--- Log opened Mon Dec 12 00:00:19 2011 | ||
-!- puneetgoyal [~puneet@115.241.195.173] has quit [Quit: Leaving] | 00:07 | |
-!- blackburn [~blackburn@188.168.5.195] has quit [Quit: Leaving.] | 00:51 | |
-!- puneetgoyal [~puneet@115.242.22.38] has joined #shogun | 04:30 | |
-!- puneetgoyal [~puneet@115.242.22.38] has quit [Ping timeout: 248 seconds] | 04:50 | |
-!- Ram108 [~amma@14.96.10.140] has joined #shogun | 05:25 | |
-!- Ram108 [~amma@14.96.10.140] has quit [Quit: Ex-Chat] | 05:36 | |
-!- blackburn [~blackburn@188.168.4.1] has joined #shogun | 07:18 | |
15SAAI18M | shogun: Soeren Sonnenburg master * r02ce3cd / src/ChangeLog : updated changelog - http://git.io/70OkIg | 08:29 |
---|---|---|
15SAAI18M | shogun: Soeren Sonnenburg master * rde7fbfd / src/NEWS : Mention osx / cygwin build fixes in NEWS. - http://git.io/shDPqQ | 08:29 |
shogun-buildbot | build #396 of cmdline_static is complete: Failure [failed git] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cmdline_static/builds/396 blamelist: sonne@debian.org | 08:38 |
shogun-buildbot | build #379 of r_static is complete: Exception [exception git] Build details are at http://www.shogun-toolbox.org/buildbot/builders/r_static/builds/379 blamelist: sonne@debian.org | 08:38 |
shogun-buildbot | build #371 of python_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/python_static/builds/371 | 08:39 |
-!- Ram108 [~amma@115.117.240.206] has joined #shogun | 08:54 | |
Ram108 | hi..... thr? | 08:55 |
sonne|work | Good morning... | 09:33 |
Ram108 | hi mr sonnenburg | 09:34 |
Ram108 | Mr Lisitsyn helped me to set up some classifiers but i could not get it to work | 09:36 |
Ram108 | i was wondering if either of u could aid me a little more | 09:36 |
sonne|work | please be more specific - just directly ask the question | 10:02 |
blackburn | Ram108: how can I help you? | 10:29 |
Ram108 | sorry for the delay | 10:37 |
Ram108 | well i have created the RealFeatures and Labels and KNN requires train and test vector sets..... how do i generate them from the dataset | 10:39 |
Ram108 | http://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data is the data set he asked me to classify | 10:40 |
-!- blackburn [~blackburn@188.168.4.1] has quit [Quit: Leaving.] | 10:42 | |
Ram108 | well hi Mr Lisitsyn hope u had a great day | 10:42 |
-!- Ram108 [~amma@115.117.240.206] has quit [Ping timeout: 244 seconds] | 12:10 | |
-!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has joined #shogun | 12:45 | |
-!- puneetgoyal [7aa0b226@gateway/web/freenode/ip.122.160.178.38] has joined #shogun | 12:50 | |
-!- puneetgoyal [7aa0b226@gateway/web/freenode/ip.122.160.178.38] has quit [Quit: Page closed] | 13:02 | |
-!- puneetgoyal [~chatzilla@115.240.104.102] has joined #shogun | 13:02 | |
-!- puneetgoyal_ [~chatzilla@115.241.130.116] has joined #shogun | 14:02 | |
-!- puneetgoyal [~chatzilla@115.240.104.102] has quit [Ping timeout: 248 seconds] | 14:02 | |
-!- puneetgoyal_ is now known as puneetgoyal | 14:02 | |
-!- Ram108 [~amma@14.96.211.205] has joined #shogun | 14:41 | |
Ram108 | hi | 14:41 |
Ram108 | Mr. Lisitsyn can u help me out with the rest of the implementation | 14:42 |
blackburn | yes, but call me more informal, I'm too young to be mr :D | 14:46 |
Ram108 | lo, :) okay sure :) | 14:46 |
blackburn | how can i help you? | 14:47 |
Ram108 | well i have created the RealFeatures and Labels. KNN requires train and test vector sets | 14:47 |
Ram108 | are test and train vectors created by handpicking some out of the data set | 14:48 |
blackburn | yes | 14:48 |
blackburn | so? | 14:48 |
blackburn | ah, it is a question? :) | 14:49 |
Ram108 | so shill i create that...... and then use KNN | 14:49 |
Ram108 | ? | 14:49 |
blackburn | yes, you have to create RealFeatures and Labels for train vectors and RealFeatures for test vectors | 14:49 |
blackburn | example: | 14:50 |
blackburn | [1,2,3] label 0 | 14:50 |
blackburn | [3,4,5] label 0 | 14:50 |
blackburn | [9,8,7] label 1 | 14:50 |
blackburn | and test set | 14:51 |
blackburn | [0,1,2] | 14:51 |
blackburn | [7,8,6] | 14:51 |
Ram108 | oh hmmm | 14:51 |
blackburn | feature matrix for train set is | 14:51 |
blackburn | [[1,3,9], | 14:51 |
blackburn | [2,4,8] | 14:51 |
blackburn | [3,5,7]] | 14:51 |
blackburn | then labels are [0,0,1] | 14:51 |
blackburn | same for test set | 14:51 |
blackburn | it should be done in this fashion | 14:52 |
Ram108 | hmmm :) | 14:52 |
Ram108 | thanks :) | 14:52 |
blackburn | then after applying you will get probably [0,1] for this data | 14:52 |
Ram108 | oh hmmm | 14:53 |
blackburn | after that you can compare labels that you've got with classifier and 'ground truth' labels | 14:53 |
Ram108 | well i have created Real features bu calling the RealFeatures() and labels by invoking the Label() method | 14:55 |
blackburn | and? | 14:55 |
Ram108 | and now i ll have to create a seperate test and train set by spooling a few samples and making then into the form vector1=[5.0,3.4,1.5,0.2] and label 0 for Iris-setosa | 14:57 |
Ram108 | likewise for all the other test and train examples | 14:58 |
blackburn | separate it before creating RealFeatures | 14:59 |
Ram108 | [[vec1],[vec2],[vec3]......] and labels [ 0, 0, 1, 1, 2, 2.... ] | 15:00 |
Ram108 | yeah i think i got it | 15:00 |
blackburn | ok | 15:00 |
Ram108 | i ll code it to u and email u the file | 15:00 |
blackburn | okay | 15:00 |
Ram108 | its just a simple modification to the old file i sent yesterday right? | 15:01 |
blackburn | something like that | 15:01 |
Ram108 | hmmm :) | 15:01 |
Ram108 | thanks a lot..... am i bugging u? | 15:01 |
Ram108 | sorry if i am.... | 15:01 |
-!- puneetgoyal [~chatzilla@115.241.130.116] has quit [Ping timeout: 248 seconds] | 15:02 | |
blackburn | no, all is ok | 15:02 |
Ram108 | :) | 15:02 |
-!- puneetgoyal [~chatzilla@115.241.130.116] has joined #shogun | 15:03 | |
puneetgoyal | hello :) can you please tell RealFeatures function? | 15:05 |
Ram108 | hi puneet | 15:06 |
puneetgoyal | what all datatype are handled in RealFeatures function? | 15:06 |
puneetgoyal | hello Ram :) | 15:06 |
Ram108 | heard u r another guy wanting to learn this toolbox out | 15:07 |
Ram108 | :P | 15:07 |
puneetgoyal | yeah...I guess you have already done this...can you please help me out with this | 15:08 |
Ram108 | well all i did was.... | 15:09 |
Ram108 | can i email u my code? | 15:09 |
Ram108 | do u have skype or sumthn? | 15:09 |
puneetgoyal | check out the other window | 15:10 |
-!- Ram108 [~amma@14.96.211.205] has quit [Ping timeout: 248 seconds] | 15:14 | |
blackburn | RealFeatures is not a function, it is a class | 15:14 |
puneetgoyal | yeah..I mean to ask what arguments does it take? blackburn ? | 15:18 |
blackburn | puneetgoyal: numpy matrix | 15:18 |
puneetgoyal | I have seen the class...it says some dotfeatures | 15:18 |
blackburn | RealFeatures is a SimpleFeatures<float64_t> | 15:19 |
blackburn | and SimpleFeatures is inherited from DotFeatures | 15:19 |
puneetgoyal | ok...but float64 is only for floating variable right? so how do we let it work for strings? | 15:20 |
blackburn | StringFeatures | 15:22 |
puneetgoyal | but it says StringFeatures is not defined...do I have to import some another file for this? | 15:23 |
sonne|work | StringCharFeatures ... | 15:23 |
puneetgoyal | ok..thank you :) | 15:28 |
puneetgoyal | and while creating a numpy matrix again we have to define a datatype...what if we have both float and character variables inside the data? we can use the object datatype..but I guess there is no class for it to convert it into feature matrix | 15:29 |
-!- Ram108 [~amma@14.99.111.107] has joined #shogun | 15:31 | |
sonne|work | puneetgoyal: create realfeatures for the float stuff and stringcharfeatures for the string stuff | 15:34 |
-!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has quit [Quit: Page closed] | 16:19 | |
Ram108 | puneeth r u thr? | 16:44 |
puneetgoyal | yup Ram108 | 16:55 |
Ram108 | call me ram | 16:55 |
puneetgoyal | oh thats just to give you a notification | 16:55 |
Ram108 | :) | 16:55 |
Ram108 | well do u want me to email u my code? | 16:55 |
Ram108 | its pretty shoddy though | 16:56 |
puneetgoyal | thanks... I figured it out...there was a silly error in my data file :P | 16:56 |
Ram108 | hmmm did u figure out the whole classification thinge? | 16:56 |
Ram108 | i havent reached till that yet | 16:57 |
Ram108 | having trouble | 16:57 |
puneetgoyal | I just needed the realfeature matrix to get going for svm classification for now | 16:57 |
Ram108 | well? could u help me out with that? | 16:57 |
puneetgoyal | yeah..sure...tell me | 16:58 |
Ram108 | well i guess i formed the RealFeature matrix and label | 16:58 |
Ram108 | and now i am supposed to generate test and train vectors | 16:58 |
Ram108 | well do u have an example for that? | 16:59 |
puneetgoyal | ok and what's the problem? | 16:59 |
Ram108 | well i am nt really clear hw that shld look like | 17:00 |
Ram108 | if i had an example i could follow it easily :) | 17:00 |
puneetgoyal | have you seen the undocumented examples? | 17:01 |
-!- ishaanmlhtr [73f21038@gateway/web/freenode/ip.115.242.16.56] has joined #shogun | 17:01 | |
Ram108 | yeah but i really want to classify http://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data | 17:02 |
puneetgoyal | using what classifier? | 17:06 |
puneetgoyal | which classifier? | 17:06 |
Ram108 | KNN | 17:07 |
puneetgoyal | there is an example classifier_knn | 17:08 |
Ram108 | yeah k thanks | 17:08 |
puneetgoyal | just replace your matrix with its matrix | 17:08 |
Ram108 | k thanks :) | 17:08 |
Ram108 | were did u exactly find it? | 17:09 |
puneetgoyal | undocumented examples | 17:10 |
Ram108 | k thanks | 17:10 |
puneetgoyal | examples/undocumented/python_modular | 17:10 |
Ram108 | thnks :) | 17:11 |
puneetgoyal | no problem :) | 17:13 |
Ram108 | :) | 17:13 |
-!- Ram108 [~amma@14.99.111.107] has quit [Quit: Ex-Chat] | 17:21 | |
-!- puneetgoyal_ [~chatzilla@115.240.38.83] has joined #shogun | 17:33 | |
-!- puneetgoyal_ [~chatzilla@115.240.38.83] has quit [Client Quit] | 17:33 | |
-!- ishaanmlhtr [73f21038@gateway/web/freenode/ip.115.242.16.56] has quit [Ping timeout: 258 seconds] | 17:33 | |
-!- puneet_ [~puneet@115.240.38.83] has joined #shogun | 17:34 | |
-!- puneetgoyal [~chatzilla@115.241.130.116] has quit [Ping timeout: 276 seconds] | 17:34 | |
-!- ishaanmlhtr [~ishaanmlh@115.241.206.137] has joined #shogun | 17:35 | |
-!- puneet_ is now known as puneetgoyal | 17:35 | |
15SAAI18M | shogun: Soeren Sonnenburg master * r8d20d53 / (4 files): updated reference documentation - http://git.io/GK46bg | 17:35 |
15SAAI18M | shogun: Soeren Sonnenburg master * ra05d132 / examples/undocumented/r_static/mkl_regression_sine.R : use 2-norm for mkl regression R example - http://git.io/06447Q | 17:35 |
15SAAI18M | shogun: Soeren Sonnenburg master * r2857ac8 / (2 files): fix set_subkernel_weights for wd* kernels - http://git.io/n2h7vg | 17:35 |
15SAAI18M | shogun: Soeren Sonnenburg master * r4870249 / src/Makefile.template : fix testsuite paths - http://git.io/JAdlyg | 17:39 |
shogun-buildbot | build #397 of cmdline_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cmdline_static/builds/397 | 17:41 |
shogun-buildbot | build #399 of cmdline_static is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cmdline_static/builds/399 blamelist: sonne@debian.org | 17:51 |
shogun-buildbot | build #382 of r_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/r_static/builds/382 | 17:54 |
shogun-buildbot | build #385 of octave_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/octave_static/builds/385 | 17:56 |
-!- Ram108 [~amma@14.96.10.254] has joined #shogun | 18:40 | |
-!- ishaanmlhtr [~ishaanmlh@115.241.206.137] has quit [Ping timeout: 255 seconds] | 18:41 | |
-!- rrenaud [~rrenaud@cpe-66-108-112-118.nyc.res.rr.com] has left #shogun [] | 19:00 | |
-!- Ram108 [~amma@14.96.10.254] has quit [Quit: Ex-Chat] | 19:57 | |
15SAAI18M | shogun: Soeren Sonnenburg master * r7a0a00f / doc/Doxyfile : Preparing for new Release shogun_1.1.0 - http://git.io/TG986w | 20:16 |
15SAAI18M | shogun: Soeren Sonnenburg shogun_1.1 * ra15dc69 / src/shogun/lib/versionstring.h : Tagging shogun_1.1.0 release - http://git.io/BImdog | 20:18 |
-!- blackburn [~blackburn@188.168.4.11] has joined #shogun | 20:28 | |
shogun-buildbot | build #375 of python_static is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/python_static/builds/375 blamelist: sonne@debian.org | 20:29 |
puneetgoyal | hey ....can I have these data files... fm_train_real.dat, fm_test_real.dat and label_train_twoclass.dat | 20:30 |
blackburn | git submodule update | 20:37 |
@sonney2k | git submodule update --init | 20:46 |
blackburn | second is better hehe | 20:48 |
blackburn | sonney2k: will you have a chance to pack a .deb? | 20:49 |
@sonney2k | hmmhh sth went wrong - the .tar.bz2 is >300M | 20:51 |
puneetgoyal | thanks :) | 20:52 |
blackburn | heh | 20:55 |
-!- puneetgoyal [~puneet@115.240.38.83] has quit [Quit: Leaving] | 21:20 | |
@sonney2k | blackburn, mind preparing release notes that you can send out to the shogun mailinglist when I give you the go | 21:31 |
@sonney2k | ? | 21:31 |
blackburn | sonney2k: you mean some announcement on release of 1.1.0? | 21:31 |
@sonney2k | yes | 21:31 |
blackburn | well I could | 21:32 |
@sonney2k | just copy what is on the website/ what I posted and add some brief summary - I guess this time it is mostly compile fixes, speedup compilation + your preproc/converter stuff | 21:32 |
blackburn | website? | 21:33 |
blackburn | what did you posted? | 21:33 |
@sonney2k | check the 1.0.0 announcement on the mailinglist | 21:35 |
blackburn | hmm ok | 21:35 |
-!- puneetgoyal [~puneet@115.240.38.83] has joined #shogun | 21:37 | |
blackburn | sonney2k: should I send it? | 21:54 |
@sonney2k | not yet | 21:55 |
@sonney2k | blackburn, can you please have a look at the .tar.bz2 I generated? | 22:14 |
@sonney2k | I had to redo it manually... | 22:14 |
@sonney2k | seems like my release script still has a bug | 22:14 |
blackburn | sonney2k: hmm sure, where can I find it? | 22:14 |
@sonney2k | blackburn, http://www.shogun-toolbox.org :-) | 22:14 |
blackburn | sonney2k: have you had experience of linking shogun and MKL? | 22:16 |
@sonney2k | I have no time for that now | 22:16 |
blackburn | I tried this morning and failed :) | 22:17 |
@sonney2k | lets get the release out and then I have to sleep ASAP | 22:17 |
blackburn | heh sure, I'm just asked whether you had | 22:17 |
@sonney2k | have to get up in ~5 hrs | 22:17 |
@sonney2k | flying to granada / nips tomorrow | 22:17 |
blackburn | wow | 22:17 |
blackburn | ehm | 22:17 |
@sonney2k | blackburn, so does it compile? | 22:17 |
blackburn | make[1]: *** No rule to make target `.depend', needed by `libshogun.so.11.0'. Stop. | 22:17 |
blackburn | sonney2k: ^ | 22:19 |
@sonney2k | blackburn, well did you run configure?? | 22:20 |
blackburn | I'm not an idiot yet :D | 22:21 |
blackburn | does it compile smoothly on your machine? | 22:22 |
@sonney2k | great then sth went wrong | 22:22 |
blackburn | sonney2k: do you know the problem? | 22:26 |
@sonney2k | no I don't understand what is going on | 22:27 |
blackburn | I faced this error before but do not know how to solve it | 22:27 |
blackburn | what is .depend? | 22:28 |
blackburn | sonney2k: btw make called in / leads to 'fatal: Not a git repository (or any of the parent directories): .git' | 22:32 |
@sonney2k | why didn't you tell me? | 22:32 |
blackburn | sonney2k: did not tell what? | 22:32 |
@sonney2k | that you have seen this error at some point | 22:32 |
blackburn | well it disappeared and I faced it I guess 10000 years ago | 22:33 |
blackburn | don't you have this problem? | 22:33 |
@sonney2k | never had | 22:33 |
@sonney2k | but now I do | 22:33 |
blackburn | haah | 22:34 |
@sonney2k | I cannot easily fix this problem... | 22:36 |
blackburn | why depend is not generated? | 22:36 |
@sonney2k | I found the problem | 22:39 |
blackburn | sonney2k: what is the problem? are you a sorcerer? | 22:40 |
@sonney2k | please try again | 22:50 |
@sonney2k | it should work now | 22:51 |
blackburn | again.. | 22:52 |
blackburn | no | 22:53 |
blackburn | sonney2k: works | 22:53 |
@sonney2k | great | 22:54 |
@sonney2k | then send out this email | 22:54 |
15SAAI18M | shogun: Soeren Sonnenburg master * r8c1f1f2 / (13 files in 3 dirs): updated reference documentation - http://git.io/QlXShw | 22:54 |
15SAAI18M | shogun: Soeren Sonnenburg master * r35ba6f1 / src/Makefile.template : fix Makefile.template for the case where no git repository is available (aka in release tarballs) - http://git.io/68-9fA | 22:54 |
15SAAI18M | shogun: Soeren Sonnenburg master * r12ea8b8 / Makefile : improve make release script - http://git.io/UFx-Kw | 22:54 |
blackburn | sonney2k: done | 22:55 |
blackburn | you can sleep with accomplished mood I guess :) | 22:55 |
@sonney2k | well we should also announce on freshmeat / mloss.org / and I should write a blog entry / g+ | 22:56 |
@sonney2k | but otherwise yes | 22:56 |
@sonney2k | I will do this tomorrow on the ride to the airport... | 22:57 |
blackburn | sonney2k: I can announce on G+ | 22:57 |
blackburn | should I? | 22:57 |
@sonney2k | so better get *some* hours of sleep now | 22:57 |
blackburn | sure, sleep well | 22:57 |
@sonney2k | sure why not - it won't hurt if I do it too later :) | 22:57 |
@sonney2k | nite | 22:57 |
blackburn | I will be short | 22:57 |
blackburn | Shogun 1.1.0 is out | 22:57 |
blackburn | :D | 22:57 |
blackburn | see you | 22:58 |
puneetgoyal | awesome! | 23:00 |
-!- alesis-novik [~alesis@5ac5713d.bb.sky.com] has joined #shogun | 23:06 | |
alesis-novik | sonney2k, I was too distracted and forgot to reply to your e-mail | 23:07 |
alesis-novik | I use the vector in GMM just for storing separate Gaussians, not as a mathematical concept(?) | 23:07 |
alesis-novik | Was that what you were asking? | 23:08 |
@sonney2k | alesis-novik, I was asking if you could change it to Dynarray or so | 23:09 |
shogun-buildbot | build #384 of r_static is complete: Failure [failed test test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/r_static/builds/384 blamelist: sonne@debian.org | 23:09 |
alesis-novik | Congratulations on getting 1.1.0 out guys! | 23:09 |
@sonney2k | alesis-novik, because I want to make vector a mathematical thing like v+u = w | 23:09 |
@sonney2k | which won't work with your gaussians | 23:09 |
alesis-novik | sonney2k, well, you can technically add Gaussians, though it wouldn't make any sense in this context | 23:10 |
alesis-novik | sonney2k, so is that still an open problem? | 23:11 |
@sonney2k | yes so please convert it then to Dynamic / Dynarray | 23:11 |
@sonney2k | yes | 23:11 |
alesis-novik | I have some time now since it's close to Christmas and research has slowed down | 23:11 |
alesis-novik | sonney2k, I'll do it over the next few days then | 23:12 |
@sonney2k | great! | 23:12 |
@sonney2k | then I can continue my SGVector effort :) | 23:12 |
alesis-novik | I'd like to have more time to work on Shogun, but lately I've been working on my own topic, so not really enough time :/ | 23:13 |
shogun-buildbot | build #388 of octave_static is complete: Failure [failed test test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/octave_static/builds/388 blamelist: sonne@debian.org | 23:17 |
@sonney2k | alright I have to sleep now | 23:18 |
@sonney2k | cu all | 23:18 |
-!- puneetgoyal [~puneet@115.240.38.83] has quit [Quit: Leaving] | 23:22 | |
-!- blackburn [~blackburn@188.168.4.11] has quit [Quit: Leaving.] | 23:53 | |
-!- abcd_ [564b21ef@gateway/web/freenode/ip.86.75.33.239] has joined #shogun | 23:56 | |
-!- abcd_ is now known as Guest1514 | 23:56 | |
-!- Guest1514 [564b21ef@gateway/web/freenode/ip.86.75.33.239] has quit [Client Quit] | 23:59 | |
--- Log closed Tue Dec 13 00:00:19 2011 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!