| --- Log opened Fri Jun 01 00:00:41 2012 | ||
| -!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 245 seconds] | 01:32 | |
| -!- blackburn [~blackburn@31.28.59.65] has left #shogun [] | 02:17 | |
| -!- av3ngr [~av3ngr@pa58-109-143-179.pa.nsw.optusnet.com.au] has joined #shogun | 03:08 | |
| -!- emrecelikten [~emrecelik@176.40.245.70] has joined #shogun | 03:33 | |
| -!- emrecelikten [~emrecelik@176.40.245.70] has quit [Client Quit] | 03:33 | |
| -!- av3ngr [~av3ngr@pa58-109-143-179.pa.nsw.optusnet.com.au] has quit [Ping timeout: 252 seconds] | 05:54 | |
| -!- av3ngr [~av3ngr@130.56.73.235] has joined #shogun | 06:29 | |
| -!- av3ngr [~av3ngr@130.56.73.235] has quit [Ping timeout: 246 seconds] | 07:20 | |
| -!- av3ngr [~av3ngr@130.56.73.235] has joined #shogun | 07:33 | |
| -!- uricamic [~uricamic@2001:718:2:1634:e851:f250:a5d5:cb86] has joined #shogun | 08:48 | |
| -!- av3ngr [~av3ngr@130.56.73.235] has quit [Quit: That's all folks!] | 09:31 | |
| -!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun | 09:32 | |
| -!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 246 seconds] | 09:37 | |
| -!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun | 09:53 | |
| -!- gsomix [~gsomix@188.168.128.108] has quit [Quit: Ex-Chat] | 10:16 | |
| -!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 10:33 | |
| -!- blackburn [~blackburn@31.28.59.65] has joined #shogun | 11:08 | |
| -!- blackburn [~blackburn@31.28.59.65] has quit [Ping timeout: 248 seconds] | 12:09 | |
| -!- nickon [~noneedtok@dD5774105.access.telenet.be] has joined #shogun | 12:27 | |
| -!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: Lost terminal] | 12:57 | |
| -!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 12:58 | |
| -!- uricamic [~uricamic@2001:718:2:1634:e851:f250:a5d5:cb86] has quit [Quit: Leaving.] | 13:20 | |
| -!- pluskid [~pluskid@111.120.44.169] has joined #shogun | 15:00 | |
| -!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 244 seconds] | 15:14 | |
| -!- pluskid [~pluskid@111.120.44.169] has quit [Ping timeout: 245 seconds] | 16:14 | |
| -!- pluskid [~pluskid@li400-235.members.linode.com] has joined #shogun | 16:14 | |
| -!- eric_ [2e1fd566@gateway/web/freenode/ip.46.31.213.102] has joined #shogun | 16:30 | |
| eric_ | hi all | 16:30 |
|---|---|---|
| -!- pluskid [~pluskid@li400-235.members.linode.com] has quit [Ping timeout: 250 seconds] | 16:51 | |
| -!- gsomix [~gsomix@85.26.235.178] has joined #shogun | 16:55 | |
| -!- blackburn [~blackburn@31.28.59.65] has joined #shogun | 17:01 | |
| -!- pluskid [~pluskid@111.120.44.169] has joined #shogun | 17:09 | |
| -!- pluskid [~pluskid@111.120.44.169] has quit [Client Quit] | 17:11 | |
| -!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 17:50 | |
| -!- eric_ [2e1fd566@gateway/web/freenode/ip.46.31.213.102] has quit [Ping timeout: 245 seconds] | 18:15 | |
| blackburn | n4nd0: hey how is it going? | 18:25 |
| -!- nickon [~noneedtok@dD5774105.access.telenet.be] has quit [Quit: ( www.nnscript.com :: NoNameScript 4.22 :: www.esnation.com )] | 18:34 | |
| n4nd0 | blackburn: ups sorry, I didn't hear it | 19:24 |
| n4nd0 | blackburn: fine and for you? | 19:24 |
| blackburn | fine some thesis things to complete still - hope to switch to slep again | 19:28 |
| n4nd0 | blackburn: good | 19:28 |
| n4nd0 | I'll be back later! | 19:29 |
| -!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 252 seconds] | 19:35 | |
| -!- gsomix [~gsomix@85.26.235.178] has quit [Ping timeout: 250 seconds] | 20:09 | |
| -!- gsomix [~gsomix@85.26.235.178] has joined #shogun | 20:32 | |
| -!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 21:08 | |
| -!- gsomix [~gsomix@85.26.235.178] has quit [Ping timeout: 245 seconds] | 21:15 | |
| -!- gsomix [~gsomix@85.26.235.178] has joined #shogun | 21:19 | |
| shogun-buildbot | build #256 of nightly_all is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_all/builds/256 | 21:31 |
| shogun-buildbot | build #167 of nightly_default is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/167 | 22:01 |
| n4nd0 | blackburn: around? | 22:12 |
| blackburn | yes | 22:12 |
| n4nd0 | blackburn: ok, I have got a doubt about the opt. problem I have to solve for SO | 22:13 |
| n4nd0 | but it might be about something general | 22:13 |
| n4nd0 | let me ask you in case you know it ok? | 22:13 |
| blackburn | wellll | 22:13 |
| blackburn | just ask but I do not promise I can help :) | 22:13 |
| n4nd0 | yeah sure, I understand ;) | 22:14 |
| blackburn | my opt. background is pretty lame | 22:14 |
| n4nd0 | http://dl.dropbox.com/u/11020840/shogun/ideas.pdf | 22:14 |
| n4nd0 | in that pdf, page 2 | 22:14 |
| n4nd0 | algorithm 1, line 11 | 22:14 |
| blackburn | aham | 22:15 |
| n4nd0 | so there we have a quadratic program | 22:15 |
| n4nd0 | so do you see the part with the sum | 22:15 |
| n4nd0 | ? | 22:15 |
| blackburn | yes | 22:15 |
| n4nd0 | so my point is that in that sum there are no unknowns | 22:16 |
| n4nd0 | since we are using the training data | 22:16 |
| n4nd0 | that is the part indexed as x_i and y_i | 22:16 |
| blackburn | what is Psi? | 22:16 |
| n4nd0 | and also results that we have previously stored in Gamma_i | 22:16 |
| blackburn | w/o () | 22:16 |
| n4nd0 | is an error | 22:17 |
| n4nd0 | Psi should go there with () | 22:17 |
| n4nd0 | and it refers to the values we have stored in Gamma_i | 22:17 |
| blackburn | (x_i,y_i)? | 22:17 |
| blackburn | okay so what is your question? | 22:17 |
| n4nd0 | ok wait a moment ... | 22:18 |
| n4nd0 | I might have realized of it :P | 22:18 |
| n4nd0 | all right | 22:19 |
| n4nd0 | it was an stupid thing I have just realized | 22:19 |
| n4nd0 | sorry :S | 22:19 |
| blackburn | heh :D | 22:19 |
| blackburn | that's always happen when you try to describe something | 22:19 |
| n4nd0 | haha yeah | 22:20 |
| n4nd0 | I forgot about the w there | 22:20 |
| n4nd0 | in the sum everything is not known | 22:20 |
| n4nd0 | since we have precisely there the optimization vector | 22:20 |
| * n4nd0 is afk | 22:25 | |
| blackburn | hmm i do not understand much in that algorithm :D | 22:27 |
| -!- CIA-113 [~CIA@cia.atheme.org] has quit [Ping timeout: 245 seconds] | 22:36 | |
| -!- CIA-9 [~CIA@cia.atheme.org] has joined #shogun | 22:46 | |
| CIA-9 | shogun: Sergey Lisitsyn master * rf100cae / src/shogun/features/StringFeatures.cpp : Added debug error instead of assert in StringFeatures - http://git.io/iKkhug | 22:51 |
| -!- gsomix [~gsomix@85.26.235.178] has quit [Ping timeout: 252 seconds] | 23:04 | |
| * n4nd0 is back | 23:12 | |
| n4nd0 | blackburn: tell me if you want to know something about the algorithm | 23:17 |
| blackburn | heh | 23:17 |
| n4nd0 | blackburn: it'll probably help to my understanding trying to explain it :D | 23:17 |
| blackburn | ok but a little bit later :) | 23:17 |
| CIA-9 | shogun: Sergey Lisitsyn master * re915d01 / (2 files): Fixed GNB zero variance handling - http://git.io/HOtiHg | 23:37 |
| -!- heiko1 [~heiko@host86-179-58-219.range86-179.btcentralplus.com] has joined #shogun | 23:46 | |
| heiko1 | blackburn, sonney2k, whats the best way to do matrix multiplication in shogun? | 23:48 |
| blackburn | heiko1: cblas_dgemm :) | 23:49 |
| blackburn | depends what is best for you | 23:49 |
| heiko1 | is there a wrapper method available? | 23:49 |
| heiko1 | just a plain multiplication | 23:49 |
| blackburn | what kind of wrapper do you need? | 23:49 |
| heiko1 | in place would be nice | 23:49 |
| blackburn | matrix-matrix? | 23:49 |
| heiko1 | sg matrix for example | 23:49 |
| heiko1 | yes | 23:49 |
| blackburn | matrix-matrix inplace?? | 23:50 |
| heiko1 | ehm | 23:50 |
| heiko1 | well | 23:50 |
| heiko1 | probably not when I think about it :) | 23:50 |
| blackburn | ok I actually don't mind a wrapper | 23:50 |
| heiko1 | oh yes, my matrices are square | 23:50 |
| heiko1 | that would be possible in place | 23:50 |
| heiko1 | right? | 23:50 |
| heiko1 | anyway, dont really care | 23:51 |
| blackburn | I can't mind this | 23:51 |
| blackburn | you would have to store matrix still | 23:51 |
| heiko1 | yes | 23:51 |
| heiko1 | is cblas_dgemm the method to use? | 23:51 |
| blackburn | I don't think inplace is possible at all | 23:51 |
| blackburn | well yes | 23:52 |
| heiko1 | its possible under certain condiutions (which are met in my case) but anyway | 23:52 |
| heiko1 | i will try to do a general wrapper | 23:52 |
| blackburn | okay actually main reason why I do not use it | 23:52 |
| blackburn | (wrapper) | 23:53 |
| blackburn | is some variety of options like transposes | 23:53 |
| blackburn | and sizes | 23:53 |
| heiko1 | mmh | 23:53 |
| blackburn | actually one may end with wrapper with parameters exactly just like dgemm | 23:53 |
| blackburn | :D | 23:53 |
| heiko1 | so you think directly using it would be better? | 23:53 |
| heiko1 | I would love to have one for code reading ease | 23:54 |
| blackburn | no I don't mind wrappers | 23:54 |
| heiko1 | kk then | 23:54 |
| blackburn | but still you would need at least two parameters (transposes) | 23:54 |
| heiko1 | ok I will include these | 23:54 |
| blackburn | ask if you are stucked with params of blas | 23:54 |
| heiko1 | ask :) | 23:55 |
| heiko1 | where is an example call in SHOGUN that I can use? | 23:55 |
| blackburn | let me describe it for you there | 23:55 |
| heiko1 | actually, seeing the blas reference | 23:56 |
| heiko1 | I think I will directly use that | 23:56 |
| blackburn | heh ok | 23:56 |
| heiko1 | you are right | 23:56 |
| blackburn | about? | 23:56 |
| heiko1 | a wrapper would have to contain all this stuff :) | 23:56 |
| blackburn | yes | 23:56 |
| blackburn | the only thing we can hide | 23:57 |
| blackburn | is LDA, LDB, LDC | 23:57 |
| blackburn | and N,M,K | 23:57 |
| -!- heiko2 [~heiko@host86-177-14-77.range86-177.btcentralplus.com] has joined #shogun | 23:59 | |
| -!- blackburn is now known as heiko3 | 23:59 | |
| --- Log closed Sat Jun 02 00:00:41 2012 | ||
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!