--- Log opened Thu Mar 08 00:00:06 2012 | ||
--- Day changed Thu Mar 08 2012 | ||
n4nd0 | I am really glad I discovered it, I think it can become very useful once I start working on my thesis :) | 00:00 |
---|---|---|
n4nd0 | sonney2k: anyway, I will check again what can I apply for in gsoc and maybe we can bring again the topic of the documentation after summer if everything goes well | 00:01 |
@sonney2k | n4nd0, of course no one hinders you to work on this - to the contrary :) | 00:02 |
n4nd0 | sonney2k: what is your opinion about making more than one gsoc application for different projects within shogun? | 00:04 |
@sonney2k | I guess more than 2 doesn't make sense | 00:05 |
@sonney2k | n4nd0, of course you can propose anything you'd like to do - this list is not exhaustive | 00:11 |
n4nd0 | sonney2k: the more I am discovering about SO, the more I like it :) | 00:12 |
n4nd0 | sonney2k: I will give a try to it, and maybe something easier like the GPs just in case | 00:13 |
n4nd0 | but SO looks pretty cool | 00:14 |
n4nd0 | there is nothing in shogun implemented yet, right? | 00:14 |
@sonney2k | n4nd0, I would say easiest task is the newton svm one - just port a few lines of matlab code | 00:19 |
@sonney2k | SO - it is about the framework - so indeed nothing like that exists though we have the argmax already in there | 00:19 |
n4nd0 | sonney2k: btw, how does the new documentation get included in the official doxygen in the web? | 00:30 |
n4nd0 | I have seen that there is nothing about Mahalanobis yet and I got curious | 00:30 |
@sonney2k | I guess I should automate this - whenever all tests go well, run make doc and put it live | 00:31 |
n4nd0 | ok | 00:32 |
@sonney2k | all of that is prepared but not done :( | 00:32 |
n4nd0 | so there is a testing framework that checks if the things that are about to enter in the project work well? | 00:32 |
@sonney2k | n4nd0, there is the examples/undocumented/python_modular directory (for example) | 00:37 |
@sonney2k | all things in there are examples illustrating things for the user but they are also used as tests | 00:37 |
@sonney2k | so there is a generator.py running these examples and dumping the output they produce | 00:38 |
@sonney2k | and there is a tester.py that checks if results are the same | 00:38 |
n4nd0 | but what are the results compared to? | 00:38 |
@sonney2k | problem is that this is not 100% enabled for everything | 00:38 |
n4nd0 | so for example, right now for the Mahalanobis distance stuff | 00:39 |
@sonney2k | n4nd0, nothing - it is not activated | 00:39 |
n4nd0 | aha ok | 00:39 |
@sonney2k | you would have to run generator.py | 00:39 |
@sonney2k | that would then store the output of mahalanobis distance stuff | 00:39 |
@sonney2k | and so tester could later compare it | 00:39 |
@sonney2k | I think this would also be an excellent gsoc task - but again I don't know about googles official policy about such endeavors | 00:40 |
@sonney2k | in any case writing just 1 example and getting regression tests for free with it is the most painless way I can think of for tests :) | 00:41 |
n4nd0 | ok | 00:42 |
n4nd0 | I was thinking that maybe the tester compared the results with the ones obtained by other software or sth like that | 00:42 |
@sonney2k | n4nd0, no - that is something we do of course when implementing $METHOD | 00:43 |
@sonney2k | but from then on it is not needed | 00:43 |
@sonney2k | anyways bed time | 00:44 |
@sonney2k | cu | 00:44 |
n4nd0 | good night | 00:45 |
n4nd0 | wiking: hey! | 00:47 |
n4nd0 | wiking: are you thinking also of applying for gsoc? | 00:47 |
n4nd0 | wiking: if you don't mind that I ask of course ;) | 00:47 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 246 seconds] | 02:24 | |
-!- wiking_ [~wiking@2002:4e17:bfc9:1:fc01:a93b:44d8:ae33] has joined #shogun | 02:34 | |
-!- wiking_ [~wiking@2002:4e17:bfc9:1:fc01:a93b:44d8:ae33] has quit [Changing host] | 02:34 | |
-!- wiking_ [~wiking@huwico/staff/wiking] has joined #shogun | 02:34 | |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Ping timeout: 245 seconds] | 02:35 | |
-!- wiking_ is now known as wiking | 02:35 | |
-!- in3xes [~in3xes@180.149.49.227] has joined #shogun | 02:46 | |
-!- in3xes [~in3xes@180.149.49.227] has quit [Quit: Leaving] | 03:09 | |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 03:51 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 07:31 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 07:40 | |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 08:14 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 08:16 | |
wiking | n4nd0: hey mate! sorry yesterday i was already sleeping.... | 10:34 |
wiking | n4nd0: yeah i'm planning to apply for gsoc | 10:34 |
wiking | it's not only a matter of which course to take... but most probably it's going to be the latent-svm implementation within shogun | 10:34 |
n4nd0 | wiking: no problem, it was already late actually :) | 10:47 |
n4nd0 | wiking: ok, I have never heard of them apart from your conversations around here with blackburn | 10:48 |
n4nd0 | complet | 10:48 |
n4nd0 | completely ignorant in that topic :S | 10:48 |
wiking | heheh no worries :P | 10:54 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 245 seconds] | 11:01 | |
-!- conor [~cronor@fb.ml.tu-berlin.de] has joined #shogun | 11:29 | |
conor | Hello! | 11:29 |
conor | how can i compile shogun with an old compiler? there is this problem that matlab interfaces must be compile with g++-4.3 and i can't find where to change it in the configure or Makefile. | 11:30 |
sonne|work | ./configure --help :) | 11:37 |
-!- n4nd0 [~nando@n179-p53.kthopen.kth.se] has joined #shogun | 11:42 | |
-!- vikram360 [~vikram360@117.192.160.160] has quit [Read error: Connection reset by peer] | 12:51 | |
conor | sonne|work: thanks that was helpful :) | 13:31 |
-!- vikram360 [~vikram360@117.192.180.45] has joined #shogun | 13:34 | |
conor | i have a weird problem. i'm used an older shogun version with the LIBLINEAR_L1R_L2LOSS_SVC classifier. there everything worked fine, but now i use the recent shogun version and use LIBLINEAR_L1R_L2LOSS_SVC again. But now i have to transpose the features and the results for w are all NaN. Does anyone know what this could be? I searched the commit messages for LIBLINEAR_L1R_L2LOSS_SVC but couldn't find anything so i don't know what changed there. | 13:35 |
sonne|work | conor: you could help figuring out when this problem happened by doing a git bisect and determining the commit which caused this | 13:43 |
conor | ok i will do this | 13:43 |
sonne|work | btw you can do git log shogun/classifier/svm/SVM_linear.cpp | 13:44 |
sonne|work | or shogun/classifier/svm/LibLinear.cpp to see the changes for these files only | 13:44 |
sonne|work | (I guess I broke it ...) | 13:45 |
-!- conor [~cronor@fb.ml.tu-berlin.de] has quit [Read error: Operation timed out] | 13:46 | |
-!- blackburn [~qdrgsm@109.226.105.25] has joined #shogun | 13:57 | |
CIA-64 | shogun: Sergey Lisitsyn master * rdb96f7c / (3 files): Added bias usage capability for crammer-singer liblinear - http://git.io/LJ3IeQ | 14:30 |
-!- vikram360 [~vikram360@117.192.180.45] has quit [Ping timeout: 245 seconds] | 14:31 | |
-!- vikram360 [~vikram360@117.192.186.135] has joined #shogun | 14:31 | |
-!- cronor [~cronor@141.23.80.206] has joined #shogun | 14:31 | |
cronor | sonne|work: i looked into that but is there a way to access the pre-May2011 logs? | 14:41 |
sonne|work | all logs since 2006 are in there | 14:42 |
sonne|work | ahh but I probably moved files around | 14:43 |
sonne|work | before that data the dir shogun/ did probably not exist | 14:44 |
sonne|work | s/data/date | 14:49 |
n4nd0 | sonne|work: hey! one question about dynamic memory allocation, if I want to create an array of SGVector | 14:55 |
n4nd0 | SG_MALLOC(SGVector< float64_t >, nvecs) | 14:56 |
n4nd0 | and later loop for each vector and do | 14:56 |
n4nd0 | SG_MALLOC(float64_t, nelems) | 14:56 |
n4nd0 | is that the right way to do it? | 14:56 |
n4nd0 | I have doubt about the second MALLOC, to allocate space for each of the vectors | 14:57 |
blackburn | n4nd0: you may use array_of_vectors[i] = SGVector<float64_t>(nelems) | 14:58 |
n4nd0 | n4nd0: hey! I didn't see you are around :) | 14:59 |
n4nd0 | ups I talked to me | 15:00 |
blackburn | yeah | 15:00 |
n4nd0 | blackburn: thank you very much :) | 15:00 |
blackburn | :) | 15:00 |
blackburn | sonne|work: +0.01% of accuracy for adding bias LOL | 15:00 |
blackburn | (I've added bias support for crammer-singer liblinear) | 15:00 |
wiking | blackburn: yoyo | 15:01 |
blackburn | wiking: hey | 15:01 |
wiking | mmm i wonder where's alex :) | 15:01 |
wiking | i supposed to have a Skype conf with him now :P | 15:01 |
blackburn | lets shout ALEX and wait him to come ;) | 15:01 |
wiking | hahahhaah | 15:01 |
wiking | maybe he won't be able to decode the irc protocoll | 15:01 |
n4nd0 | blackburn: what about later to free that stuff, would SG_FREE(array_of_vectors) do the trick without having to call free_vector for every one of them? | 15:03 |
blackburn | n4nd0: it won't, you would have to free it | 15:03 |
n4nd0 | n4nd0: ook, thank you again :) | 15:04 |
blackburn | blackburn: no problem | 15:04 |
blackburn | :D | 15:04 |
CIA-64 | shogun: Sergey Lisitsyn master * r0c71dba / src/shogun/classifier/svm/MulticlassLibLinear.cpp : Added bias usage capability for crammer-singer liblinear - http://git.io/iKQ-hw | 15:05 |
blackburn | wiking: did you try multiclassliblinear already? | 15:06 |
wiking | haven't had time yet :( | 15:06 |
wiking | sorry man | 15:06 |
blackburn | no need to sorry, I mean you were waiting for it :) | 15:06 |
-!- cronor [~cronor@141.23.80.206] has quit [Quit: cronor] | 15:11 | |
wiking | heheh he has shown up \o/ | 15:19 |
n4nd0 | :) | 15:22 |
wiking | sonney2k: here? | 15:37 |
sonne|work | wiking: yes | 15:47 |
sonne|work | whats up? | 15:47 |
vikram360 | sonney2k : There seems to be no information about how or why C5.0 is better than C4.5. Is there any paper in particular or any resource that you know about that elicits the superiority? | 16:42 |
-!- n4nd0 [~nando@n179-p53.kthopen.kth.se] has quit [Ping timeout: 244 seconds] | 16:53 | |
@sonney2k | blackburn, how did you add the bias term to multiclass liblinear? | 17:02 |
blackburn | sonney2k: I'd say virtually | 17:03 |
blackburn | w = [w, b] , x = [x, 1] | 17:03 |
@sonney2k | so you added some constant? | 17:03 |
@sonney2k | without modifying features? | 17:03 |
blackburn | yes | 17:03 |
blackburn | attached b to w and 1 to x | 17:03 |
@sonney2k | ok | 17:03 |
blackburn | virtually | 17:03 |
blackburn | :) | 17:04 |
@sonney2k | how? | 17:04 |
@sonney2k | ok | 17:04 |
blackburn | sonney2k: just replaced dotproducts in MCSVM solver | 17:04 |
@sonney2k | that is one of the problems that they modify data from time to time | 17:04 |
blackburn | pretty useless though | 17:04 |
@sonney2k | IIRC L1 svm had that problem | 17:04 |
@sonney2k | depends on the problem | 17:05 |
blackburn | sonney2k: how can it be true that bias is so useless? | 17:05 |
@sonney2k | I have to leave the train... | 17:05 |
@sonney2k | cu | 17:05 |
blackburn | I never thought it is not really required for high-dim tasks | 17:05 |
blackburn | ok | 17:05 |
vikram360 | :( | 17:06 |
-!- cronor [~cronor@141.23.80.206] has joined #shogun | 18:11 | |
-!- vikram360 [~vikram360@117.192.186.135] has quit [Read error: Connection reset by peer] | 18:18 | |
-!- vikram360 [~vikram360@117.192.186.135] has joined #shogun | 18:28 | |
@sonney2k | blackburn, well it is a *regularized* bias so not really the the one you have in standard SVM | 18:41 |
@sonney2k | so one more dim more doesn't really matter much | 18:41 |
blackburn | sonney2k: I guess it is the one vojtech did in his 2002 paper | 18:41 |
blackburn | for kernel bsvm | 18:41 |
@sonney2k | I haven't read his paper | 18:41 |
@sonney2k | but almost everyone does this 'hack' | 18:42 |
@sonney2k | ocas / liblinear ... | 18:42 |
blackburn | sonney2k: min (||w||^2+b^2) + C \sum \xi | 18:42 |
blackburn | sonney2k: I'm nearly finished with mc ocas! | 18:42 |
blackburn | took a little while | 18:42 |
@sonney2k | very good | 18:43 |
@sonney2k | now we only need to get good old ocas back to work :) | 18:43 |
blackburn | sonney2k: yeah probably we need to | 18:44 |
blackburn | sonney2k: what do you think about that idea you suggested? | 18:44 |
@sonney2k | there is certainly enough other stuff too... ohh well | 18:44 |
@sonney2k | well I would say we don't have to rush | 18:45 |
blackburn | deadline is coming, but ideas are not concerned I guess? | 18:45 |
@sonney2k | have to leave train now | 18:45 |
blackburn | damn do you move in train whole day long? ;) | 18:46 |
blackburn | do you work in siberia? :) | 18:46 |
@sonney2k | Re | 18:51 |
@sonney2k | well I think one can even do spontaneous things | 18:51 |
@sonney2k | so if some student has some nice idea that he wants to persue - that woudl be possible as well | 18:52 |
@sonney2k | but I certainly would love tighther python integration | 18:52 |
@sonney2k | in the end that is the interface I use mostly | 18:53 |
blackburn | sonney2k: others are pretty useless for me as well | 18:59 |
@sonney2k | not useless - I but I just use them very rarely | 19:00 |
blackburn | do you use it? :) | 19:01 |
blackburn | I have never used neither java nor ruby :) | 19:01 |
blackburn | sonney2k: btw java integration is pretty interesting thing as well | 19:02 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 19:02 | |
blackburn | sonney2k: directory arrangement is still crappy unclear for me :( | 19:11 |
-!- cronor [~cronor@141.23.80.206] has quit [Quit: cronor] | 19:30 | |
-!- cronor [~cronor@141.23.80.206] has joined #shogun | 19:32 | |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 19:41 | |
-!- cronor [~cronor@141.23.80.206] has quit [Quit: cronor] | 20:03 | |
blackburn | sonney2k: done with mc ocas - got 21.6% accuracy :D | 20:05 |
blackburn | going to diff libocas | 20:05 |
-!- cronor [~cronor@141.23.80.206] has joined #shogun | 20:18 | |
-!- cronor [~cronor@141.23.80.206] has quit [Quit: cronor] | 20:24 | |
blackburn | ah I see, maxtraintime :) | 20:54 |
blackburn | sonney2k: btw I checked libocas and libqp - all looks ok | 20:54 |
blackburn | I do not understand why you ->ed it to int32_t/float64_t, but everything else is ok | 20:55 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 246 seconds] | 21:05 | |
-!- cronor [~cronor@e178170108.adsl.alicedsl.de] has joined #shogun | 21:21 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 21:39 | |
CIA-64 | shogun: Sergey Lisitsyn master * rec86ec8 / (2 files): Updated multiclass machine base class - http://git.io/QgS8CA | 22:42 |
CIA-64 | shogun: Sergey Lisitsyn master * r229943e / src/shogun/classifier/svm/libocas.cpp : Fixes for MC OCAS - http://git.io/yfEKbQ | 22:42 |
CIA-64 | shogun: Sergey Lisitsyn master * r7e1d354 / (5 files in 3 dirs): Introduced multiclass OCAS - http://git.io/HnBINg | 22:42 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Read error: Operation timed out] | 22:55 | |
-!- wiking [~wiking@78-23-191-201.access.telenet.be] has joined #shogun | 23:07 | |
-!- wiking [~wiking@78-23-191-201.access.telenet.be] has quit [Changing host] | 23:07 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 23:07 | |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 23:15 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 23:15 | |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Ping timeout: 260 seconds] | 23:20 | |
-!- blackburn [~qdrgsm@109.226.105.25] has quit [Ping timeout: 276 seconds] | 23:22 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 23:36 | |
--- Log closed Fri Mar 09 00:00:19 2012 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!