--- Log opened Thu Feb 16 00:00:19 2012 | ||
-!- n4nd0 [~n4nd0@s83-179-44-135.cust.tele2.se] has quit [Quit: Leaving] | 00:20 | |
CIA-18 | shogun: Soeren Sonnenburg master * r3649f3b / (31 files in 9 dirs): | 03:25 |
---|---|---|
CIA-18 | shogun: Merge pull request #367 from karlnapf/master | 03:25 |
CIA-18 | shogun: A draft for training on fixed kernel matrices/data in general - http://git.io/qv89yQ | 03:25 |
CIA-18 | shogun: Soeren Sonnenburg master * r3e3db13 / (33 files in 14 dirs): | 03:25 |
CIA-18 | shogun: add linear least squares and ridge regression | 03:25 |
CIA-18 | shogun: This add linear ridge regression and a convenience class | 03:25 |
CIA-18 | shogun: CLeastSquaresRegression calling CLinearRidgeRegression with | 03:25 |
CIA-18 | shogun: regularization parameter tau=0. To not cause confusion KRR is | 03:25 |
CIA-18 | shogun: renamed to KernelRidgeRegression throughout examples/code. - http://git.io/pQJ0OA | 03:25 |
CIA-18 | shogun: Soeren Sonnenburg master * r0338587 / examples/undocumented/python_modular/regression_linear_ridge_modular.py : remove unused gaussian kernel from example - http://git.io/Sf2o5w | 03:25 |
CIA-18 | shogun: Soeren Sonnenburg master * rc612a4c / (26 files in 7 dirs): | 03:28 |
CIA-18 | shogun: Merge pull request #373 from karlnapf/master | 03:28 |
CIA-18 | shogun: a bunch of improvements (+27 more commits...) - http://git.io/0IzzXg | 03:28 |
-!- shogun-buildbot_ [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting] | 03:35 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 03:36 | |
-!- nageshp [~nagesh@triband-mum-120.61.0.77.mtnl.net.in] has joined #shogun | 04:06 | |
-!- nageshp [~nagesh@triband-mum-120.61.0.77.mtnl.net.in] has left #shogun [] | 04:12 | |
shogun-buildbot | build #316 of ruby_modular is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/ruby_modular/builds/316 | 05:08 |
-!- dfrx [~f-x@inet-hqmc05-o.oracle.com] has joined #shogun | 05:36 | |
shogun-buildbot | build #344 of octave_modular is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/octave_modular/builds/344 blamelist: viktor.gal@maeth.com, heiko.strathmann@gmail.com, fernando.iglesiasg@gmail.com, slianoglou@gmail.com, blackburn91@gmail.com, sonne@debian.org | 06:25 |
-!- n4nd0 [82ed3d7d@gateway/web/freenode/ip.130.237.61.125] has joined #shogun | 08:30 | |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 08:51 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 09:02 | |
CIA-18 | shogun: Soeren Sonnenburg master * r32bfe20 / examples/undocumented/octave_modular/regression_kernel_ridge_modular.m : fix missing rename KRR -> KernelRidgeRegression - http://git.io/rcpScw | 09:24 |
-!- n4nd0 [82ed3d7d@gateway/web/freenode/ip.130.237.61.125] has quit [Quit: Page closed] | 10:02 | |
shogun-buildbot | build #345 of octave_modular is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/octave_modular/builds/345 | 10:03 |
wiking | doh, why SGVector does not have an iterator class around it... ;( | 11:24 |
sonne|work | wiking: SGVector used to be just a float ptr + length for storing data... | 11:26 |
wiking | well i just kind of like wish there'd be an iterator struct for it now :P | 11:27 |
sonne|work | I wish we introduced proper reference counts in SGVector & co | 11:29 |
sonne|work | need that now but at some point 'every byte counted' | 11:29 |
-!- n4nd0 [~n4nd0@n145-p102.kthopen.kth.se] has joined #shogun | 12:10 | |
-!- dfrx [~f-x@inet-hqmc05-o.oracle.com] has quit [Quit: Leaving.] | 12:14 | |
wiking | sonne|work: are there any rules of adding BSD licensed code to shogun? | 13:00 |
wiking | i mean most of the code is derived from a bsd licensed file that i'd like to commit soon, that's why i'm asking | 13:01 |
sonne|work | wiking: anything GPL compatible is fine | 13:24 |
wiking | cool | 13:24 |
-!- stratis [9232285a@gateway/web/freenode/ip.146.50.40.90] has joined #shogun | 14:14 | |
stratis | hi!!! | 14:14 |
stratis | to all of you | 14:14 |
stratis | i m a new user of shogun | 14:14 |
stratis | and I am working with the matlab version | 14:14 |
stratis | so, i cant seem to find the function for calculating the area and the precision-recall curve | 14:15 |
stratis | or at least how to get the return distances from the svm | 14:15 |
stratis | i have a trained a multi-class svm | 14:16 |
sonne|work | stratis: hi - the perf measures are only available in modular interfaces | 14:16 |
stratis | oh ok | 14:16 |
stratis | so | 14:16 |
stratis | i have a multi class svm | 14:16 |
stratis | can i at least some how get | 14:16 |
sonne|work | stratis: IIRC you can do get_classifier with an index | 14:16 |
stratis | the distances for the samples | 14:16 |
sonne|work | sg('get_classifier',0) or so | 14:16 |
stratis | ok i ll try that | 14:17 |
stratis | when i try | 14:20 |
stratis | sg('get_classifier',0) | 14:20 |
stratis | it says | 14:20 |
stratis | [bias, weights]=sg('get_classifier', [index in case of MultiClassSVM]) | 14:20 |
stratis | the same when I don't use an index | 14:20 |
stratis | how can I see the variables stored in "sg" ? | 14:20 |
sonne|work | well you need to do | 14:24 |
sonne|work | [bias, weights]=sg('get_classifier',0) | 14:24 |
sonne|work | :) | 14:24 |
-!- stratis [9232285a@gateway/web/freenode/ip.146.50.40.90] has quit [Ping timeout: 245 seconds] | 14:27 | |
-!- n4nd0 [~n4nd0@n145-p102.kthopen.kth.se] has quit [Quit: Leaving] | 15:29 | |
wiking | mmm why isn't it possible to define a class that has no default constructor? | 15:48 |
sonne|work | wiking: for serialization we need that | 16:07 |
wiking | mmm that'll require then some amount of error checking now :P | 16:23 |
wiking | sonne|work: if i define an enum within a class would that be revealed to the modular API? | 17:05 |
-!- dfrx [~f-x@inet-hqmc04-o.oracle.com] has joined #shogun | 18:06 | |
-!- n4nd0 [~n4nd0@s83-179-44-135.cust.tele2.se] has joined #shogun | 19:20 | |
-!- dfrx [~f-x@inet-hqmc04-o.oracle.com] has left #shogun [] | 19:49 | |
@sonney2k | wiking, no nested classes /enums etc won't work | 20:14 |
wiking | how can i have constants ? | 20:15 |
wiking | that one could use from modular as well | 20:15 |
@sonney2k | use enums | 20:22 |
@sonney2k | just not nested within a class | 20:22 |
wiking | mmmhm | 20:27 |
wiking | and how do i call it in python | 20:27 |
wiking | because i've tried to call for LibSVM LIBSVM_C_SVC | 20:28 |
wiking | but python was crying that it's not defined | 20:28 |
wiking | doh, what's this: TypeError: not a numpy vector of appropriate type | 20:38 |
wiking | :> | 20:38 |
@sonney2k | wiking, not sure what you are trying to do | 21:04 |
@sonney2k | just use the enum as usual - it is just an int also in python | 21:04 |
wiking | well just that i'm creating a new preprocessor | 21:04 |
wiking | and i need to use some enums | 21:04 |
wiking | and i was wondering if those enums are going to be defined in the modular part as well | 21:04 |
wiking | since now as far as i can see | 21:05 |
wiking | i can just use in python only the corresponding int value of the enum | 21:05 |
wiking | but not the macro for it | 21:05 |
wiking | so in case of LibSVM there's a 2 element enum | 21:05 |
wiking | and i was trying in python modular part whether i could use LIBSVM_C_SVC | 21:05 |
wiking | or how is it possible at all to use that enum | 21:06 |
shogun-buildbot | build #151 of nightly_none is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_none/builds/151 | 21:08 |
@sonney2k | wiking, you should just see the enum in python_modular | 21:09 |
@sonney2k | exaclty like in C++ | 21:09 |
wiking | really | 21:24 |
wiking | and you would call it like LIBSVM_C_SVC | 21:24 |
wiking | because if you try print LIBSVM_C_SVC | 21:24 |
wiking | it'll say that no such thing exist... in python | 21:24 |
shogun-buildbot | build #150 of nightly_all is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_all/builds/150 | 21:25 |
@sonney2k | wiking, yes really - look at some of the examples, e.g. classifier_liblinear_modular.py | 21:33 |
@sonney2k | L2R_L2LOSS_SVC_DUAL is an enum | 21:33 |
@sonney2k | alright n8 | 21:33 |
shogun-buildbot | build #137 of nightly_default is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/137 | 21:54 |
-!- blackburn [~qdrgsm@188.168.5.29] has joined #shogun | 21:55 | |
wiking | blackburn: maaan | 22:11 |
blackburn | yeah? | 22:11 |
wiking | i'm about to finish the homkermap | 22:11 |
blackburn | huh | 22:11 |
blackburn | you are fast | 22:11 |
wiking | just trying to test some shit with it | 22:11 |
wiking | because for sure i've fucked up some stuff | 22:11 |
wiking | but yeah i still need to clean the api | 22:11 |
wiking | because now the constructor is like this: CHomogeneousKernelMap (HomogeneousKernelType kernel, HomogeneousKernelMapWindowType wType, float64_t gamma, uint64_t order, float64_t period); | 22:12 |
wiking | btw what's the consensus | 22:12 |
wiking | for the parameters of a class | 22:12 |
blackburn | hmm | 22:12 |
wiking | there should be getter/setter function pairs... ? for each of them? | 22:13 |
blackburn | yeah would be nice | 22:13 |
wiking | aha | 22:13 |
wiking | btw | 22:13 |
wiking | m_parameters->add thingy | 22:14 |
wiking | is for? | 22:14 |
wiking | i didn't get that yet... | 22:14 |
blackburn | serialization stuff | 22:14 |
wiking | when does serialization happen? | 22:14 |
wiking | so when is a class written to file? | 22:14 |
blackburn | hmm when you call it in python | 22:14 |
blackburn | yes | 22:14 |
blackburn | or octave or etc | 22:14 |
wiking | aha so modular | 22:15 |
blackburn | sure, static sucks | 22:15 |
wiking | so each private variables should be serialized, right? :) | 22:15 |
blackburn | sure | 22:16 |
wiking | mmhm | 22:16 |
blackburn | well I don't think it is so necessary for onw | 22:16 |
blackburn | about your constructor | 22:16 |
wiking | well i'm trying to test the thing in python | 22:16 |
blackburn | because now the constructor is like this: CHomogeneousKernelMap (HomogeneousKernelType kernel, HomogeneousKernelMapWindowType wType, float64_t gamma, uint64_t order, float64_t period); | 22:16 |
wiking | so i'll have to do it | 22:16 |
blackburn | is HomogenousKernelType an enum? | 22:16 |
wiking | yep | 22:16 |
blackburn | same for window type? | 22:16 |
wiking | yep | 22:17 |
wiking | if i call m_parameters->add(&m_table, "", ""); | 22:17 |
wiking | where SGVector<float64_t> m_table; | 22:17 |
wiking | will it know how to serialize it/ | 22:17 |
wiking | ? | 22:17 |
blackburn | yes | 22:17 |
blackburn | add_vector IIRC | 22:17 |
blackburn | or simply add if it works | 22:17 |
wiking | ok let's see | 22:21 |
wiking | recompiling ;) | 22:21 |
wiking | hahaha | 22:21 |
wiking | how do you serialize an enum? :) | 22:21 |
wiking | should i cast it into int? | 22:21 |
blackburn | ehmm | 22:21 |
wiking | SG_ADD((machine_int_t*) &liblinear_solver_type, "liblinear_solver_type", | 22:22 |
wiking | "Type of LibLinear solver.", MS_NOT_AVAILABLE); | 22:22 |
wiking | that's in liblinear | 22:22 |
blackburn | yeah | 22:22 |
blackburn | do the same | 22:22 |
wiking | doh | 22:24 |
wiking | add_vector is not working as i thought it would | 22:24 |
wiking | ok now... | 22:26 |
wiking | recompiling :P | 22:26 |
wiking | ok one more question | 22:26 |
wiking | so in an SGMatrix | 22:26 |
wiking | the number of rows are the number of examples? or that the number of columns? | 22:27 |
blackburn | re | 22:35 |
blackburn | wiking: ehmm what is the matrix you are talking about? | 22:36 |
wiking | implementing the apply_to_feature_matrix | 22:36 |
wiking | so there i'm returning a SGMatrix<float64_t> | 22:36 |
wiking | and now i have | 22:36 |
wiking | CSimpleFeatures<float64_t>* simple_features = (CSimpleFeatures<float64_t>*)features; | 22:36 |
wiking | int32_t num_vectors = simple_features->get_num_vectors (); | 22:36 |
wiking | int32_t num_features = simple_features->get_num_features (); | 22:36 |
wiking | SGMatrix<float64_t> result (num_vectors, num_features*(2*m_order+1)); | 22:36 |
wiking | and i wonder if the arguments of result should be swapped or not | 22:37 |
blackburn | feature matrix contains feature vectors row-wise | 22:37 |
blackburn | so dim = rows count | 22:37 |
blackburn | and number of vectors = cols count | 22:37 |
wiking | ahaaa | 22:38 |
wiking | so it should b swapped in that case | 22:38 |
wiking | or wait | 22:38 |
blackburn | yes, should be | 22:38 |
wiking | yey recompile ;P | 22:40 |
wiking | fuckthisshiat | 22:40 |
blackburn | did you simply integrate their code? | 22:42 |
blackburn | btw I've read paper recently, now understand things much better | 22:43 |
wiking | yep basically it's a wrapping | 22:46 |
wiking | i just want to finish it up | 22:47 |
wiking | because then i can start adding new mappings | 22:47 |
wiking | or at least to test new mappings | 22:47 |
wiking | for other kernels.. | 22:47 |
blackburn | yeah | 22:48 |
blackburn | wiking: btw thanks for your efforts | 22:57 |
blackburn | okay power off now :) | 22:58 |
blackburn | see you | 22:58 |
-!- blackburn [~qdrgsm@188.168.5.29] has quit [Quit: Leaving.] | 22:58 | |
--- Log closed Fri Feb 17 00:00:19 2012 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!