--- Log opened Sat May 28 00:00:38 2011 | ||
CIA-32 | shogun: Soeren Sonnenburg master * r8381b60 / src/configure : at least on debian based systems ruby libs are named libruby<VER>*.so* - http://bit.ly/lqKxF4 | 00:06 |
---|---|---|
serialhex | sonney2k, i'm updating my shogun now, will test the ruby_modular interface when it's built | 00:59 |
serialhex | fscking slow-ass computer!!! | 02:02 |
serialhex | sonney2k, my ruby doesn't work with your change, so we'll have to find a medium somewhere in between, it gives me a '/usr/bin/ld cant find -lruby1.8' error (i'm trying different configs & stuff and lost the exact msg, but i'll post it when i try that again) | 06:04 |
serialhex | also, i was having trouble loading modules, and i came across this: http://goo.gl/sHdbK and i went into each of the modules. like it says, %module names should be lowercase, and since all of the modules (classifying, clustering, distance, etc...) are capitalized, we should do some non-break-backwards-compatibility thing like add either a test to see whether we're compiling ruby or not in the Foo.i file, or in the ruby | 06:10 |
serialhex | compile opt include the "-module foo" the first is probably easier & more direct | 06:10 |
serialhex | my days have recently gotten very busy, as i *just* got a job (yay me!) so while i have some ideas on how to fix this stuff i might not get the time to actually do so... | 06:11 |
serialhex | ...and i just noticed sploving is on! | 06:12 |
serialhex | ...well, afk actually :P | 06:12 |
serialhex | and i *really* wish my computer was faster, 20+ min to compile shogun is just sad! but i got to sleep, so i'll see what i can find out tomorrow, and let sploving & sonney2k know what i find... take care! | 06:16 |
@sonney2k | serialhex, congrats on the job! | 07:30 |
@sonney2k | serialhex, I am doing the conditional rename for ruby now | 07:35 |
@sonney2k | serialhex, regarding configure - we will need a test to compile a minimal ruby extension? do you have an example for that? | 07:39 |
-!- blackburn [~qdrgsm@188.168.2.233] has joined #shogun | 12:49 | |
blackburn | ehh.. CMath::display_matrix is column major routine | 13:37 |
blackburn | sonney2k: why it is? | 13:37 |
blackburn | omg these matrices is rotating in LLE all the time | 13:41 |
blackburn | damn it | 13:41 |
-!- blackburn [~qdrgsm@188.168.2.233] has quit [Ping timeout: 258 seconds] | 17:45 | |
-!- blackburn [~qdrgsm@188.168.4.77] has joined #shogun | 19:45 | |
blackburn | sonney2k: reported ;) | 20:15 |
@bettyboo | ;> | 20:15 |
@sonney2k | blackburn, all matrices are assumed to be column by column | 20:39 |
@sonney2k | that is the fortran way (tm) | 20:39 |
@sonney2k | blackburn, thnanks | 20:39 |
blackburn | sonney2k: I had a problem with set_feature_matrix | 20:39 |
blackburn | well I have eigenvectors as ROWS of matrix | 20:40 |
blackburn | then I have to choose m of them | 20:40 |
blackburn | and columns of that matrix is feature matrix | 20:40 |
blackburn | everything seems to be ok, right? | 20:40 |
blackburn | but when I set this feature matrix and print it in python - I have eh.. something like transposed | 20:41 |
blackburn | right sizes - but not that order I think | 20:41 |
@sonney2k | blackburn, it could of course be that I introduced bugs in the new typemaps | 20:44 |
@sonney2k | it should be like 1 column == 1 vector | 20:44 |
blackburn | for the moment I just i<->j | 20:44 |
blackburn | I don't know exactly | 20:44 |
blackburn | these matrix issues makes me mad | 20:44 |
@sonney2k | and matrix == mulltiple columns concatenated | 20:44 |
@sonney2k | does that match your assumption | 20:45 |
@sonney2k | ? | 20:45 |
@sonney2k | if so try to set a feature matrix and get it afterwards | 20:45 |
blackburn | in python? | 20:45 |
@sonney2k | then you will see if it is still the same | 20:46 |
blackburn | sonney2k: have you seen pics in my report? looks like it doing the things right, but again I'm not confident with it | 20:47 |
blackburn | anyway the possible 'error' could be only in this issue - changing feature matrix | 20:47 |
@sonney2k | blackburn, I suggest to sleep one night over this and then check again | 21:07 |
@sonney2k | I don't know... | 21:07 |
alesis-novik | sonney2k, I assume code maintainability is more important than slight increase in performance. | 21:11 |
@sonney2k | alesis-novik, depends... | 21:11 |
alesis-novik | As in using Gaussian class with GMM instead of having means and covariances in the GMM class itself | 21:11 |
@sonney2k | usually yes though | 21:11 |
@sonney2k | alesis-novik, how does having them in the gaussian class possibly slow down things? | 21:12 |
@sonney2k | I would keep them there... | 21:13 |
alesis-novik | Well, moving stuff around will take a bit of time, but like I said, nothing major. Will need to completely redo the Gaussian code though. | 21:35 |
alesis-novik | But since Gaussian isn't used anywhere else as far as I know, it won't cause any problems | 21:35 |
@sonney2k | alesis-novik, don't you have covariance matrix / mean vector in gaussian too? what else is missing? | 21:47 |
alesis-novik | yeah, but I need to represent the covariance depending on the type used (as svd or diagonal or spherical). | 21:50 |
alesis-novik | So the covariance field is effectively gone | 21:51 |
blackburn | sonney2k: could you remind me the name of a tool for compile caching? | 22:31 |
serialhex | hey blackburn, i think it's cmake | 23:06 |
blackburn | hi, serialhex | 23:06 |
serialhex | (i know, i'm like a half hour after the Q but whatev! i just got homw :P ) | 23:06 |
serialhex | how are ya? | 23:06 |
blackburn | fine, but sitting at home all day long and having a sexual time with some dim reduction things :D | 23:08 |
blackburn | and you? | 23:08 |
serialhex | really? you're having sex with some dim reduction stuff?? :P | 23:08 |
serialhex | i'm doing well, just got done with my last day of training for my new job!!! | 23:09 |
* serialhex sees blackburn's spiffy 3d graph!!! | 23:09 | |
blackburn | what is your new job? | 23:09 |
blackburn | serialhex: yeah it could show that method is working | 23:10 |
blackburn | or not | 23:10 |
blackburn | I don't know :D | 23:10 |
blackburn | I have to check if neighborhood is 'saved' in 2d | 23:10 |
blackburn | damn borg book about MDS have 612 pages | 23:10 |
serialhex | yeah, that souds like it might work | 23:10 |
serialhex | DAMN!!! | 23:11 |
blackburn | seems they have wrote almost everything that they know | 23:11 |
serialhex | well my new job is really kind of cool, i'm selling really spiffy high quality knives!! | 23:11 |
serialhex | yes it does! | 23:11 |
blackburn | selling knives sounds very profitable | 23:13 |
serialhex | i sometimes dont like books like that because theres no "this is what you *REALLY REALLY* need to know about this stuff" section in the beginning... so you dont have to grok the whole book | 23:13 |
blackburn | at least you can threat someone with your goods :D | 23:13 |
serialhex | but still, it looks impressive on a shelf!! :D | 23:13 |
serialhex | oh yeah!! | 23:13 |
blackburn | 'damn man i'll show you how sharp it is' :D | 23:14 |
blackburn | hrshsh | 23:14 |
serialhex | dude, you've probably never dealt with a cocanut in russia, but while i was in jamaica i had one and it took the dude 3 whacks with a machete to break open the sucker! | 23:14 |
blackburn | hehe | 23:14 |
serialhex | with the cleaver that my boss (whos really hot and really awesome) let me use, i split that thing in HALF with one swing!!! | 23:15 |
serialhex | it was freaking awesome!!! | 23:15 |
serialhex | so yeah, i think it's going to be a lot of fun and probably very profitable! | 23:16 |
blackburn | nice to you :) | 23:16 |
serialhex | tho i'm still working on some of the ruby things, but i gotta make it work first!! :( | 23:17 |
@sonney2k | serialhex, I hope you can recommend me a good knive now :) | 23:17 |
serialhex | hellz yeah!!! i can probably sell em to ya if you like, they're really spiffy!!! | 23:18 |
@sonney2k | all the ones I have here are just blunt | 23:18 |
@sonney2k | serialhex, made in germany ;-) | 23:18 |
serialhex | well i'm not 100% sure if it'll work outside the US but they have a forever guarantee so you can get them sharpened whenever for the price of shipping | 23:19 |
serialhex | theres a lot more to em, but i'll talk to my manager and see what i can do :D | 23:19 |
blackburn | sonney2k: at least you have good cars :D | 23:19 |
@bettyboo | lolomat | 23:20 |
serialhex | yes indeed!! german cars rock!!! | 23:20 |
serialhex | anyway, i'm going to work on a ruby example tonight, and some more probably sunday a bit, tho it's still not compiling for me | 23:21 |
@sonney2k | blackburn, I am riding a bike carrying to kids now | 23:21 |
@sonney2k | in some kind of trailer | 23:21 |
serialhex | HAHAHA!!! | 23:21 |
@sonney2k | serialhex, yeah we need a test | 23:21 |
serialhex | having fun sonney2k?? | 23:21 |
@sonney2k | serialhex, sure | 23:21 |
@sonney2k | serialhex, some minimal ruby extension test... | 23:21 |
serialhex | do you want something that loads up all the modules? | 23:21 |
@sonney2k | if you have that we could really check if things compile | 23:22 |
serialhex | ok, i'll see what i can cook up | 23:22 |
@sonney2k | serialhex, ahh and btw, I tried lowercasing all module names but then the swig wrapped stuff stopped to work completely | 23:22 |
@sonney2k | I couldn't import e.g. features nor Features | 23:22 |
serialhex | hmm... thats odd... | 23:22 |
@sonney2k | err require | 23:22 |
@sonney2k | serialhex, not sure if this is related to a pre 2.0 swig version here | 23:23 |
serialhex | well i couldnt do that before either | 23:23 |
serialhex | idk... | 23:23 |
* serialhex should probably find his old copy of vmware and install it | 23:23 | |
@sonney2k | alesis-novik, I am not an expert in GMM - EM so ask vojtech what he thinks what is better | 23:24 |
serialhex | i also can't link it with the -lruby1.8 thing, i'll gist my error msg now... | 23:24 |
@sonney2k | serialhex, just edit .config an replace the -lruby1.8 with -lruby | 23:25 |
@sonney2k | (occurs twice) | 23:25 |
alesis-novik | sonney2k, I think this is more an aesthetic issue than anything else | 23:25 |
serialhex | ok... i'll do that for now then | 23:25 |
@sonney2k | alesis-novik, I just don't know what EM for GMM will involve | 23:25 |
blackburn | sonney2k: have you heard about some dim reduction for strings? just recalled that shogun is highly biased to it :) | 23:26 |
alesis-novik | ok, I'll ask Vojtech for his opinion | 23:26 |
@sonney2k | alesis-novik, in the end vojtech and you decide what is reasonable - only if the code later looks really weird I will give comments :) | 23:27 |
@sonney2k | blackburn, nope ... well if they are kernel based then I guess you can use kpca | 23:27 |
@sonney2k | blackburn, and shogun is not *that* biased towards strings | 23:28 |
blackburn | I just thinking about what to add to my project | 23:28 |
alesis-novik | sonney2k, well, that's one point of writing this software: making it readable and manageable :) | 23:28 |
@sonney2k | heh | 23:28 |
@sonney2k | blackburn, whatever you like :) | 23:28 |
@sonney2k | blackburn, btw I would love to do the CClassifier -> CMachine rename | 23:29 |
@sonney2k | any objections? | 23:29 |
blackburn | it takes no difference for me | 23:29 |
blackburn | but sounds better, yes | 23:30 |
blackburn | my worries is related to preproc 'framework' | 23:30 |
@sonney2k | blackburn, in which sense? | 23:30 |
@sonney2k | btw, serialhex we have one module name PreProc | 23:31 |
serialhex | ok | 23:31 |
blackburn | as I said earlier, mds and lle doesn't make sense at 'learning' | 23:31 |
@sonney2k | I guess that should then at least be called just Preprocessors | 23:31 |
serialhex | ok, sounds good | 23:31 |
blackburn | so init() in preproc and apply_to_feature_matrix() doesn't make sense at all | 23:31 |
blackburn | calling init for one feature matrix and apply to another - makes no useful embedding | 23:32 |
blackburn | *be back in 10 minutes | 23:32 |
@sonney2k | blackburn, I don't see a problem | 23:32 |
@sonney2k | blackburn, you just don't do anything in init() | 23:32 |
@sonney2k | and always do the embedding in apply* | 23:32 |
@sonney2k | to the current fm | 23:33 |
@sonney2k | serialhex, ok will rename it then | 23:33 |
serialhex | oh, sonney2k, how were you doing the %module renaming? something like #ifdef SWIGRUBY; %module foo; #endif; ?? | 23:33 |
@sonney2k | serialhex, yes | 23:33 |
serialhex | ok, i'll test that out then sonney2k, and see how it works on my end | 23:33 |
@sonney2k | serialhex, this is the patch https://gist.github.com/997253 | 23:35 |
serialhex | sweet!!! thanks sonney2k, i'm compiling now but i'll get this applied once it's done (and recompile :P ) | 23:36 |
@sonney2k | serialhex, blackburn: shall I name the PreProc module Preprocessor or Preprocessors? | 23:37 |
@sonney2k | I tend towards Preprocessor | 23:37 |
serialhex | that works for me | 23:37 |
blackburn | sonney2k: hehe i don't understand why I didn't choose this way (with do nothing in init) | 23:39 |
@sonney2k | brain twists... | 23:39 |
blackburn | sonney2k: do we really need 'pre'? | 23:39 |
blackburn | Preprocessor is better for me | 23:40 |
@sonney2k | blackburn, you mean Processor? | 23:41 |
@sonney2k | nah | 23:41 |
blackburn | may be | 23:41 |
blackburn | but it sounds not so 'sensefully', I agree now :) | 23:41 |
@sonney2k | should I also rename the preproc dir to preprocessor ? | 23:41 |
@sonney2k | and the PreProc class to Preprocessor? | 23:42 |
blackburn | makes sense | 23:42 |
@sonney2k | doing ... | 23:42 |
blackburn | PreProc is strange camel case | 23:42 |
blackburn | caught an error a few times because of Preproc<->PreProc | 23:43 |
blackburn | sonney2k: I think in preproc there should be a class for dim.reduction :) | 23:43 |
blackburn | may be I'll add one later | 23:43 |
blackburn | does it makes sense to you? | 23:44 |
blackburn | anyway I'm going to bed now, tomorrow will be learning some shity game theory and so on, damn exams | 23:45 |
blackburn | see you | 23:45 |
-!- blackburn [~qdrgsm@188.168.4.77] has quit [Quit: Leaving.] | 23:45 | |
@sonney2k | best of luck... | 23:46 |
--- Log closed Sun May 29 00:00:41 2011 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!