--- Log opened Tue May 15 00:00:40 2012 | ||
-!- karlnapf [~heiko@host86-174-150-108.range86-174.btcentralplus.com] has joined #shogun | 00:10 | |
-!- blackburn [~qdrgsm@188.122.250.167] has left #shogun [] | 00:10 | |
-!- karlnapf [~heiko@host86-174-150-108.range86-174.btcentralplus.com] has left #shogun [] | 00:10 | |
-!- ckwidmer [~chris@HSI-KBW-046-005-237-106.hsi8.kabel-badenwuerttemberg.de] has quit [Ping timeout: 272 seconds] | 00:39 | |
-!- cronor [~cronor@e178176184.adsl.alicedsl.de] has quit [Quit: cronor] | 01:33 | |
-!- av3ngr [av3ngr@nat/redhat/x-yqokwqfxztgfhxwp] has joined #shogun | 01:55 | |
--- Log closed Tue May 15 02:05:45 2012 | ||
--- Log opened Tue May 15 02:05:51 2012 | ||
-!- shogun-toolbox [~shogun@7nn.de] has joined #shogun | 02:05 | |
-!- Irssi: #shogun: Total of 8 nicks [1 ops, 0 halfops, 0 voices, 7 normal] | 02:05 | |
-!- Irssi: Join to #shogun was synced in 6 secs | 02:05 | |
-!- in3xes [~in3xes@106.78.49.114] has quit [Ping timeout: 260 seconds] | 03:31 | |
-!- in3xes [~in3xes@210.212.58.168] has joined #shogun | 03:44 | |
-!- in3xes_ [~in3xes@180.149.49.230] has joined #shogun | 04:00 | |
-!- in3xes [~in3xes@210.212.58.168] has quit [Ping timeout: 265 seconds] | 04:04 | |
-!- in3xes_ [~in3xes@180.149.49.230] has quit [Quit: Leaving] | 04:08 | |
-!- Francis_Chan [~Adium@59.108.60.58] has joined #shogun | 06:27 | |
-!- gsomix_ [~gsomix@83.234.54.38] has joined #shogun | 07:12 | |
-!- gsomix_ is now known as gsomix | 08:01 | |
gsomix | hi all | 08:01 |
---|---|---|
sonne|work | hey hey | 08:34 |
-!- uricamic [~uricamic@2001:718:2:1634:8568:fb14:615c:1667] has joined #shogun | 08:41 | |
-!- Francis_Chan [~Adium@59.108.60.58] has quit [Quit: Leaving.] | 08:52 | |
sonne|work | gsomix: now master of optics? | 09:01 |
gsomix | sonne|work, haha. :) in future, maybe | 09:02 |
-!- cronor [~cronor@e178176184.adsl.alicedsl.de] has joined #shogun | 09:18 | |
-!- cronor [~cronor@e178176184.adsl.alicedsl.de] has quit [Quit: cronor] | 09:24 | |
sonne|work | gsomix: so do you have another exam? | 09:35 |
sonne|work | gsomix: or can you finish the dynarray stuff? | 09:35 |
gsomix | sonne|work, exams will be in June. I just studied yesterday material for future research. | 09:39 |
sonne|work | I see | 09:40 |
gsomix | sonne|work, about dynarray. I need only add some doc's in code. | 09:41 |
sonne|work | gsomix: so you finished the dynamicobjectarray stuff too? | 09:41 |
gsomix | sonney2k, yep, yep. | 09:41 |
sonne|work | great! | 09:41 |
sonne|work | I would rather merge it now and you add the doc during the day or so | 09:42 |
gsomix | sonne|work, ok, just a sec. | 09:44 |
gsomix | sonne|work, please, check it. | 10:01 |
gsomix | need to go at uni | 10:28 |
gsomix | cu | 10:28 |
-!- gsomix [~gsomix@83.234.54.38] has quit [Quit: Ex-Chat] | 10:28 | |
-!- n4nd0 [~nando@n173-p203.kthopen.kth.se] has joined #shogun | 10:47 | |
-!- av3ngr [av3ngr@nat/redhat/x-yqokwqfxztgfhxwp] has quit [Quit: That's all folks!] | 11:28 | |
-!- uricamic [~uricamic@2001:718:2:1634:8568:fb14:615c:1667] has quit [Quit: Leaving.] | 12:30 | |
-!- Francis_Chan [~Adium@58.194.224.109] has joined #shogun | 12:45 | |
-!- n4nd0 [~nando@n173-p203.kthopen.kth.se] has quit [Ping timeout: 256 seconds] | 12:49 | |
-!- n4nd0 [~nando@n173-p203.kthopen.kth.se] has joined #shogun | 12:55 | |
sonne|work | n4nd0: regarding dotfeatures - so you would derive from dot features and just put the labels inside them and implement new dense_dot etc operations using labels, input features and whatever Phi is wanted - right? | 13:08 |
n4nd0 | sonne|work: hey! sorry I didn't notice the bell rang | 13:44 |
n4nd0 | yes, that's exactly my idea | 13:44 |
n4nd0 | sonne|work: do you think it is ok? | 13:44 |
-!- Marty28 [~marty@cable-158-181-77-81.cust.telecolumbus.net] has joined #shogun | 13:46 | |
sonne|work | n4nd0: yes - makes a lot of sense though I would derive 'StructuredDotFeatures' or so first (that have a StructuredLabel) and then from that the other classes | 13:46 |
n4nd0 | sonne|work: all right | 13:47 |
n4nd0 | sonne|work: thank you, by the way, did you have time to take a look to the PR? | 13:48 |
sonne|work | no not yet | 13:48 |
n4nd0 | ok :) | 13:49 |
n4nd0 | maybe someone can help me to make a decision, I need to select a course for next Fall | 13:50 |
n4nd0 | I was looking into optimization courses | 13:50 |
n4nd0 | and there is this one, Applied Linear Optimization, I don't know if the methods they teach there are interesting in machine learning | 13:50 |
n4nd0 | some of these methods are simplex, interior methods, stochastic programming, Lagrangian relaxation, integer programming, subgradient | 13:51 |
n4nd0 | any idea?? | 13:51 |
sonne|work | yes they are... | 13:52 |
n4nd0 | sonne|work: cool, thank you :) | 13:54 |
n4nd0 | I will take that course then | 13:54 |
-!- Marty28 [~marty@cable-158-181-77-81.cust.telecolumbus.net] has quit [Quit: Colloquy for iPad - http://colloquy.mobi] | 14:00 | |
-!- wiking [~wiking@c-98-231-171-175.hsd1.md.comcast.net] has joined #shogun | 14:49 | |
-!- wiking [~wiking@c-98-231-171-175.hsd1.md.comcast.net] has quit [Changing host] | 14:49 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 14:49 | |
n4nd0 | hey wiking, how are you? | 15:07 |
wiking | n4nd0: hey... a bit jet lag :))0 | 15:07 |
n4nd0 | wiking: oh yeah, I remember something about a conference | 15:08 |
n4nd0 | is that right? | 15:08 |
wiking | yep | 15:08 |
wiking | this week | 15:08 |
n4nd0 | where is it? | 15:08 |
wiking | in baltimore | 15:09 |
n4nd0 | nice | 15:09 |
n4nd0 | are you presenting something there? | 15:09 |
wiking | yep | 15:21 |
wiking | one of my paper | 15:21 |
-!- gsomix [~gsomix@83.234.54.38] has joined #shogun | 15:23 | |
gsomix | hey, hey | 15:23 |
n4nd0 | wiking: I wish you luck with it ;) | 15:24 |
n4nd0 | hey gsomix | 15:24 |
gsomix | I got the first primitive task related to research. | 15:27 |
n4nd0 | wiking: I pushed some SO stuff to the repo, let me know if you have any comment related to your latent features part | 15:27 |
n4nd0 | gsomix: cool! what is it about? | 15:27 |
wiking | aaah cool | 15:28 |
wiking | i'll take a look at it now thnx for mentioning | 15:28 |
n4nd0 | wiking: I think we could provide another type of pointer function in StructuredModel for the latent features part | 15:28 |
n4nd0 | wiking: what do you think? | 15:28 |
sonne|work | n4nd0: ahh btw would it be possible to not use function pointers? | 15:29 |
sonne|work | but instead some say CArgMax class or so | 15:29 |
gsomix | n4nd0, optics. calculation of the optical element that gives a uniformly illuminated circle. | 15:29 |
gsomix | sonne|work, what about patch? | 15:30 |
sonne|work | gsomix: did you fix the SG_ADD stuff? | 15:30 |
n4nd0 | sonne|work: mmm we could do it like that but I liked the idea that wiking suggested since in that way the user that wants to use SO wouldn't need to create any class but rather just define some functions | 15:30 |
gsomix | sonne|work, give me five minutes | 15:31 |
sonne|work | n4nd0: I don't see the difference... | 15:32 |
sonne|work | except that I don't like function ptrs... | 15:33 |
gsomix | sonney2k, what about add_vector? | 15:33 |
-!- hoijui [~hoijui@141.23.65.251] has joined #shogun | 15:33 | |
sonne|work | gsomix: IIRC that should work too | 15:34 |
n4nd0 | sonne|work: so your plan would be to provide a CArgMax base class and extend it for every different argmax method? | 15:36 |
gsomix | sonne|work, ok. but add_vector have num_of_elements argument. just another SG_ADD for num? | 15:37 |
sonne|work | n4nd0: yes like CLossFunction | 15:37 |
sonne|work | gsomix: errm I don't know by heart | 15:38 |
n4nd0 | sonne|work: ok ... why don't you like function pointers? :) | 15:38 |
sonne|work | gsomix: if it doesn't work we stick with SG_ADD | 15:39 |
sonne|work | n4nd0: hard to read | 15:39 |
gsomix | sonne|work, I see. | 15:40 |
gsomix | sorry | 15:40 |
sonne|work | n4nd0: have a look at Lossfunction - it is very easy to understand - have a function ptr somewhere and you need heavy documentation... | 15:40 |
-!- hoijui [~hoijui@141.23.65.251] has quit [Quit: Leaving] | 15:44 | |
sonne|work | n4nd0: in addition, some argmax functions are *very* hefty so you want to split them up into several small functions | 15:44 |
n4nd0 | sonne|work: I think you have just saved me some work with the CLossFunction class :) | 15:45 |
sonne|work | heh | 15:46 |
n4nd0 | I defined a new one for that but it looks like that one is the same we need in SO | 15:46 |
wiking | sonney2k: the difference is huge | 15:48 |
wiking | sonney2k: a lot of code duplication for no use | 15:48 |
wiking | sonney2k: if u take a class inheritance then you already have to take care of ctors, dtors | 15:49 |
wiking | and of course the actual implementation of the overridden function | 15:49 |
wiking | and all this because you want a different way to calculate a given function | 15:50 |
wiking | n4nd0: just checking your commits | 15:51 |
sonne|work | wiking: but that is not just a small function | 15:53 |
wiking | well for me an argmax function in case of latent variable is about 5 lines | 15:54 |
wiking | and PSI is quite the same | 15:54 |
sonne|work | it is sth. like viterbi etc | 15:54 |
sonne|work | so it can be HUGE | 15:54 |
wiking | and? | 15:54 |
wiking | i still don't see the point that depending on a size of a function | 15:55 |
sonne|work | wiking: you will need ~5 helper functions to keep it tractable | 15:55 |
wiking | that would justify to wrap the whole thing into a class | 15:55 |
sonne|work | and these are better suited to be in a class | 15:55 |
sonne|work | wiking: it is the same rationale why we have CLossFunction as class | 15:55 |
wiking | n4nd0: commented | 15:57 |
n4nd0 | wiking: thank you | 15:58 |
-!- puffin444 [62e3926e@gateway/web/freenode/ip.98.227.146.110] has joined #shogun | 16:47 | |
-!- puffin444 [62e3926e@gateway/web/freenode/ip.98.227.146.110] has quit [Client Quit] | 16:51 | |
@sonney2k | gsomix, can I merge the request now? | 16:59 |
CIA-113 | shogun: Soeren Sonnenburg master * ra71d99e / (18 files in 6 dirs): Merge pull request #523 from gsomix/Arrays (+6 more commits...) - http://git.io/rgwtvQ | 17:01 |
@sonney2k | gsomix, merged | 17:01 |
@sonney2k | gsomix, ahh btw I just realize you did lots of casts like foo = (CPlifBase*)ptr; ... would you please use a white space after the ) ? | 17:02 |
@sonney2k | thanks | 17:02 |
-!- n4nd0 [~nando@n173-p203.kthopen.kth.se] has quit [Quit: leaving] | 17:03 | |
gsomix | sonney2k, ok. | 17:04 |
@sonney2k | gsomix, please update this patch with these small changes | 17:05 |
CIA-113 | shogun: Soeren Sonnenburg master * rc54f3e6 / (19 files in 3 dirs): Merge pull request #524 from pluskid/multiclass-ecoc (+10 more commits...) - http://git.io/-YN6Cw | 17:05 |
@sonney2k | gsomix, btw what else did we have in the queue? | 17:05 |
@sonney2k | I mean on your TODO list? | 17:06 |
gsomix | sonney2k, save/loads, titled features with CMap and consts for methods in features. | 17:06 |
gsomix | sonney2k, btw, blackburn have some bloody task for me. | 17:07 |
@sonney2k | gsomix, ahh load/save... it seems like we need them sometimes | 17:07 |
@sonney2k | gsomix, ignore him :D | 17:07 |
@sonney2k | gsomix, so the load/save stuff is on hold until further notice | 17:07 |
@sonney2k | we realized that it is nice to be able to load e.g. old svm light ascii files etc | 17:07 |
@sonney2k | or ascii features etc | 17:07 |
@sonney2k | however usually not any other Machine - so removing the functions there was ok | 17:08 |
@sonney2k | gsomix, what is consts for methods in features? | 17:08 |
gsomix | sonney2k, just add 'const' keywords. | 17:08 |
gsomix | for methods, that should be const. | 17:09 |
* sonney2k considers const useless | 17:09 | |
@sonney2k | anyway | 17:09 |
@sonney2k | feel free | 17:09 |
@sonney2k | gsomix, I think after this we should do some bigger task | 17:09 |
@sonney2k | gsomix, what do you prefer the swig director classes stuff? | 17:09 |
@sonney2k | or the x-validation typemaps? | 17:10 |
@sonney2k | all heavy stuff... | 17:10 |
gsomix | sonney2k, director classes, I think. | 17:10 |
@sonney2k | I attempted the first one but failed | 17:10 |
@sonney2k | heh | 17:10 |
@sonney2k | so in this case I would suggest you | 17:10 |
@sonney2k | 1) write some small example C++ class with just one virtual set/get method for some integer | 17:11 |
@sonney2k | 2) wrap this class with swig to python | 17:11 |
@sonney2k | 3) test if you can get/set stuff from python | 17:11 |
@sonney2k | 4) enable directors | 17:11 |
@sonney2k | 5) overload the C++ set function within python | 17:12 |
@sonney2k | 6) test if it works :) | 17:12 |
@sonney2k | gsomix, does that make sense to you? | 17:12 |
@sonney2k | if it works you have to figure out why it crashes when done from shogun.... | 17:12 |
gsomix | sonney2k, yep. I would do it anyway. | 17:13 |
@sonney2k | gsomix, the x-validation task is probably easier | 17:14 |
@sonney2k | it would mean to create typemaps for some more complex python objects and to construct some model selection stuff based on this | 17:15 |
@sonney2k | gsomix, I have to talk to heiko how we want it to look like but it will be some list/dictionaries/tuples to parse no more | 17:16 |
@sonney2k | gsomix, ahh note I will be away for some time in june so we need sth where you know what to do | 17:17 |
@sonney2k | gsomix, in worst case you have to start to remove ASSERT() and replace them with meaningful error msgs (SG_ERROR("...")) when you get stuck :) | 17:18 |
@sonney2k | gsomix, the director stuff is probably most interesting/difficult, the x-validation typemaps most useful for shogun people | 17:19 |
gsomix | sonney2k, ok. I need some time for planning. | 17:20 |
gsomix | because I have exams in early June. | 17:20 |
@sonney2k | got to leave train | 17:22 |
@sonney2k | cu | 17:22 |
-!- Francis_Chan [~Adium@58.194.224.109] has quit [Ping timeout: 272 seconds] | 17:43 | |
-!- blackburn [~qdrgsm@188.122.250.167] has joined #shogun | 17:45 | |
-!- sonne|work [~sonnenbu@194.78.35.195] has quit [Ping timeout: 256 seconds] | 18:03 | |
-!- Netsplit *.net <-> *.split quits: @sonney2k | 18:07 | |
-!- Netsplit over, joins: @sonney2k | 18:13 | |
blackburn | sonney2k: ignore me?! I declare war :D | 18:14 |
-!- sonne|work [~sonnenbu@194.78.35.195] has joined #shogun | 18:18 | |
-!- blackburn [~qdrgsm@188.122.250.167] has quit [Quit: Leaving.] | 18:18 | |
-!- blackburn [~blackburn@188.122.250.167] has joined #shogun | 18:20 | |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 18:29 | |
-!- wiking [~wiking@c-98-231-171-175.hsd1.md.comcast.net] has joined #shogun | 18:32 | |
-!- wiking [~wiking@c-98-231-171-175.hsd1.md.comcast.net] has quit [Changing host] | 18:32 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 18:32 | |
shogun-buildbot | build #540 of python_modular is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/python_modular/builds/540 blamelist: gsomix@gmail.com | 18:39 |
gsomix | shogun-buildbot, bebebe. | 18:40 |
-!- blackburn [~blackburn@188.122.250.167] has quit [Quit: Leaving.] | 18:43 | |
-!- blackburn [~blackburn@188.122.250.167] has joined #shogun | 18:44 | |
-!- blackburn [~blackburn@188.122.250.167] has quit [Quit: Leaving.] | 19:27 | |
-!- blackburn [~blackburn@188.122.250.167] has joined #shogun | 19:29 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 19:48 | |
shogun-buildbot | build #541 of python_modular is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/python_modular/builds/541 | 19:58 |
-!- puffin444 [62e3926e@gateway/web/freenode/ip.98.227.146.110] has joined #shogun | 20:00 | |
puffin444 | Hey | 20:00 |
blackburn | hey | 20:05 |
blackburn | puffin444: how is it going? | 20:05 |
puffin444 | It's going well so far. I hope to make a presentation about my project next Wednesday. | 20:05 |
n4nd0 | puffin444: the diagrams in your mail look nice by the way ;) | 20:06 |
puffin444 | Thanks! I hope the design makes sense. | 20:06 |
puffin444 | blackburn, do you have any advice on presenting shogun? | 20:07 |
blackburn | in general? | 20:07 |
n4nd0 | puffin444: have you seen some of the videolectures of shogun? | 20:08 |
n4nd0 | they could be a source of inspiration | 20:08 |
puffin444 | blackburn: Yes. I seem to recall that you once made a similar presentation, yes? | 20:09 |
puffin444 | n4ndo: Sure. Where are these videolectures? | 20:09 |
blackburn | no I didn't :) | 20:09 |
puffin444 | blackburn: Maybe it was sonney2k then. | 20:09 |
blackburn | yes probably | 20:09 |
n4nd0 | puffin444: there is one here by master S?ren http://videolectures.net/icml2010_sonnenburg_shog/ | 20:10 |
n4nd0 | I think there are other videos as well in videolectures.net | 20:11 |
blackburn | I have some suggestions though | 20:11 |
puffin444 | n4ndo: Thanks, this looks helpful | 20:11 |
n4nd0 | puffin444: and there are some other links at shogun's homepage, just look for videolecture | 20:11 |
n4nd0 | puffin444: no problem, just take into account that some of the things there may be a bit outdated | 20:12 |
n4nd0 | blackburn: right ^? | 20:12 |
puffin444 | What are your suggestions? | 20:13 |
blackburn | n4nd0: yes can be outdate | 20:14 |
blackburn | d | 20:14 |
blackburn | puffin444: stress that we have a lot of interfaces and we still tend to be largescale | 20:14 |
puffin444 | blackburn: Will do. | 20:15 |
blackburn | you can also claim it is used not only for bioinf | 20:15 |
blackburn | I use it for kind of computer vision task (traffic sign recognition) | 20:15 |
puffin444 | I still don't have a time length for the presentation though. They call them "lightning talks", so I suspect they are short. | 20:15 |
puffin444 | Absolutely. I will mention examples of applications. | 20:16 |
blackburn | it is important because for a while shogun was really biased to bioinf | 20:16 |
blackburn | but for now we have no active bioinformatician developers | 20:17 |
puffin444 | I'll try to come up with a rough collection of slides in the next few days or so and put them on the mailing list for review. | 20:18 |
blackburn | cool | 20:18 |
blackburn | another idea could be to describe how cool shogun will be after this gsoc :) | 20:19 |
puffin444 | I'll definitely emphasize that :) | 20:19 |
blackburn | puffin444: I can clarify what are our tasks about if you need it | 20:19 |
-!- gsomix [~gsomix@83.234.54.38] has quit [Ping timeout: 248 seconds] | 20:20 | |
puffin444 | I am also unsure about my audience. I don't know how much they know about Statistics/Machine Learning vs. Computer Programming. | 20:20 |
blackburn | I would rather assume they are totally lame in ML | 20:21 |
puffin444 | Frankly the host hasn't given me much information yet, so it looks like I'll have to pry it out of him :) | 20:21 |
puffin444 | Oh and I have a question about the ModelSelection framework | 20:24 |
blackburn | so what is the question? ;) | 20:26 |
puffin444 | For hyperparameter learning I'm trying to minimize a log negative marginal likelihood function | 20:26 |
puffin444 | But it appears that modelselection uses a crossvalidation class to optimize quantites such as accuracy, precision, etc. | 20:26 |
puffin444 | My fear is that right now it would not make sense to use the ModelSelection framework for hyperparameter learning. Should I modify the framework to incorporate more general evalution? | 20:27 |
blackburn | puffin444: so you need no xval here? | 20:29 |
puffin444 | I do not believe so. I am not sure it makes sense to use cross validation to minimize the lnml function. | 20:29 |
puffin444 | I'll see what Oliver says about too. | 20:30 |
blackburn | puffin444: it looks like you can implement your own modelselection interface based class w/o using crossvalidation | 20:31 |
puffin444 | Are you suggesting a separate interface for use only with GPs? | 20:32 |
blackburn | puffin444: no I mean all you need is to implement a class with your own select_model() | 20:33 |
blackburn | under this ModelSelection interface | 20:33 |
blackburn | wouldn't it work? | 20:34 |
puffin444 | Yes. I'll take a look at doing that. | 20:34 |
-!- gsomix [~gsomix@188.168.2.22] has joined #shogun | 20:34 | |
blackburn | it could be a little confusing that we have null crossvalidation here but I think it is ok | 20:35 |
puffin444 | I'll try to update the diagrams then with an addtional modelselection interface. | 20:37 |
-!- blackburn [~blackburn@188.122.250.167] has quit [Quit: Leaving.] | 20:43 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 252 seconds] | 20:54 | |
-!- puffin444 [62e3926e@gateway/web/freenode/ip.98.227.146.110] has left #shogun [] | 21:03 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting] | 21:20 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 21:20 | |
CIA-113 | shogun: Soeren Sonnenburg master * rc47b63e / src/shogun/lib/SGReferencedData.h : unconditionally use refcounting for SGVector/Matrix - http://git.io/fCQLhw | 21:34 |
-!- Marty28 [~marty@cable-158-181-77-81.cust.telecolumbus.net] has joined #shogun | 21:42 | |
CIA-113 | shogun: Evgeniy Andreev master * rd7c7e59 / (4 files in 3 dirs): fixed codestyle and doc's - http://git.io/__1EXQ | 21:42 |
CIA-113 | shogun: Soeren Sonnenburg master * r2730c73 / (4 files in 3 dirs): Merge pull request #528 from gsomix/Arrays - http://git.io/9sY37w | 21:42 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 21:49 | |
@sonney2k | n4nd0, it is just confusing to have these unrelated fixes in the patch... you could change them back and use git commit --amend | 22:00 |
@sonney2k | and prepare a separate patch | 22:00 |
n4nd0 | sonney2k: ok I will take a look into how to do this now | 22:01 |
@sonney2k | n4nd0, btw I hope you get more feedback | 22:01 |
n4nd0 | sonney2k: thank you very much for asking the others about it :) | 22:01 |
n4nd0 | sonney2k: about the function pointers thing, after the discussion with wiking, are you sure that you want the argmax following the strategy used in CLoss? | 22:02 |
-!- Marty28 [~marty@cable-158-181-77-81.cust.telecolumbus.net] has quit [Quit: Colloquy for iPad - http://colloquy.mobi] | 22:02 | |
@sonney2k | n4nd0, yes | 22:03 |
@sonney2k | look at structure/DynProg.{cpp,h} for an argmax for segmentations | 22:03 |
@sonney2k | it is ~2.5kLoC | 22:04 |
n4nd0 | a lot | 22:04 |
n4nd0 | it looks reasonable to have classes for that yeah, fut I king of liked the other idea too | 22:04 |
n4nd0 | look comfortable | 22:04 |
n4nd0 | it looks* | 22:04 |
n4nd0 | sonney2k: so I guess that for the delta loss the same, no function pointers | 22:05 |
n4nd0 | I will ask Nico about that tomorrow, but I think that there is no need for StructuredLoss, I understand that its functionality is the same of the CLoss you pointed out before for me | 22:06 |
@sonney2k | n4nd0, yeah nico is not so familiar with shogun internals | 22:07 |
@sonney2k | ...though an SO expert | 22:07 |
n4nd0 | he's helping me a lot with his indications | 22:10 |
n4nd0 | I make backups of the IRC conversations just in case :D | 22:11 |
@sonney2k | n4nd0, too bad that you are not discussing here on the channel... | 22:18 |
@sonney2k | you really should... | 22:18 |
n4nd0 | sonney2k: I thought of it but wouldn't that be a little like spamming the chat? I mean we would take over control of it for a while | 22:19 |
n4nd0 | but for me it is ok to do it ... | 22:19 |
@sonney2k | n4nd0, no it is what this channel is for | 22:20 |
n4nd0 | all right | 22:20 |
@sonney2k | I also only discuss things with gsomix in public here... | 22:20 |
@sonney2k | so all my screwups are documented :D | 22:21 |
@sonney2k | so I highly prefer many people to screw up here such that I can hide mine in the mass of others :) | 22:23 |
* gsomix woke up. | 22:24 | |
gsomix | sonney2k, I prefer a director classes as next task. | 22:24 |
n4nd0 | sonney2k: I like your strategy to hide screw ups :P | 22:26 |
@sonney2k | gsomix, all right then director classes it is | 22:26 |
@sonney2k | gsomix, just make sure you have some result within the next week - otherwise I won't be able to help if this all fails... | 22:27 |
@sonney2k | that means attempted 1-6) | 22:27 |
@sonney2k | n4nd0, that is why no one wanted to post on mloss.org initially :D | 22:27 |
gsomix | sonney2k, >> that means attempted 1-6) | could not parse. what do you mean? | 22:31 |
gsomix | good night guys | 22:38 |
@sonney2k | gsomix, the plan I outlined above | 22:39 |
n4nd0 | good night gsomix | 22:39 |
@sonney2k | <sonney2k> 1) write some small example C++ class with just one virtual set/get method for some integer | 22:39 |
@sonney2k | <sonney2k> 2) wrap this class with swig to python | 22:39 |
@sonney2k | <sonney2k> 3) test if you can get/set stuff from python | 22:39 |
@sonney2k | <sonney2k> 4) enable directors | 22:39 |
@sonney2k | <sonney2k> 5) overload the C++ set function within python | 22:39 |
@sonney2k | <sonney2k> 6) test if it works :) | 22:39 |
gsomix | sonney2k, aaa, I see... z-z-z | 22:40 |
* sonney2k ZZZZZzzzzzz... | 22:44 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has left #shogun [] | 23:33 | |
--- Log closed Wed May 16 00:00:40 2012 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!