--- Log opened Wed Nov 30 00:00:59 2011 | ||
blackburn | uh I fell asleep before :D | 08:29 |
---|---|---|
sonne|work | we should really add ignores | 08:49 |
sonne|work | as much as possible | 08:49 |
sonne|work | err morning blackburn | 08:50 |
blackburn | morning | 08:50 |
blackburn | sonne|work: ignores for what? | 08:50 |
sonne|work | unused stuff / non-useful stuff | 08:51 |
sonne|work | like free_feature_* | 08:51 |
blackburn | hmm sure | 08:51 |
sonne|work | I already started but I guess there are like 1000 times more | 08:51 |
-!- blackburn [~blackburn@188.168.2.74] has quit [Quit: Leaving.] | 09:21 | |
-!- sonne|work [~sonnenbu@194.78.35.195] has left #shogun [] | 09:44 | |
CIA-33 | shogun: Soeren Sonnenburg master * rafa4852 / (4 files): add option to get corresponding bias-thresholds together with PRC / ROC curve - http://git.io/VdnLFQ | 13:53 |
-!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has joined #shogun | 13:59 | |
CIA-33 | shogun: Soeren Sonnenburg master * r646ff20 / examples/undocumented/python_modular/evaluation_thresholds_modular.py : add example for thresholds - http://git.io/OYmhYw | 14:04 |
blackburn | sonney2k: what a trick: [fp,tp]=evaluator.get_ROC() | 14:05 |
-!- in3xes [~in3xes@180.149.49.230] has joined #shogun | 14:26 | |
-!- sonne|work [~sonnenbu@194.78.35.195] has joined #shogun | 15:07 | |
blackburn | sonne|work: our paper with Chris is on the way to be finished in max 2 weeks I hope | 15:36 |
sonne|work | finally... | 15:36 |
blackburn | sonne|work: well I was really stucked with a lot of tasks | 15:36 |
blackburn | and now Chris helped to write some text | 15:37 |
blackburn | sonne|work: and recall I did a lot of programming things :) | 15:37 |
sonne|work | you mean you do a lot of things? | 15:37 |
blackburn | sonne|work: well.. there are a lot of code I have done? | 15:41 |
blackburn | and I have some pretty examples | 15:41 |
blackburn | later I will integrate some of them with dataset downloaders I guess | 15:42 |
sonne|work | data set downloaders? | 16:28 |
sonne|work | you mean contributions to shogun? | 16:28 |
sonne|work | I guess I still don't understand | 16:28 |
sonne|work | in any case I think we are going twice as fast as before with shogun development - but we need a new release ASAP now :) | 16:29 |
blackburn | sonne|work: my examples use data I can't integrate into shogun | 16:29 |
blackburn | sonne|work: we've got strange misunderstanding :) | 16:30 |
blackburn | sonne|work: I thought just after GSoC 'race' started shogun is on 10x faster development :) | 16:31 |
sonne|work | yeah but you/me were doing much more back then not to forget all the others :) | 16:31 |
sonne|work | blackburn: btw I added the threshold thing to evalutions | 16:32 |
sonne|work | one can now directly fetch the threshold needed to e.g. get a false positive rate of 1% | 16:33 |
sonne|work | very convenient | 16:33 |
blackburn | sonne|work: I do not understand how to use it but I have seen you did it | 16:33 |
sonne|work | just evaluator.get_thresholds() | 16:34 |
sonne|work | then you have for every tp/fp pair the coresponding threshold | 16:34 |
blackburn | sonne|work: what is threshold? | 16:34 |
sonne|work | the bias | 16:35 |
sonne|work | f(x)>threshold means predict +1 | 16:35 |
sonne|work | -1 otherwise | 16:35 |
blackburn | ah | 16:35 |
blackburn | I see | 16:35 |
blackburn | you should tell me everything in that way while I am such a dumbass | 16:36 |
blackburn | :D | 16:36 |
sonne|work | I don't think you are | 16:37 |
blackburn | believe me :) | 16:37 |
sonne|work | trust me ;) | 16:37 |
blackburn | sonne|work: are you still satisfied with java programming? | 16:38 |
sonne|work | for data analysis it is not very convenient - I am orders of magnitude faster in python | 16:42 |
sonne|work | but otherwise it is very convenient with eclipse at least | 16:43 |
blackburn | I'm thinking to change job in next year and choose one without java :D | 16:45 |
sonne|work | well at least no memory leaks - but I don't like ThisLongFunctionNamesHelpTotallyInUnderstandingWhatever | 16:45 |
blackburn | sonne|work: EE is way too crappy.. | 16:47 |
blackburn | SQL + XML + Java code mixed up | 16:47 |
sonne|work | I am not that deep into these things | 16:51 |
blackburn | sonne|work: I am happy for you and you should be too :) | 16:56 |
blackburn | 'jedem das seine' and I think I would like to do something different, like you do :) | 16:57 |
blackburn | hmm btw is it a negative phrase for you? | 16:58 |
blackburn | oh! it's time to go to gym again | 17:05 |
blackburn | see you | 17:05 |
-!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has quit [Quit: Page closed] | 17:05 | |
-!- in3xes [~in3xes@180.149.49.230] has quit [Quit: Leaving] | 17:45 | |
-!- blackburn [~blackburn@188.168.2.110] has joined #shogun | 19:04 | |
CIA-33 | shogun: Soeren Sonnenburg master * r1760090 / src/shogun/kernel/MultitaskKernelTreeNormalizer.h : coding style fixes - http://git.io/xfifiA | 20:42 |
CIA-33 | shogun: Soeren Sonnenburg master * r053b232 / src/interfaces/python_modular/swig_typemaps.i : add get_sparse_feature_vector support in python typemaps - http://git.io/nXkZHw | 20:42 |
CIA-33 | shogun: Soeren Sonnenburg master * rd47946c / src/interfaces/modular/Library.i : define sparse feature vector template - http://git.io/JljxVw | 20:42 |
CIA-33 | shogun: Soeren Sonnenburg master * r44581da / src/interfaces/modular/modshogun_ignores.i : ignore a number of irrelevant functions - http://git.io/NaeQ0Q | 20:42 |
CIA-33 | shogun: Soeren Sonnenburg master * r23d1853 / src/interfaces/modular/modshogun_ignores.i : fix syntax error in ignores - http://git.io/3PYvCg | 20:49 |
shogun-buildbot | build #70 of nightly_all is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_all/builds/70 | 20:49 |
shogun-buildbot | build #343 of octave_static is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/octave_static/builds/343 blamelist: sonne@debian.org | 21:13 |
blackburn | sonney2k: how significantly did it speedup compilation? | 21:21 |
@sonney2k | blackburn, not significantly but it will remove potentially dangerous/unneeded functions | 21:59 |
blackburn | sure, I started these things before | 22:00 |
@sonney2k | feel free to continue | 22:06 |
blackburn | I don't mind if I was jobless :D | 22:11 |
blackburn | I have pretty hard next two days.. | 22:11 |
@sonney2k | blackburn, how can I get you fired ;-) | 22:33 |
blackburn | sonney2k: well unfortunately I have no time not because of job | 22:34 |
blackburn | but because of studies | 22:34 |
@sonney2k | then it is better to not fire you | 22:34 |
@sonney2k | better do your homework | 22:34 |
blackburn | sonney2k: I have no homework, i'm too old for that :D | 22:35 |
blackburn | sonney2k: I have one crazy semestral job on numerical methods in mathematical physics | 22:35 |
blackburn | that takes a lot of time but will be finished next week | 22:35 |
-!- blackburn [~blackburn@188.168.2.110] has quit [Quit: Leaving.] | 23:46 | |
--- Log closed Thu Dec 01 00:00:59 2011 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!