--- Log opened Fri Jun 15 00:00:41 2012 | ||
-!- blackburn2 [5bde8018@gateway/web/freenode/ip.91.222.128.24] has joined #shogun | 00:08 | |
-!- blackburn2 [5bde8018@gateway/web/freenode/ip.91.222.128.24] has quit [Client Quit] | 00:08 | |
-!- blackburn [5bde8018@gateway/web/freenode/ip.91.222.128.24] has quit [Quit: Page closed] | 00:09 | |
-!- vikram360 [~vikram360@117.192.161.149] has joined #shogun | 00:18 | |
-!- vikram360 [~vikram360@117.192.161.149] has quit [Ping timeout: 244 seconds] | 00:25 | |
-!- cronor [~cronor@e178178115.adsl.alicedsl.de] has quit [Read error: Connection reset by peer] | 00:25 | |
-!- vikram360 [~vikram360@117.192.165.99] has joined #shogun | 00:25 | |
-!- cronor [~cronor@e178178115.adsl.alicedsl.de] has joined #shogun | 00:26 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 00:41 | |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 256 seconds] | 00:45 | |
-!- romi_ [~mizobe@187.74.0.229] has quit [Remote host closed the connection] | 01:04 | |
-!- cronor_ [~cronor@e178178115.adsl.alicedsl.de] has joined #shogun | 01:06 | |
-!- cronor [~cronor@e178178115.adsl.alicedsl.de] has quit [Read error: Connection reset by peer] | 01:06 | |
-!- cronor_ is now known as cronor | 01:06 | |
-!- vikram360 [~vikram360@117.192.165.99] has quit [Read error: Connection reset by peer] | 02:03 | |
-!- vikram360 [~vikram360@117.192.165.99] has joined #shogun | 02:20 | |
-!- cronor_ [~cronor@g230120109.adsl.alicedsl.de] has joined #shogun | 02:39 | |
-!- cronor [~cronor@e178178115.adsl.alicedsl.de] has quit [Ping timeout: 244 seconds] | 02:41 | |
-!- cronor_ is now known as cronor | 02:41 | |
-!- vikram360 [~vikram360@117.192.165.99] has quit [Ping timeout: 244 seconds] | 02:52 | |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 03:08 | |
-!- gsomix [~gsomix@178.45.81.48] has joined #shogun | 06:38 | |
-!- cronor [~cronor@g230120109.adsl.alicedsl.de] has quit [Quit: cronor] | 06:59 | |
-!- cronor [~cronor@g230120109.adsl.alicedsl.de] has joined #shogun | 07:01 | |
-!- cronor_ [~cronor@g230120109.adsl.alicedsl.de] has joined #shogun | 07:25 | |
-!- cronor [~cronor@g230120109.adsl.alicedsl.de] has quit [Read error: Connection reset by peer] | 07:25 | |
-!- cronor_ is now known as cronor | 07:25 | |
-!- uricamic [~uricamic@2001:718:2:1634:f9c1:bb08:2f9e:5eb9] has joined #shogun | 08:57 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 09:30 | |
uricamic | hi n4nd0 | 09:32 |
---|---|---|
-!- cronor [~cronor@g230120109.adsl.alicedsl.de] has quit [Quit: cronor] | 09:38 | |
n4nd0 | uricamic: hey | 09:40 |
uricamic | n4nd0: when do you think your changes in SO framework will be pulled to shogun? I want to write a simple example using the BMRM solver, but I think it will be better to do this after your new changes will be accepted | 09:42 |
n4nd0 | uricamic: I am not completely sure, I did the pull request and sonney2k told me to make some changes, they were not very important though | 09:43 |
n4nd0 | the problem is that currently my multiclass example doesn't work good | 09:44 |
uricamic | n4nd0: any ideas what could be wrong? | 09:44 |
n4nd0 | not many, just that there must be a bug somewhere | 09:45 |
uricamic | in the solver? | 09:45 |
n4nd0 | I guess so | 09:46 |
n4nd0 | there is a thing that bugs me a bit | 09:46 |
n4nd0 | right now I use the same data for training and testing | 09:46 |
n4nd0 | just to make it easier | 09:46 |
uricamic | ok | 09:47 |
n4nd0 | and all the examples of one class are classified as another class | 09:47 |
n4nd0 | but still, the values of the slacks after training for those examples are 0 | 09:47 |
n4nd0 | I think that tells me that something is going wrong | 09:48 |
uricamic | so the classes are swapped? | 09:48 |
n4nd0 | not really | 09:48 |
n4nd0 | it is that some of the classes used for training are never classified | 09:49 |
uricamic | I see | 09:49 |
n4nd0 | I going to compare now how the weight vectors learned in so and in one-vs-rest look like | 09:51 |
n4nd0 | I expect that they should be somewhat similar | 09:51 |
uricamic | yep | 09:54 |
uricamic | and you are now working on multiclass example with how many classes? | 09:54 |
n4nd0 | I change the number of classes as a parameter | 09:56 |
n4nd0 | I am focusing more in three because for two classes looks generally ok | 09:56 |
uricamic | so for 2 classes it works and the problem is only when the number of classes is >2 ? | 09:57 |
n4nd0 | yes, apparently | 10:00 |
uricamic | I see, that is somehow strange | 10:02 |
uricamic | do you have the source codes for this pushed in your fork of shogun? | 10:03 |
n4nd0 | yes | 10:03 |
n4nd0 | let me give you the link | 10:04 |
n4nd0 | https://github.com/iglesias/shogun/blob/so/src/shogun/so/PrimalMosekSOSVM.cpp | 10:04 |
n4nd0 | there in train_machine | 10:04 |
uricamic | thanks, I will look at it | 10:05 |
n4nd0 | https://github.com/iglesias/shogun/blob/so/src/shogun/so/MulticlassModel.cpp | 10:05 |
n4nd0 | and also in the mathematics/mosek methods used | 10:05 |
uricamic | ok | 10:06 |
-!- wiking [~wiking@78-23-189-112.access.telenet.be] has joined #shogun | 10:07 | |
-!- wiking [~wiking@78-23-189-112.access.telenet.be] has quit [Changing host] | 10:07 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 10:07 | |
-!- cronor [~cronor@141.23.80.206] has joined #shogun | 10:12 | |
-!- cronor [~cronor@141.23.80.206] has quit [Client Quit] | 10:16 | |
-!- cronor [~cronor@fb.ml.tu-berlin.de] has joined #shogun | 10:29 | |
-!- vikram360 [~vikram360@117.192.173.202] has joined #shogun | 11:33 | |
-!- vikram360 [~vikram360@117.192.173.202] has quit [Ping timeout: 244 seconds] | 11:44 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 252 seconds] | 11:55 | |
-!- gsomix_ [~gsomix@109.169.252.185] has joined #shogun | 12:33 | |
-!- gsomix [~gsomix@178.45.81.48] has quit [Ping timeout: 245 seconds] | 12:35 | |
-!- blackburn1 [~blackburn@62.106.111.221] has joined #shogun | 13:55 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 14:49 | |
n4nd0 | uricamic: hey, how is it going? | 14:57 |
uricamic | n4nd0: hi | 14:57 |
uricamic | I have looked on your code, but haven't found anything wrong | 14:58 |
uricamic | but I haven't found the example putting things together | 14:58 |
uricamic | have you commited it to your fork on github? | 14:59 |
n4nd0 | no | 14:59 |
n4nd0 | I am going to do that now | 14:59 |
uricamic | ok | 15:00 |
n4nd0 | uricamic: did you find something that looked weird or you didn't really get it=? | 15:00 |
uricamic | in the MulticlassModel everything looked clear | 15:00 |
uricamic | I will try to combine your up-to-date version of the SO framework and produce the example with the BMRM method and then to compare your solver with mine | 15:02 |
n4nd0 | ok | 15:03 |
n4nd0 | about the example | 15:03 |
n4nd0 | if you would like to test it in your machine, note that you need mosek for that | 15:03 |
n4nd0 | and to use mosek within shogun you'd have to change by hand some parameters in your .config | 15:04 |
n4nd0 | but I don't think you really need to test it | 15:04 |
uricamic | ok | 15:04 |
n4nd0 | I mean, you have done enough taking a look to the core ;) | 15:04 |
n4nd0 | code* | 15:04 |
uricamic | but maybe then it will be easier to locate the source of the problem | 15:05 |
n4nd0 | ok, as you like | 15:06 |
n4nd0 | in any case, if you plan to run it, just tell me to describe you what parameters you need to change to make it run | 15:06 |
uricamic | ok | 15:07 |
uricamic | first I will have to get my solver working | 15:07 |
n4nd0 | sure | 15:08 |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun | 16:04 | |
-!- uricamic [~uricamic@2001:718:2:1634:f9c1:bb08:2f9e:5eb9] has quit [Quit: Leaving.] | 16:11 | |
-!- cronor [~cronor@fb.ml.tu-berlin.de] has quit [Ping timeout: 252 seconds] | 16:18 | |
-!- vikram360 [~vikram360@117.192.173.202] has joined #shogun | 17:38 | |
-!- emrecelikten [~emre@176.40.254.205] has quit [Remote host closed the connection] | 17:47 | |
-!- cronor [~cronor@g230120109.adsl.alicedsl.de] has joined #shogun | 18:02 | |
-!- gsomix__ [~gsomix@109.169.254.147] has joined #shogun | 18:37 | |
-!- gsomix_ [~gsomix@109.169.252.185] has quit [Read error: Connection reset by peer] | 18:37 | |
-!- ckwidmer [8ca3fe9d@gateway/web/freenode/ip.140.163.254.157] has quit [Ping timeout: 245 seconds] | 18:46 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 18:47 | |
-!- gsomix__ [~gsomix@109.169.254.147] has quit [Ping timeout: 240 seconds] | 18:49 | |
-!- emrecelikten [~emre@176.40.254.205] has joined #shogun | 18:52 | |
-!- ckwidmer [8ca3fe9d@gateway/web/freenode/ip.140.163.254.157] has joined #shogun | 19:25 | |
-!- ckwidmer [8ca3fe9d@gateway/web/freenode/ip.140.163.254.157] has quit [Ping timeout: 245 seconds] | 19:30 | |
-!- ckwidmer [8ca3fe9d@gateway/web/freenode/ip.140.163.254.157] has joined #shogun | 20:19 | |
-!- puffin444 [62e3926e@gateway/web/freenode/ip.98.227.146.110] has joined #shogun | 20:28 | |
CIA-18 | shogun: Christian Widmer master * r5423ead / (2 files in 2 dirs): included std_vector typemap while avoiding java compile error - http://git.io/krs4Ng | 20:43 |
CIA-18 | shogun: Christian Widmer master * r6b71ab4 / (2 files): replaced SGMatrix with SGSparseMatrix to handle the case of sparse task similarities - http://git.io/f1Cvwg | 20:43 |
CIA-18 | shogun: Sergey Lisitsyn master * r861c682 / (4 files in 3 dirs): Merge pull request #586 from cwidmer/master - http://git.io/YlXy-w | 20:43 |
shogun-buildbot_ | build #844 of octave_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/octave_static/builds/844 | 21:07 |
shogun-buildbot_ | build #955 of cmdline_static is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cmdline_static/builds/955 blamelist: cwidmer@tuebingen.mpg.de, blackburn91@gmail.com | 21:17 |
-!- vikram360 [~vikram360@117.192.173.202] has quit [Ping timeout: 265 seconds] | 21:19 | |
-!- vikram360 [~vikram360@117.192.166.163] has joined #shogun | 21:20 | |
-!- vikram360 [~vikram360@117.192.166.163] has quit [Ping timeout: 244 seconds] | 21:46 | |
shogun-buildbot_ | build #934 of r_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/r_static/builds/934 | 21:57 |
CIA-18 | shogun: Sergey Lisitsyn master * r6b605bf / (4 files in 4 dirs): Made SparseMatrix SGReferencedData - http://git.io/Iy6_Wg | 22:15 |
-!- cronor_ [~cronor@g230120109.adsl.alicedsl.de] has joined #shogun | 22:19 | |
-!- cronor [~cronor@g230120109.adsl.alicedsl.de] has quit [Read error: Connection reset by peer] | 22:19 | |
-!- cronor_ is now known as cronor | 22:19 | |
-!- romi_ [~mizobe@187.74.0.229] has joined #shogun | 22:22 | |
-!- blackburn1 is now known as blackburn | 22:26 | |
shogun-buildbot_ | build #956 of cmdline_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cmdline_static/builds/956 | 22:32 |
-!- ckwidmer_ [8ca3fe9d@gateway/web/freenode/ip.140.163.254.157] has joined #shogun | 22:50 | |
-!- ckwidmer [8ca3fe9d@gateway/web/freenode/ip.140.163.254.157] has quit [Ping timeout: 245 seconds] | 22:52 | |
-!- ckwidmer_ [8ca3fe9d@gateway/web/freenode/ip.140.163.254.157] has quit [Ping timeout: 245 seconds] | 22:55 | |
-!- emrecelikten [~emre@176.40.254.205] has quit [Quit: Leaving.] | 22:55 | |
shogun-buildbot_ | build #627 of octave_modular is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/octave_modular/builds/627 blamelist: blackburn91@gmail.com | 22:58 |
-!- ckwidmer [8ca3fe9d@gateway/web/freenode/ip.140.163.254.157] has joined #shogun | 23:07 | |
shogun-buildbot_ | build #602 of lua_modular is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/lua_modular/builds/602 | 23:13 |
--- Log closed Sat Jun 16 00:00:41 2012 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!