--- Log opened Mon May 02 00:00:41 2011 | ||
-!- f-x [~gen@180.149.49.227] has quit [Quit: leaving] | 00:03 | |
-!- blackburn [~qdrgsm@188.122.250.155] has quit [Quit: Leaving.] | 00:28 | |
-!- ameerkat [~ameerkat@184-98-140-155.phnx.qwest.net] has quit [Ping timeout: 246 seconds] | 00:32 | |
-!- ameerkat [~ameerkat@184-98-140-155.phnx.qwest.net] has joined #shogun | 03:19 | |
-!- ameerkat [~ameerkat@184-98-140-155.phnx.qwest.net] has quit [Ping timeout: 252 seconds] | 04:32 | |
-!- ameerkat [~ameerkat@184-98-140-155.phnx.qwest.net] has joined #shogun | 07:15 | |
-!- heiko [~heiko@infole-06.uni-duisburg.de] has joined #shogun | 09:46 | |
-!- blackburn [~qdrgsm@188.122.250.155] has joined #shogun | 10:00 | |
-!- ameerkat [~ameerkat@184-98-140-155.phnx.qwest.net] has quit [Ping timeout: 260 seconds] | 10:01 | |
heiko | sonney2k, are you there? | 10:27 |
---|---|---|
@sonney2k | heiko, now yes | 11:19 |
heiko | hey | 11:19 |
@sonney2k | ho :) | 11:19 |
@bettyboo | sonney2k, weird | 11:20 |
heiko | ;) had a question, but i think i solved it | 11:20 |
@sonney2k | heiko, very good :) | 11:20 |
heiko | with this set_from_parameters stuff ... | 11:20 |
heiko | I am currently doing it like this | 11:20 |
heiko | added a function for get_size and get_num_elements to TSGDataType | 11:21 |
heiko | and in set_from_parameters, if the primitive type is not PT_SGOBJECT I simple do a memcpy of the m_parameter | 11:21 |
heiko | and if its an SGObject, I also just copy the pointers but unref the old ones before and ref the new ones | 11:22 |
heiko | I found a lot of warnings for unhandled CT_NDARRAY types | 11:23 |
@sonney2k | well ref before and unref when you are done right? | 11:23 |
@sonney2k | otherwise you see crashes - but apart from that fine | 11:24 |
@sonney2k | the errors for CT_NDARRARY are 'OK' support for that needs to be implemented in hdf5 | 11:25 |
heiko | ok | 11:25 |
heiko | perhaps I still have a little problem with the referenceing: | 11:26 |
heiko | when I for example create a new kernel with a parameter that is registered. | 11:27 |
heiko | then I call set_from_parameters on its m_parameters member | 11:27 |
heiko | there, the pointer to an SG_OBJECT is overwritten | 11:27 |
heiko | then its gone, so I have to unref the old one before, right? | 11:27 |
heiko | I mean the SGObject is still there but the pointer is gone | 11:28 |
@sonney2k | heiko, yes you are right, first unref the old one and then put the one there but ref it | 11:28 |
heiko | ok, thats what i did | 11:29 |
heiko | I will write a test for it | 11:29 |
heiko | btw when we write tests, should they have to be included in shogun? for make tests? | 11:29 |
-!- sploving [~root@124.16.139.196] has joined #shogun | 11:31 | |
sploving | hello sonney2k | 11:32 |
sploving | are you around? | 11:32 |
@sonney2k | heiko, write an example in examples/undocumentd/python_modular in the same style | 11:32 |
@sonney2k | (as there are examples already) | 11:32 |
@sonney2k | that should then be an example on how to use things and simultaneously function as a test | 11:33 |
@sonney2k | sploving, yes | 11:33 |
@sonney2k | did you make progress on the java typemaps? | 11:33 |
sploving | sonney2k, I change the typemape file, to make: double[] y; | 11:34 |
sploving | x.get_labels(y); | 11:34 |
sploving | but it failed: y is not initializated | 11:35 |
@sonney2k | sploving, hmmhh still not ideal - do you know if this is the only possible way? | 11:35 |
@sonney2k | I mean double[]y = x.get_labels(); would be nicer | 11:35 |
sploving | I tried that too. but the wrapper file function is like this: void F() | 11:36 |
sploving | it should be jlong F() | 11:37 |
sploving | like this | 11:37 |
sploving | I have not found the typemap example in java like this typemap(in, inputnums=0) | 11:40 |
heiko | current git has a compile error | 11:59 |
heiko | CMath::calcroc call, this function is not there | 12:00 |
blackburn | eh.. | 12:02 |
blackburn | working on my | 12:03 |
blackburn | heiko: may be 'make clean' will fix it | 12:04 |
heiko | ok will try | 12:05 |
blackburn | heiko: calcroc is called somewhere? | 12:05 |
heiko | yes | 12:05 |
blackburn | where? | 12:05 |
heiko | in CGUIMath:48 | 12:05 |
blackburn | oh | 12:05 |
blackburn | heiko: thank you! will fix it now | 12:06 |
heiko | ok :) | 12:06 |
blackburn | heiko: it is just my fault because I removed calcroc but not using libshogunui | 12:07 |
heiko | ok, I see. | 12:07 |
heiko | why not using shogunui? | 12:08 |
heiko | still does not compile | 12:08 |
blackburn | still? | 12:08 |
heiko | make: no rule to create target PerformanceMeasures.h needed by Evaluation_wrap.cxx | 12:08 |
heiko | python modular | 12:08 |
blackburn | hm.. PerformanceMeasures is removed from evaluation | 12:09 |
heiko | mmmh | 12:09 |
blackburn | did you made clean? | 12:09 |
heiko | ah perhaps that is fixed by the make clean | 12:09 |
heiko | sorry, old error before make clean after commenting out the line | 12:09 |
blackburn | yeah it should because there was some 'refactoring' on evaluation :) | 12:09 |
heiko | ok | 12:09 |
heiko | compiles | 12:09 |
heiko | ok :) | 12:10 |
@bettyboo | ha ha heiko | 12:10 |
blackburn | I'm stucked with GUIMath :) | 12:10 |
blackburn | it uses some mess to evaluate results | 12:10 |
heiko | some mess? | 12:12 |
blackburn | I mean it is a bit complicated and etc | 12:12 |
blackburn | but I think I will fix it now | 12:13 |
blackburn | sonney2k: around? | 12:13 |
@sonney2k | heiko, well do a make distclean and then only configure --interfaces=libshogun,python_modular | 12:16 |
@sonney2k | (for now) | 12:16 |
@sonney2k | blackburn, just remove that roc crap from libshogunui GUIMath or so | 12:16 |
blackburn | ok | 12:16 |
blackburn | sonney2k: should I add evaluation usage there later? | 12:17 |
@sonney2k | blackburn, I don't use it - so I wouldn't waste time :)( | 12:20 |
blackburn | sonney2k: is there anybody using it? | 12:20 |
@sonney2k | blackburn, not that I know :) | 12:21 |
blackburn | hehe | 12:21 |
@sonney2k | sploving, I would suggest to ask on the swig mailinglist then | 12:21 |
blackburn | sonney2k: could you please comment out lines 41-67 (implementation of evaluate_results) in GUIMath.cpp? it will fix the problem | 12:26 |
@sonney2k | ^^heiko - do that for now :) | 12:26 |
heiko | already done here :) | 12:27 |
blackburn | sonney2k: you suggested to translate doc to russian using google translate. look at the first sentence: :D | 12:31 |
blackburn | SHOGUN это большая машина масштаба обучения инструментов с акцентом наособенности опорных векторов (SVM). | 12:31 |
blackburn | translating.. | 12:31 |
blackburn | SHOGUN is a big machine scale learning toolbox with accent to features of support vectors | 12:32 |
blackburn | :D | 12:32 |
blackburn | yeah google translate rocks | 12:32 |
@sonney2k | :) | 12:33 |
-!- heiko [~heiko@infole-06.uni-duisburg.de] has left #shogun [] | 12:58 | |
CIA-90 | shogun: Soeren Sonnenburg master * r5904d7e / src/libshogun/kernel/Kernel.cpp : fix bessel warning - http://bit.ly/luxves | 13:48 |
CIA-90 | shogun: Soeren Sonnenburg master * r48a49be / (3 files in 2 dirs): ppwm fixes in compute w - http://bit.ly/kx8xpj | 13:48 |
sploving | sonney2k, sent the email | 13:58 |
sploving | to swig | 13:58 |
@sonney2k | sploving, in the meantime you can add your lua example to the examples and could do some other typemaps for lua... | 14:00 |
sploving | sonney2k, no I should finish one then another | 14:01 |
sploving | I am taking a look at ant | 14:01 |
@sonney2k | sploving, or that | 14:01 |
@sonney2k | just don't idle | 14:01 |
sploving | yeap | 14:02 |
* blackburn made a new branch for LLE | 14:14 | |
blackburn | hehe git is cool | 14:14 |
-!- alesis-novik [~alesis@188.74.87.84] has joined #shogun | 14:43 | |
alesis-novik | Greetings | 14:43 |
blackburn | hi | 14:44 |
alesis-novik | 4 down, 4 to go. There was no EM in this years exam. Damn. | 14:44 |
blackburn | nice to hear that | 14:45 |
blackburn | done halfly :) | 14:45 |
alesis-novik | I just want it to be over now, so I can get on with my GSoC project | 14:46 |
blackburn | alesis-novik: same shit here | 14:46 |
-!- sploving [~root@124.16.139.196] has quit [Remote host closed the connection] | 14:48 | |
CIA-90 | shogun: Soeren Sonnenburg master * r5e61a00 / (3 files in 2 dirs): PPWMs are really in logspace + malloc instead of new - http://bit.ly/kdZ0Sf | 14:51 |
-!- serialhex [~quassel@99-101-149-136.lightspeed.wepbfl.sbcglobal.net] has joined #shogun | 15:58 | |
-!- serialhex [~quassel@99-101-149-136.lightspeed.wepbfl.sbcglobal.net] has quit [Remote host closed the connection] | 16:08 | |
-!- sploving [~root@124.16.139.196] has joined #shogun | 16:11 | |
-!- blackburn [~qdrgsm@188.122.250.155] has quit [Quit: Leaving.] | 16:12 | |
-!- heiko [~heiko@infole-06.uni-duisburg.de] has joined #shogun | 16:29 | |
-!- serialhex01 [~androirc@99-101-149-136.lightspeed.wepbfl.sbcglobal.net] has joined #shogun | 16:33 | |
serialhex01 | anyone here?? | 16:42 |
serialhex01 | i guess not | 16:43 |
serialhex01 | sonney2k & sploving: my computer has been down & i realized that i wasnt studying the right stuff before, computer is still down but ive got the docs i need on my tablet, so im not *completely* inneffectual. will post more updates when i can. (and i wish this thing had a real keyboard!!!) | 16:48 |
@sonney2k | very good :) | 16:49 |
serialhex01 | sonney2k i read in the log you were looking for me saturday i believe... anything important??? | 16:51 |
@sonney2k | serialhex01, I wanted to say that it is not clear whether sploving volunteered to do c# when you do ruby - as you were telling ... | 16:52 |
sploving | sonney2k, sorry. I would not do that. as I never used that before | 16:53 |
@sonney2k | sploving, I know - but you haven't used ruby either - so not a big difference :) | 16:53 |
serialhex01 | ahh, ok, a misunderstanding on my part then... sorry | 16:53 |
@sonney2k | but lets see first if serialhex01 can deal with ruby | 16:54 |
sploving | no. ruby is a scripting language, it is easy to learn | 16:54 |
@sonney2k | sploving, c# is *very similar* to java - in particular the swig interface is. | 16:54 |
@sonney2k | my minimal example is exactly working like in java so I don't think it will be very different typemap wise either | 16:55 |
serialhex01 | afaik c# is similar to c++ & java... | 16:55 |
sploving | sonney2k, I know that. but when debug, it may take long time | 16:56 |
sploving | so I just support what i write in the proposal | 16:57 |
serialhex01 | why would you need to debug? everyone writes perfect code once they start it dont they?? :P | 16:57 |
* serialhex01 hates this damn tablet keyboard!!!!!!! | 16:58 | |
sploving | serialhex01, you know, I am just a student, not an expert | 17:00 |
@sonney2k | sploving, serialhex01 was making a joke | 17:01 |
serialhex01 | im just messing with you sploving, i can almost guarante my code will be worse than yours to begin with :P | 17:02 |
sploving | sonnwy2k, i have built up the ant env | 17:02 |
sploving | it works well | 17:02 |
@sonney2k | sploving, don't worry lets do java first and then see how it is going - c# - it is not that difficult | 17:02 |
sploving | but I think I need change the configure and typemap to generate all the .java file into the shogun package | 17:02 |
serialhex01 | my relative inexpierence is why i only applied for ruby | 17:03 |
serialhex01 | well, im off. ill be back (eventually) | 17:05 |
@sonney2k | sploving, feel free... | 17:05 |
-!- serialhex01 [~androirc@99-101-149-136.lightspeed.wepbfl.sbcglobal.net] has left #shogun ["AndroIRC"] | 17:06 | |
@sonney2k | submit the ant (without absoute paths and changes you had to do for the package stuff) as git pull request... | 17:06 |
@sonney2k | sploving, anyway no idea why you worry so much about c# etc you managed to get some java code to work in 1 day too... yes not ideal yet but that will work sooner or later | 17:07 |
@sonney2k | and in any case if you get stuck it makes a lot of sense to discuss with mikio and me or if you need a break from java work on the language you know best first (that was lua right) | 17:08 |
sploving | sonney2k, I want to support the new language well/complemently, not half-work. If I took time to support other language, I will have no much time to support java/lua/ruby well | 17:14 |
sploving | sonnwy2k, another thing is idiom. every language has its idiom | 17:15 |
sploving | when wrap the c++ funtion, we should make it as idiomic as possible | 17:15 |
sploving | so i think quality is more important | 17:16 |
@sonney2k | sploving, as I said start with java and then we will go from there | 17:17 |
sploving | sonney2k, I have to go now | 17:20 |
sploving | how many days are you away from tommorrow? | 17:20 |
@sonney2k | about 2 months but I will try to be online as much as time permits | 17:21 |
@sonney2k | sploving, and mikio is available anywasy :) | 17:21 |
sploving | 2 months? so long? I just thought some weeks | 17:22 |
sploving | ok. I know. I will use the mailing list to ask question. | 17:22 |
sploving | bye | 17:22 |
@sonney2k | yes exaclty... | 17:22 |
@sonney2k | l8r | 17:22 |
-!- sploving [~root@124.16.139.196] has left #shogun [] | 17:23 | |
@sonney2k | hmmhh root@ ? not good | 17:23 |
@sonney2k | next time someone tell sploving that he should not irc as root | 17:23 |
-!- ameerkat [~ameerkat@184-98-140-155.phnx.qwest.net] has joined #shogun | 18:30 | |
-!- heiko [~heiko@infole-06.uni-duisburg.de] has quit [Quit: Leaving.] | 19:09 | |
-!- Netsplit *.net <-> *.split quits: @bettyboo | 19:11 | |
-!- Netsplit over, joins: @bettyboo | 19:13 | |
-!- blackburn [~qdrgsm@188.168.5.113] has joined #shogun | 20:05 | |
* sonney2k enjoys this nice quite evening | 21:09 | |
@sonney2k | quiet and nice actually :) | 21:11 |
-!- ameerkat [~ameerkat@184-98-140-155.phnx.qwest.net] has quit [Ping timeout: 252 seconds] | 22:26 | |
blackburn | sonney2k: yeah very quiet :) | 22:53 |
-!- f-x [b49531e3@gateway/web/freenode/ip.180.149.49.227] has joined #shogun | 22:58 | |
-!- f-x [b49531e3@gateway/web/freenode/ip.180.149.49.227] has quit [Ping timeout: 252 seconds] | 23:04 | |
@sonney2k | very very quiet :D | 23:05 |
* alesis-novik shouts | 23:05 | |
alesis-novik | better? | 23:05 |
@sonney2k | heh | 23:06 |
@sonney2k | awesome shouting experience | 23:07 |
alesis-novik | I know, right? | 23:14 |
alesis-novik | Going to discuss details of EM with Vojtech tomorrow | 23:15 |
blackburn | that mendeley is pretty good but its interface is intricated sometimes! | 23:16 |
blackburn | sonney2k: do you use mendeley? | 23:16 |
@sonney2k | what is mendeley? | 23:16 |
@sonney2k | ok google told me now | 23:17 |
blackburn | http://www.mendeley.com/ | 23:17 |
blackburn | yeah, it's a service for managing papers etc | 23:17 |
blackburn | chris suggested me to use it for managing papers I will use in my project | 23:18 |
alesis-novik | I generally use citeulike | 23:18 |
@sonney2k | I don't use any tools... | 23:18 |
alesis-novik | I find exporting to bibtex useful | 23:21 |
-!- blackburn [~qdrgsm@188.168.5.113] has left #shogun [] | 23:29 | |
--- Log closed Tue May 03 00:00:42 2011 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!