| --- Log opened Fri May 03 00:00:30 2013 | ||
| -!- cameron__ [~quassel@global-2-11.nat.csx.cam.ac.uk] has joined #shogun | 00:38 | |
| -!- cameron__ [~quassel@global-2-11.nat.csx.cam.ac.uk] has quit [Ping timeout: 276 seconds] | 01:25 | |
| -!- cameron__ [~quassel@global-2-11.nat.csx.cam.ac.uk] has joined #shogun | 01:28 | |
| -!- cameron__ [~quassel@global-2-11.nat.csx.cam.ac.uk] has quit [Remote host closed the connection] | 01:30 | |
| -!- sids_aquarius [~sidi_@li400-124.members.linode.com] has joined #shogun | 01:31 | |
| -!- sids_aquarius [~sidi_@li400-124.members.linode.com] has left #shogun [] | 01:34 | |
| -!- iglesiasg [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 01:55 | |
| -!- cesarfm [b2c73bf3@gateway/web/freenode/ip.178.199.59.243] has joined #shogun | 03:13 | |
| cesarfm | Hello, I made another pull request for the Group lasso example, now in the develop branch | 03:16 |
|---|---|---|
| -!- rgirish28 [~rgirish28@61.17.226.13] has joined #shogun | 04:11 | |
| shogun-buildbot | build #378 of nightly_default is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/378 | 04:11 |
| -!- rgirish28 [~rgirish28@61.17.226.13] has left #shogun [] | 04:14 | |
| -!- cesarfm [b2c73bf3@gateway/web/freenode/ip.178.199.59.243] has quit [Ping timeout: 245 seconds] | 04:37 | |
| -!- rgirish28 [~rgirish28@61.17.226.13] has joined #shogun | 04:37 | |
| -!- rgirish28 [~rgirish28@61.17.226.13] has quit [Ping timeout: 276 seconds] | 06:54 | |
| -!- hushell [43bd6474@gateway/web/freenode/ip.67.189.100.116] has joined #shogun | 07:07 | |
| -!- rgirish28 [~rgirish28@111.92.71.249] has joined #shogun | 07:13 | |
| -!- rgirish28 [~rgirish28@111.92.71.249] has quit [Quit: rgirish28] | 07:56 | |
| -!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has joined #shogun | 08:02 | |
| @sonney2k | wiking, pong again :) | 08:20 |
| @sonney2k | lisitsyn, happy? 34 now :) | 08:20 |
| gsomix | good morning | 08:55 |
| -!- iglesiasg [~nando@n165-p239.kthopen.kth.se] has joined #shogun | 09:03 | |
| -!- mode/#shogun [+o iglesiasg] by ChanServ | 09:04 | |
| -!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 09:07 | |
| sonne|work | gsomix: morning! how is your csv reader doing :) | 09:18 |
| gsomix | sonne|work, hm, I'll do it little later - doing some python3 related stuff | 09:20 |
| sonne|work | ahh ok | 09:20 |
| -!- hushell [43bd6474@gateway/web/freenode/ip.67.189.100.116] has quit [Ping timeout: 245 seconds] | 09:21 | |
| gsomix | sonne|work, btw, about csv. should I add it into shogun at once? or just develop separately? | 09:25 |
| sonne|work | gsomix: well a line reader function first in CFile or CAsciiFile | 09:26 |
| sonne|work | but you can do as you wish | 09:26 |
| gsomix | ok | 09:26 |
| -!- hushell [43bd6474@gateway/web/freenode/ip.67.189.100.116] has joined #shogun | 09:35 | |
| -!- makokal [~bokal@pD9EF09DC.dip0.t-ipconnect.de] has quit [Quit: makokal] | 09:39 | |
| -!- monalisag [6eeaec2c@gateway/web/freenode/ip.110.234.236.44] has joined #shogun | 09:43 | |
| monalisag | Hello, I wanted to know some details about the project Large Scale Learning & Dimension reduction using hashing. | 09:45 |
| monalisag | In the project descrption it isnt mentioned but is this method anyway related to locality sensitive hashing ? | 09:45 |
| -!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 10:01 | |
| -!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 10:14 | |
| sonne|work | monalisag: a little bit maybe. | 10:15 |
| sonne|work | monalisag: it is not about designing a hash though | 10:18 |
| sonne|work | but more to just use a fast one to reduce #dims | 10:18 |
| sonne|work | and one is totally ignorant about hash collisions etc\ | 10:19 |
| monalisag | part of my work for this project would be to explore the kind of hashing function we can use ? | 10:23 |
| monalisag | or i have to use the already implemented CHash class for this purpose ? | 10:23 |
| sonne|work | monalisag: it would only involve using the hash function 'MurmurHash3' | 10:24 |
| sonne|work | which is already in CHash yes | 10:24 |
| sonne|work | monalisag: it is more about defining certain type of DotFeature derived classes that use this compression | 10:25 |
| sonne|work | e.g. you could think of n-grams and bag of words (text documents) that you could hash | 10:25 |
| sonne|work | or e.g. quadratic (or higher order) polynomial features | 10:26 |
| sonne|work | monalisag: so one can do real large scale *non-linear* learning by computing features on-the-fly | 10:27 |
| sonne|work | for example I have used this to detect splice sites on DNA | 10:27 |
| monalisag | got it. thanks :) | 10:28 |
| monalisag | one more thing. Should I look into Vowpal Wabbit and see how they use this on-the-fly hashing trick ? | 10:29 |
| sonne|work | so one could do fancy applications with it w/o requiring lots of memory etc | 10:29 |
| sonne|work | monalisag: no need to do this | 10:29 |
| sonne|work | shogun does this too | 10:29 |
| sonne|work | as soon as you implement the functions in CDotFeatures | 10:29 |
| sonne|work | vw & liblinear & ocas & sgd etc will use the on-the-fly features | 10:30 |
| monalisag | for the tokenizing purposes can i take inspiration from scikit or some other library which performs similar job ? | 10:35 |
| -!- hushell [43bd6474@gateway/web/freenode/ip.67.189.100.116] has quit [Quit: Page closed] | 10:36 | |
| -!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has quit [Quit: Leaving] | 10:39 | |
| sonne|work | monalisag: it has to be C/C++ code though. Recall that this is called billions of times | 10:51 |
| monalisag | I would be implementing the code myself but I would like to explore how they(other libraries) have made their code efficient in cases where the text corpuses are huge. | 10:52 |
| sonne|work | monalisag: sure. It is not limited to text though. But we will have to focus on a few domains otherwise this is getting too much. | 10:55 |
| sonne|work | maybe with some nice application | 10:55 |
| -!- makokal [~bokal@eduroam-pool3-097.wlan.uni-bremen.de] has joined #shogun | 10:57 | |
| -!- cameron__ [~quassel@global-2-1.nat.csx.cam.ac.uk] has joined #shogun | 11:15 | |
| -!- thoralf [~thoralf@enki.zib.de] has joined #shogun | 11:24 | |
| thoralf | Hey there. | 11:24 |
| sonne|work | thoralf: hey | 11:25 |
| thoralf | sonne|work: I just committed my hack - but I did not check if it has side effects. | 11:26 |
| thoralf | sonne|work: Is there some kind of unit or regression testing? | 11:26 |
| sonne|work | thoralf: yes. Might be though that we don't have any for streaming features | 11:26 |
| sonne|work | in any case this will either work or crash(tm) | 11:27 |
| thoralf | And how can someone run the tests locally? I didn't find anything in the documentation... | 11:30 |
| -!- cameron__ [~quassel@global-2-1.nat.csx.cam.ac.uk] has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.] | 11:31 | |
| -!- goldbug [~quassel@global-2-1.nat.csx.cam.ac.uk] has joined #shogun | 11:31 | |
| -!- goldbug [~quassel@global-2-1.nat.csx.cam.ac.uk] has quit [Client Quit] | 11:32 | |
| sonne|work | thoralf: make tests | 11:34 |
| sonne|work | thoralf: and they are run for your PR too | 11:34 |
| sonne|work | thoralf: https://travis-ci.org/shogun-toolbox/shogun/builds/6846696 | 11:34 |
| -!- abinash [01261262@gateway/web/freenode/ip.1.38.18.98] has joined #shogun | 11:36 | |
| -!- abinash [01261262@gateway/web/freenode/ip.1.38.18.98] has quit [Ping timeout: 245 seconds] | 11:40 | |
| monalisag | hey, could you redirect me to a source where I can read about quadritic feature hashing ? | 11:47 |
| sonne|work | monalisag: it is a one-liner: hash(idx) and idx follows some enumeration of the features, e.g. when you have x_i i=0,1,2 you could have x_0, x_1, x_3 and x_0 x_1, x_0 x_2, x_1 x_2, x_1 x_3 | 11:50 |
| sonne|work | scratch the last :) there is no x_3 :) | 11:51 |
| -!- goldbug [~quassel@global-2-1.nat.csx.cam.ac.uk] has joined #shogun | 11:57 | |
| -!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has joined #shogun | 12:00 | |
| -!- goldbug [~quassel@global-2-1.nat.csx.cam.ac.uk] has quit [Remote host closed the connection] | 12:08 | |
| -!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 12:14 | |
| -!- sijin [~smuxi@144.214.222.109] has quit [Read error: Connection reset by peer] | 12:19 | |
| monalisag | sonne|work: will the text tokenizer and classifier code be added to shogun or it would just used as a use case for testing the implemenation of CHashedNGramDotFeatures ? | 12:21 |
| -!- goldbug [~quassel@global-2-1.nat.csx.cam.ac.uk] has joined #shogun | 12:31 | |
| -!- iglesiasg [~nando@n165-p239.kthopen.kth.se] has quit [Ping timeout: 264 seconds] | 12:35 | |
| goldbug | just submitted my gsoc proposal | 12:40 |
| goldbug | :) | 12:40 |
| -!- goldbug [~quassel@global-2-1.nat.csx.cam.ac.uk] has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.] | 12:48 | |
| -!- goldbug [~quassel@global-2-1.nat.csx.cam.ac.uk] has joined #shogun | 12:48 | |
| -!- iglesiasg [~nando@n165-p239.kthopen.kth.se] has joined #shogun | 12:49 | |
| -!- mode/#shogun [+o iglesiasg] by ChanServ | 12:49 | |
| -!- erlenda [~erlenda@fw-oslo.intelcom.no] has joined #shogun | 12:51 | |
| -!- van51 [~van51@athedsl-401908.home.otenet.gr] has joined #shogun | 12:59 | |
| -!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has quit [Quit: Leaving] | 13:08 | |
| -!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 13:10 | |
| @wiking | sonne|work: ping | 13:13 |
| monalisag | sonne|work: I have submitted my proposal. Please have a look. | 13:15 |
| sonne|work | wiking: do you have a python3 machine? | 13:39 |
| @iglesiasg | sonne|work: python3 working here if I might be able to help :) | 13:40 |
| @wiking | sonne|work: sorry no | 13:41 |
| sonne|work | iglesiasg: we need some python3 enabled buildbot | 13:41 |
| @iglesiasg | sonne|work: aah ok I see, I meant my laptop so I don't think we should use it as buildbot :) | 13:41 |
| -!- monalisag [6eeaec2c@gateway/web/freenode/ip.110.234.236.44] has quit [Quit: Page closed] | 14:13 | |
| -!- ozan [~ozan@ee-167-250.eee.metu.edu.tr] has joined #shogun | 14:25 | |
| ozan | I am currently writing a proposal for GSOC 2013 | 14:27 |
| ozan | already submitted a patch today https://github.com/shogun-toolbox/shogun/pull/1046 | 14:27 |
| ozan | and contacted with potential mentor (Vojtech) | 14:28 |
| ozan | is there anything else I need to do after submitting my proposal to GSOC 2013 | 14:28 |
| @iglesiasg | ozan: I think your PR is against master, isn't it? | 14:29 |
| ozan | yes should I change ? | 14:29 |
| ozan | is there any specific branch for that | 14:29 |
| -!- goldbug [~quassel@global-2-1.nat.csx.cam.ac.uk] has quit [Ping timeout: 256 seconds] | 14:30 | |
| @iglesiasg | ozan: see the comment here for example https://github.com/shogun-toolbox/shogun/pull/1044 | 14:31 |
| ozan | and is there any other QDA example/application that I can test | 14:31 |
| @iglesiasg | ozan: there ara QDA examples in examples/undocumented/python_modular and libshogun too | 14:31 |
| ozan | I already used them | 14:32 |
| ozan | ok I am fixing | 14:32 |
| @iglesiasg | ozan: I took a quick look to your PR | 14:32 |
| @iglesiasg | ozan: one comment is that the idea of using Eigen was not because of using their structures but mainly to avoid loops in the computations and so on | 14:33 |
| @iglesiasg | ozan: maybe I am wrong, but it looks to me like you have kept most of the loops | 14:33 |
| ozan | except the one of them yes I keep them all | 14:33 |
| ozan | issue is that I first wanted to get working example | 14:34 |
| @iglesiasg | ozan: sure, good idea | 14:34 |
| ozan | after finishing my application to GSOC 2013 I am planning to solve that loops | 14:34 |
| ozan | since deadline is today I wanted to have working patch :) | 14:34 |
| @iglesiasg | ok | 14:34 |
| ozan | I think I can solve loops and prepare a new pull request next week | 14:35 |
| -!- van51 [~van51@athedsl-401908.home.otenet.gr] has quit [Quit: Leaving.] | 14:35 | |
| @iglesiasg | ozan: aham ok | 14:36 |
| sonne|work | iglesiasg: hmmhh 37 already | 14:36 |
| @iglesiasg | ozan: there is also another PR someone sent with Eigen for QDA, you might want to take a look | 14:37 |
| @iglesiasg | sonne|work: how many do you guess at the end of the day? | 14:37 |
| sonne|work | I was estimating 30 | 14:37 |
| sonne|work | and yesterday 35 | 14:37 |
| @iglesiasg | haha ok no more estimates then | 14:37 |
| sonne|work | now maybe 40 | 14:38 |
| @iglesiasg | I think 40 for sure | 14:38 |
| sonne|work | IDK | 14:38 |
| sonne|work | we had 48 last year | 14:38 |
| sonne|work | so almost no difference then | 14:38 |
| @iglesiasg | good | 14:38 |
| ozan | I have not seen that PR yet. I can merge my code with the PR you mentioned | 14:38 |
| @iglesiasg | sonne|work: it would be pretty good if we get the same #slots | 14:38 |
| sonne|work | yeah but we cannot count on it | 14:39 |
| @iglesiasg | ozan: https://github.com/shogun-toolbox/shogun/pull/1028 | 14:39 |
| @iglesiasg | sonne|work: we shall see | 14:39 |
| -!- goldbug [~quassel@global-2-11.nat.csx.cam.ac.uk] has joined #shogun | 14:41 | |
| ozan | I closed my PR on master request | 14:41 |
| -!- nube [~rho@49.244.7.158] has joined #shogun | 14:43 | |
| ozan | Finally submitted my propsal :) | 15:00 |
| ozan | I have another question should I be enrolled as a student during GSOC | 15:00 |
| ozan | ? | 15:00 |
| ozan | I am finishing my master in may and starting my PhD in september. Is it a problem ? | 15:00 |
| @iglesiasg | ozan: I think it is better if you check GSoC's FAQ or the many mails that have been regarding that topic in GSoC's ML | 15:01 |
| -!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has joined #shogun | 15:22 | |
| -!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 15:32 | |
| shogun-notifier- | shogun: Thoralf Klein :master * 875e256 / src/shogun/features/streaming/StreamingSparseFeatures.cpp: https://github.com/shogun-toolbox/shogun/commit/875e2565b13bf6cb6fba0b5ab1272b968d0c79eb | 15:32 |
| shogun-notifier- | shogun: Fixes #660: Hack to fix double-free (simply setting features to NULL). | 15:32 |
| shogun-notifier- | shogun: Soeren Sonnenburg :master * 3a6d04e / src/shogun/features/streaming/StreamingSparseFeatures.cpp: https://github.com/shogun-toolbox/shogun/commit/3a6d04ea7af155e0b0779ef825a6f4f4d336ef3a | 15:32 |
| shogun-notifier- | shogun: Merge pull request #1047 from tklein23/master | 15:32 |
| shogun-notifier- | shogun: | 15:32 |
| shogun-notifier- | shogun: Small hack to fix double-free in StreamingSparseFeatures | 15:32 |
| -!- monalisag [6eeaec2c@gateway/web/freenode/ip.110.234.236.44] has joined #shogun | 15:33 | |
| shogun-notifier- | shogun: Evgeniy Andreev :develop * 21a894f / src/interfaces/python_modular/protocols_helper.i: https://github.com/shogun-toolbox/shogun/commit/21a894fd2605d8f262a02482ebe577837d7f78a6 | 15:34 |
| shogun-notifier- | shogun: fixed compilation with python3 bindings | 15:34 |
| shogun-notifier- | shogun: Evgeniy Andreev :develop * bc3e9ba / / (20 files): https://github.com/shogun-toolbox/shogun/commit/bc3e9ba2d971d0372730478689ceef2b180e1d54 | 15:34 |
| shogun-notifier- | shogun: small fixes in examples for python3 compatibility | 15:34 |
| shogun-notifier- | shogun: Soeren Sonnenburg :develop * 2a04b18 / / (21 files): https://github.com/shogun-toolbox/shogun/commit/2a04b184641fd16f21e1b1b119a3f182390c03d9 | 15:34 |
| shogun-notifier- | shogun: Merge pull request #1048 from gsomix/feature/python3_bugfix | 15:34 |
| shogun-notifier- | shogun: | 15:34 |
| shogun-notifier- | shogun: Pack of fixes for python3 modular interface | 15:34 |
| -!- monalisag [6eeaec2c@gateway/web/freenode/ip.110.234.236.44] has quit [Client Quit] | 15:36 | |
| shogun-buildbot | build #904 of bsd1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/904 blamelist: Thoralf Klein <thoralf.klein@zib.de> | 15:38 |
| -!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has quit [Quit: Leaving] | 15:43 | |
| @lisitsyn | sonne|work: why do these assholes want our domains? | 15:44 |
| sonne|work | thoralf: argh my bad - next time please do the PR against the development branch | 15:44 |
| sonne|work | lisitsyn: strictly speaking they are not ours | 15:44 |
| sonne|work | but everyone can get them | 15:44 |
| @lisitsyn | sonne|work: yes of course | 15:45 |
| @lisitsyn | sonne|work: but why? | 15:45 |
| -!- monalisag [6eeaec2c@gateway/web/freenode/ip.110.234.236.44] has joined #shogun | 15:45 | |
| shogun-buildbot | build #905 of bsd1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/905 | 15:45 |
| shogun-buildbot | build #906 of bsd1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/906 blamelist: Soeren Sonnenburg <sonne@debian.org> | 15:46 |
| -!- monalisag_ [6eeaec2c@gateway/web/freenode/ip.110.234.236.44] has joined #shogun | 15:46 | |
| -!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has joined #shogun | 15:47 | |
| @lisitsyn | sonne|work: and we've got 40! | 15:48 |
| @lisitsyn | :D | 15:48 |
| sonne|work | already? | 15:49 |
| sonne|work | hmmhh | 15:49 |
| sonne|work | so number is about the same then | 15:49 |
| sonne|work | 83% | 15:49 |
| shogun-buildbot | build #739 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/739 | 15:50 |
| shogun-buildbot | build #740 of cyg1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/740 blamelist: Evgeniy Andreev <gsomix@gmail.com>, Soeren Sonnenburg <sonne@debian.org> | 15:50 |
| -!- cesarfm [528247df@gateway/web/freenode/ip.82.130.71.223] has joined #shogun | 15:52 | |
| gsomix | shogun-buildbot, I love you too, dear | 15:53 |
| gsomix | lisitsyn, sonne|work 5 hours, 6 minutes remaining. there is time to write another proposal... and one more... and | 15:54 |
| sonne|work | gsomix: cyg1 is in berserk mode | 15:55 |
| -!- monalisag [6eeaec2c@gateway/web/freenode/ip.110.234.236.44] has quit [Quit: Page closed] | 15:57 | |
| -!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 15:57 | |
| shogun-buildbot | build #741 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/741 | 16:04 |
| -!- iglesiasg [~nando@n165-p239.kthopen.kth.se] has quit [Quit: leaving] | 16:10 | |
| -!- monalisag_ [6eeaec2c@gateway/web/freenode/ip.110.234.236.44] has quit [Quit: Page closed] | 16:10 | |
| shogun-buildbot | build #1039 of deb3 - modular_interfaces is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1039 | 16:14 |
| -!- mikhailBelous [~quassel@37.98.242.99] has joined #shogun | 16:15 | |
| gsomix | shogun-buildbot, goow work | 16:17 |
| gsomix | *good | 16:17 |
| -!- erlenda [~erlenda@fw-oslo.intelcom.no] has quit [Quit: Leaving] | 16:18 | |
| @lisitsyn | sonne|work: and 41 | 16:22 |
| sonne|work | I would prefer 42 | 16:22 |
| sonne|work | but ok qpsize in SVMLight is set to 41 | 16:22 |
| sonne|work | simply because of some theorem that could prove convergence for odd numbers :D | 16:23 |
| @wiking | sonne|work: ok got a sec? | 16:27 |
| sonne|work | sec yes | 16:27 |
| @wiking | sonne|work: ok | 16:27 |
| @wiking | sonne|work: what should we do with the PRNG try | 16:28 |
| @wiking | keep it or drop it? | 16:28 |
| sonne|work | test it and merge it | 16:28 |
| @wiking | since it did not really fix our integration problem :( | 16:28 |
| sonne|work | not all of it | 16:28 |
| @wiking | sonne|work: ok then before doing that i'll add dSFMT | 16:28 |
| @wiking | so that we can generate random float vectors fast | 16:28 |
| sonne|work | but even having a re-entrant rng is very worth having! | 16:28 |
| @wiking | my only concern is | 16:28 |
| @wiking | that for each random generation whether it's only a number or an array of randoms | 16:29 |
| @wiking | we have an SG_REF/SG_UNREF pair | 16:29 |
| @wiking | which is a bit of an overhead | 16:29 |
| sonne|work | wiking: btw could you please merge master into develop? I accidentally merged one commit into master... | 16:29 |
| @wiking | maybe i should just use it w/o the get_global_rand() and define it as extern | 16:29 |
| @wiking | sonne|work: ok i'll do that as soon as i have some time today | 16:30 |
| @wiking | there's some other stuff on my todo list | 16:30 |
| sonne|work | wiking: yeah that is sth we cannot afford | 16:30 |
| @wiking | like get a 10 minutes sleep :) | 16:30 |
| @wiking | sonne|work: you mean SG_REF/UNREF ? | 16:30 |
| sonne|work | c'mon sleep is overrated | 16:30 |
| @wiking | sonne|work: haha lol yeah | 16:30 |
| sonne|work | yes it should need no ref/unref | 16:30 |
| @wiking | ok | 16:30 |
| @wiking | i'll fix that one | 16:30 |
| @wiking | and as u suggested | 16:31 |
| -!- mikhailBelous [~quassel@37.98.242.99] has quit [Remote host closed the connection] | 16:31 | |
| @wiking | i'll add a non-thread safe api as well | 16:31 |
| sonne|work | yes | 16:31 |
| @wiking | oooor | 16:31 |
| @wiking | i mean to be honest | 16:31 |
| @wiking | we could have 2 SFMT structs in the clas | 16:31 |
| @wiking | class | 16:31 |
| @wiking | and then we dont need to reinit it | 16:31 |
| @wiking | it's not thaaaat big of a struct that it would start eating the memory like crazy | 16:32 |
| @wiking | :P | 16:32 |
| @wiking | 1 sfmt for 32 bits and 1 sfmt for 64 bits | 16:32 |
| @wiking | sonne|work: your 2 cents on this one? | 16:32 |
| @wiking | basically sfmt_32 and sfmt_64. inited once and that's all | 16:35 |
| @wiking | and have dsfmt for the doubles | 16:35 |
| sonne|work | yeah why not. but you would init them with the same seed for the non-thread safe api | 16:35 |
| @wiking | sonne|work: yep | 16:36 |
| @wiking | everything the same | 16:36 |
| @wiking | just that we dont have to reinit each time when we change state | 16:36 |
| -!- nube [~rho@49.244.7.158] has quit [Quit: Leaving.] | 16:37 | |
| shogun-buildbot | build #1040 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/1040 blamelist: Evgeniy Andreev <gsomix@gmail.com>, Soeren Sonnenburg <sonne@debian.org> | 16:37 |
| sonne|work | wiking: fine with me. if possible lets have one way to seed all 3 rng's | 16:37 |
| sonne|work | the int/float/double one | 16:37 |
| sonne|work | I recall hunting bugs when still doing matlab for hours | 16:38 |
| @wiking | yep | 16:38 |
| sonne|work | until I figured that randn and rand have a different init/seed | 16:38 |
| sonne|work | and everyone who knows better can still use the thread-safe functions | 16:39 |
| sonne|work | though we then have a hard time doing anything reproducible | 16:39 |
| @lisitsyn | sonne|work: 42 | 16:41 |
| -!- nube [~rho@49.244.7.158] has joined #shogun | 16:41 | |
| sonne|work | now lads&gents please leave it at that :D | 16:43 |
| -!- gsomix_ [~gsomix@88.200.243.155] has joined #shogun | 16:47 | |
| -!- gsomix [~gsomix@95.67.166.72] has quit [Read error: Connection reset by peer] | 16:47 | |
| -!- gsomix_ is now known as gsomix | 16:47 | |
| -!- goldbug [~quassel@global-2-11.nat.csx.cam.ac.uk] has quit [Ping timeout: 245 seconds] | 16:49 | |
| @wiking | :> | 16:53 |
| @lisitsyn | sonne|work: 43 | 16:53 |
| * sonne|work senses spam | 16:55 | |
| -!- sonne|work [~sonnenbu@sams-office-nat.tomtomgroup.com] has quit [Quit: Leaving.] | 16:55 | |
| shogun-buildbot | build #1041 of deb3 - modular_interfaces is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1041 | 16:55 |
| -!- goldbug [~quassel@global-2-11.nat.csx.cam.ac.uk] has joined #shogun | 16:56 | |
| -!- mona2302 [4fbdb06e@gateway/web/freenode/ip.79.189.176.110] has joined #shogun | 17:01 | |
| -!- nube [~rho@49.244.7.158] has quit [Quit: Leaving.] | 17:04 | |
| shogun-notifier- | shogun: Thoralf Klein :develop * 875e256 / src/shogun/features/streaming/StreamingSparseFeatures.cpp: https://github.com/shogun-toolbox/shogun/commit/875e2565b13bf6cb6fba0b5ab1272b968d0c79eb | 17:16 |
| shogun-notifier- | shogun: Fixes #660: Hack to fix double-free (simply setting features to NULL). | 17:16 |
| shogun-notifier- | shogun: Soeren Sonnenburg :develop * 3a6d04e / src/shogun/features/streaming/StreamingSparseFeatures.cpp: https://github.com/shogun-toolbox/shogun/commit/3a6d04ea7af155e0b0779ef825a6f4f4d336ef3a | 17:16 |
| shogun-notifier- | shogun: Merge pull request #1047 from tklein23/master | 17:16 |
| shogun-notifier- | shogun: | 17:16 |
| shogun-notifier- | shogun: Small hack to fix double-free in StreamingSparseFeatures | 17:16 |
| shogun-notifier- | shogun: Soeren Sonnenburg :develop * e97bd41 / src/shogun/features/streaming/StreamingSparseFeatures.cpp: https://github.com/shogun-toolbox/shogun/commit/e97bd4190016fbfab6916221630394294c531ba8 | 17:16 |
| shogun-notifier- | shogun: merge accidental commit to master containing streamingfeatures fix into develop | 17:16 |
| -!- goldbug [~quassel@global-2-11.nat.csx.cam.ac.uk] has quit [Ping timeout: 255 seconds] | 17:17 | |
| shogun-buildbot | build #907 of bsd1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/907 blamelist: Thoralf Klein <thoralf.klein@zib.de> | 17:18 |
| shogun-buildbot | build #742 of cyg1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/742 blamelist: Thoralf Klein <thoralf.klein@zib.de> | 17:19 |
| -!- ozan [~ozan@ee-167-250.eee.metu.edu.tr] has quit [Quit: Leaving] | 17:26 | |
| shogun-buildbot | build #743 of cyg1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/743 blamelist: Soeren Sonnenburg <sonne@debian.org> | 17:27 |
| shogun-buildbot | build #908 of bsd1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/908 | 17:32 |
| -!- monalisag [6eeaec2c@gateway/web/freenode/ip.110.234.236.44] has joined #shogun | 17:33 | |
| -!- makokal [~bokal@eduroam-pool3-097.wlan.uni-bremen.de] has quit [Quit: makokal] | 17:34 | |
| -!- monalisag [6eeaec2c@gateway/web/freenode/ip.110.234.236.44] has quit [Client Quit] | 17:36 | |
| -!- Hao [d8a55f4d@gateway/web/freenode/ip.216.165.95.77] has joined #shogun | 17:44 | |
| Hao | This is a report of "Set yourself up" part in proposal of GSOC 2013. Name: Hao Liu. Email: haoliu@nyu.edu | 17:46 |
| gsomix | Hao, hi | 17:49 |
| Hao | Hi:-) | 17:50 |
| Hao | I am interested in Large Scale Learning: loglinear time kernel expansion (aka Fast Food) project. I am still preparing my proposal, hope it's not too late. | 17:54 |
| -!- mikhailBelous [~quassel@37.98.242.99] has joined #shogun | 17:59 | |
| gsomix | Hao, so, there are 3 hours before deadline. :) | 18:01 |
| gsomix | If you have any questions about project, you should contact sonney2k directly - it's Soeren Sonnenburg. | 18:01 |
| -!- goldbug [~quassel@global-2-11.nat.csx.cam.ac.uk] has joined #shogun | 18:01 | |
| mona2302 | Hi, I'm a beginner user of Shogun Can I maybe create some R-program for my entry task? I want to answer for one of entry tasks. | 18:06 |
| @lisitsyn | 45 | 18:10 |
| @lisitsyn | 46 | 18:15 |
| shogun-buildbot | build #1043 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/1043 blamelist: Soeren Sonnenburg <sonne@debian.org> | 18:16 |
| gsomix | shogun-buildbot, dance | 18:17 |
| shogun-buildbot | <(^.^<) | 18:17 |
| shogun-buildbot | <(^.^)> | 18:17 |
| shogun-buildbot | (>^.^)> | 18:17 |
| shogun-buildbot | (7^.^)7 | 18:17 |
| shogun-buildbot | (>^.^<) | 18:17 |
| gsomix | cool | 18:17 |
| @wiking | :> | 18:18 |
| -!- van51 [~van51@athedsl-401908.home.otenet.gr] has joined #shogun | 18:23 | |
| Hao | let me try this. | 18:24 |
| Hao | shogun-buildbot, dance | 18:24 |
| -!- Hao [d8a55f4d@gateway/web/freenode/ip.216.165.95.77] has quit [Quit: Page closed] | 18:25 | |
| -!- Hao [d8a55f4a@gateway/web/freenode/ip.216.165.95.74] has joined #shogun | 18:26 | |
| -!- mona2302 [4fbdb06e@gateway/web/freenode/ip.79.189.176.110] has quit [Ping timeout: 245 seconds] | 18:28 | |
| -!- nube [~rho@49.244.66.144] has joined #shogun | 18:35 | |
| -!- van511 [~van51@athedsl-318776.home.otenet.gr] has joined #shogun | 18:36 | |
| @lisitsyn | 47 | 18:37 |
| -!- van51 [~van51@athedsl-401908.home.otenet.gr] has quit [Ping timeout: 252 seconds] | 18:38 | |
| -!- van511 [~van51@athedsl-318776.home.otenet.gr] has left #shogun [] | 18:39 | |
| -!- cesarfm [528247df@gateway/web/freenode/ip.82.130.71.223] has quit [Ping timeout: 245 seconds] | 18:43 | |
| -!- van511 [~van51@athedsl-318776.home.otenet.gr] has joined #shogun | 18:44 | |
| -!- mikhailBelous [~quassel@37.98.242.99] has quit [Remote host closed the connection] | 18:58 | |
| -!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 18:59 | |
| -!- mikhailBelous [~quassel@37.98.242.99] has joined #shogun | 18:59 | |
| -!- mikhailBelous [~quassel@37.98.242.99] has quit [Remote host closed the connection] | 19:00 | |
| @lisitsyn | sonney2k: michael schumacher in the mentors list himself! | 19:10 |
| @lisitsyn | :D | 19:10 |
| @lisitsyn | 48 finally | 19:32 |
| gsomix | shogun-buildbot, dance | 19:32 |
| shogun-buildbot | <(^.^<) | 19:32 |
| shogun-buildbot | <(^.^)> | 19:32 |
| shogun-buildbot | (>^.^)> | 19:32 |
| shogun-buildbot | (7^.^)7 | 19:32 |
| shogun-buildbot | (>^.^<) | 19:32 |
| -!- abinayam [3d0c1363@gateway/web/freenode/ip.61.12.19.99] has joined #shogun | 19:42 | |
| -!- nube [~rho@49.244.66.144] has quit [Quit: Leaving.] | 19:44 | |
| -!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 19:45 | |
| -!- Hao [d8a55f4a@gateway/web/freenode/ip.216.165.95.74] has quit [Ping timeout: 245 seconds] | 19:45 | |
| -!- abinayam [3d0c1363@gateway/web/freenode/ip.61.12.19.99] has quit [Ping timeout: 245 seconds] | 19:46 | |
| @lisitsyn | 49 | 19:47 |
| -!- abinayam [3d0c1363@gateway/web/freenode/ip.61.12.19.99] has joined #shogun | 19:52 | |
| abinayam | Hi | 19:52 |
| gsomix | shogun-buildbot, don't dance, bot, don't dance... | 19:52 |
| -!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 19:53 | |
| abinayam | I have submitted my proposal on implementing robust classification methods. Is anyone out there with the same choice? | 19:53 |
| shogun-notifier- | shogun: Cesar Fuentes :develop * b372a7f / examples/undocumented/python_modular/graphical/group_lasso.py: https://github.com/shogun-toolbox/shogun/commit/b372a7fa8f22436bba1718232c76bdfe9a8863ae | 19:56 |
| shogun-notifier- | shogun: FeatureBlockLogisticRegression example | 19:56 |
| shogun-notifier- | shogun: Sergey Lisitsyn :develop * 62eeb15 / examples/undocumented/python_modular/graphical/group_lasso.py: https://github.com/shogun-toolbox/shogun/commit/62eeb15227ab59c2f8bdd3a08ddf253ab0019683 | 19:56 |
| shogun-notifier- | shogun: Merge pull request #1045 from cesarfm/develop | 19:56 |
| shogun-notifier- | shogun: | 19:56 |
| shogun-notifier- | shogun: Group lasso example | 19:56 |
| -!- abinayam [3d0c1363@gateway/web/freenode/ip.61.12.19.99] has quit [Ping timeout: 245 seconds] | 19:57 | |
| vgorbati | lisitsyn: hi, here? | 20:06 |
| @lisitsyn | vgorbati: hey, yes | 20:06 |
| vgorbati | lisitsyn: I have finally started working on implementation, and I am looking through your code to find how to construct matrices of distances and angles to neighbors. It looks like using SparseTriplet's is the way? | 20:07 |
| @lisitsyn | vgorbati: yes that's the fastest way | 20:08 |
| vgorbati | lisitsyn: ok, then what I can't get, is why are you reserving space for (k+1)^2 values per row? I thought that there would only k-values (1 for each neighbor) per row.. | 20:10 |
| @lisitsyn | vgorbati: in lle? | 20:10 |
| vgorbati | lisitsyn: yes | 20:10 |
| @lisitsyn | vgorbati: if we put these weights per row | 20:11 |
| @lisitsyn | we get the W matrix | 20:11 |
| @lisitsyn | but later we need WW' | 20:11 |
| @lisitsyn | so I just form WW' here | 20:11 |
| vgorbati | lisitsyn: got it. btw, then it looks like there is a small typo in locally_linear.hpp at line 31: should be sparse_triplets.reserve((k*k+2*k+1)*(end-begin)); | 20:12 |
| shogun-buildbot | build #744 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/744 | 20:12 |
| @lisitsyn | yes 2 is missed it seems | 20:13 |
| @lisitsyn | thanks! | 20:13 |
| @sonney2k | lisitsyn, yes germans can only code and drive :D | 20:13 |
| @sonney2k | lisitsyn, woa 50?! | 20:14 |
| @sonney2k | more than last year? didn't expect that/ | 20:14 |
| @lisitsyn | sonney2k: yeah 50 | 20:14 |
| @lisitsyn | sonney2k: better for us I guess :) | 20:14 |
| -!- goldbug [~quassel@global-2-11.nat.csx.cam.ac.uk] has quit [Ping timeout: 252 seconds] | 20:15 | |
| gsomix | >> 44 minutes remaining | 20:15 |
| @sonney2k | lisitsyn, as in likely more slots yes but c'mon where is that rush coming from | 20:15 |
| @lisitsyn | sonney2k: some of them are kind of desperate | 20:16 |
| gsomix | waiting for many last-femtosecond-proposals | 20:16 |
| @sonney2k | lisitsyn, my theory now is that people since they are only allowed to apply for 5 (instead of 20) projects, are more carefully selecting projects | 20:17 |
| vgorbati | lisitsyn: it's my pleasure:) I am also wondering, is the '#pragma omp parallel' this 'OpenMP' parallelism that is mentioned on ideas page? | 20:17 |
| @sonney2k | and not do it in the very beginning | 20:17 |
| @lisitsyn | vgorbati: yes that's openmp | 20:17 |
| @lisitsyn | sonney2k: yeah they don't spread | 20:18 |
| vgorbati | lisitsyn: I guess I should also use it? | 20:18 |
| @lisitsyn | it is not like fishing now | 20:18 |
| @sonney2k | lisitsyn, btw do you know how one can stop the running openmp threads? | 20:18 |
| @lisitsyn | vgorbati: serial first! | 20:18 |
| vgorbati | lisitsyn: ok, better for me) | 20:18 |
| @lisitsyn | sonney2k: from the outer code? | 20:18 |
| @sonney2k | lisitsyn, yes | 20:18 |
| @lisitsyn | vgorbati: *never ever* write parallel code without serial ;) | 20:18 |
| @lisitsyn | sonney2k: ehmm some flag may be | 20:19 |
| vgorbati | lisitsyn: got it, sensei:) I just thought that the problem is quite the same | 20:19 |
| @lisitsyn | vgorbati: openmp can be surprising hard to debug sometimes | 20:20 |
| @lisitsyn | well once you get used to it it is easy though | 20:20 |
| @lisitsyn | sonney2k: any way to do that with pthreads? | 20:20 |
| @lisitsyn | sonney2k: it doesn't sound like a good practice for me actually | 20:21 |
| -!- van511 [~van51@athedsl-318776.home.otenet.gr] has left #shogun ["QUIT :Leaving."] | 20:25 | |
| @lisitsyn | sonney2k: 51 | 20:25 |
| shogun-buildbot | build #1044 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/1044 blamelist: Cesar Fuentes <cfuentesm85@gmail.com> | 20:27 |
| @lisitsyn | half an hour left! | 20:27 |
| -!- goldbug [~quassel@global-2-11.nat.csx.cam.ac.uk] has joined #shogun | 20:29 | |
| @sonney2k | lisitsyn, when you train some shogun method you can cancel any method by sending it the URG signal | 20:29 |
| @sonney2k | or press ctrl+c | 20:29 |
| gsomix | sonney2k, btw buildbot spams me in personal chat. | 20:30 |
| @lisitsyn | sonney2k: we have a double player! | 20:30 |
| @lisitsyn | pickle27 is doing stuff for scikits :) | 20:30 |
| @sonney2k | I used this quite a bit when doing research | 20:30 |
| @sonney2k | lisitsyn, how do you know? | 20:30 |
| @lisitsyn | sonney2k: got it - should be solveable | 20:30 |
| @lisitsyn | sonney2k: https://github.com/scikit-learn/scikit-learn/pull/1885 | 20:30 |
| gsomix | sonney2k, can I perform "notify off" in personal chat? | 20:31 |
| gsomix | shogun-buildbot, notify list | 20:31 |
| shogun-buildbot | The following events are being notified: ['failure', 'successToFailure', 'failureToSuccess'] | 20:31 |
| @lisitsyn | gsomix: youve got to change your nickname then.. | 20:31 |
| @lisitsyn | hah | 20:31 |
| gsomix | sonney2k, nevermind... it's all ok now | 20:31 |
| gsomix | lisitsyn, hey, my nickname is registered | 20:32 |
| @sonney2k | lisitsyn, I have no idea how this could work with openmp | 20:33 |
| @lisitsyn | 52 | 20:33 |
| @sonney2k | lisitsyn, looks like it is not possible http://stackoverflow.com/questions/8265191/force-closing-a-thread-using-openmp-in-c | 20:34 |
| @sonney2k | but hey what we could do is install a signal handler | 20:35 |
| @sonney2k | set some global variable | 20:35 |
| @lisitsyn | sonney2k: yeah nothing really impossible | 20:35 |
| @sonney2k | and then poll this variable | 20:35 |
| @sonney2k | the only problem with the signal handler is that it is not allowed to run in thread context | 20:35 |
| @sonney2k | so it must run in the main thread | 20:35 |
| @sonney2k | and it is also not clear what happens if the other threads finish before we get user input | 20:36 |
| * sonney2k gets drunk | 20:36 | |
| @sonney2k | B52 | 20:36 |
| @lisitsyn | haha | 20:36 |
| @lisitsyn | 53 | 20:37 |
| @sonney2k | there goes my drink | 20:38 |
| @lisitsyn | ehm last one is ehhm | 20:38 |
| @sonney2k | lisitsyn, please prepare an email to mentors | 20:39 |
| @lisitsyn | sonney2k: yeah will do | 20:39 |
| @sonney2k | once the deadline is passed - we should have at least 2 people scoring each student | 20:39 |
| @sonney2k | for students we intend to accept better >3 | 20:39 |
| @sonney2k | and if a mentor wants to accept a student he should give him some practical task to be really sure and also press the 'I want to mentor this' | 20:40 |
| @lisitsyn | 54 | 20:46 |
| @sonney2k | lisitsyn, c'mon is it you under disguise? | 20:47 |
| @lisitsyn | hah no | 20:47 |
| @sonney2k | sonney2k, tell them they should become pretty sure within next week... | 20:48 |
| @sonney2k | this number of proposals is insane ... so it will be an acceptance rate of <20% | 20:49 |
| @lisitsyn | sonney2k: ten more minutes! | 20:50 |
| * sonney2k looks up the script for plotting stats | 20:54 | |
| @lisitsyn | 55 | 20:56 |
| @lisitsyn | 56 | 20:57 |
| shogun-buildbot | build #1045 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/1045 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 20:57 |
| gsomix | last minute! | 20:59 |
| @lisitsyn | that's it | 21:00 |
| gsomix | lisitsyn, so what? 56? | 21:05 |
| @lisitsyn | gsomix: yes | 21:05 |
| @lisitsyn | here comes a deadline mail | 21:08 |
| @lisitsyn | oops | 21:08 |
| @lisitsyn | 57 | 21:08 |
| @lisitsyn | I missed one | 21:09 |
| -!- makokal [~bokal@212.201.44.246] has joined #shogun | 21:09 | |
| @lisitsyn | crazy | 21:10 |
| @lisitsyn | sonney2k: is that more than first year? | 21:10 |
| gsomix | you are popular, guys, hehe | 21:11 |
| -!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 21:12 | |
| gsomix | sonney2k, http://researchinprogress.tumblr.com/post/32893888233/reviewing-conference-papers | 21:16 |
| naywhayare | the quality of applicants is... variable | 21:19 |
| naywhayare | I asked if they had worked with an open source project before... one applicant responded: | 21:19 |
| naywhayare | I have worked with many open source solutions wile creating my projects.Some of them are:MS SQL Server | 21:19 |
| gsomix | O_O | 21:19 |
| naywhayare | that was my reaction | 21:20 |
| @lisitsyn | naywhayare: hah nice | 21:21 |
| @lisitsyn | naywhayare: how many did you get? | 21:21 |
| @sonney2k | lisitsyn, https://dl.dropboxusercontent.com/u/31480471/gsoc2013-applications.png | 21:21 |
| @lisitsyn | crazy curve | 21:22 |
| @sonney2k | lisitsyn, compare to last year http://shogun-toolbox.org/page/Events/gsoc2012 | 21:22 |
| @lisitsyn | quite similar? | 21:22 |
| naywhayare | lisitsyn: 41 | 21:25 |
| @lisitsyn | naywhayare: quite a lot too | 21:25 |
| naywhayare | the majority came in during the past hour | 21:25 |
| naywhayare | the curve probably looks about the same as what sonney2k posted | 21:26 |
| @sonney2k | lisitsyn, not at all | 21:26 |
| @sonney2k | last year was 48 | 21:26 |
| @lisitsyn | naywhayare: our last minute proposals are like spam | 21:27 |
| naywhayare | many posit that they can solve extremely difficult publication-worthy problems without any supporting evidence whatsoever... | 21:27 |
| @sonney2k | lisitsyn, exactly. In our first year we made the mistake not to require a code contribution | 21:28 |
| @sonney2k | so we had 70 or so proposals | 21:28 |
| @lisitsyn | haha | 21:28 |
| @sonney2k | kind of exponential growth | 21:29 |
| naywhayare | I might try requiring a patch next year, but them I imagine I'd have to sort through numerous patches that probably don't even work | 21:29 |
| naywhayare | s/them/then/ | 21:29 |
| @sonney2k | naywhayare, yes that's the price | 21:29 |
| @sonney2k | but you know then who is good | 21:29 |
| @sonney2k | naywhayare, btw you still have until may 22 to decide | 21:29 |
| @sonney2k | so you should really require from every student you consider some kind of code contribution | 21:30 |
| @sonney2k | this way you become very sure if the student can do the job | 21:30 |
| -!- makokal [~bokal@212.201.44.246] has quit [Quit: makokal] | 21:31 | |
| naywhayare | don't I have to have my projected slot allocations done by sunday though? | 21:31 |
| @lisitsyn | naywhayare: yeah you'd better have to | 21:31 |
| @lisitsyn | naywhayare: but you'll have time to finally settle it | 21:32 |
| naywhayare | okay, that's good | 21:33 |
| naywhayare | I definitely have a few students who I'd like to get in further touch with, so I am glad I have a few weeks to do that | 21:33 |
| @lisitsyn | naywhayare: what is the number of slots you'd like to have? | 21:35 |
| naywhayare | it depends on what my mentors want. we could do up to 6 but I've been told to not expect more than 1 or 2 during the first year | 21:36 |
| naywhayare | so in that sense it depends more on google than the other mentors... | 21:36 |
| @lisitsyn | sonney2k: I am going to send a welcome letters | 21:43 |
| @lisitsyn | for the workshop | 21:43 |
| @lisitsyn | sonney2k: y/n? | 21:43 |
| @sonney2k | naywhayare, yeah first year is usually no more than 2 | 21:47 |
| @sonney2k | so you have to make tough choices | 21:47 |
| @sonney2k | in very rare circumstances up to 5 | 21:48 |
| @sonney2k | but still <10% acceptance... | 21:48 |
| naywhayare | yeah, 2 is about what I'm expecting. based on the ratings it looks like we have 2 or 3 standout students and 7 or 8 who could be reasonable | 21:48 |
| naywhayare | anyway, I have to run for now... thanks for the advice :) | 21:48 |
| @sonney2k | naywhayare, np | 21:48 |
| -!- makokal [~bokal@pD9EF09DC.dip0.t-ipconnect.de] has joined #shogun | 21:53 | |
| @sonney2k | lisitsyn, please send out the email to the mentors first | 21:54 |
| @lisitsyn | sonney2k: yeah will do | 21:54 |
| @lisitsyn | sonney2k: workshop mails - y/n? | 21:55 |
| @sonney2k | lisitsyn, yes but let me fix them first | 21:56 |
| @lisitsyn | sonney2k: let me send you draft then | 21:56 |
| @lisitsyn | sonney2k: in your inbox | 21:57 |
| -!- goldbug [~quassel@global-2-11.nat.csx.cam.ac.uk] has quit [Remote host closed the connection] | 22:03 | |
| -!- iglesiasg [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 22:03 | |
| -!- mode/#shogun [+o iglesiasg] by ChanServ | 22:03 | |
| @iglesiasg | 56 applications OMG! | 22:03 |
| @sonney2k | 57 even | 22:04 |
| @sonney2k | so much about my estimate of 20 | 22:04 |
| @iglesiasg | hehe | 22:04 |
| @iglesiasg | sonney2k: how many applicants? | 22:05 |
| @lisitsyn | iglesiasg: you mean different persons? | 22:05 |
| @lisitsyn | iglesiasg: 5 double applications | 22:06 |
| @lisitsyn | so 52 different persons | 22:06 |
| @iglesiasg | lisitsyn: oh impressive | 22:06 |
| @iglesiasg | sonney2k: I will wait for your post and the graph showing that exponential curve :) | 22:07 |
| gsomix | iglesiasg, hi | 22:11 |
| @iglesiasg | gsomix: hello hello | 22:11 |
| @lisitsyn | sonney2k: done | 22:13 |
| gsomix | sonney2k, cool, there is patch for lua52 support. btw how many people who use lua for ML? | 22:15 |
| -!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has quit [Quit: Leaving] | 22:20 | |
| @iglesiasg | oh just saw the graph, crazy | 22:20 |
| @sonney2k | gsomix, well torch is written in lua | 22:22 |
| @sonney2k | so I know a couple of big shots that use it | 22:22 |
| gsomix | sonney2k, ok, thanks | 22:24 |
| -!- iglesiasg [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 22:31 | |
| @sonney2k | lisitsyn, argh argh argh | 22:37 |
| @lisitsyn | sonney2k: ? | 22:37 |
| @sonney2k | I again overlooked that the PR was against master | 22:37 |
| shogun-notifier- | shogun: Ingvaras Merkys :master * fd4ada9 / examples/generate_documented.sh,src/interfaces/lua_modular/swig_typemaps.i: https://github.com/shogun-toolbox/shogun/commit/fd4ada911b0cbfc9a4bd1846cc6d18533435fe8c | 22:37 |
| shogun-notifier- | shogun: lua52 fix | 22:37 |
| shogun-notifier- | shogun: Soeren Sonnenburg :master * 255beac / examples/generate_documented.sh,src/interfaces/lua_modular/swig_typemaps.i: https://github.com/shogun-toolbox/shogun/commit/255beac14c3952ab4443388aaa55ad673bc3745b | 22:37 |
| shogun-notifier- | shogun: Merge pull request #1056 from crunchiness/master | 22:37 |
| shogun-notifier- | shogun: | 22:37 |
| shogun-notifier- | shogun: Fixed interface for Lua 5.2 | 22:37 |
| @lisitsyn | sonney2k: you evil! | 22:37 |
| @sonney2k | lisitsyn, is there any better way than git merge master? | 22:38 |
| shogun-buildbot | build #911 of bsd1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/911 blamelist: Ingvaras Merkys <ingvaras@gmail.com> | 22:38 |
| @lisitsyn | sonney2k: I don't think so | 22:38 |
| * sonney2k bangs his head against a wall | 22:38 | |
| @lisitsyn | sonney2k: what is wrong with merge? | 22:38 |
| shogun-buildbot | build #746 of cyg1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/746 blamelist: Ingvaras Merkys <ingvaras@gmail.com> | 22:39 |
| shogun-notifier- | shogun: Ingvaras Merkys :develop * fd4ada9 / examples/generate_documented.sh,src/interfaces/lua_modular/swig_typemaps.i: https://github.com/shogun-toolbox/shogun/commit/fd4ada911b0cbfc9a4bd1846cc6d18533435fe8c | 22:41 |
| shogun-notifier- | shogun: lua52 fix | 22:41 |
| shogun-notifier- | shogun: Soeren Sonnenburg :develop * 255beac / examples/generate_documented.sh,src/interfaces/lua_modular/swig_typemaps.i: https://github.com/shogun-toolbox/shogun/commit/255beac14c3952ab4443388aaa55ad673bc3745b | 22:41 |
| shogun-notifier- | shogun: Merge pull request #1056 from crunchiness/master | 22:41 |
| shogun-notifier- | shogun: | 22:41 |
| shogun-notifier- | shogun: Fixed interface for Lua 5.2 | 22:41 |
| shogun-notifier- | shogun: Soeren Sonnenburg :develop * 72f954a / examples/generate_documented.sh,src/interfaces/lua_modular/swig_typemaps.i: https://github.com/shogun-toolbox/shogun/commit/72f954a7d02bddef7deb2a8228f20fe95eb24a4c | 22:41 |
| shogun-notifier- | shogun: Merge accidentally accepted PR against master into develop | 22:41 |
| shogun-buildbot | build #912 of bsd1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/912 blamelist: Soeren Sonnenburg <sonne@debian.org> | 22:46 |
| -!- makokal [~bokal@pD9EF09DC.dip0.t-ipconnect.de] has quit [Quit: makokal] | 22:50 | |
| -!- travis-ci [~travis-ci@ec2-107-22-69-126.compute-1.amazonaws.com] has joined #shogun | 22:59 | |
| 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/6865358 | 22:59 |
| -!- travis-ci [~travis-ci@ec2-107-22-69-126.compute-1.amazonaws.com] has left #shogun [] | 22:59 | |
| shogun-buildbot | build #747 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/747 | 23:00 |
| shogun-buildbot | build #748 of cyg1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/748 blamelist: Ingvaras Merkys <ingvaras@gmail.com>, Soeren Sonnenburg <sonne@debian.org> | 23:07 |
| shogun-buildbot | build #913 of bsd1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/913 | 23:13 |
| shogun-buildbot | build #1046 of deb3 - modular_interfaces is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1046 | 23:14 |
| shogun-buildbot | build #1047 of deb3 - modular_interfaces is complete: Failure [failed compile lua_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/1047 blamelist: Soeren Sonnenburg <sonne@debian.org> | 23:24 |
| -!- cesarfm [b2c73bf3@gateway/web/freenode/ip.178.199.59.243] has joined #shogun | 23:28 | |
| gsomix | nite | 23:37 |
| shogun-buildbot | build #1048 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/1048 blamelist: Ingvaras Merkys <ingvaras@gmail.com>, Soeren Sonnenburg <sonne@debian.org> | 23:53 |
| --- Log closed Sat May 04 00:00:30 2013 | ||
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!
