] has joined #shogun | 18:35 | |
-!- kprah_ [67157f4c@gateway/web/freenode/ip.103.21.127.76] has quit [Ping timeout: 245 seconds] | 18:35 | |
kprah____ | wiking: am i right? | 18:35 |
@wiking | kprah____: have u seen my messages? | 18:36 |
@wiking | if not then i suggest i'll write an issue and u check it out there | 18:36 |
kprah____ | yeah, sure. But where am I supposed to find these issues? | 18:36 |
-!- kprah__ [67157f4c@gateway/web/freenode/ip.103.21.127.76] has quit [Ping timeout: 245 seconds] | 18:37 | |
@wiking | u suppose to solve this issue :) | 18:37 |
-!- kprah [67157f4c@gateway/web/freenode/ip.103.21.127.76] has joined #shogun | 18:37 | |
-!- kprah___ [67157d4c@gateway/web/freenode/ip.103.21.125.76] has quit [Ping timeout: 245 seconds] | 18:38 | |
kprah | wiking: sorry, but you didn't reply. | 18:39 |
@wiking | kprah: i think i did | 18:39 |
@wiking | maybe during your bouncing | 18:40 |
@wiking | you've missed some of the stuff i've written | 18:40 |
@wiking | anyways | 18:40 |
@wiking | here's an entrance task | 18:40 |
@wiking | https://github.com/shogun-toolbox/shogun/issues/1969 | 18:40 |
kprah | thanks | 18:40 |
@wiking | why the fuck the analyzers are not being fired...:S | 18:40 |
-!- kprah____ [67157f4c@gateway/web/freenode/ip.103.21.127.76] has quit [Ping timeout: 245 seconds] | 18:41 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting] | 18:43 | |
-!- shogun-buildbot_ [~shogun-bu@7nn.de] has joined #shogun | 18:43 | |
-!- kprah_ [67157e4c@gateway/web/freenode/ip.103.21.126.76] has joined #shogun | 18:43 | |
@wiking | shogun-buildbot_: force build --branch=develop 'debian wheezy - memcheck' | 18:43 |
shogun-buildbot_ | build #212 forced | 18:43 |
shogun-buildbot_ | I'll give a shout when the build finishes | 18:43 |
-!- kprah [67157f4c@gateway/web/freenode/ip.103.21.127.76] has quit [Ping timeout: 245 seconds] | 18:44 | |
-!- kprah_ [67157e4c@gateway/web/freenode/ip.103.21.126.76] has quit [Ping timeout: 245 seconds] | 18:47 | |
thoralf | wiking, lisitsyn: I missed how we can fix the issue Heiko had. | 18:53 |
thoralf | lisitsyn: Regarding the formatting issues: I used "bcpp" for this - If I fix this by-hand, then it will reverted the next time I call bcpp. | 18:54 |
lisitsyn | thoralf: no idea :D | 18:55 |
-!- KJainGSoC [783bc6ca@gateway/web/freenode/ip.120.59.198.202] has quit [Ping timeout: 245 seconds] | 19:09 | |
-!- thoralf [~thoralf@p4FEC89F4.dip0.t-ipconnect.de] has quit [Quit: Konversation terminated!] | 19:19 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 19:43 | |
-!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has joined #shogun | 19:44 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 19:44 | |
-!- newcastle [~newcastle@193.106.31.74] has joined #shogun | 19:46 | |
-!- cameo54321 [2729ef55@gateway/web/freenode/ip.39.41.239.85] has quit [Ping timeout: 245 seconds] | 19:54 | |
lisitsyn | besser82: going to install f20 in a few minutes ;) | 19:56 |
lisitsyn | you better hope I like it or I sue you!! :D | 19:56 |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 19:58 | |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * c6eb04c / doc/md/LICENSE_SVMlight.md: https://github.com/shogun-toolbox/shogun/commit/c6eb04c7c72b6f88a91cea501469663ff9eb4579 | 19:58 |
shogun-notifier- | shogun: remove german umlaut | 19:58 |
@sonney2k | lisitsyn, at least you no longer run around with your AK47 - besser you are pretty lucky ;) | 19:58 |
-!- newcastle [~newcastle@193.106.31.74] has quit [K-Lined] | 19:58 | |
lisitsyn | sonney2k: ohh welcome back :D | 19:59 |
lisitsyn | you are apparently pretty rare here | 19:59 |
@sonney2k | lisitsyn, you know I am just a senior manager now | 20:06 |
lisitsyn | sonney2k: no, I don't know :) | 20:06 |
@sonney2k | so I don't need to do any footwork ;) | 20:07 |
lisitsyn | sonney2k: yeah HeikoS and iglesiasg doing a lot of communication now | 20:07 |
@sonney2k | lisitsyn, I cannot say how glad I am that they / you all do this ;) | 20:09 |
-!- lisitsyn [~lisitsyn@109-226-71-150.clients.tlt.100megabit.ru] has quit [Quit: Leaving.] | 20:27 | |
-!- lisitsyn [~lisitsyn@109-226-71-150.clients.tlt.100megabit.ru] has joined #shogun | 20:41 | |
lisitsyn | besser82: can't control display brightness! :D | 20:41 |
lisitsyn | unacceptablee!!!!!! | 20:41 |
shogun-buildbot_ | build #212 of debian wheezy - memcheck is complete: Failure [failed memory check] Build details are at http://buildbot.shogun-toolbox.org/builders/debian%20wheezy%20-%20memcheck/builds/212 | 20:47 |
-!- sonne|osx [~sonne@e179064197.adsl.alicedsl.de] has quit [Quit: sonne|osx] | 20:48 | |
@iglesiasg | lisitsyn, apart from brightness, how does it feel? | 20:51 |
lisitsyn | iglesiasg: immediately rebooted back to ubuntu :D | 20:52 |
@iglesiasg | hahaha | 20:52 |
@wiking | :D | 20:59 |
@wiking | 20:40 < lisitsyn> besser82: can't control display brightness! :D | 20:59 |
@wiking | 20:41 < lisitsyn> unacceptablee!!!!!! | 20:59 |
@wiking | :D | 20:59 |
lisitsyn | wiking: as you like to say it is 2014 | 20:59 |
@wiking | well fuck it is | 21:00 |
lisitsyn | satellites are evolving around the earth | 21:00 |
lisitsyn | and I still can't control brightness :D | 21:01 |
@wiking | lisitsyn: brightness is overrated :D | 21:14 |
-!- tnosovich [~tnosovich@91.217.162.238] has joined #shogun | 21:37 | |
-!- tomy18 [~alex@141.85.0.106] has joined #shogun | 21:40 | |
-!- tomy18 [~alex@141.85.0.106] has quit [Quit: Konversation terminated!] | 21:49 | |
-!- thoralf [~thoralf@46.114.5.77] has joined #shogun | 22:14 | |
thoralf | Hey shogis. | 22:15 |
thoralf | https://github.com/shogun-toolbox/shogun/pull/1954 <-- can someone have a look on the PR of Piros? He gets double-frees -- My suspect is using "exit_shogun()", since we don't use it in any of the python modular scripts... | 22:26 |
thoralf | Any ideas? | 22:26 |
@iglesiasg | I am off for today thoralf, sorry | 22:31 |
@iglesiasg | too tired already | 22:31 |
thoralf | No problem. | 22:31 |
@iglesiasg | I will read it tomorrow if it is still unsolved | 22:31 |
thoralf | Just think if exit_shogun() is neccessary in python modular. ;) | 22:32 |
thoralf | Because I don't know. The rest is up to Piros. | 22:32 |
@iglesiasg | I have never seen/used it in a Shogun program from a modular interface | 22:32 |
@iglesiasg | so I don't think it is necessary at all | 22:32 |
@iglesiasg | good night! | 22:33 |
-!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has quit [Quit: Leaving] | 22:33 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 22:57 | |
-!- kislay [~kislay@93.114.45.13] has quit [Ping timeout: 240 seconds] | 23:07 | |
-!- thoralf [~thoralf@46.114.5.77] has quit [Ping timeout: 265 seconds] | 23:15 | |
shogun-buildbot_ | build #213 of debian wheezy - memcheck is complete: Failure [failed memory check] Build details are at http://buildbot.shogun-toolbox.org/builders/debian%20wheezy%20-%20memcheck/builds/213 blamelist: Soeren Sonnenburg <sonne@debian.org> | 23:20 |
-!- HeikoS [~heiko@ip-176-199-255-21.unitymediagroup.de] has joined #shogun | 23:37 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 23:37 | |
-!- HeikoS [~heiko@ip-176-199-255-21.unitymediagroup.de] has quit [Client Quit] | 23:38 | |
-!- HeikoS [~heiko@ip-176-199-255-21.unitymediagroup.de] has joined #shogun | 23:39 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 23:40 | |
-!- lambday [67157f54@gateway/web/freenode/ip.103.21.127.84] has joined #shogun | 23:44 | |
lambday | HeikoS: hi! | 23:47 |
@HeikoS | lambday: heya | 23:48 |
lambday | HeikoS: ummm I have a doubt | 23:48 |
lambday | HeikoS: in the B-test paper, | 23:48 |
@HeikoS | lambday: yeah shoot | 23:49 |
lambday | they show that the mmd estimates converges to Gaussian - for both under alternative hypothesis (eq. 7, similar to the one for linear time mmd) and null hypothesis (eq. 10, similar to the one when we compute centered kernel matrices eigenvalues) | 23:50 |
lambday | HeikoS: which one should we have? | 23:50 |
@HeikoS | let me check | 23:50 |
@HeikoS | lambday: null distribution | 23:55 |
@HeikoS | thats the one we want | 23:55 |
@HeikoS | so 10 | 23:55 |
@HeikoS | the one with zero mean | 23:55 |
lambday | HeikoS: umm... but then we have to compute the whole kernel matrix, no? | 23:55 |
lambday | I mean, for streaming data | 23:56 |
lambday | it would be problematic | 23:56 |
@HeikoS | should not | 23:56 |
@HeikoS | why do you think that? | 23:56 |
@HeikoS | just checking what C is | 23:56 |
lambday | just above eq 10 | 23:57 |
@HeikoS | lambday: ah I see. so wait | 23:57 |
@HeikoS | these are population expressions | 23:57 |
@HeikoS | but we can estimate them in linear time | 23:57 |
@HeikoS | that what we have done in the kernel selection paper | 23:57 |
@HeikoS | there is a proof that if one constructs a test with a linear time estimate for the variance, this is correct | 23:57 |
@HeikoS | you dont need to udnerstand the proof (in fact, I dont really ;) | 23:57 |
@HeikoS | lambday: but the estimation is very similar | 23:58 |
@HeikoS | lambday: just an online variance estimate of the terms that are averaged over | 23:58 |
lambday | HeikoS: similar to the way linear mmd does it? estimate variance | 23:58 |
@HeikoS | lambday: so the code should be almost the same as for the linear time mmd | 23:58 |
@HeikoS | yeah | 23:58 |
lambday | so that fits eq 7, doens't it | 23:58 |
lambday | oh no | 23:58 |
lambday | the mean is different | 23:58 |
lambday | there we have 0 mean | 23:58 |
@HeikoS | lambday: how do they do it in their codes? constructing the threshold | 23:59 |
@HeikoS | lambday: wait ill check | 23:59 |
lambday | HeikoS: var_est = (1 / m2) * cov(hh') | 23:59 |
--- Log closed Mon Mar 10 00:00:05 2014 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!