--- Log opened Wed Jul 25 00:00:05 2012 | ||
--- Day changed Wed Jul 25 2012 | ||
yoo | dense, string, any .. | 00:00 |
---|---|---|
blackburn | well in case of dense | 00:00 |
blackburn | features->get_feature_matrix().display_matrix("feature_matrix"); | 00:00 |
yoo | allright | 00:00 |
yoo | thanks | 00:00 |
yoo | I have my own function then for printing string features | 00:00 |
blackburn | string is a little more complex - you iterate from 0 to features->get_num_vectors() and get and print strings inside | 00:01 |
yoo | yes I have that | 00:01 |
yoo | just wanna to understand to way iterator works but I loosing my french in that | 00:01 |
yoo | by the way .. can you explain me the purpose of CMulticlassMultipleOutput ? | 00:03 |
blackburn | yoo: heh | 00:03 |
blackburn | yes | 00:03 |
blackburn | to support top k outputs | 00:03 |
blackburn | i.e. if you have 50 classes and you want to store top 5 predictions | 00:04 |
yoo | allright | 00:04 |
yoo | I am suffering of the lack of c++ example towards to python modular ^^ | 00:04 |
blackburn | yoo: feel free to contribute ;) | 00:04 |
blackburn | yoo: it is much easier to use shogun from python | 00:05 |
yoo | sure I did a little bit with mkl and cross valid example :p | 00:05 |
yoo | blackburn: yes but lots of little modification have to be made on c++ | 00:05 |
blackburn | I am open for good ideas | 00:06 |
yoo | blackburn: for example in crossvalid and multiclass: cannot get the average confusion matrix after many crossvalid runs .. | 00:06 |
yoo | I tried to add this but thats not trivial to consider multiclasslabels for base crossvalidation object | 00:07 |
yoo | see what I mean ? | 00:07 |
blackburn | yeah | 00:07 |
yoo | sorry its late my english is going down .. | 00:07 |
blackburn | isn't it native for you? | 00:07 |
yoo | no | 00:08 |
blackburn | I thought it is :) | 00:08 |
yoo | I would like to ass more evaluation for multiclass | 00:08 |
yoo | -ss +dd :p | 00:08 |
blackburn | yeah we can ass some | 00:08 |
yoo | =) | 00:09 |
blackburn | as for confusion matrix | 00:09 |
yoo | confusion matrix after Xvalidation and EER for intra class could be fun )? | 00:09 |
blackburn | I think one way is to extend it using msoutput | 00:09 |
blackburn | what is EER? | 00:09 |
yoo | equal error rate | 00:09 |
-!- blackburn [~blackburn@109.226.92.17] has left #shogun [] | 00:10 | |
yoo | for each class in multiclass evaluation | 00:10 |
-!- blackburn [~blackburn@109.226.92.17] has joined #shogun | 00:10 | |
blackburn | oops | 00:10 |
blackburn | yoo: how to compute it? | 00:11 |
-!- needsch175 [~user@ip-176-198-229-75.unitymediagroup.de] has quit [Ping timeout: 245 seconds] | 00:11 | |
yoo | the same way you compute ROC curve | 00:13 |
blackburn | so ROC one vs rest? | 00:13 |
yoo | yep | 00:13 |
blackburn | I wanted to add that too | 00:14 |
yoo | =) | 00:14 |
yoo | nice | 00:14 |
yoo | I ll be back tommorow to discuss this if you want some help | 00:14 |
yoo | msoutput should be the right way | 00:15 |
blackburn | it is pretty easy actually | 00:15 |
blackburn | I think | 00:15 |
blackburn | I'll do that a little later | 00:15 |
blackburn | ahhh | 00:15 |
yoo | I will follow this to better understand shougunlib | 00:15 |
blackburn | one problem is here though | 00:15 |
blackburn | let me check | 00:15 |
yoo | btw I never succeed in passing msoutput in evaluate function without strange outputs ... | 00:16 |
yoo | and seg fault but maybe I mentioned it already before | 00:16 |
blackburn | do you have that example still? | 00:16 |
yoo | which one ? | 00:16 |
blackburn | segfaulting one | 00:16 |
yoo | ah | 00:17 |
yoo | let me check | 00:17 |
yoo | I just use this one http://pastebin.com/c7wKje0H | 00:18 |
yoo | and add msoutput object in evaluate | 00:18 |
yoo | think to change the inputs | 00:18 |
blackburn | yoo: hmm it works here | 00:22 |
blackburn | ahh | 00:22 |
blackburn | okay my bad probably | 00:27 |
blackburn | yoo: will commit a fix in a few mins | 00:29 |
yoo | ) | 00:29 |
yoo | good night then, see you tomorrow | 00:29 |
blackburn | yoo: what is you are applying shogun for? | 00:30 |
yoo | machine learning =) | 00:30 |
blackburn | heh | 00:30 |
blackburn | okay good night | 00:30 |
yoo | school project in action recognition | 00:30 |
yoo | apply to face expression mostly | 00:30 |
blackburn | school like high school? | 00:30 |
blackburn | :) | 00:31 |
yoo | yes I am 13 y old | 00:31 |
blackburn | no way | 00:31 |
yoo | coded my first neural nets at 11 | 00:31 |
yoo | =) | 00:31 |
yoo | see ya, good night | 00:31 |
blackburn | hard to believe but you would become youngest shogun user ever :D | 00:32 |
yoo | just joking | 00:32 |
yoo | + | 00:32 |
blackburn | exactly | 00:32 |
-!- yoo [575b08cb@gateway/web/freenode/ip.87.91.8.203] has quit [Quit: Page closed] | 00:32 | |
wiking | blackburn: :DDDD | 00:34 |
CIA-18 | shogun: Sergey Lisitsyn master * r8e7d14c / src/shogun/modelselection/ModelSelectionOutput.cpp : Added support of other than linear machines in modelselectionoutput - http://git.io/Wjbm7A | 00:40 |
-!- blackburn [~blackburn@109.226.92.17] has quit [Quit: Leaving.] | 00:41 | |
-!- romi_ [~mizobe@189-18-41-111.dsl.telesp.net.br] has quit [Quit: Leaving] | 01:15 | |
-!- alexlovesdata [~binder@e178022176.adsl.alicedsl.de] has left #shogun [] | 01:47 | |
shogun-buildbot | build #33 of nightly_none is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_none/builds/33 | 03:02 |
shogun-buildbot | build #39 of nightly_default is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/39 | 03:46 |
shogun-buildbot | build #29 of nightly_all is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_all/builds/29 | 04:17 |
-!- n4nd0 [~nando@h121n3-ksg-a11.ias.bredband.telia.com] has joined #shogun | 07:00 | |
-!- n4nd0 [~nando@h121n3-ksg-a11.ias.bredband.telia.com] has quit [Quit: leaving] | 08:35 | |
-!- uricamic [~uricamic@2001:718:2:1634:3de0:9a42:a489:6401] has joined #shogun | 08:46 | |
-!- pluskid [~pluskid@li411-226.members.linode.com] has joined #shogun | 09:06 | |
-!- uricamic1 [~uricamic@cmpgw-27.felk.cvut.cz] has joined #shogun | 09:36 | |
-!- uricamic [~uricamic@2001:718:2:1634:3de0:9a42:a489:6401] has quit [Quit: Leaving.] | 09:36 | |
-!- uricamic1 [~uricamic@cmpgw-27.felk.cvut.cz] has quit [Quit: Leaving.] | 10:01 | |
-!- uricamic [~uricamic@2001:718:2:1634:3de0:9a42:a489:6401] has joined #shogun | 10:04 | |
-!- needsch [~user@ip-176-198-229-75.unitymediagroup.de] has joined #shogun | 10:34 | |
-!- bern4rd [53368e31@gateway/web/freenode/ip.83.54.142.49] has joined #shogun | 10:46 | |
-!- bern4rd [53368e31@gateway/web/freenode/ip.83.54.142.49] has quit [Quit: Page closed] | 11:16 | |
-!- n4nd0 [~nando@h121n3-ksg-a11.ias.bredband.telia.com] has joined #shogun | 11:18 | |
n4nd0 | someone available to help me out? | 11:28 |
koen-shogun | the question would be... with what? | 11:29 |
n4nd0 | koen-shogun: no idea about why this seg fault happens http://pastebin.com/t6vPxi9T | 11:31 |
pluskid | n4nd0: it might be caused by some memory error, wild pointers or anything terrible... maybe run with valgrind might give some hint? | 11:35 |
n4nd0 | pluskid: it doesn't seg fault with valgrind | 11:36 |
n4nd0 | pluskid: do you think that reading valgrind's trace can give me a clue about it in any case? | 11:36 |
pluskid | does valgrind report any memory leak or something? | 11:36 |
pluskid | if valgrind says some pointers are used before initialized, or some memory are written out of bound, etc. it might be the cause of the problem | 11:37 |
n4nd0 | I have some Conditional jump or move depends on uninitialised value(s) and Invalid write of size ... | 11:38 |
n4nd0 | I am going to investigate | 11:38 |
pluskid | n4nd0: good luck then, I'll be right back (for supper :)) | 11:38 |
n4nd0 | pluskid: thank you! I don't have the habit to use valgrind to check seg faults, but to look for memory leaks | 11:39 |
-!- needsch [~user@ip-176-198-229-75.unitymediagroup.de] has quit [Ping timeout: 250 seconds] | 11:51 | |
n4nd0 | pluskid: problem detected and solved :) | 11:54 |
droopy | cool :^) | 11:54 |
n4nd0 | a couple of errors doing arithmetic with indices | 11:54 |
n4nd0 | pluskid: thank you again! | 11:54 |
pluskid | n4nd0: cool! | 11:55 |
koen-shogun | what you can do to find uninitialized memory: during compilation define -DMALLOC_PERTURB_=69 | 11:59 |
koen-shogun | this will initialize all memory allocated using malloc with 69, instead of being uninitialized | 12:00 |
koen-shogun | can shake out some memory bugs quite easily | 12:00 |
n4nd0 | koen-shogun: thank you for that | 12:07 |
n4nd0 | this one was a problem related to writing out of bounds though | 12:07 |
n4nd0 | I will be back later, bye! | 12:08 |
-!- n4nd0 [~nando@h121n3-ksg-a11.ias.bredband.telia.com] has quit [Read error: Operation timed out] | 12:12 | |
wiking | blackburn https://github.com/shogun-toolbox/shogun/commit/8e7d14c92c4be9e6eea5170847861e8acf539b97#commitcomment-1626918 | 12:45 |
wiking | oh you are not here | 12:45 |
-!- gsomix [~gsomix@188.168.128.107] has joined #shogun | 13:34 | |
gsomix | good day | 13:34 |
gsomix | sonney2k, around? | 13:34 |
wiking | heiko ? | 13:48 |
wiking | !seen heiko | 13:48 |
droopy | heiko (heiko@dhcp-184-16.internal.eduroam.ucl.ac.uk) was last seen parting #shogun 18 hours, 9 minutes ago stating "{}". | 13:48 |
wiking | mmm | 13:48 |
wiking | thnx droopy | 13:48 |
droopy | kk | 13:49 |
-!- yoo [575b08cb@gateway/web/freenode/ip.87.91.8.203] has joined #shogun | 14:09 | |
yoo | hi all | 14:09 |
droopy | hiho | 14:09 |
gsomix | droopy, mimimi. | 14:10 |
droopy | gsomix: http://www.beer-club.info/beer-brands/german-beer-brands.html | 14:10 |
gsomix | cool! | 14:11 |
-!- needsch [~user@ip-176-198-229-75.unitymediagroup.de] has joined #shogun | 15:29 | |
-!- gsomix [~gsomix@188.168.128.107] has quit [Read error: Operation timed out] | 15:33 | |
-!- bern4rd [53368e31@gateway/web/freenode/ip.83.54.142.49] has joined #shogun | 15:41 | |
-!- heiko1 [~heiko@host86-183-73-3.range86-183.btcentralplus.com] has joined #shogun | 15:54 | |
-!- yoo [575b08cb@gateway/web/freenode/ip.87.91.8.203] has quit [Quit: Page closed] | 16:06 | |
-!- pluskid [~pluskid@li411-226.members.linode.com] has quit [Quit: Leaving] | 16:40 | |
-!- uricamic [~uricamic@2001:718:2:1634:3de0:9a42:a489:6401] has quit [Quit: Leaving.] | 16:41 | |
CIA-18 | shogun: Heiko Strathmann master * rd8d1158 / (16 files in 3 dirs): changed class structure a bit to avoid redundancy and allowing easier - http://git.io/hI6vgg | 17:00 |
CIA-18 | shogun: Heiko Strathmann master * rfac0f66 / examples/undocumented/python_modular/statistics_quadratic_time_mmd.py : comment changes - http://git.io/jirHuQ | 17:00 |
CIA-18 | shogun: Heiko Strathmann master * rf3b1c98 / (17 files in 4 dirs): Merge pull request #674 from karlnapf/master - http://git.io/Cy9D0Q | 17:00 |
CIA-18 | shogun: Heiko Strathmann master * rfa25275 / examples/undocumented/python_modular/statistics_hsic.py : initial commit - http://git.io/3P-1Aw | 17:08 |
CIA-18 | shogun: Heiko Strathmann master * ra88feed / examples/undocumented/python_modular/statistics_hsic.py : Merge pull request #675 from karlnapf/master - http://git.io/Zp9clw | 17:08 |
-!- bern4rd [53368e31@gateway/web/freenode/ip.83.54.142.49] has quit [Quit: Page closed] | 17:13 | |
-!- n4nd0 [~nando@h121n3-ksg-a11.ias.bredband.telia.com] has joined #shogun | 17:34 | |
-!- n4nd0 [~nando@h121n3-ksg-a11.ias.bredband.telia.com] has quit [Client Quit] | 17:34 | |
-!- blackburn [~blackburn@109.226.92.17] has joined #shogun | 17:44 | |
-!- tiger_eye [~tiger_eye@rrcs-24-106-116-254.central.biz.rr.com] has joined #shogun | 17:55 | |
CIA-18 | shogun: Heiko Strathmann master * rb427525 / (2 files): added inverse gamma cdf - http://git.io/bgzYeg | 17:56 |
CIA-18 | shogun: Heiko Strathmann master * r8f57540 / examples/undocumented/libshogun/statistics.cpp : added test for inverse gamma cdf - http://git.io/GO5_QA | 17:56 |
CIA-18 | shogun: Heiko Strathmann master * radc555a / (4 files): added threshold computation in a better way - http://git.io/TeMaFg | 17:56 |
CIA-18 | shogun: Heiko Strathmann master * r0f8b2a9 / (7 files in 3 dirs): Merge pull request #677 from karlnapf/master - http://git.io/ycTGkQ | 17:56 |
-!- heiko1 [~heiko@host86-183-73-3.range86-183.btcentralplus.com] has left #shogun [] | 17:58 | |
-!- yoo [575b08cb@gateway/web/freenode/ip.87.91.8.203] has joined #shogun | 18:03 | |
-!- yoo [575b08cb@gateway/web/freenode/ip.87.91.8.203] has quit [Ping timeout: 245 seconds] | 18:10 | |
CIA-18 | shogun: Heiko Strathmann master * r6ef226b / (2 files): renamed inverse_normal_distribution to inverse_normal_cdf - http://git.io/T-D-9A | 18:21 |
CIA-18 | shogun: Heiko Strathmann master * ra32c81b / src/shogun/statistics/HSIC.cpp : added comments - http://git.io/8GKh_g | 18:21 |
CIA-18 | shogun: Heiko Strathmann master * r83f385e / src/shogun/statistics/LinearTimeMMD.cpp : minor checks and enabling threshold computations - http://git.io/bSHV-A | 18:21 |
CIA-18 | shogun: Heiko Strathmann master * r5b3d566 / (4 files in 2 dirs): Merge pull request #678 from karlnapf/master - http://git.io/4oExmw | 18:21 |
blackburn | wiking: ping | 18:42 |
blackburn | I just compared asm outputs of eigen3, loop and blas | 18:50 |
blackburn | totally crazy | 18:50 |
blackburn | eigen3 one in particular | 18:50 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Ping timeout: 255 seconds] | 18:53 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 18:53 | |
shogun-buildbot | build #179 of deb2 - static_interfaces is complete: Failure [failed test octave_static] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/179 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 19:38 |
wiking | vl | 19:52 |
wiking | blackburn: resutl | 19:52 |
wiking | ? | 19:52 |
blackburn | wiking: I compared on vectors of length 1 000 000 | 19:53 |
blackburn | computed average or 20000 runs | 19:53 |
blackburn | blas takes 2251 clock in average here | 19:53 |
blackburn | basic loop 2234 | 19:54 |
blackburn | eigen takes 1934 | 19:54 |
blackburn | I run it a few times and results are similar so I expect eigen is really faster | 20:00 |
-!- n4nd0 [~nando@h121n3-ksg-a11.ias.bredband.telia.com] has joined #shogun | 20:41 | |
n4nd0 | I found this article interesting: http://www.cccblog.org/2012/07/20/first-graphlab-workshop-on-large-scale-machine-learning/ | 20:49 |
CIA-18 | shogun: Sergey Lisitsyn master * r086ee0a / (3 files in 2 dirs): Added custom evaluations support for model selection output - http://git.io/JAS3QA | 21:08 |
shogun-buildbot | build #180 of deb2 - static_interfaces is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/180 | 21:16 |
-!- needsch [~user@ip-176-198-229-75.unitymediagroup.de] has quit [Ping timeout: 244 seconds] | 21:58 | |
-!- needsch175 [~user@ip-176-198-229-75.unitymediagroup.de] has joined #shogun | 22:00 | |
-!- heiko [~heiko@host86-174-151-17.range86-174.btcentralplus.com] has joined #shogun | 22:01 | |
-!- heiko [~heiko@host86-174-151-17.range86-174.btcentralplus.com] has left #shogun [] | 22:01 | |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 264 seconds] | 22:32 | |
wiking | blackburn: why wouldn't it throw | 23:00 |
wiking | it throws :) | 23:00 |
blackburn | wiking: well at least it works ;) | 23:01 |
@sonney2k | blackburn, did you run your blas benchmark on optimized blas? | 23:01 |
blackburn | sonney2k: atlas | 23:01 |
@sonney2k | or just the crap that gets installed? | 23:01 |
@sonney2k | well atlas selfcompiled? | 23:02 |
blackburn | sonney2k: installed atlas | 23:02 |
@sonney2k | or the standard atlas | 23:02 |
@sonney2k | ok then | 23:02 |
@sonney2k | then it is the same as basic loop | 23:02 |
blackburn | sonney2k: let me share code | 23:02 |
@sonney2k | you need to compile blas | 23:02 |
@sonney2k | yes let me try this | 23:02 |
blackburn | furthermore may be you will fix something in test that can be unfair | 23:03 |
@sonney2k | what did you compute btw? | 23:03 |
blackburn | sonney2k: dot | 23:03 |
blackburn | https://gist.github.com/3178664 | 23:03 |
blackburn | sonney2k: first one is to power up my cpu :D :D | 23:05 |
blackburn | we have no step back actually - I already told eigen team we use they stuff :D | 23:10 |
blackburn | #define EIGEN_YES_I_KNOW_SPARSE_MODULE_IS_NOT_STABLE_YET | 23:14 |
blackburn | lol | 23:14 |
@sonney2k | blackburn, arg I recently updated atlas and don't will have to compile it again... | 23:19 |
@sonney2k | that can take ages | 23:19 |
@sonney2k | blackburn, current output | 23:19 |
@sonney2k | elapsed 2973.500000 clocks | 23:19 |
@sonney2k | blas elapsed 3006.500000 clocks | 23:19 |
@sonney2k | loop elapsed 3031.500000 clocks | 23:19 |
@sonney2k | eigen elapsed 3006.000000 clocks | 23:19 |
blackburn | well pretty same | 23:19 |
blackburn | they all are | 23:19 |
@sonney2k | yeah no difference | 23:20 |
@sonney2k | I think dot is not the best op | 23:20 |
blackburn | sonney2k: I measured it because we heavy rely on it | 23:20 |
@sonney2k | sth like matrix matrix mul or vector woudl be better | 23:20 |
@sonney2k | blackburn, but measuring time diff with clock() is also not so good | 23:20 |
blackburn | matrix matrix is rare here in shogun | 23:20 |
blackburn | sonney2k: gettimeofday? | 23:20 |
@sonney2k | true | 23:21 |
blackburn | I failed to recall the best way | 23:21 |
@sonney2k | yes | 23:21 |
@sonney2k | ahh I totally forgot about renameing Ctime functions | 23:21 |
blackburn | sonney2k: after we add -I/usr/include/eigen3 I have a lot of warnings | 23:22 |
blackburn | any idea why they haven't been there before? | 23:22 |
-!- n4nd0 [~nando@h121n3-ksg-a11.ias.bredband.telia.com] has quit [Quit: leaving] | 23:47 | |
--- Log closed Thu Jul 26 00:00:17 2012 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!