Open in new window / Try shogun cloud
--- Log opened Mon May 23 00:00:24 2011
blackburnhe, never heard about this measure00:02
serialhexyeah, i think it's a US-only measure, as it was created during prohibition (when we temporarily outlawed alc thinking it would make ppl better, and only ended up making big $$$ for gangsters and the like)00:05
serialhexit has something to do with how much alc is in a giant vat of moonshine (corn whisky)00:05
serialhexhmm...  sonney2k, is there a way to have to dirs in the (for instance) "gcc -c example_wrap.c -I/usr/local/lib/ruby/1.6/i686-linux" line??  cause for some fscking stupid reason config.h isnt where ruby.h expects it to be...00:08
serialhexmy idea is to basically find both of those files relative to the current ruby & then include them in that line00:09
serialhexnvm... i think i may have found the answer...00:13
blackburngoing to sleep some00:15
blackburnserialhex: see ya00:15
serialhexok, nite blackburn00:15
serialhex:D bettyboo!!00:15
@bettybooserialhex: x==y will cause theta/000:15
blackburnshe divided by zero00:16
blackburnget outta here :D00:16
blackburnbe right back in 12 hrs :D00:16
-!- blackburn [~qdrgsm@] has quit [Quit: Leaving.]00:16
-!- serialhex [] has quit [Read error: Connection reset by peer]07:58
-!- serialhex [] has joined #shogun08:12
-!- blackburn [~qdrgsm@] has joined #shogun08:59
blackburnlet the coding begins :D09:00
-!- hackered [~sabrina@] has joined #shogun09:19
-!- hackered [~sabrina@] has left #shogun []09:19
-!- heiko [] has joined #shogun09:21
-!- blackburn [~qdrgsm@] has quit [Ping timeout: 240 seconds]09:43
-!- heiko [] has quit [Quit: Leaving.]12:17
-!- ameerkat [] has quit [Ping timeout: 255 seconds]12:26
-!- ChanServ changed the topic of #shogun to: Shogun Machine Learning Toolbox | We participate in GSoC with 5 projects - let the games begin - all hail the code! | Channel logs
-!- blackburn [~qdrgsm@] has joined #shogun12:39
-!- blackburn1 [~qdrgsm@] has joined #shogun13:57
-!- blackburn [~qdrgsm@] has quit [Ping timeout: 260 seconds]13:57
blackburn1damn where is my mentor :)19:03
-!- serialhex [] has quit [Read error: Connection reset by peer]20:44
@sonney2kblackburn1, not that you don't know how to spent your time ;-)21:16
blackburn1sonney2k: what?21:51
-!- blackburn1 is now known as blackburn21:52
blackburndon't understand this offence :)21:55
@sonney2kblackburn, kudos definitely no offence!22:45
blackburnanyway we planned to have a short meeting tonight22:49
blackburndon't you think it would be *slow*?22:52
@sonney2kblackburn, if you don't do the ASSERT stuff it will be as fast as accessing memory directly22:53
@sonney2kahh and define it inline - can help...22:53
CIA-31shogun: Baozeng Ding master * r2341f5f / (2 files in 2 dirs): add ruby vector typemap support -
CIA-31shogun: Baozeng Ding master * r386b0cd / (2 files in 2 dirs): add ruby matrix typemap support -
CIA-31shogun: Baozeng Ding master * rba381a2 / (2 files): remove init_shogun/exit_shogun -
blackburnsonney2k: should we assert it?22:55
blackburnsome safety/speed tradeoff22:55
@sonney2kblackburn, I would suggest two functions - one that does safety checks and one that does not or alternatively have some define to disable /enable the checks22:56
@sonney2kI mean when developing it helps a lot to have error checks22:57
@sonney2klater on you don't want those if you excessively use these functions22:57
blackburnsonney2k: but if there would be check it could be slow too23:04
blackburnsonney2k: is there any way to exclude say get_(...) method from modular?23:05
blackburnI mean it is ok we would use it internally (if we know what we are doing) but it is bad for modular interfaces23:05
@sonney2kblackburn, yes %ignore23:06
blackburnsonney2k: do you like this way of doing it?23:07
@sonney2kblackburn, I don't see an alternative23:08
@sonney2kwe cannot sacrifice speed23:08
@sonney2knot that much23:08
blackburnsonney2k: yeap23:08
blackburnso, set() for safe and set_() for unsafe23:09
@sonney2kthe only alternative I see is to add some kind of block operations that then do all in one go and are fast23:09
@sonney2knot sure if we can always do this though23:09
@sonney2kI think python uses set and SET23:10
blackburnsonney2k: but it is ok if class have method set23:11
@sonney2kblackburn, yes - the only problem will be that java modular will create a set / get function too23:12
@sonney2kbut just check if I am right23:12
blackburnwhat is the problem with java modular?23:13
-!- epps [~epps@] has joined #shogun23:14
-!- epps [~epps@] has quit [Changing host]23:14
-!- epps [~epps@unaffiliated/epps] has joined #shogun23:14
@sonney2kblackburn, I think for public arrays swig generates a set / get function for every public variable23:15
blackburnah I see23:16
blackburnI forgot that SGMatrix is struct :D23:17
blackburn*well so why I am doing it*23:18
blackburnanyway matrix.matrix[i*matrix.num_rows+j] will work as good as get..23:18
blackburneh I'd rather work on mds :D23:19
@bettybooblackburn, funny23:19
blackburnnot so23:19
@sonney2kblackburn, well give it a try23:23
@sonney2kif it compiles with java_modular I would happily add this23:24
blackburnsonney2k: I renamed it to get_value anyway :)23:25
-!- ameerkat [] has joined #shogun23:26
@sonney2ka much better name anyway :)23:37
blackburnsth this way23:43
@sonney2kblackburn, i think it should be [i,j] -> [row,col]23:53
-!- epps [~epps@unaffiliated/epps] has quit [Quit: Leaving]23:59
--- Log closed Tue May 24 00:00:13 2011