Open in new window / Try shogun cloud
--- Log opened Sat Jun 18 00:00:55 2011
-!- f-x [~user@] has joined #shogun00:04
-!- blackburn [~blackburn@] has quit [Quit: Leaving.]00:28
-!- blackburn [~blackburn@] has joined #shogun00:29
-!- blackburn [~blackburn@] has quit [Quit: Leaving.]00:38
-!- f-x [~user@] has quit [Remote host closed the connection]01:28
-!- in3xes [~in3xes@] has left #shogun ["Leaving"]06:15
-!- in3xes [~in3xes@] has joined #shogun06:15
-!- blackburn [~blackburn@] has joined #shogun12:18
-!- in3xes [~in3xes@] has quit [Remote host closed the connection]12:33
-!- in3xes [~in3xes@] has joined #shogun12:41
-!- in3xes_ [~in3xes@] has joined #shogun13:07
-!- in3xes_ is now known as inxes13:07
-!- inxes is now known as in3xes_13:08
-!- in3xes [~in3xes@] has quit [Ping timeout: 240 seconds]13:09
-!- in3xes_ is now known as in3xes13:11
-!- blackburn [~blackburn@] has quit [Ping timeout: 240 seconds]15:41
-!- f-x [~user@] has joined #shogun17:21
-!- in3xes_ [~in3xes@] has joined #shogun18:04
-!- in3xes [~in3xes@] has quit [Ping timeout: 244 seconds]18:08
-!- in3xes_ is now known as in3xes19:20
-!- f-x [~user@] has quit [Ping timeout: 260 seconds]20:10
-!- blackburn [~blackburn@] has joined #shogun20:22
-!- f-x [~user@] has joined #shogun20:32
-!- f-x [~user@] has quit [Ping timeout: 260 seconds]20:59
-!- f-x [~user@] has joined #shogun21:26
-!- blackburn1 [~blackburn@] has joined #shogun21:49
-!- blackburn [~blackburn@] has quit [Read error: Connection reset by peer]21:49
CIA-32shogun: Sergey Lisitsyn master * rc58720e / (3 files): Some isomap refactoring -
CIA-32shogun: Sergey Lisitsyn master * r462b0c0 / src/libshogun/preprocessor/LocallyLinearEmbedding.cpp : Use DPOSV instead of DGESV for linear system -
CIA-32shogun: Sergey Lisitsyn master * r42afb17 / (13 files): Added examples for new preprocessors. Fixed names for existing -
CIA-32shogun: Soeren Sonnenburg master * r0b0ffa7 / src/libshogun/preprocessor/LocallyLinearEmbedding.cpp : Merge branch 'lle' of git:// -
CIA-32shogun: Soeren Sonnenburg master * r2e114b5 / (13 files): Merge branch 'master' of git:// -
@sonney2kblackburn1, could you please not use random data but load any of the toy data in your example scripts?22:35
@sonney2kor save the random data and make me upload it22:35
blackburn1sonney2k: no problem22:36
blackburn1in a minute22:36
blackburn1sonney2k: done22:46
blackburn1you already merged..22:47
blackburn1I've done amend commit22:47
@sonney2kcan you do the merge22:50
blackburn1I'll try22:50
blackburn1messed it up! damn22:53
blackburn1sonney2k: it is ok now23:03
CIA-32shogun: Sergey Lisitsyn master * re8143c5 / (3 files): Fixed examples -
@sonney2kblackburn1, btw regarding the bool vs enum issue - I became a big fan of using enums not boolean...23:04
blackburn1sonney2k: just worried of usage in interfaces23:05
@sonney2kblackburn1, it is just more readable when someone writes  optimizer(E_MAXIMIZE) instead of optimizer(true)23:05
blackburn1if it is ok - sure I'm for that too23:05
@sonney2kblackburn1, people just have to import the constants too23:05
blackburn1my opinion is if it is non-binary then better boolean23:06
blackburn1vice versa23:06
blackburn1I mean if there are many options - it is better of course - we simply can't do this with boolean23:07
@sonney2kblackburn1, yeah I used to do that as well but I no longer like it23:07
@sonney2ksimply because one cannot see from the line optimizer(true) what it will do23:07
* sonney2k checks the weekly reports23:09
blackburn1I did :P23:10
@sonney2kwe need a better scheme with some subject rules like weekly report #xx it becomes messy now23:15
@sonney2kI think f-x and sploving are still missing23:16
blackburn1sonney2k: alesis too23:23
blackburn1it is just sunday now, may be they will report today23:24
@sonney2kblackburn1, ahh sploving did already23:25
@sonney2kbtw did you read his mail23:25
@sonney2kwe need to return SGVector / SGMatrix stuff in evaluation23:25
@sonney2kotherwise the interface will not recognize23:26
blackburn1oh I didn't23:26
blackburn1could you please say date?23:26
@sonney2kblackburn1, yesterday 11:5123:29
@sonney2kweekly report :)23:29
blackburn1I see23:33
blackburn1will change23:34
blackburn1sonney2k: should get_ROC return a copy of roc graph?23:35
@sonney2kif possible never copy23:43
blackburn1I don't remember why I used copy23:43
blackburn1ah, I guess because of typemaps23:43
@sonney2kblackburn1, yes that was the default before23:45
@sonney2kfor these swig compat functions23:45
CIA-32shogun: Soeren Sonnenburg master * r3c48b36 / src/libshogun/preprocessor/LocallyLinearEmbedding.cpp :23:49
CIA-32shogun: add a dummy return statement - even though it never reached it fixes a23:49
CIA-32shogun: compiler warning -
CIA-32shogun: Soeren Sonnenburg master * r1b54f6b / examples/undocumented/java_modular/ : don't output anything in check script -
CIA-32shogun: Soeren Sonnenburg master * rb0c6415 / (2 files): add SGMatrix<float64_t> support to custom kernel constructor -
blackburn1sonney2k: not yet, fixing graphical examples23:50
@sonney2kblackburn1, ? how is that related?23:50
blackburn1sonney2k: well I have stored roc graph row-major23:50
blackburn1changing it23:50
@sonney2kahh the old problem :)23:51
blackburn1ah twisted mind23:51
blackburn1for column major: 2 columns23:51
blackburn12*j and 2*j+1?23:51
blackburn1btw this example was wrong anyway23:58
blackburn1classify() is used23:58
--- Log closed Sun Jun 19 00:00:27 2011