--- Log opened Sat May 05 00:00:37 2012 | ||
-!- n4nd0 [02893bbe@gateway/web/freenode/ip.2.137.59.190] has quit [Quit: Page closed] | 00:07 | |
-!- cronor_ [~cronor@e177095229.adsl.alicedsl.de] has joined #shogun | 00:52 | |
-!- cronor [~cronor@e177095229.adsl.alicedsl.de] has quit [Read error: Connection reset by peer] | 00:52 | |
-!- cronor_ is now known as cronor | 00:52 | |
-!- cronor [~cronor@e177095229.adsl.alicedsl.de] has quit [Client Quit] | 00:57 | |
-!- PhilTillet [~Philippe@npasserelle10.minet.net] has joined #shogun | 01:15 | |
-!- PhilTillet [~Philippe@npasserelle10.minet.net] has quit [Remote host closed the connection] | 01:19 | |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 03:07 | |
-!- wiking [~wiking@78-23-189-112.access.telenet.be] has joined #shogun | 03:13 | |
-!- wiking [~wiking@78-23-189-112.access.telenet.be] has quit [Changing host] | 03:13 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 03:13 | |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Client Quit] | 03:15 | |
@sonney2k | blackburn I hope you will see a big one from me tonight too... I changed sooo much SGVector related... | 06:04 |
---|---|---|
-!- gsomix [~gsomix@188.168.5.6] has quit [Quit: Ex-Chat] | 08:11 | |
-!- gsomix [~gsomix@188.168.5.6] has joined #shogun | 08:11 | |
-!- gsomix [~gsomix@188.168.5.6] has quit [Remote host closed the connection] | 08:21 | |
-!- vikram360 [~vikram360@117.192.190.158] has quit [Ping timeout: 244 seconds] | 08:47 | |
-!- vikram360 [~vikram360@117.192.190.158] has joined #shogun | 09:39 | |
-!- blackburn [~qdrgsm@188.168.2.65] has joined #shogun | 09:55 | |
-!- wiking [~wiking@78-23-189-112.access.telenet.be] has joined #shogun | 10:00 | |
-!- wiking [~wiking@78-23-189-112.access.telenet.be] has quit [Changing host] | 10:00 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 10:00 | |
blackburn | wiking: hey | 10:15 |
blackburn | here? | 10:15 |
wiking | yes yes | 10:16 |
-!- vikram360 [~vikram360@117.192.190.158] has quit [Ping timeout: 252 seconds] | 10:39 | |
blackburn | wiking: could you send me clef data somehow? | 11:04 |
wiking | blackburn: ah yeah | 11:04 |
wiking | i have to redownload as they changed something the other day | 11:05 |
blackburn | wiking: they require some shity papers to sign :D | 11:05 |
wiking | but i'll make it available for you | 11:05 |
wiking | yeye no worries | 11:05 |
blackburn | I guess you did it already | 11:05 |
wiking | yes | 11:05 |
blackburn | do you use dropbox? | 11:05 |
wiking | mmm my drop is full :))) | 11:06 |
wiking | and i really don't want to buy | 11:06 |
wiking | i'll put it on the web for ya | 11:06 |
-!- vikram360 [~vikram360@117.192.190.158] has joined #shogun | 11:06 | |
blackburn | wiking: ok thanks | 11:06 |
blackburn | wiking: bad I can't expand your dropbox with shared folder :( | 11:07 |
wiking | eheheh noup | 11:07 |
blackburn | would be nice hehe | 11:08 |
wiking | blackburn: i've made some space on it | 11:09 |
wiking | what's your db account ? | 11:10 |
blackburn | wiking: how large is data btw? | 11:10 |
wiking | it's around 65mb | 11:10 |
wiking | the training set | 11:10 |
blackburn | heh | 11:10 |
blackburn | small! | 11:10 |
blackburn | blackburn91@gmail.com | 11:10 |
wiking | yep it's just 1000 pics | 11:10 |
wiking | blackburn: invitation sent... but it's still uploading... | 11:12 |
blackburn | wiking: thanks! | 11:12 |
wiking | nw | 11:12 |
wiking | can i ask maybe actually a favour | 11:12 |
wiking | HOG the pictures? | 11:12 |
wiking | i mean generate hog features of it? | 11:13 |
blackburn | wiking: do you mean my favor | 11:16 |
blackburn | ? | 11:16 |
wiking | yes | 11:16 |
blackburn | let me see pictures first ;) | 11:16 |
wiking | u don't want to :p | 11:20 |
wiking | some of them are really scary :) | 11:21 |
blackburn | wiking: damn they are really varying | 11:22 |
blackburn | wiking: I don't think HOGs can help out there.. | 11:23 |
wiking | yeah i was guessing so | 11:23 |
wiking | but it's good to have it | 11:23 |
wiking | as well | 11:24 |
blackburn | some classes are really really crazy | 11:24 |
blackburn | wiking: did you use MKL on top of that?? | 11:25 |
wiking | nono | 11:25 |
wiking | i used BoW | 11:25 |
wiking | with sift | 11:25 |
blackburn | yeah BoW is the only thing that could work here | 11:26 |
blackburn | wiking: and it was something like of 90% accuracy? | 11:26 |
wiking | not with this | 11:26 |
wiking | :))) | 11:26 |
wiking | last year | 11:26 |
wiking | this year this is fucking crazy | 11:26 |
blackburn | oh shit I opened DVOR class | 11:26 |
wiking | hahahahah | 11:26 |
wiking | doooont! | 11:27 |
blackburn | DFFE#$@#$ | 11:27 |
blackburn | WTF | 11:27 |
wiking | not interested :) | 11:27 |
blackburn | oh damn | 11:27 |
blackburn | shit | 11:27 |
blackburn | wtf | 11:27 |
blackburn | FUUUUUUUU | 11:27 |
wiking | told you | 11:28 |
blackburn | I will not sleep anymore | 11:28 |
blackburn | DVDM is nice too.. | 11:28 |
blackburn | wiking: I guess this competition is for ones who can to not vomit whole day long seeing this pictures | 11:29 |
wiking | :))) | 11:29 |
blackburn | wiking: text is to be extracted somehow there.. | 11:29 |
blackburn | wiking: fmri is much more clean and not vomitting :D | 11:30 |
wiking | heheheh | 11:30 |
CIA-113 | shogun: Sergey Lisitsyn master * re6525c7 / src/shogun/machine/KernelMachine.cpp : Workaround for kernel machine segfault - http://git.io/4aImSg | 12:02 |
-!- blackburn [~qdrgsm@188.168.2.65] has quit [Read error: Operation timed out] | 13:00 | |
-!- vikram360 [~vikram360@117.192.190.158] has quit [Ping timeout: 252 seconds] | 16:56 | |
-!- vikram360 [~vikram360@117.192.190.221] has joined #shogun | 16:57 | |
-!- karlnapf [~heiko@host86-176-1-187.range86-176.btcentralplus.com] has joined #shogun | 17:38 | |
karlnapf | anybody here? | 17:41 |
karlnapf | my current git doesnt compile, some hash -set related problems | 17:44 |
karlnapf | sonney2k, do you have a clue where that might come from | 17:47 |
-!- blackburn [5bde8018@gateway/web/freenode/ip.91.222.128.24] has joined #shogun | 17:48 | |
karlnapf | blackburn, hi | 17:51 |
karlnapf | just cloned current git, does not compile | 17:51 |
blackburn | karlnapf: hey | 17:51 |
blackburn | karlnapf: really ? | 17:51 |
karlnapf | yes | 17:51 |
karlnapf | my options are | 17:51 |
blackburn | feel free to fix ;) | 17:52 |
karlnapf | --enable-debug --disable-optimization --enable-trace ... | 17:52 |
karlnapf | just asking whether you knew about that | 17:52 |
blackburn | compiles here hmm | 17:52 |
karlnapf | which switches are you using? | 17:52 |
blackburn | karlnapf: opt one only | 17:53 |
blackburn | let me check with yours | 17:53 |
blackburn | karlnapf: python_modular? | 17:53 |
karlnapf | yes | 17:53 |
karlnapf | but fails before the module | 17:53 |
karlnapf | works without any switches | 17:54 |
blackburn | karlnapf: uh right | 17:55 |
blackburn | trace mallocs is broken | 17:55 |
karlnapf | is the problem obvious to you? if yes I wont get into it :) ... | 17:57 |
karlnapf | if no, I will fix it | 17:57 |
blackburn | karlnapf: not obvious but I'll try to take care | 17:57 |
karlnapf | ok then :) | 17:58 |
karlnapf | Will fix my own bugs instead then :) | 17:58 |
blackburn | karlnapf: strange bug actually | 17:59 |
karlnapf | blackburn, I was wondering whether we should add an assert that zero blocks are allocated in the exit_shogun() method | 17:59 |
karlnapf | because some examples contain memleaks | 17:59 |
karlnapf | but they are hidden | 17:59 |
blackburn | karlnapf: hmmm I don't mind it | 17:59 |
karlnapf | if we would enforce not leaking if trace-mallocs is activated | 17:59 |
karlnapf | that would automatically find all of them in the libshogun examples | 18:00 |
blackburn | karlnapf: make check-valgrind does the job btw | 18:00 |
blackburn | in libshogun examples | 18:00 |
karlnapf | oh, really? | 18:00 |
karlnapf | ok did not know that | 18:00 |
karlnapf | :) | 18:01 |
blackburn | karlnapf: but not trace mallocs | 18:01 |
blackburn | runs valgrind on each example | 18:01 |
karlnapf | but valgrind is better | 18:01 |
karlnapf | although slower :) | 18:01 |
blackburn | and outputs error count | 18:01 |
blackburn | I have added it some day | 18:01 |
karlnapf | blackburn, how does the ref-counting for vectors work? | 18:02 |
karlnapf | do I call ref method | 18:02 |
karlnapf | or do I use SG_REF? | 18:02 |
blackburn | karlnapf: hmmm actually it relies on magic now | 18:03 |
karlnapf | meaning? | 18:03 |
karlnapf | I dont have to do stuff? | 18:03 |
blackburn | karlnapf: copy constructor and constructor increases ref | 18:03 |
karlnapf | and unref? | 18:03 |
blackburn | destructor unrefs | 18:03 |
karlnapf | ah | 18:03 |
karlnapf | nice thing | 18:03 |
karlnapf | will play around with that | 18:03 |
blackburn | karlnapf: e.g. if you have function that has SGVector as parameter | 18:04 |
blackburn | func (SGVector vec) | 18:04 |
karlnapf | yes get it | 18:04 |
blackburn | not const & | 18:04 |
karlnapf | what does const& do? | 18:04 |
blackburn | it calls copy constructor and increases ref for local scope | 18:04 |
karlnapf | oh its a reference | 18:04 |
karlnapf | kk | 18:04 |
blackburn | not "const SGVector& vec" | 18:04 |
blackburn | karlnapf: sounds pretty cool actually | 18:05 |
karlnapf | yes indeed | 18:06 |
karlnapf | Will try to incorparate that into the subset stuff | 18:06 |
karlnapf | currently, everything is copied | 18:06 |
karlnapf | that sucks | 18:06 |
blackburn | karlnapf: ohoh | 18:07 |
karlnapf | whats the matter? | 18:08 |
blackburn | karlnapf: I think we will die in that hell.. :D | 18:08 |
karlnapf | ? | 18:08 |
blackburn | karlnapf: painful transition | 18:08 |
karlnapf | with all the vectors? | 18:08 |
blackburn | yeah | 18:09 |
karlnapf | I think its less bad than last year :) | 18:09 |
karlnapf | although we could have done that properly from the beginning :) | 18:09 |
karlnapf | but thats how it works ;) | 18:09 |
CIA-113 | shogun: Heiko Strathmann master * r04e2696 / src/shogun/mathematics/Math.h : added copyright and a permute_vector method - http://git.io/X7Gzzg | 18:10 |
CIA-113 | shogun: Heiko Strathmann master * r49cccd0 / (7 files): initial sketch for new statistical testing framework - http://git.io/aVptHg | 18:10 |
CIA-113 | shogun: Heiko Strathmann master * r20e72f4 / src/shogun/statistics/LinearTimeMMD.cpp : new SGVector system changes - http://git.io/e4nMlA | 18:10 |
CIA-113 | shogun: Heiko Strathmann master * rf432068 / (8 files in 2 dirs): | 18:10 |
CIA-113 | shogun: Merge pull request #494 from karlnapf/master | 18:10 |
CIA-113 | shogun: Suggestion for a statistical testing framework - http://git.io/s8tP-w | 18:10 |
blackburn | karlnapf: that bug is really really strange.. | 18:12 |
blackburn | would be strange if it caused by inline | 18:13 |
-!- gsomix [~gsomix@85.26.234.149] has joined #shogun | 18:45 | |
-!- blackburn [5bde8018@gateway/web/freenode/ip.91.222.128.24] has quit [Ping timeout: 245 seconds] | 18:49 | |
gsomix | hi all | 18:49 |
gsomix | karlnapf, hey, what's wrong with hashset? | 18:53 |
gsomix | karlnapf, ah, okay. =___= | 18:56 |
karlnapf | gsomix ,hi :) | 18:58 |
karlnapf | you got it :) | 18:58 |
karlnapf | gsomix, are all the memory errors in make test caused by the SGVector transition? | 19:10 |
gsomix | karlnapf, I don't know really. I'm checking now for my bugs. | 19:14 |
@sonney2k | karlnapf, git clean -dfx | 19:21 |
karlnapf | sonney2k, thanks I keep forgetting that | 19:21 |
@sonney2k | gsomix, nothing wrong with it | 19:21 |
karlnapf | sonney2k, about this SGVector<char> thing | 19:21 |
@sonney2k | people always forget to do it when things change | 19:21 |
karlnapf | I will replace it via a non-sgvector alternative | 19:22 |
karlnapf | otherwise I will have to dig into generich classes which is messy | 19:22 |
gsomix | sonney2k, huh? I found some bugs. | 19:22 |
@sonney2k | gsomix, which? | 19:22 |
@sonney2k | karlnapf, can you be more concrete? | 19:23 |
@sonney2k | void ptr and length or what? | 19:23 |
karlnapf | sonney2k, yes | 19:23 |
karlnapf | and actually | 19:23 |
karlnapf | not registering parameters | 19:23 |
karlnapf | for these parameter trees its not necessary anyway | 19:23 |
@sonney2k | well ok | 19:24 |
gsomix | sonney2k, wait a minute, please | 19:24 |
karlnapf | perhaps override the serialize method | 19:24 |
* sonney2k continues with some intrusive changes all over the place | 19:24 | |
karlnapf | sonney2k, even fater git clean -dfx I get all these errors | 19:41 |
karlnapf | I made make distclean and the git clean | 19:42 |
karlnapf | just valgrinded, its SGVector related, will check whether configure switches change something | 19:44 |
@sonney2k | karlnapf, which error do you get? | 19:46 |
karlnapf | sonney2k, let me check it a bit more, give me a moment | 19:47 |
karlnapf | currently this spooky one: | 19:47 |
karlnapf | vex amd64->IR: unhandled instruction bytes: 0xC5 0xF9 0x57 0xC0 0xE8 0x63 | 19:47 |
karlnapf | 64 bit stuff perhaps? | 19:47 |
CIA-113 | shogun: Soeren Sonnenburg master * r1f419bb / examples/undocumented/libshogun/statistics.cpp : slightly simplify example - http://git.io/6Kh4zw | 19:49 |
CIA-113 | shogun: Soeren Sonnenburg master * reaf5c09 / (17 files in 8 dirs): | 19:49 |
CIA-113 | shogun: continued sgvector transition | 19:49 |
CIA-113 | shogun: - convert linear machines to use SGVector w instead of double ptr | 19:49 |
CIA-113 | shogun: - remove various unneedef SG_FREEs - http://git.io/Rn4Ksw | 19:49 |
@sonney2k | karlnapf, when you run things with valgrind always use --enable-debug --disable-optimization | 19:49 |
@sonney2k | karlnapf, but yes you will still see several crashers due to sgvector | 19:50 |
karlnapf | oh yes of course thats this error | 19:50 |
karlnapf | but this wasnt the one I meant | 19:51 |
karlnapf | ah ok | 19:51 |
karlnapf | alright then | 19:51 |
@sonney2k | karlnapf, and it will get worse | 19:52 |
@sonney2k | once sgvector is done | 19:52 |
@sonney2k | sgmatrix will follow | 19:52 |
karlnapf | ohoh | 19:52 |
@sonney2k | sgsparse amtrix | 19:52 |
@sonney2k | matrix | 19:52 |
@sonney2k | sgstringlist | 19:52 |
@sonney2k | but one by one | 19:53 |
karlnapf | mh, I should wait with any changes then | 19:53 |
@sonney2k | and sgvector is most used so it is not that bad | 19:53 |
karlnapf | almost half of the tests are failing here | 19:53 |
@sonney2k | which tests? | 19:53 |
karlnapf | libshogun | 19:53 |
@sonney2k | huh? | 19:53 |
karlnapf | in a freshly cloned repo | 19:53 |
@sonney2k | only one fails here | 19:53 |
@sonney2k | and I am currently fixing it | 19:53 |
karlnapf | classifier_mklmulticlass, classifier_qda, classifier_multiclasslinearmachine, and more | 19:55 |
karlnapf | gotta go now, will be back in 30 mins | 19:55 |
@sonney2k | k | 19:56 |
gsomix | sonney2k, segfaults with --enable-trace-mallocs. in hashset. | 19:56 |
@sonney2k | gsomix, heh | 19:57 |
@sonney2k | could be endlessly looping | 19:57 |
gsomix | sonney2k, what do you mean? | 20:04 |
gsomix | ah, I got it | 20:05 |
@sonney2k | gsomix, you have to check how it was done before to avoid this | 20:07 |
@sonney2k | karlnapf, in CKernelMachine - will indices work already? | 20:13 |
@sonney2k | and what is params.indices=SGVector<index_t>(0) in apply() supposed to do | 20:13 |
@sonney2k | ? | 20:13 |
karlnapf | sonney2k, let me have a look | 20:14 |
karlnapf | which line which file? | 20:15 |
@sonney2k | shogun/machine/KernelMachine.cpp | 20:15 |
karlnapf | k got it | 20:15 |
@sonney2k | 304 | 20:15 |
@sonney2k | (it is disabled right now) | 20:15 |
karlnapf | initialise vector with 0 | 20:16 |
karlnapf | the 0 probably sneaked in | 20:16 |
karlnapf | remove it | 20:16 |
karlnapf | but the vector has to be NULL | 20:16 |
karlnapf | sorry not 0, but NULL | 20:16 |
karlnapf | so SGVector<index_t>() will do it | 20:16 |
@sonney2k | karlnapf, so removing it will work as good | 20:17 |
@sonney2k | (default constructor does NULL already) | 20:17 |
karlnapf | k | 20:17 |
karlnapf | yes true | 20:17 |
karlnapf | sonney2k, any clue why I could have all these memory errors? | 20:18 |
karlnapf | and you dont? | 20:18 |
CIA-113 | shogun: Soeren Sonnenburg master * r052e3c3 / src/shogun/classifier/mkl/MKLMulticlass.cpp : fix mkl multiclass sgvector bug - http://git.io/TwCp8w | 20:19 |
CIA-113 | shogun: Soeren Sonnenburg master * r39d9382 / src/shogun/machine/KernelMachine.cpp : remove unneeded initialization in KernelMachine - http://git.io/ub_zUg | 20:19 |
@sonney2k | karlnapf, old shogun? | 20:19 |
@sonney2k | karlnapf, btw I have one more question | 20:19 |
@sonney2k | I am getting a new error in training error: 1.005422 | 20:21 |
@sonney2k | cross_validation estimate: | 20:21 |
@sonney2k | [2449.555420,2462.500676] with alpha=0.050000, mean=2456.028048 | 20:21 |
@sonney2k | [ERROR] assertion result.mean<2 failed in file evaluation_cross_validation_regression.cpp line 103 | 20:21 |
@sonney2k | but no memory errors or anything | 20:22 |
karlnapf | mmh | 20:22 |
@sonney2k | valgrind errors I mean | 20:22 |
karlnapf | these values sound like uninitialised memory | 20:22 |
karlnapf | the assertion is there to ensure that the value is small | 20:22 |
karlnapf | so valgrind errors? | 20:22 |
karlnapf | strange | 20:22 |
@sonney2k | karlnapf, so that value should be around what you get for training error? | 20:23 |
karlnapf | I mean the estimate is obviously wrong | 20:23 |
karlnapf | no, but the data is linear seperable, | 20:24 |
karlnapf | let me have a look at the example | 20:24 |
karlnapf | oh no, not lin. sep. | 20:24 |
@sonney2k | in this transition I am doing a few intrusive changes that can change anything | 20:25 |
@sonney2k | one is that get_labels() etc never copies | 20:25 |
@sonney2k | and same with CAlphabet | 20:25 |
@sonney2k | (we used to always copy there) | 20:25 |
karlnapf | that might be a problem for cross-validation ... | 20:26 |
@sonney2k | so if this is assumed somehwere then -> kaboom | 20:26 |
karlnapf | perhaps something with the subsets | 20:26 |
@sonney2k | but you call get_labels_copy | 20:26 |
karlnapf | because get_labels with a subset HAS to copy | 20:26 |
karlnapf | a | 20:26 |
karlnapf | mmh | 20:26 |
karlnapf | that is probably the reason for the large x-val error and to mem-error: wrong labels | 20:27 |
@sonney2k | also get_int_labels will copy | 20:30 |
karlnapf | oh there is at least one bug | 20:31 |
@sonney2k | yes? | 20:31 |
karlnapf | get_int_labels | 20:32 |
karlnapf | oh wait | 20:32 |
karlnapf | I might be wrong | 20:32 |
karlnapf | no sorry | 20:33 |
karlnapf | wrong alert | 20:33 |
karlnapf | did the error occur when you changed that get_labels thing? or before? | 20:34 |
@sonney2k | no idea | 20:34 |
@sonney2k | I guess after change | 20:35 |
@sonney2k | but still I this is more tricky | 20:35 |
@sonney2k | I guess I will continue to try to fix the errors I see in the more elaborate python_modular exmaples | 20:35 |
@sonney2k | maybe it is a side effect | 20:35 |
karlnapf | yes, you could remove the assertion for now | 20:36 |
@sonney2k | well not remove - I will just work on python_modular and revisit it later | 20:36 |
karlnapf | I guess its something with the splitting strategy producing wrong labels | 20:36 |
karlnapf | ok | 20:36 |
karlnapf | I will try to complete the ModelSelectionparameters thing now | 20:37 |
@sonney2k | k | 20:40 |
gsomix | sonney2k, fixed | 20:43 |
CIA-113 | shogun: Soeren Sonnenburg master * r3ceca27 / (5 files in 2 dirs): polish get_labels_copy() and drop load/save from CLabels - http://git.io/YRIo2A | 20:44 |
@sonney2k | gsomix, where? | 20:44 |
gsomix | sonney2k, in CSet. | 20:45 |
gsomix | just added use_sg_mallocs flag. | 20:45 |
gsomix | I'm testing now... with --enable-trace-mallocs now. =___= | 20:46 |
karlnapf | sonney2k, gsomix, there are some memory leaks in the model-selection stuff | 20:46 |
karlnapf | modelselection_grid_search_kernel.cpp for example | 20:46 |
karlnapf | mmh, kind of hard to fix as long as there is no trace mallocs, Ill check in later when this works and try to find out whats going on | 20:48 |
karlnapf | Cannot fix the other thing because I dont know whether I cause mem-leaks | 20:48 |
-!- blackburn [~qdrgsm@83.234.54.222] has joined #shogun | 20:49 | |
karlnapf | is the trace-mallocs working again? | 20:49 |
@sonney2k | karlnapf, gsomix found the issue - so very soon | 20:49 |
blackburn | sorry I get a little bit ill | 20:49 |
gsomix | blackburn, grustno | 20:50 |
@sonney2k | blackburn, working too hard... | 20:50 |
blackburn | sonney2k: hardly work rather hehe | 20:50 |
gsomix | blackburn, I also got ill. | 20:51 |
blackburn | ok according to logs gsomix you will take care of it? | 20:51 |
@sonney2k | basically all multiclass stuff crashes | 20:51 |
@sonney2k | gsomix - once this is fixed - please remove all the load() and save() crap ... seems to be in many many classes not only CMachine derived ones... | 20:51 |
blackburn | sonney2k: I know the reason probably | 20:52 |
@sonney2k | wow - quite many python examples work already | 20:52 |
blackburn | sonney2k: issue is mainly related with linearmachine's w | 20:52 |
@sonney2k | gtg | 20:52 |
gsomix | sonney2k, okay. Added to my todo list. :) | 20:52 |
@sonney2k | will be back in ~20 min | 20:52 |
gsomix | karlnapf, ping | 20:58 |
karlnapf | gsomix, hi | 20:59 |
karlnapf | got it? | 20:59 |
gsomix | karlnapf, yep. can you review and merge it? | 20:59 |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 20:59 | |
blackburn | wow I never expected | 21:00 |
karlnapf | will do | 21:00 |
karlnapf | gsomix, does it compile with trace-mallocs without adding more errors in tests? | 21:01 |
blackburn | karlnapf: set is used only in memory blocks counting ;) | 21:02 |
gsomix | karlnapf, yep, it seems. | 21:02 |
karlnapf | patch looks nice, I dont know what you talked about with soeren about this default constructor, if you agreed on that I iwll merge | 21:03 |
gsomix | karlnapf, just a second | 21:09 |
karlnapf | k | 21:11 |
gsomix | karlnapf, done | 21:20 |
karlnapf | gsomix, did you yet push? | 21:20 |
gsomix | karlnapf, yep | 21:22 |
karlnapf | k | 21:22 |
CIA-113 | shogun: Evgeniy Andreev master * r3fc0b13 / (5 files in 3 dirs): fixes for using non-shogun mallocs/callocs - http://git.io/M3ua6Q | 21:22 |
CIA-113 | shogun: Evgeniy Andreev master * r1786947 / (src/shogun/lib/memory.cpp src/shogun/lib/memory.h): default constructor is not needed - http://git.io/5apqVQ | 21:22 |
CIA-113 | shogun: Heiko Strathmann master * r64d63bd / (4 files in 3 dirs): | 21:22 |
CIA-113 | shogun: Merge pull request #498 from gsomix/CSet | 21:22 |
CIA-113 | shogun: Fixes in CSet - http://git.io/vmS7qQ | 21:22 |
gsomix | karlnapf, check it. I need a feedback. :) | 21:22 |
karlnapf | compiling ... | 21:23 |
blackburn | gsomix: what are your ongoing tasks? | 21:23 |
gsomix | blackburn, hmm, Arrays -> DynArray. | 21:23 |
gsomix | map in titled features | 21:23 |
gsomix | and load/save deletions | 21:24 |
blackburn | ok no need for other tasks then | 21:24 |
-!- vikram360 [~vikram360@117.192.190.221] has quit [Read error: Connection timed out] | 21:27 | |
-!- vikram360 [~vikram360@117.192.190.221] has joined #shogun | 21:27 | |
blackburn | karlnapf: | 21:28 |
blackburn | ==17052== Invalid read of size 8 | 21:28 |
blackburn | ==17052== at 0x4C2A327: memcpy (mc_replace_strmem.c:635) | 21:28 |
blackburn | ==17052== by 0x53088AF: shogun::CSubset::CSubset(shogun::SGVector<int>) (Subset.cpp:29) | 21:28 |
blackburn | ==17052== by 0x52DC6C5: shogun::CSubsetStack::add_subset(shogun::SGVector<int>) (SubsetStack.cpp:105) | 21:28 |
blackburn | ==17052== by 0x52959FD: shogun::CLabels::add_subset(shogun::SGVector<int>) (Labels.cpp:257) | 21:28 |
blackburn | something crazy is going here heh | 21:28 |
karlnapf | gsomix, ehm lots of errors with trace-mallocs | 21:28 |
karlnapf | in the libshogun tests | 21:28 |
karlnapf | without trace-mallocs runs ok | 21:29 |
gsomix | karlnapf, what tests? | 21:29 |
gsomix | *which | 21:30 |
karlnapf | classifier_libsvm is the first one | 21:31 |
karlnapf | classifier_minimal_svm | 21:31 |
karlnapf | gaussian_naive_bayes does not terminate | 21:31 |
gsomix | karlnapf, okay, thanks. | 21:32 |
blackburn | for brave souls who want to solve multiclass machines - the error is in copy constructor of linear machine | 21:32 |
karlnapf | gsomix, do the tests run for you? | 21:33 |
blackburn | it is being double freed for some reason | 21:33 |
blackburn | ah and also subsets are b0rken | 21:33 |
karlnapf | blackburn, yes these are probably because of SGVector stuff | 21:33 |
karlnapf | I will review that once the other stuff works | 21:33 |
blackburn | karlnapf: exactly because of it :) | 21:34 |
blackburn | I just told what I have found out while checking valgrind | 21:34 |
blackburn | have to rest now - my cold is too annoying to fix code right now | 21:34 |
karlnapf | hah blackburn, same here | 21:35 |
blackburn | karlnapf: are you cold too? | 21:35 |
karlnapf | flew | 21:35 |
karlnapf | since almost two weeks now, annoying | 21:35 |
blackburn | ah I hope you get rid of it soon | 21:36 |
blackburn | however after I saw DVOR category of medical CLEF I can't complain anymore | 21:36 |
blackburn | :D | 21:36 |
-!- vikram360 [~vikram360@117.192.190.221] has quit [Read error: Operation timed out] | 21:37 | |
-!- vikram360 [~vikram360@117.192.190.221] has joined #shogun | 21:37 | |
karlnapf | ehm what? | 21:37 |
blackburn | karlnapf: in the medical CLEF benchmark there are a lot of nice photos | 21:38 |
blackburn | I lost my sleep today probably seeing it :D | 21:38 |
karlnapf | I guess I dont wanna see that | 21:39 |
blackburn | karlnapf: yeah that's why I didn't send it to you ;) | 21:39 |
blackburn | ok see you tomorrow | 21:39 |
karlnapf | k blackburn, take care | 21:39 |
-!- blackburn [~qdrgsm@83.234.54.222] has quit [Quit: Leaving.] | 21:40 | |
karlnapf | gsomix, tell me when do sent another path ;) | 21:41 |
karlnapf | will do some cooking now | 21:41 |
gsomix | karlnapf, okay. | 21:41 |
gsomix | it seems, I need asleep a little. | 21:41 |
-!- wiking [~wiking@78-23-189-112.access.telenet.be] has joined #shogun | 21:41 | |
-!- wiking [~wiking@78-23-189-112.access.telenet.be] has quit [Changing host] | 21:41 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 21:41 | |
gsomix | T___t | 21:41 |
karlnapf | gsomix, dont forget to take a break from time to time ;) | 21:42 |
gsomix | good night, guys | 21:49 |
-!- Marty28 [~marty@cable-158-181-78-199.cust.telecolumbus.net] has joined #shogun | 21:50 | |
-!- gsomix [~gsomix@85.26.234.149] has quit [Ping timeout: 276 seconds] | 21:54 | |
@sonney2k | Re | 21:54 |
@sonney2k | I guess it is best pluskid has a look at multiclass errors | 21:58 |
-!- Marty28 [~marty@cable-158-181-78-199.cust.telecolumbus.net] has quit [Quit: Colloquy for iPad - http://colloquy.mobi] | 22:00 | |
-!- karlnapf [~heiko@host86-176-1-187.range86-176.btcentralplus.com] has quit [Ping timeout: 272 seconds] | 22:03 | |
CIA-113 | shogun: Soeren Sonnenburg master * re8c646b / src/shogun/clustering/Hierarchical.cpp : fix crasher due to double free in hierarchical - http://git.io/Of9VPg | 22:05 |
CIA-113 | shogun: Soeren Sonnenburg master * r6b5dcbc / src/shogun/regression/KernelRidgeRegression.h : move enum to upper class level - http://git.io/wNyc2g | 22:05 |
-!- emrecelikten [~emrecelik@82.222.102.86] has joined #shogun | 22:08 | |
CIA-113 | shogun: Soeren Sonnenburg master * r21b11be / src/shogun/distributions/LinearHMM.cpp : fix double free in linearhmm - http://git.io/nvbxXg | 22:08 |
CIA-113 | shogun: Soeren Sonnenburg master * rd847327 / src/shogun/kernel/AUCKernel.cpp : fix double free in auc kernel - http://git.io/4d7cDA | 23:12 |
CIA-113 | shogun: Soeren Sonnenburg master * rb27af96 / (3 files in 2 dirs): fix double free errors in mkl / combined kernel / wds kernel - http://git.io/AekvJA | 23:12 |
CIA-113 | shogun: Soeren Sonnenburg master * ra576554 / (5 files in 2 dirs): | 23:24 |
CIA-113 | shogun: re-introduce load/save to labels | 23:24 |
CIA-113 | shogun: it does make sense to have these functions to be able to load/save | 23:24 |
CIA-113 | shogun: labels in standard file formats - http://git.io/xD1U0g | 23:24 |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 23:34 | |
CIA-113 | shogun: Soeren Sonnenburg master * rea70748 / src/shogun/transfer/domain_adaptation/DomainAdaptationSVM.cpp : fix double free in da svm - http://git.io/RHsIRA | 23:35 |
--- Log closed Sun May 06 00:00:37 2012 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!