--- Log opened Thu Sep 27 00:00:17 2012 | ||
-!- in3xes_ [~in3xes@122.174.90.177] has quit [Ping timeout: 252 seconds] | 01:36 | |
shogun-buildbot | build #115 of nightly_default is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/115 | 03:51 |
---|---|---|
-!- blackburn [~blackburn@188.168.14.95] has joined #shogun | 07:43 | |
-!- hoijui [~hoijui@dslb-088-074-107-101.pools.arcor-ip.net] has joined #shogun | 08:06 | |
-!- blackburn [~blackburn@188.168.14.95] has quit [Quit: Leaving.] | 08:33 | |
-!- hoijui [~hoijui@dslb-088-074-107-101.pools.arcor-ip.net] has quit [Ping timeout: 265 seconds] | 12:40 | |
-!- heiko1 [~heiko@p4FED538D.dip.t-dialin.net] has joined #shogun | 13:55 | |
-!- heiko1 [~heiko@p4FED538D.dip.t-dialin.net] has quit [Ping timeout: 245 seconds] | 14:29 | |
-!- blackburn [~blackburn@188.168.4.185] has joined #shogun | 14:46 | |
-!- heiko1 [~heiko@p4FED538D.dip.t-dialin.net] has joined #shogun | 16:00 | |
-!- heiko1 [~heiko@p4FED538D.dip.t-dialin.net] has quit [Ping timeout: 248 seconds] | 16:24 | |
-!- heiko1 [~heiko@p4FED538D.dip.t-dialin.net] has joined #shogun | 16:37 | |
heiko1 | blackburn around? | 16:57 |
blackburn | heiko1: hey, yes | 16:57 |
heiko1 | I have another problem with c++ (many these days) | 16:57 |
heiko1 | I want to make a subclass of a templated class | 16:57 |
heiko1 | in particular, subclass of CSTreamingDeanseFeatures<T> | 16:58 |
heiko1 | template <class T> | 16:58 |
heiko1 | class CStreamingDenseDataGenerator: public CStreamingDenseFeatures<T> | 16:58 |
heiko1 | but that doesnt work | 16:58 |
heiko1 | a) I cannot access members of CStreamingDenseFeatures in methods of CStreamingDenseDataGenerator<T> | 16:58 |
heiko1 | b) shogun doesnt recognise the new class as generic so compiling class_list.cpp fails | 16:58 |
heiko1 | any ideas? | 16:58 |
heiko1 | blackburn, weird problem | 16:59 |
blackburn | hmm | 16:59 |
blackburn | yeah pretty weird | 17:01 |
heiko1 | man no idea :)( | 17:03 |
heiko1 | :( | 17:03 |
-!- ptizoom [55d25a64@gateway/web/freenode/ip.85.210.90.100] has joined #shogun | 17:03 | |
blackburn | heiko1: about a) - does it compile? | 17:04 |
heiko1 | yes | 17:04 |
heiko1 | only b) fails to compile | 17:05 |
heiko1 | oh sorry | 17:05 |
heiko1 | no | 17:05 |
heiko1 | complains that the member elements are not known | 17:05 |
heiko1 | ah but I can avoid this using | 17:05 |
heiko1 | CStreamingDenseDataGenerator<T>::member_name | 17:05 |
heiko1 | still b) | 17:05 |
heiko1 | in class list, generic classes are instanciated via this switch thing | 17:06 |
heiko1 | that is automatically generated for generic classes | 17:06 |
heiko1 | but for this subclass, htis doesnt happen | 17:06 |
blackburn | did you instantiate your new class? | 17:07 |
blackburn | in cpp | 17:07 |
ptizoom | hi, I am looking into associating PDL and Shogun (by replacing numpy). I am progressing slowly...I have some framework (which is not yet compiling), but is it possible to have a development git-branch ? | 17:09 |
blackburn | ptizoom: I am unsure I got what you mean | 17:10 |
blackburn | dev git branch of shogun? | 17:11 |
ptizoom | blackburn: is shogun under the git? | 17:11 |
blackburn | well, yes | 17:12 |
blackburn | but you may commit everything you want in your fork | 17:12 |
heiko1 | blackburn, how do you mean that? | 17:13 |
ptizoom | is it easy to get a writting branch from the project? | 17:13 |
heiko1 | ptizoom, that takes a while, but as blackburn said, you can just push everything into your fork and then send pull-requests. We then review these, give feedback, and evtl. merge | 17:14 |
heiko1 | blackburn, its an error in the class_list.py script | 17:17 |
heiko1 | I tuned it manually and now it compiles | 17:17 |
blackburn | ptizoom: yeah pretty easy | 17:20 |
blackburn | ah heiko1 answered | 17:20 |
blackburn | :) | 17:20 |
blackburn | heiko1: how does class_list.py check whether it is a template? | 17:21 |
heiko1 | blackburn, currently checking | 17:21 |
ptizoom | heiko1: cool | 17:22 |
heiko1 | blackburn, is there any problem with subclassing templated classes? | 17:37 |
heiko1 | I not get weird errors | 17:38 |
heiko1 | for example when trying to create an instance I get | 17:38 |
heiko1 | undefined reference to `shogun::CStreamingDenseDataGenerator<double>::CStreamingDenseDataGenerator()' | 17:38 |
heiko1 | which is totally weird since the definition is there | 17:38 |
heiko1 | why is c++ alway such a horror when it comes to certain things | 17:39 |
blackburn | heh | 17:40 |
heiko1 | this is so annoying, I am wasting my time with this stupid stuff where all I want is a generic subclass | 17:40 |
heiko1 | cant be that difficult | 17:40 |
blackburn | so it fails to find implementation? | 17:40 |
heiko1 | iyes | 17:40 |
blackburn | can you paste the code somewhere? | 17:41 |
heiko1 | yes, will do | 17:41 |
heiko1 | http://pastebin.com/QSqaGBte | 17:42 |
heiko1 | http://pastebin.com/FRCFrbMc | 17:43 |
heiko1 | first error you get should be class_list.cpp | 17:43 |
heiko1 | I think its related to pure virtual functions | 17:45 |
heiko1 | and that inheritance of templated classes doesn not give superclasses elements | 17:46 |
heiko1 | (which would explain the pure virtual methods and the non-accessability of superclass members) | 17:46 |
heiko1 | blackburn, any ideas? | 17:47 |
blackburn | a minute | 17:47 |
heiko1 | blackburn, leaving now, if you would find something that was amazing :) | 17:59 |
blackburn | heiko1: sorry got distracted | 17:59 |
heiko1 | no worries | 17:59 |
blackburn | let me check now, I'll let you know | 17:59 |
heiko1 | cool, thx | 18:00 |
heiko1 | ill check back in 15 mins in case something is missing to test it | 18:00 |
blackburn | heiko1: I think you have to instantiate using | 18:01 |
blackburn | template class CStreamingDenseDataGenerator<bool>; | 18:01 |
blackburn | etc | 18:01 |
blackburn | in .cpp | 18:02 |
blackburn | see DenseFeatures.cpp for example | 18:02 |
-!- audy [~audy@heyaudy.com] has quit [Changing host] | 18:11 | |
-!- audy [~audy@unaffiliated/audy] has joined #shogun | 18:11 | |
heiko1 | blackburn, no that doesnt work | 18:16 |
heiko1 | have you tried it? | 18:17 |
heiko1 | blackburn, compiler complains about non-pure virtual methods | 18:17 |
blackburn | hmm do you have some pure virtual methodds around? | 18:18 |
blackburn | I can't see any | 18:19 |
heiko1 | no | 18:20 |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun | 18:20 | |
heiko1 | it complains about all the pure virtual methods from CStreamingFetaures that are imeplemented in CStreamingDenseFeatures | 18:21 |
heiko1 | but these are for some reason not visible in the subclass | 18:21 |
blackburn | heiko1: did you try forward declaration of streaming dense feats? | 18:23 |
heiko1 | blackburn, what is that? | 18:23 |
blackburn | template <class ST> class CStreamingDenseFeatures; | 18:24 |
blackburn | before definition in *.h | 18:24 |
heiko1 | ah sorry ok, Ill try | 18:24 |
blackburn | yeah C++ is pain.. | 18:25 |
heiko1 | blackburn, no doesnt change anything | 18:25 |
blackburn | damn | 18:25 |
heiko1 | but makes sense since the header is included above | 18:25 |
heiko1 | yeah man, c++ and OO is so annoying sometimes | 18:25 |
heiko1 | no interfaces, no generics | 18:25 |
heiko1 | I need these all the time :D | 18:25 |
heiko1 | blackburn, Ill update the listings, have you tried to compile them? | 18:26 |
blackburn | no, not yet | 18:26 |
blackburn | let me try then | 18:26 |
heiko1 | would be nice ofyou | 18:26 |
blackburn | base/class_list.cpp:414:103: error: expected type-specifier before 'CStreamingDenseDataGenerator' | 18:29 |
blackburn | is that you get? | 18:29 |
heiko1 | yes | 18:30 |
heiko1 | you can get around that | 18:30 |
heiko1 | via searching for CStreamingDenseDataGenerator in class_list.cpp | 18:30 |
heiko1 | and then cut the line and replace it by a block that is similar to the other templated classes, e.g. CDenseFeatures | 18:31 |
heiko1 | with a switch over the types | 18:31 |
blackburn | heh yeah but we should rather fix class_list thing | 18:31 |
heiko1 | no | 18:32 |
blackburn | http://www.youtube.com/watch?v=_cG7ZVBXQII&feature=player_detailpage watch that video meanwhile | 18:32 |
blackburn | :D | 18:32 |
heiko1 | it doesnt solve the problem untimately | 18:32 |
heiko1 | cut the line | 18:32 |
heiko1 | static CSGObject* __new_CStreamingDenseDataGenerator(EPrimitiveType g) { return g == PT_NOT_GENERIC? new CStreamingDenseDataGenerator(): NULL; } | 18:32 |
heiko1 | and replace it by a block that you can find from line 753 ... | 18:32 |
heiko1 | lol the video is great .) | 18:33 |
blackburn | but why does it treat your class as not templated? | 18:33 |
heiko1 | I think its because of the templates doing something wrong | 18:34 |
heiko1 | blackburn, there is another problem even before the class_list | 18:35 |
heiko1 | if you add | 18:35 |
heiko1 | template class CStreamingDenseDataGenerator<bool>; in the last line of the .cpp file | 18:36 |
heiko1 | it doesnt compile already | 18:36 |
blackburn | lol! you've got to make it one line | 18:36 |
heiko1 | due to these pure virtual function | 18:36 |
heiko1 | ? | 18:36 |
blackburn | 32 template <class T> class CStreamingDenseDataGenerator: public CStreamingDenseFeatures<T> | 18:36 |
blackburn | this way | 18:36 |
blackburn | not two liner | 18:36 |
blackburn | class_list solved then | 18:37 |
heiko1 | what??? | 18:37 |
heiko1 | argh | 18:37 |
blackburn | okay lets get to undefined reference problem | 18:37 |
heiko1 | since python is line based | 18:37 |
blackburn | not a python but the class_list.cpp.py thing | 18:37 |
heiko1 | do you now also get the undefined references? | 18:37 |
blackburn | yeah | 18:38 |
heiko1 | These can be solved by adding the lines | 18:38 |
blackburn | work work sleep then work | 18:38 |
blackburn | :D | 18:38 |
heiko1 | template class CStreamingDenseDataGenerator<bool>; | 18:38 |
heiko1 | etc | 18:38 |
heiko1 | but then another error appears | 18:38 |
blackburn | okay let me get to next error in our funny journey | 18:38 |
heiko1 | what do you think of this one? | 18:39 |
blackburn | features/StreamingDenseDataGenerator.cpp:16:35: error: type 'shogun::CStreamingDenseDataGenerator<bool>' is not a direct base of 'shogun::CStreamingDenseDataGenerator<bool>' | 18:39 |
heiko1 | yeah | 18:39 |
heiko1 | that | 18:39 |
blackburn | hhhaha | 18:40 |
blackburn | okay do not blame C++, it is you :) | 18:40 |
blackburn | check your constructor carefully | 18:40 |
blackburn | you kinda calling it itself again | 18:40 |
blackburn | however other errors stay :D | 18:41 |
heiko1 | aaaa | 18:41 |
heiko1 | no | 18:41 |
heiko1 | i already fixed the others | 18:41 |
heiko1 | copy paste -.- | 18:41 |
blackburn | hehe | 18:41 |
heiko1 | man | 18:41 |
heiko1 | thanks a lot | 18:41 |
heiko1 | guess i have been starring to the monitor too much | 18:41 |
blackburn | lets dance north korean style | 18:41 |
blackburn | kim jong style | 18:41 |
blackburn | :D | 18:41 |
heiko1 | would never have found this | 18:41 |
heiko1 | yeah, go :D | 18:41 |
heiko1 | yep works now | 18:42 |
blackburn | I absolutely love this country btw, would love to visit they | 18:42 |
blackburn | but too scary | 18:42 |
blackburn | :D | 18:42 |
heiko1 | why do you love it? | 18:42 |
blackburn | it is like they freezed soviet union in 60s | 18:43 |
blackburn | so funny and crazy | 18:43 |
heiko1 | soyou like that retroish style :) | 18:43 |
heiko1 | you should get a lomography camera then :) | 18:44 |
blackburn | hah one friend of mine has a few | 18:44 |
heiko1 | http://microsites.lomography.com/stores/gallery-stores/east-london | 18:44 |
heiko1 | my girlfriend has a a few | 18:44 |
heiko1 | they make cool phoos | 18:44 |
heiko1 | tos | 18:44 |
blackburn | his photos look like this: http://cs9220.userapi.com/u596778/146498181/z_8d525488.jpg | 18:45 |
heiko1 | yeah | 18:45 |
heiko1 | ohoh | 18:45 |
heiko1 | there are more errors | 18:45 |
heiko1 | warnings: | 18:45 |
heiko1 | inline-Funktion >>shogun::EFeatureType shogun::CStreamingDenseFeatures<T>::get_feature_type() const [with T = long double]<< verwendet, aber nirgendwo definiert | 18:45 |
heiko1 | bla used but not defined | 18:45 |
blackburn | hahahha | 18:46 |
blackburn | haha | 18:46 |
blackburn | that's why you didn't paste | 18:46 |
blackburn | errors | 18:46 |
blackburn | :D | 18:46 |
heiko1 | and these are exactly the pure virtual methods that are imeplemented in CStreamingDenseFeatures | 18:46 |
heiko1 | you got it :) | 18:46 |
blackburn | hmm | 18:46 |
blackburn | aber nirgendwo definiert! | 18:46 |
blackburn | :D | 18:46 |
heiko1 | thats: but nowhere defined :) | 18:47 |
blackburn | I know only what GmbH is | 18:47 |
blackburn | gesselschaft mit beschraenkter haftung! | 18:47 |
heiko1 | what is also strange: have a look at CStreamingDenseFeatures | 18:48 |
blackburn | obschestvo s ogranichennoy otvetstvennostyu | 18:48 |
heiko1 | there is an element current_sgvector | 18:48 |
blackburn | not that shorter in russian, lol | 18:48 |
heiko1 | but I cannot access that from my new class, compile error | 18:48 |
heiko1 | Russion looks mess | 18:48 |
heiko1 | y | 18:48 |
heiko1 | and its beschr?????nkter ) | 18:48 |
blackburn | yeah, I thought ae works too | 18:48 |
heiko1 | yeah it does | 18:49 |
blackburn | lets see updated code then | 18:49 |
heiko1 | I will create another pastebin | 18:50 |
blackburn | heiko1: btw I am now developing the ideal machine learning environment (TM) | 18:50 |
heiko1 | http://pastebin.com/JiM2S6vU | 18:50 |
blackburn | check one idea | 18:50 |
heiko1 | http://pastebin.com/mMhQVMtG | 18:51 |
heiko1 | Ill check | 18:51 |
blackburn | heiko1: https://gist.github.com/3795078 | 18:51 |
heiko1 | btw I renamed the variable to current_vector in my code-base | 18:51 |
blackburn | briefly - all the things that happen to 'context' variables are going to be serialized | 18:51 |
blackburn | but that's all transparent and you don't care about that | 18:52 |
heiko1 | but thats python only or? | 18:52 |
blackburn | yeah sure | 18:53 |
blackburn | such madness can't happen in any normal language | 18:53 |
heiko1 | :) | 18:53 |
blackburn | do you like the idea? | 18:53 |
blackburn | my other extension is | 18:54 |
blackburn | with operation('reading files') as context:context.train_data = numpy.loadtxt('train.csv',delimiter=",",skiprows=1)context.test_data = numpy.loadtxt('test.csv',delimiter=",",skiprows=1) | 18:54 |
blackburn | argh | 18:54 |
blackburn | with operation('reading files') as context: | 18:54 |
blackburn | context.train_data = numpy.loadtxt('train.csv',delimiter=",",skiprows=1) | 18:54 |
blackburn | context.test_data = numpy.loadtxt('test.csv',delimiter=",",skiprows=1) | 18:54 |
blackburn | context.train_data is going to be saved and then serialized to file | 18:54 |
blackburn | after with block | 18:54 |
blackburn | I am thinking about a way to load it from serialized automatically as well | 18:55 |
heiko1 | yeah I mean, thats all the magic of pickle | 18:55 |
heiko1 | nice, like it | 18:55 |
blackburn | yes, it is easy | 18:55 |
blackburn | the thing I am trying to reach | 18:55 |
heiko1 | though its not really shogun-ish since only works in one binding | 18:55 |
blackburn | is transparency of such operations like measuring time | 18:55 |
blackburn | or serialization | 18:56 |
heiko1 | transparency? | 18:56 |
blackburn | or rather implicit way | 18:56 |
blackburn | I mean I want to save everything without calling anything | 18:56 |
heiko1 | yeah, thats great use | 18:56 |
blackburn | have you seen patterns for ML article? | 18:57 |
heiko1 | no | 18:57 |
heiko1 | where ? | 18:57 |
blackburn | http://arkitus.com/PRML/ | 18:57 |
heiko1 | doing that all the time anyway :) | 18:58 |
blackburn | yes, thats the spirit | 18:58 |
blackburn | features/streaming/StreamingDenseDataGenerator.cpp:75:68: error: invalid conversion from 'const char*' to 'int32_t {aka int}' [-fpermissive] | 18:58 |
blackburn | ../shogun/base/Parameter.h:206:21: error: initializing argument 1 of 'shogun::TParameter* shogun::Parameter::get_parameter(int32_t)' [-fpermis | 18:58 |
blackburn | that's what I get | 18:59 |
heiko1 | ah year | 18:59 |
heiko1 | I added stuff | 19:00 |
heiko1 | wait, I will push that quicklly | 19:00 |
blackburn | heiko1: btw are you using vim? | 19:00 |
heiko1 | no, why? | 19:00 |
blackburn | just curious | 19:00 |
blackburn | I argue with my team lead at job every day :D | 19:00 |
heiko1 | blackburn, you can just comment that line 75 | 19:01 |
blackburn | I am a vim guy, he says sth like GUI rocks | 19:01 |
heiko1 | no I actually use eclipse | 19:01 |
heiko1 | like browsing around in the code by mouse | 19:01 |
blackburn | I see | 19:01 |
blackburn | looks like it appears to compile then | 19:02 |
heiko1 | yeah | 19:02 |
heiko1 | but change something and compile again: | 19:02 |
heiko1 | warnings | 19:02 |
blackburn | yeah, I get quite a lot warnings | 19:03 |
heiko1 | these are strange | 19:03 |
heiko1 | because the methods are all implemented in superclass | 19:03 |
heiko1 | (in fact, thats why I derived) | 19:03 |
blackburn | probably we have some other glitch around there | 19:04 |
blackburn | gangnam style! | 19:05 |
blackburn | woop woop woop woop | 19:05 |
heiko1 | ? | 19:06 |
blackburn | :D | 19:06 |
heiko1 | code works apart from these warnings btw :) | 19:07 |
blackburn | can't see nothing wrong | 19:07 |
blackburn | walking kim jong style | 19:07 |
heiko1 | :D | 19:10 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 19:22 | |
n4nd0 | wiking: around? | 19:31 |
blackburn | n4nd0 entered room kim jong style | 19:31 |
n4nd0 | haha why? | 19:32 |
blackburn | http://www.youtube.com/watch?v=_cG7ZVBXQII&feature=player_detailpage | 19:32 |
n4nd0 | blackburn: any news about your interview? | 19:40 |
blackburn | n4nd0: no, nothing yet | 19:40 |
blackburn | they have to come up with a reject message :D | 19:41 |
n4nd0 | I bet you don't believe that ;) | 19:41 |
blackburn | no I really believe I am not that cool :) | 19:43 |
-!- heiko1 [~heiko@p4FED538D.dip.t-dialin.net] has left #shogun [] | 19:49 | |
wiking | n4nd0: yes | 21:14 |
n4nd0 | wiking: how is it going using the SO framework? | 21:15 |
n4nd0 | blackburn: do you know if the post-gsoc post came eventually to life? | 21:20 |
n4nd0 | post-gsoc post haha | 21:20 |
n4nd0 | funny name | 21:20 |
blackburn | you mean t-shirt thing? | 21:20 |
blackburn | iirc it should arrive before 30th | 21:20 |
n4nd0 | no no | 21:21 |
n4nd0 | the post | 21:22 |
n4nd0 | sonney2k asked for help for a post and I sent him by mail one | 21:22 |
n4nd0 | a couple of weeks ago he said he'd started fixing it | 21:22 |
wiking | n4nd0: i've got convergence with bmrm | 21:22 |
blackburn | ahh | 21:24 |
blackburn | heh | 21:24 |
blackburn | well no idea | 21:24 |
n4nd0 | wiking: really? I think you should actually get the same solution as with PrimalMosek | 21:28 |
wiking | yeah | 22:08 |
wiking | my problem is/was really the memory consumption | 22:09 |
wiking | of mosek | 22:09 |
n4nd0 | oh sorry I misunderstood. for some strange reason I thought I read that it didn't converge with bmrm | 22:18 |
n4nd0 | I don't know if I screw sth up with my code using mosek | 22:18 |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 240 seconds] | 22:26 | |
-!- blackburn [~blackburn@188.168.4.185] has quit [Quit: Leaving.] | 22:43 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 248 seconds] | 23:02 | |
--- Log closed Fri Sep 28 00:00:17 2012 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!