--- Log opened Sat Aug 04 00:00:17 2012 | ||
shogun-buildbot_ | build #253 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/253 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 00:22 |
---|---|---|
CIA-18 | shogun: Sergey Lisitsyn master * r3d62578 / (4 files in 2 dirs): Revert "drop GMM - it abuses SGVector to store Gaussians" - https://github.com/shogun-toolbox/shogun/commit/3d625786d9fbcd8273b197a0e11744bd81da42db | 00:34 |
CIA-18 | shogun: Sergey Lisitsyn master * ra79170d / (4 files in 2 dirs): Get GMM back to work again - https://github.com/shogun-toolbox/shogun/commit/a79170d593d99745a6ce8cbcd5f88cdf064c83b3 | 00:34 |
shogun-buildbot_ | build #244 of deb1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/244 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 00:36 |
shogun-buildbot_ | build #245 of deb1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/245 | 00:36 |
shogun-buildbot_ | build #254 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/254 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 01:10 |
shogun-buildbot_ | build #210 of deb2 - static_interfaces is complete: Failure [failed test cmdline_static] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/210 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 01:12 |
CIA-18 | shogun: Sergey Lisitsyn master * ra686d8c / (6 files in 4 dirs): Introduced Kernel Mean Matching - https://github.com/shogun-toolbox/shogun/commit/a686d8c75783f49c6ada03ea3204fd74a999dd75 | 02:09 |
CIA-18 | shogun: Sergey Lisitsyn master * rebf738c / examples/undocumented/python_modular/statistics_kmm.py : Added KMM example - https://github.com/shogun-toolbox/shogun/commit/ebf738c3fca7d6524c1beecd9bac6048c139514a | 02:10 |
CIA-18 | shogun: Sergey Lisitsyn master * r9bf02e0 / src/shogun/regression/gp/InferenceMethod.cpp : Fixed inference method crasher - https://github.com/shogun-toolbox/shogun/commit/9bf02e0995f378926b1a9a5799f61aaf9faa0820 | 02:14 |
shogun-buildbot_ | build #255 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/255 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 02:33 |
shogun-buildbot_ | build #211 of deb2 - static_interfaces is complete: Failure [failed test cmdline_static] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/211 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 02:37 |
blackburn | ??? | 02:38 |
-!- blackburn [~blackburn@109.226.80.43] has quit [Quit: Leaving.] | 02:49 | |
shogun-buildbot_ | build #256 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/256 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 02:53 |
shogun-buildbot_ | build #212 of deb2 - static_interfaces is complete: Failure [failed test cmdline_static] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/212 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 02:55 |
shogun-buildbot_ | build #43 of nightly_none is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_none/builds/43 | 03:10 |
shogun-buildbot_ | build #49 of nightly_default is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/49 | 03:33 |
shogun-buildbot_ | build #39 of nightly_all is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_all/builds/39 | 03:56 |
-!- something_someth [~tejaswidp@117.202.17.244] has joined #shogun | 07:03 | |
-!- something_someth [~tejaswidp@117.202.17.244] has left #shogun ["Leaving"] | 07:04 | |
-!- zxtx [~zv@75-59-237-124.lightspeed.sntcca.sbcglobal.net] has joined #shogun | 07:21 | |
-!- zxtx [~zv@75-59-237-124.lightspeed.sntcca.sbcglobal.net] has quit [Ping timeout: 272 seconds] | 08:14 | |
-!- pluskid [~pluskid@108.171.196.83] has joined #shogun | 08:36 | |
pluskid | oops! it seems Euclidean is typo-ed as Euclidian in shogun... | 10:31 |
-!- n4nd0 [~nando@114.Red-2-137-6.dynamicIP.rima-tde.net] has joined #shogun | 12:46 | |
n4nd0 | :-O I cannot believe my eyes | 12:47 |
n4nd0 | sonney2k is not online! | 12:47 |
-!- blackburn [~blackburn@109.226.80.43] has joined #shogun | 13:25 | |
n4nd0 | blackburn: hey! | 13:35 |
blackburn | n4nd0: hey | 13:35 |
n4nd0 | where is the captain?! | 13:35 |
blackburn | how was your flight? | 13:35 |
n4nd0 | blackburn: it was fine, no problem | 13:36 |
blackburn | hmm no idea there was some net split I guess | 13:36 |
n4nd0 | the plne was half empty | 13:36 |
n4nd0 | I had a full row (three seats) just for me :) | 13:36 |
n4nd0 | you know, I had never really thought of it before | 13:36 |
n4nd0 | but Spain is warm .... pretty warm | 13:37 |
n4nd0 | :D | 13:37 |
blackburn | heh | 13:38 |
-!- n4nd0 [~nando@114.Red-2-137-6.dynamicIP.rima-tde.net] has quit [Quit: leaving] | 14:04 | |
CIA-18 | shogun: Sergey Lisitsyn master * rb65dad3 / (2 files in 2 dirs): Removed size() of sgvector - https://github.com/shogun-toolbox/shogun/commit/b65dad36802264fc5cb63ded2436831f23178ebd | 14:05 |
blackburn | pluskid: I can add primal and dual formulation of crammer-singer (I derived it once) to the multiclass part | 14:05 |
pluskid | blackburn: cool! | 14:06 |
pluskid | I'm about to asking you about that :D | 14:06 |
blackburn | pluskid: do you think we should have detailed derivations of such stuff? | 14:07 |
pluskid | so a netsplit cause half of the plane into another space? :D | 14:07 |
blackburn | actually in papers and books they are usually do not derive it | 14:08 |
blackburn | yes probably | 14:08 |
pluskid | blackburn: I don't know the crammer-singer derivation, I think we'd better only add short derivations | 14:08 |
pluskid | blackburn: btw, I saw the skeleton of the "essentials" part, you are not going to finish all those before you die, haha :D | 14:08 |
blackburn | why? I am pretty interested in that stuff | 14:09 |
pluskid | blackburn: I mean there are soooooo many contents | 14:09 |
blackburn | well 10-20 pages :D | 14:09 |
pluskid | no, you should be able to write 600 pages for that part :D | 14:10 |
pluskid | like vapnik's book | 14:10 |
blackburn | lets see :) | 14:10 |
shogun-buildbot_ | build #248 of deb1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/248 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 14:10 |
blackburn | vapnik's EDBED is ~250 pages long IIRC so we should get into that :) | 14:11 |
CIA-18 | shogun: Sergey Lisitsyn master * rc07a8a1 / src/shogun/statistics/KernelMeanMatching.cpp : Fixed wrong size() call - https://github.com/shogun-toolbox/shogun/commit/c07a8a1f62cbd4b5897d8eafdf8b098ef06a59ce | 14:12 |
blackburn | pluskid: so yuo are going to MIT? | 14:13 |
pluskid | blackburn: right after GSoC | 14:13 |
blackburn | pluskid: what are you going to research there? | 14:13 |
pluskid | machine learning :D | 14:13 |
pluskid | specific topic not decided yet | 14:14 |
blackburn | I hope you will manage to use shogun and extend it through research | 14:14 |
pluskid | yeah, hope so :) | 14:15 |
shogun-buildbot_ | build #249 of deb1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/249 | 14:17 |
blackburn | pluskid: I see you are interested in VC too :D | 14:17 |
pluskid | blackburn: VC you mean VC-theory? | 14:19 |
blackburn | yeah | 14:19 |
pluskid | blackburn: yeah, I'm pretty interested in that :D | 14:19 |
blackburn | pluskid: however google translates your blog pretty bad | 14:21 |
pluskid | blackburn: haha, let me try it | 14:21 |
blackburn | pluskid: well actually sometimes pretty well | 14:23 |
blackburn | but sometimes I find it hard to parse :) | 14:23 |
pluskid | oops, beautiful margin-notes and equations disappeared :D | 14:24 |
pluskid | one of my classmate actually use google translate literally for an assignment of paper translation (into chinese) | 14:24 |
pluskid | and the teacher didn't notice it | 14:25 |
pluskid | (though I guess the teacher never looked at the assignments) | 14:25 |
blackburn | heh | 14:26 |
pluskid | ahh... I tried to read some fragments, and feel completely unreadable... | 14:26 |
pluskid | but google translate is the state of the art, right? | 14:27 |
blackburn | no idea about that actually | 14:28 |
blackburn | CIA should has something better to translate secret chinese docs :D | 14:29 |
blackburn | argh have* | 14:29 |
pluskid | no no | 14:29 |
pluskid | they must have human translators | 14:30 |
shogun-buildbot_ | build #257 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/257 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 14:38 |
shogun-buildbot_ | build #213 of deb2 - static_interfaces is complete: Failure [failed test cmdline_static] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/213 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 14:41 |
-!- blackburn [~blackburn@109.226.80.43] has quit [Quit: Leaving.] | 14:58 | |
-!- n4nd0 [~nando@114.Red-2-137-6.dynamicIP.rima-tde.net] has joined #shogun | 15:37 | |
-!- n4nd0 [~nando@114.Red-2-137-6.dynamicIP.rima-tde.net] has quit [Ping timeout: 272 seconds] | 16:38 | |
-!- n4nd0 [~nando@114.Red-2-137-6.dynamicIP.rima-tde.net] has joined #shogun | 16:40 | |
-!- pluskid [~pluskid@108.171.196.83] has quit [Quit: Leaving] | 16:42 | |
-!- n4nd0 [~nando@114.Red-2-137-6.dynamicIP.rima-tde.net] has quit [Quit: leaving] | 18:15 | |
-!- sonney2k [~shogun@7nn.de] has joined #shogun | 18:42 | |
-!- mode/#shogun [+o sonney2k] by ChanServ | 18:42 | |
-!- blackburn [~blackburn@109.226.80.43] has joined #shogun | 19:07 | |
@sonney2k | blackburn, why did you remove size() from sgvector? | 19:22 |
blackburn | sonney2k: it conflicts with csharp stuff | 19:22 |
blackburn | I'd prefer not to touch csharp :) | 19:23 |
blackburn | could break things | 19:23 |
blackburn | but actually I'd like to have stl-like interface in vector | 19:23 |
blackburn | sonney2k: btw gmm is back | 19:25 |
@sonney2k | blackburn, ok so we can just put it back and rename it for c# then | 19:26 |
blackburn | sonney2k: yeah but I don't want to touch csharp typemaps actually :) | 19:26 |
CIA-18 | shogun: Soeren Sonnenburg master * rd7a2b11 / src/Makefile.template : fix path to regression tests in makefile - https://github.com/shogun-toolbox/shogun/commit/d7a2b116bb3fefb75ccff12817919d49c15d0ecb | 19:27 |
CIA-18 | shogun: Soeren Sonnenburg master * r18b4d74 / src/shogun/base/Parameter.cpp : add '\n' at end of error msg - https://github.com/shogun-toolbox/shogun/commit/18b4d74636d1bb95dc3e5f5f898cced4ba441435 | 19:27 |
@sonney2k | blackburn, seen that - nice work! | 19:27 |
blackburn | sonney2k: but with std::vector actually :D | 19:27 |
blackburn | instead of sgvector | 19:27 |
blackburn | however could be changed | 19:27 |
@sonney2k | seen that too and I guess I will convert it to dynarray at some point :-] | 19:27 |
blackburn | should be easy actually | 19:28 |
blackburn | if dynarray supports [] | 19:28 |
@sonney2k | it does | 19:28 |
@sonney2k | however std:: is a nogo certainly for header files - for the rest I guess I will have to live with whatever you do... | 19:29 |
@sonney2k | blackburn, I hope the buildbot gets happy again... | 19:29 |
blackburn | sonney2k: why not to put into headers? | 19:30 |
blackburn | sonney2k: well for serialized fields yes for sure | 19:30 |
blackburn | sonney2k: do you want to write a makefile for tutorial? ;) | 19:32 |
@sonney2k | compile time++ | 19:32 |
blackburn | ahh | 19:32 |
@sonney2k | blackburn, why me? | 19:32 |
blackburn | sonney2k: I am lame with it still (crying) | 19:33 |
@sonney2k | blackburn, I first have to take care of the PRs | 19:34 |
@sonney2k | blackburn, is n4nd0's PR OK? | 19:34 |
@sonney2k | blackburn, but we shouldn't merge anything before we get the buildbot back to green | 19:34 |
blackburn | hmm let me check it again | 19:34 |
blackburn | isn't it all green now? | 19:34 |
@sonney2k | nothing is | 19:35 |
blackburn | sonney2k: no, n4nd0 should get USE_MOSEK guards back | 19:35 |
blackburn | not to merge it right now | 19:35 |
@sonney2k | blackburn, err? I think he did? | 19:35 |
blackburn | ahh | 19:35 |
blackburn | right | 19:35 |
blackburn | sure | 19:35 |
blackburn | I was confused with no commit after | 19:36 |
blackburn | sonney2k: http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/213/steps/test%20cmdline_static/logs/stdio that fails for no real reason | 19:37 |
@sonney2k | blackburn, well due to the test restructuring of wiking/heiko | 19:37 |
@sonney2k | I tried to fix it just now | 19:37 |
blackburn | I see | 19:38 |
@sonney2k | blackburn, do you understand the MatrixFeatures n4nd0 introduced? | 19:38 |
@sonney2k | I dont' understand the doc | 19:38 |
blackburn | sonney2k: well yeah a little | 19:38 |
blackburn | sonney2k: feature vector = matrix | 19:38 |
@sonney2k | why such a weird name? | 19:42 |
blackburn | sonney2k: because it contains a set of matrices | 19:42 |
@sonney2k | I guess underneath it is treated as vector again | 19:42 |
@sonney2k | blackburn, no I mean feature vector | 19:43 |
@sonney2k | I think he should also rather use sth like NDArray to set the matrices | 19:43 |
@sonney2k | no idea how to do it for matrices of different size though | 19:43 |
@sonney2k | (and we have no typemaps to work with these features...) | 19:44 |
shogun-buildbot_ | build #258 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/258 blamelist: Soeren Sonnenburg <sonne@debian.org> | 19:46 |
blackburn | yeah it is a problem | 19:46 |
blackburn | ehm | 19:46 |
blackburn | why does it faill still? | 19:46 |
shogun-buildbot_ | build #214 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/214 | 19:49 |
@sonney2k | blackburn, there is this one typemap I have no idea how to fix for c# | 19:51 |
@sonney2k | blackburn, when you have some method that gets >1 SGVector or so as argument | 19:51 |
@sonney2k | then the build will fail for c# | 19:51 |
blackburn | oooh | 19:51 |
blackburn | interesting | 19:51 |
@sonney2k | so the only fix I have is to %ignore that method for c# | 19:51 |
@sonney2k | blackburn, I guess you didn't revert that typemap %ignore thing... | 19:52 |
@sonney2k | it should have been in interfaces/modular/modshogun_ignores.i | 19:52 |
@sonney2k | in the #ifdef SWIGCSHARP section | 19:52 |
blackburn | okay I see | 19:53 |
blackburn | will add it then | 19:53 |
blackburn | sonney2k: so I can get size back then | 19:54 |
@sonney2k | hehe | 19:54 |
-!- n4nd0 [~nando@114.Red-2-137-6.dynamicIP.rima-tde.net] has joined #shogun | 19:55 | |
n4nd0 | hi sonney2k | 19:55 |
n4nd0 | let me ask you, how do you prefer that I introduce the changes you have commented in the pull request? | 19:56 |
n4nd0 | new commit or ammending? | 19:56 |
n4nd0 | I remember once you said something about amend better | 19:56 |
@sonney2k | blackburn, please also have a look at pluskid's PR again - he explained why he needs a kernel factory | 19:56 |
n4nd0 | however, that would require to force the push and I am not really confident whether it is good to do that | 19:56 |
@sonney2k | blackburn, in principle he needs to *create* kernels of a certain type | 19:57 |
@sonney2k | though I dont' really understand why | 19:57 |
@sonney2k | n4nd0, amend is better ... there are certainly enough commits already in this PR | 19:58 |
@sonney2k | n4nd0, btw good that you are here | 19:58 |
@sonney2k | n4nd0, I don't understand the matrix features doc | 19:58 |
blackburn | that kind of factory could be replaced by copy constructor I think | 19:58 |
blackburn | or some clone() | 19:58 |
n4nd0 | sonney2k: probably the documentation about it is not very elaborated :S | 19:59 |
@sonney2k | blackburn, I mean if pluskid needs to create a new kernel then it will have no arguments set! | 19:59 |
@sonney2k | so it cannot give reasonable results !? | 19:59 |
@sonney2k | I must be missing sth | 20:00 |
@sonney2k | n4nd0, so please explain again :D | 20:00 |
n4nd0 | sonney2k: there is a type in the doc now that I read it again | 20:00 |
@sonney2k | typo? | 20:00 |
n4nd0 | sonney2k: it is caused by the first variable length matrix features I created, and later I went to MatrixFeatures | 20:00 |
n4nd0 | yeah, typo, sorry | 20:00 |
n4nd0 | sonney2k: however ammending you won't be able to note the new changes | 20:03 |
@sonney2k | n4nd0, your PR is just too big | 20:04 |
shogun-buildbot_ | build #259 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/259 blamelist: Soeren Sonnenburg <sonne@debian.org> | 20:04 |
@sonney2k | why didn't you just send a PR for matrixfeautres? | 20:04 |
n4nd0 | sorry about that :( | 20:04 |
@sonney2k | and then others for isolated changes? | 20:05 |
n4nd0 | because I couldn't develop first the whole MatrixFeatures and later the HM-SVM, etc | 20:05 |
-!- zxtx [~zv@75-59-237-124.lightspeed.sntcca.sbcglobal.net] has joined #shogun | 20:07 | |
CIA-18 | shogun: Sergey Lisitsyn master * r4efd4e9 / (2 files in 2 dirs): Restored size(), ignored cast operators and ignored KMM constructor in csharp - https://github.com/shogun-toolbox/shogun/commit/4efd4e966518e928daac51e48d215e95e24f3d1b | 20:18 |
-!- zxtx [~zv@75-59-237-124.lightspeed.sntcca.sbcglobal.net] has quit [Read error: Operation timed out] | 20:20 | |
@sonney2k | n4nd0, what is a twostate model? | 20:22 |
n4nd0 | sonney2k: I have introduced the changes you told me in github and modify the MatrixFeatures doc, please tell me if now is more comprehensible | 20:22 |
n4nd0 | sonney2k: are you familiarized with hmsvm toolbox? | 20:23 |
@sonney2k | n4nd0, no - I know the theory / concepts though | 20:23 |
n4nd0 | sonney2k: ok, so the TwoState model is used for the state representation of the markov model | 20:25 |
n4nd0 | sonney2k: the TwoState model is the simplest example | 20:25 |
n4nd0 | sonney2k: it is needed because the labels do not represent directly the states in the model | 20:25 |
n4nd0 | though the labels are strictly related with the states | 20:26 |
shogun-buildbot_ | build #260 of deb3 - modular_interfaces is complete: Failure [failed compile python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/260 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 20:26 |
blackburn | :( | 20:26 |
@sonney2k | n4nd0, so a toy example right? | 20:28 |
CIA-18 | shogun: Sergey Lisitsyn master * rf513ae0 / src/interfaces/modular/modshogun_ignores.i : Fixed syntax error in includes - https://github.com/shogun-toolbox/shogun/commit/f513ae0cc314956dc39a5a7f8b66639f062e44aa | 20:28 |
n4nd0 | sonney2k: not really | 20:29 |
n4nd0 | it covers any HM-SVM application where the labels can take two different values | 20:30 |
n4nd0 | sonney2k: do you understand what I mean? | 20:33 |
@sonney2k | yes | 20:35 |
@sonney2k | n4nd0, so it could be used for sth like background vs. foreground in images etc right? | 20:35 |
@sonney2k | image segmentation I mean | 20:35 |
n4nd0 | sonney2k: yes, for example | 20:35 |
n4nd0 | yes | 20:36 |
@sonney2k | n4nd0, btw everything currently only works with mosek right? | 20:36 |
n4nd0 | sonney2k: yes, but according to what wiking and uricamic told me yesterday | 20:36 |
n4nd0 | if I define a risk function for the HMSVM model | 20:37 |
blackburn | explain! how can I represent image as markov model? | 20:37 |
n4nd0 | I could use uricamic bundle methods, that is really cool | 20:37 |
n4nd0 | blackburn: it is not really an image what you can represent as a markov model | 20:37 |
n4nd0 | blackburn: because there is no time variance there | 20:37 |
n4nd0 | blackburn: but for example imagine a video sequence | 20:37 |
blackburn | okay sequence, what's next? | 20:38 |
n4nd0 | blackburn: take a window of the video sequence | 20:38 |
n4nd0 | you could model that as an markov model | 20:38 |
blackburn | ehmmmmm | 20:38 |
@sonney2k | n4nd0, that should really have a high priority then - otherwise the code you added cannot be used :/ | 20:38 |
n4nd0 | say I want to know whether this region is in foreground or backbround | 20:38 |
blackburn | window like some area of image right? | 20:38 |
n4nd0 | blackburn: sure | 20:39 |
blackburn | so you just go through all image | 20:39 |
n4nd0 | maybe you could do that to | 20:39 |
n4nd0 | but what I was thinking is | 20:39 |
blackburn | hmm I see | 20:39 |
n4nd0 | you fix your window | 20:39 |
n4nd0 | and run the video sequence | 20:39 |
blackburn | but actually I agree with soeren we should make it usable not only with mosek | 20:40 |
n4nd0 | for each frame (or image) you model if that window is in background or foreground | 20:40 |
n4nd0 | blackburn, sonney2k: I agree with you guys | 20:40 |
blackburn | okay so you get something like | 20:40 |
blackburn | fore - back - fore - back - back | 20:40 |
blackburn | for frames | 20:40 |
blackburn | right? | 20:40 |
n4nd0 | yes | 20:40 |
n4nd0 | those would be the hidden states | 20:41 |
@sonney2k | n4nd0, I still don't get the description | 20:41 |
n4nd0 | sonney2k: of MatrixFeatures? | 20:41 |
@sonney2k | for matrix feats | 20:41 |
@sonney2k | yes | 20:41 |
@sonney2k | n4nd0, can you just explain here please? | 20:41 |
n4nd0 | sonney2k: sure | 20:41 |
n4nd0 | so with dot features we have things like | 20:42 |
n4nd0 | x1 = [1, 0, -1, 4, 5] | 20:42 |
n4nd0 | x2 = [-4, 9, 7, 6, 5] | 20:42 |
n4nd0 | x3 = ... | 20:42 |
n4nd0 | each xi is a feature vector | 20:42 |
n4nd0 | sonney2k: ok? | 20:42 |
n4nd0 | the feature 1 of x1 takes the value 1, of x2 takes the value -4 | 20:43 |
CIA-18 | shogun: Soeren Sonnenburg master * r4ec77d8 / src/interfaces/modular/modshogun_ignores.i : fix %ignore - SGVector is of type index_t not float... - https://github.com/shogun-toolbox/shogun/commit/4ec77d80012857ffdeb8eae2c6314699f160a683 | 20:44 |
@sonney2k | n4nd0, yes | 20:44 |
@sonney2k | blackburn, I hope this commit fixes the build finally... | 20:45 |
blackburn | argh | 20:45 |
blackburn | damn | 20:45 |
blackburn | sure index t | 20:45 |
blackburn | thnask | 20:45 |
n4nd0 | in matrix features each xi is a matrix | 20:46 |
n4nd0 | x1 = [ 1, 0, 4; | 20:46 |
n4nd0 | 2, -1, 3] | 20:46 |
n4nd0 | x2 = [ 4, -3, 2, 5; | 20:46 |
n4nd0 | -6, 5, 0, -3] | 20:46 |
n4nd0 | again each xi is a feature vector | 20:46 |
n4nd0 | I don't really like to say feature vector because now they are represented as matrices | 20:46 |
n4nd0 | I like to use more example | 20:46 |
n4nd0 | anyway | 20:46 |
n4nd0 | now | 20:46 |
n4nd0 | the feature 1 of x1 is [1 0 4], and of x2 [4 -3 2 5] | 20:47 |
@sonney2k | n4nd0, ok so you basically have matrices as examples right? | 20:48 |
@sonney2k | do they all have the same size? | 20:48 |
n4nd0 | NOTE: in MatrixFeatures each of the features are the columns, while in this toy example they are the rows | 20:48 |
@sonney2k | I mean same nr of rows/cols? | 20:48 |
n4nd0 | sonney2k: now about the sizes | 20:48 |
n4nd0 | they must have the same number of features | 20:48 |
@sonney2k | features? | 20:48 |
n4nd0 | BUT the features don't need to have the same length among examples | 20:48 |
n4nd0 | yes | 20:48 |
@sonney2k | == rows? | 20:49 |
n4nd0 | in my example ^, the same number of rows | 20:49 |
n4nd0 | in MatrixFeatures, the same number of columns | 20:49 |
n4nd0 | but as you see above, the other dimension is not the same for all the examples | 20:50 |
n4nd0 | sonney2k: am I making it clear or not? | 20:50 |
@sonney2k | n4nd0, why do they have to have the same number of columns? | 20:51 |
n4nd0 | there is another point of view that may make more sense | 20:51 |
n4nd0 | sonney2k: where? in the example above or in MatrixFeatures? | 20:51 |
@sonney2k | it is not crystal clear yet | 20:51 |
@sonney2k | in MatrixFeatures | 20:51 |
@sonney2k | n4nd0, I first thought it is basiclly an array of matrices of same size | 20:51 |
n4nd0 | it is almost that | 20:52 |
shogun-buildbot_ | build #261 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/261 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 20:52 |
n4nd0 | it is an array of matrices with the same number of columns | 20:52 |
@sonney2k | now you say it is basically an array of matrices where the matrices have the same number of columns | 20:52 |
n4nd0 | yes, it is that | 20:52 |
@sonney2k | that is much much easier to understand | 20:52 |
@sonney2k | but still doesn't explain why this restriction is needed! | 20:52 |
@sonney2k | and what one can do with this | 20:52 |
n4nd0 | the restriction is needed because you need to have the same number of features for all the examples | 20:53 |
n4nd0 | imagine dotfeatures where x1 has 4 elements and x2 three | 20:53 |
@sonney2k | the term features creates a parse error for me here | 20:53 |
n4nd0 | yeah | 20:53 |
n4nd0 | features != feature vector | 20:54 |
n4nd0 | so what I said | 20:55 |
n4nd0 | in DotFeatures | 20:55 |
n4nd0 | all the examples have the same number of features | 20:55 |
n4nd0 | (otherwise we are talking about StringFeatures) | 20:55 |
n4nd0 | in my MatrixFeatures this is still the same, same number of features | 20:55 |
n4nd0 | therefore, the restriction of the same number of columns arises | 20:56 |
n4nd0 | 2) about what can one do with this: train HM-SVMs :) | 20:56 |
n4nd0 | sonney2k: closer to crystal clear? | 20:57 |
@sonney2k | n4nd0, closer but what do the different nr of rows stand for? | 21:00 |
@sonney2k | n4nd0, I am currently thinking if it would not be better to have a general matrix features type with that restriction removed (or at least optional) | 21:00 |
n4nd0 | sonney2k: sure it could be | 21:01 |
n4nd0 | I don't think that actually I coded this restriction in the MatrixFeatures | 21:01 |
n4nd0 | although I use it in the HM-SVM | 21:02 |
n4nd0 | because there the restriction exists | 21:02 |
n4nd0 | but I am sure about this | 21:03 |
n4nd0 | *not* sure | 21:05 |
n4nd0 | sonney2k: so, are we ok now or should I explain something else? | 21:09 |
@sonney2k | n4nd0, I understand that now - though the doc is not too clear | 21:12 |
@sonney2k | but I would really like to have these matrix features to be more general | 21:12 |
@sonney2k | otherwise we need to have one more class | 21:12 |
n4nd0 | sonney2k: yes, you are right | 21:12 |
n4nd0 | sonney2k: I am opening an issue in github for me, ok? | 21:13 |
@sonney2k | one thing that is not so easy to understand to me | 21:13 |
n4nd0 | tell me | 21:13 |
@sonney2k | momplz | 21:13 |
@sonney2k | n4nd0, ok back | 21:21 |
@sonney2k | n4nd0, how do we feed data into the matrixfeatures? | 21:21 |
@sonney2k | I mean I could imagine we use SGNDArray to put in matrices of same size | 21:22 |
-!- zxtx [~zv@75-59-237-124.lightspeed.sntcca.sbcglobal.net] has joined #shogun | 21:22 | |
n4nd0 | sonney2k: one way you can currently do is feeding matrix by matrix | 21:22 |
@sonney2k | n4nd0, I guess you might not see the problem but it is related to how all the interfaces currently work - we have typemaps aka conversion routines that convert e.g. a numpy array into shogun's dense features | 21:23 |
n4nd0 | sonney2k: if you don't want to restrict that the matrices are of the same size, then I think that an SGNDArray is not appropriate | 21:23 |
@sonney2k | n4nd0, so I guess one way to do it would be to put in a list of numpy matrices | 21:23 |
n4nd0 | yes | 21:23 |
@sonney2k | but we could as well have a new SG* data type for that then | 21:24 |
@sonney2k | e.g. for strings we have SGStringList | 21:24 |
@sonney2k | so we could have SGMatrixList | 21:24 |
n4nd0 | yeah, I think that makes sense | 21:25 |
shogun-buildbot_ | build #262 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/262 blamelist: Soeren Sonnenburg <sonne@debian.org> | 21:25 |
n4nd0 | sonney2k: do you think it would be better having an SGMatrixList <--- SGReferencedData rather than the SGMatrix*? | 21:25 |
-!- zxtx [~zv@75-59-237-124.lightspeed.sntcca.sbcglobal.net] has quit [Ping timeout: 244 seconds] | 21:29 | |
@sonney2k | n4nd0, yes | 21:29 |
@sonney2k | n4nd0, we do this already for SGSparseMatrix | 21:31 |
@sonney2k | ...and will do it for stringfeautres too so this would be the consistent way | 21:31 |
@sonney2k | blackburn, ^ seen the buildbot is still unhappy | 21:32 |
@sonney2k | blackburn, new error though | 21:33 |
blackburn | sonney2k: I do not understand it | 21:33 |
blackburn | tests related again as I can see | 21:33 |
@sonney2k | blackburn, yeah and wiking hides somewhere and doesn't want to fix the bugs he introduced :D | 21:33 |
@sonney2k | n4nd0, can you do these changes (introducing SGMatrixList?) | 21:35 |
n4nd0 | sonney2k: is "in the long run" a valid answer? :-O | 21:36 |
n4nd0 | I'd prefer to get the plifs, coffin, try uricamic's bundle methods first | 21:36 |
@sonney2k | n4nd0, c'mon this will take you 4 hours max | 21:38 |
@sonney2k | gtg now sry | 21:38 |
-!- n4nd0 [~nando@114.Red-2-137-6.dynamicIP.rima-tde.net] has quit [Quit: leaving] | 21:49 | |
-!- zxtx [~zv@75-59-237-124.lightspeed.sntcca.sbcglobal.net] has joined #shogun | 22:36 | |
--- Log closed Sun Aug 05 00:00:17 2012 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!