IRC logs of #shogun for Tuesday, 2017-07-11

« Monday, 2017-07-10 Index Wednesday, 2017-07-12 »
--- Log opened Tue Jul 11 00:00:50 2017
micmnwiking: lol you sleep like 4 hours? XD00:22
-!- sukey1 [~nodebot@ks312251.kimsufi.com] has quit [Remote host closed the connection]03:40
-!- sukey [~nodebot@ks312251.kimsufi.com] has joined #shogun03:40
-!- mode/#shogun [+o sukey] by ChanServ03:40
-shogun-buildbot:#shogun- Build nightly_all #14 is complete: Failure [failed test (failure)] - http://buildbot.shogun-toolbox.org:8080/#builders/22/builds/1403:52
@sukey[https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3920 merged by vigsterkr08:03
@sukey[https://github.com/shogun-toolbox/shogun] vigsterkr pushed 2 commits:08:03
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/a26f9f142925af720d081550b6ea749532d03b2208:03
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/ed1bb04b5aa0483616b38be02f81a5e6ce97d79f08:03
-!- mikeling [uid89706@gateway/web/irccloud.com/x-jathdrlvlcdzbkyj] has quit [Quit: Connection closed for inactivity]08:48
-!- geektoni [c1cdd252@gateway/web/freenode/ip.193.205.210.82] has joined #shogun09:11
-!- geektoni_ [c1cdd252@gateway/web/freenode/ip.193.205.210.82] has joined #shogun09:26
-!- geektoni [c1cdd252@gateway/web/freenode/ip.193.205.210.82] has quit [Ping timeout: 260 seconds]09:27
-!- iglesiasg [~iglesiasg@217.119.234.214] has joined #shogun09:46
-!- mode/#shogun [+o iglesiasg] by ChanServ09:46
@iglesiasgqué pasa09:47
lisitsynwhat are reasonable answers? never taught yet10:12
@wikingnada10:15
@wikingbien10:15
@wikingmierda10:15
@wikingsukey, flip10:15
@sukey(ﾉಥДಥ)ﾉ︵┻━┻･/10:15
@wikingany of these are fine :)10:15
@wikingjoder :)10:16
lisitsynwiking: mierda!10:26
@iglesiasgshit is happening?!10:26
lisitsynhaha10:28
@sukey[https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3906 synchronized by MikeLing10:28
@iglesiasgany hint why swig would be including a class that has been removed?10:31
lisitsynremoved how?10:32
@iglesiasgit is not in libshogun.so, but dist-packages/_modshogun.so has it10:32
@wikingig10:32
@iglesiasgremove the header and implementation file10:32
@wikingtemplate10:32
@wiking?10:32
@iglesiasgno, it is not a template - I double check10:32
@iglesiasgclass CSumOne : public CDensePreprocessor<float64_t>10:33
@sukey[https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3919 synchronized by geektoni10:33
@iglesiasgI also removed the include and the rename from the Preprocessor.i10:34
@wikinghave you checked10:35
@wikingthat the in the build/src/interface.... the right verison of Preprocessor.i10:35
@wikingis there10:35
@wikingi.e. it's updated?10:35
@wikingbut i mean10:36
@wikingfuck10:36
@wikingi need to work10:36
@wikingso you just do it10:36
@iglesiasg?10:36
@wikingdunno :d10:37
@wikingthat's in short10:37
@iglesiasgtake it easy10:37
@iglesiasgI am just having this doubt, tried a couple of things and run out of ideas for now10:37
@iglesiasgso I asked for help10:37
@iglesiasgI am not telling none to fix nothing10:37
micmniglesiasg: did you try deleting the old swig build dir? (happened to me :p)10:40
lisitsynwiking: don't worry be happy (c) v.v. putin10:40
lisitsyn:P10:40
@wikinghahah10:41
@wikingthat was so fucking good10:41
@wikinghttps://www.youtube.com/watch?v=yfLx6X6461E10:42
@wiking:D10:42
@wiking[16:37]  * moriarty (~dtijd@unaffiliated/moriarty) has joined #tensorflow10:42
@wiking[16:37]  <Trioxin> need help: https://hastebin.com/itakumibiw.sql10:42
@wiking[16:38]  <moriarty> That will be 10 bitcoins thank you10:42
@wiking:DDDDDDDDDDDDDDDDDDDDDDDDDD10:42
@wikingtoo good10:42
@iglesiasgmicmn: I removed the complete build directory10:43
@iglesiasgfound it!11:19
@iglesiasg'import modshogun' no no no11:19
@iglesiasg'import shogun' xD11:19
lisitsynimport shogun!!!!!!!!!!!!!11:20
@iglesiasgoh yeah11:20
@wiking:)11:21
@wikingcrazy biatch :)11:21
lisitsynwiking: should've been swagshogun11:25
@wikingimport !@#$shogun11:26
@wikingthug life11:26
lisitsynimport if you can11:26
lisitsynfucker11:26
lisitsynwiking: it should also curse quite a lot11:26
@wikingyes11:26
@wikingbazdmeg11:26
@wikingevery 2nd __get__11:26
lisitsynwiking: if you don't do anything it should shout at you like what are you waiting for11:26
@wikingin the python interface11:26
@wiking:)11:26
@sukey[https://github.com/shogun-toolbox/shogun] vigsterkr pushed 2 commits:11:29
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/706872f70a813288ae5a6c4e9763b385c854a35611:29
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/da08025401ee9d42405bdc9f45dfe63e725634f511:29
-!- HeikoS [~heiko@untrust-out.swc.ucl.ac.uk] has joined #shogun11:37
-!- mode/#shogun [+o HeikoS] by ChanServ11:37
@HeikoSmicmn: hi!11:39
micmnHeikoS: hi, do you have an answer for me? :P11:40
@HeikoSmicmn: answer to what? :D11:40
@HeikoSswitching on/off the gpl?11:40
micmnyeah11:40
@HeikoSi see11:40
micmnmy last mail11:40
@HeikoSso there is two mechanics11:41
@HeikoS1.) the src/gpl folder.11:41
@HeikoS2.) the cmake switch11:41
@HeikoSthe gpl folder contents can be deleted11:41
@HeikoSbut then the switch (currently) also has to be turned off11:42
@HeikoS(because of the guards in unit tests)11:42
@HeikoSI will later remove the switch and set USE_GPL_SHOGUN based on the presence of the gpl folder11:42
@HeikoSso for now, set the cmake switch to off, and delete gpl folder to enable a full BSD build11:42
@HeikoSone thing11:42
@wikingHeikoS, i wouldnt do auto switching based on USE_GPL_SHOGUN11:43
@HeikoSthe build currently even fails if the folder is present and the switch is turned off (trained_model)11:43
@wikingi mean baed on a presence of a folder or not11:43
@wikinglet it fail11:43
@wikingif the folder is not and USE_GPL_SHOGUN11:43
@HeikoSwiking: whatever, we are not there yet11:43
@HeikoSyeah also ok11:43
@HeikoSpoint is11:43
@wikingi'm just saying11:43
@HeikoSbuild is broken :)11:43
@HeikoSwiking: yeah ok I take the point11:43
@HeikoSso the switch stays .11:44
micmnyeah the problem is just that11:44
@HeikoSmicmn: so can just guard the parts in the test that include files from gpl folder11:44
@HeikoSwith USE_GPL_SHOGUN11:44
micmnyep that was my question, if I could rely on USE_GPL_SHOGUN or not11:44
@HeikoSmicmn: so if the folder is removed, the test actually compiles11:45
micmnor it'll get removed sooner or later11:45
@HeikoSbut oif the folder is there and the switch is OFF then it doesnt11:45
@HeikoSyeah you can rely on USE_GPL_SHOGUN11:45
@HeikoSbut not in the classes in the src/gpl folder11:45
@HeikoSno guards are necessary in the folder11:45
@HeikoSin the files in the folder I mean11:45
micmnyeah yeah I get it11:45
@HeikoScool!11:45
@sukey[https://github.com/shogun-toolbox/shogun] New commit https://github.com/shogun-toolbox/shogun/commit/4950bfa9a47003988a971a3f4c1840d6529d6d9a by karlnapf11:46
@HeikoSshogun-buildbot: force build 'nighly_bsd_license'11:47
-shogun-buildbot:#shogun- no such builder 'nighly_bsd_license'11:47
@HeikoSwiking: ah cool the webinterface works again for force building?11:47
@wiking?11:47
@wikingwhat?11:47
@HeikoSI just pressed the button "rebuild"11:49
@HeikoSon buildbot website11:49
@HeikoSand now a build is queued11:49
@wikingah cool11:51
@wiking:D11:51
@wikingdidnt know11:51
@wiking:DDDDDDDDD11:51
@sukey[https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3919 synchronized by geektoni11:56
-shogun-buildbot:#shogun- Build osx1 - libshogun #77 is complete: Failure [failed test (failure)] - http://buildbot.shogun-toolbox.org:8080/#builders/25/builds/7712:03
Trixiswiking: any idea what could be causing this error? "MKL.cpp line 901: Assertion R >= 0 failed!" the code at that location doesnt seem to make it immediately obvious what R is...12:07
@wikingmmmm12:07
@wikingTrixis, for( p=0; p<num_kernels; ++p )12:15
@wikingthi sis my line 901 in MKL.cpp12:15
Trixiswiking: odd. sec12:15
@wikingok anyhow couple of lines above12:15
@wikingi'm wondeirng hat might have happened here12:15
@wikingmmm12:16
@wikingconst float64_t R = CMath::sqrt( preR / mkl_norm ) * epsRegul;12:16
@wikingand afaik neighter of the vars in elements in sqrt are complex12:16
@wikingonly way is that R is negative12:16
Trixishttp://www.shogun-toolbox.org/api/latest/MKL_8cpp_source.html seems to be line 901 in this12:16
@wikingthat epsRegul is negative12:16
@wikingright?12:16
@wikingbut then again12:17
@wikingconst float64_t epsRegul = 0.01;  // fraction of root mean squared deviation12:17
@wiking:D12:17
Trixisyeah12:17
@wikingand i dont see anywhere else epsRegul being defined differently12:17
@wikingin compute_optimal_betas_elasticnet12:17
Trixisi mean it could be NaN as well, right?12:18
@wikingmmm12:18
Trixisalthough im not sure how NaN behaves with comparisons in c++12:18
@wikingyou mean preR / mkl_norm12:18
@wiking?12:18
@wikingi mean whatever f(NaN ) = NaN afaik12:18
Trixisyes12:19
@wikingand float64_t preR;12:19
TrixisNaN is unordered as per specification12:19
Trixisso12:19
Trixisyoud have false12:19
Trixison any comparison12:19
@wikingyeah but12:19
@wikingah yeah12:19
@wikingok12:19
@wikingso you get somehow there a nan?12:20
Trixisi'd assume so it seems the only realistic scenario?12:20
@wikingok12:20
@wikingso12:20
@wiking++ the logging level12:20
@wikingof shogun12:20
@wikingset_loglevel(12:20
@wikingsince there are many debugging info there12:20
@wikingprinted12:20
@wikinglike12:21
@wikingSG_PRINT("MKL-direct: R = %e\n", R )12:21
@wikingso you should have an idea wtf is happening there12:21
Trixisactually no, cant be nan, forgot sqrt of nefative doesnt return nan but throws an error12:21
@wikingyeah but12:22
@wikingsqrt(NaN)12:22
@wiking?12:22
Trixishm true12:24
@wikingTrixis, from so: " What's the square root of -9? It's an imaginary number, but sqrt doesn't support imaginary results, so the result is nan (not a number). It's a domain error. The solution: Don't pass negative arguments to sqrt."12:24
@wikingso actually sqrt(R < 0) = NaN12:24
Trixisyeah12:24
Trixisoh12:24
Trixisright12:24
@wikingwould be great to have finally a patch on MKL12:25
@wikingCMath::isnan(R)12:25
@wiking:D12:25
@wikingand print a different error12:25
@wiking:P12:25
-!- geektoni_ [c1cdd252@gateway/web/freenode/ip.193.205.210.82] has quit [Ping timeout: 260 seconds]12:25
Trixisim trying to find out right now how to increase the logging level via modular interfaces12:26
@wikingso every SGObject12:26
@wikingso like MKL12:26
Trixisyeah12:27
@wikinghas get_global_io()12:27
@wikingMKL.get_global_io().set_loglevel(MSG_DEBUG);12:27
@wikingfor example12:27
Trixisso swig doesnt seem to be generating full interfaces for the SGIO object, there's only "SWIGTYPE_p_shogun__SGIO" which is has no methods12:30
-!- travis-ci [~travis-ci@ec2-54-144-106-53.compute-1.amazonaws.com] has joined #shogun12:30
travis-ciit's MikeLing's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: https://travis-ci.org/shogun-toolbox/shogun/builds/25232860412:30
-!- travis-ci [~travis-ci@ec2-54-144-106-53.compute-1.amazonaws.com] has left #shogun []12:30
Trixisnvm managed to get the output using init_shogun_with_defaults12:32
Trixisyeah R is NaN12:32
Trixishttps://pastebin.com/SMLNSMkn this is the output12:32
Trixisthat Z doesnt look good either lol12:34
@HeikoSTrixis, wiking probably some initialization problem12:35
@wikingTrixis, PRs are MOST welcome12:35
@wikingespecially for MKL12:35
@wiking:D12:35
@sukey[https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3923 opened by micmn12:52
micmnHeikoS: ^12:53
Trixiswiking: judging by the code id assume beta[] is all 0s12:53
@HeikoSmicmn: chceking12:54
@HeikoSmicmn: so what about other guards?12:54
@HeikoSmicmn: this works for GPL12:54
@HeikoSbut what if a class is guarded otherwise?12:55
Trixisthat way you'd get Z to inf, then on line 878 you'd get 0*inf which indeed is a NaN. The only issue is i dont see why it would be 0 :\12:55
micmnHeikoS: other guards are already handled12:55
@HeikoSah sorry where?12:55
@HeikoSline91?12:55
micmndef is_guarded(include, defined_guards):12:55
micmnyes12:56
@HeikoShttps://github.com/micmn/shogun/blob/aaa9c760e44d7ee1780da3eceb8a1661de1168be/tests/unit/base/trained_model_serialization_unittest.cc.py#L2512:56
@HeikoSmicmn: mmmmh12:56
@HeikoSmicmn: couldnt we also use ctags here?12:56
@HeikoScant that be used to find out whether a class is defined or not?12:56
micmnuhm no12:56
@HeikoSI dont like this parsing preprocesor code too much :(12:56
@HeikoSmicmn: I wonder whether there is a ctags option that you can pass all the #defines so that it ignores certain classes if they are guarded?12:57
micmnchecking12:58
micmnbut I don't think so12:58
@HeikoSmicmn: if not, it is fine as well .... just would be a bit nicer12:58
@HeikoShttps://github.com/micmn/shogun/blob/aaa9c760e44d7ee1780da3eceb8a1661de1168be/tests/unit/base/trained_model_serialization_unittest.cc.py#L3612:58
@HeikoSmicmn: this is quite nice12:58
@HeikoSmicmn: Ok i will wait for CI and then merge12:59
@HeikoSmicmn: can also be improved later on13:00
@HeikoSmicmn: thanks for the patch13:00
@HeikoSI think I broke the nightly_bsd_license tests with my last commit, but at least all GPL code is now in that folder and shogun compiles13:01
micmnHeikoS: can't find anything anyway I already checked when writing the script and resort to that as last option, I think that if we want to make it cleaner we should use clang's frontend13:04
@HeikoSok13:07
@HeikoSmicmn: nevermind then13:07
@HeikoSI think this might cause problems at some point .... but maybe it doesnt :)13:07
micmnmmm it'll do for sure :)13:09
lisitsynHeikoS: can I use some MMD-thingy to reweight dataset to match distribution?13:10
@HeikoSlisitsyn: can you explain13:10
@HeikoSlisitsyn: I am about to get lunch though13:10
lisitsynHeikoS: I have biased train and my test is everything else13:10
@HeikoSbut yes I guess!13:10
lisitsyn:D13:10
lisitsynok13:10
lisitsyncan you give me any reference later?13:10
@HeikoSill chekc log in an hour13:10
lisitsynkk13:10
@HeikoS"A kernel two sample test" JMLR13:10
@HeikoSbut I mean this is just nonparametic two-sample testing13:11
lisitsynalthough I don't need to test13:11
@HeikoSlisitsyn: explain problem, and I can help13:11
lisitsynI need to reweight them13:11
lisitsynor do something to make my train unbiased13:11
@HeikoSso that they are more similar?13:11
lisitsynyes13:11
@HeikoSso you want a distance measure on distributions?13:11
lisitsynI want to change distribution of train to match the distribution of test13:11
@HeikoSi.e. try to find a re-weighting such that distance is smaller13:11
Trixiswiking: that way you'd get Z to inf, then on line 878 you'd get 0*inf which indeed is a NaN. The only way beta can be all 0 if sumw is all zero. guess ill have to try to get the c++ code into a debugger, or use another way to track the sumw computation13:11
lisitsynyes13:12
@HeikoSsure that can be done13:12
@HeikoSsimplest idea:13:12
@HeikoSchoose a couple of weighting schemes13:12
@HeikoSrun MMD test between test and train for each13:12
@HeikoSpick the one with largest p-value13:12
@HeikoSthere is details, but something like this has been done before13:13
@HeikoSi.e. when training GANs :D13:13
@HeikoSok lunch13:13
@HeikoStalk later13:13
lisitsynsurer13:13
lisitsynthanks a lot13:13
Trixiswiking: ... which would be because the svm returns 0 as the number of support vectors -- i made a typo and used SVRLight instead of SVMLight. checks out. I might make a PR later on adding an if() to check if betas/sumw are all 0 and print a meaningful error later on.13:18
Trixisis parallel computation enabled by default in shogun?13:57
@wikingTrixis, should be13:58
@wikingTrixis, but you can query that as well13:58
@wikingnote that not everything is using parallel computation ;)13:58
@HeikoSlisitsyn: back14:01
Trixiswiking: yeah gotcha, i dont expect everything running in parallel, lol14:02
Trixiswiking: at the very least the matrix computations should, lol14:03
Trixishm the training time looks reasonable for a low dimensional training sample where log10(n) = ~4, about 19minutes14:06
@sukey[https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3923 merged by karlnapf14:25
@sukey[https://github.com/shogun-toolbox/shogun] karlnapf pushed 2 commits:14:25
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/aaa9c760e44d7ee1780da3eceb8a1661de1168be14:25
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/f027fb3a4bcb8dc42156826974ce9e53c1ef46ca14:25
-shogun-buildbot:#shogun- Build osx1 - libshogun #78 is complete: Success [build successful] - http://buildbot.shogun-toolbox.org:8080/#builders/25/builds/7814:30
@HeikoSshogun-buildbot: force build nightly_bsd_licensy14:33
-shogun-buildbot:#shogun- no such builder 'nightly_bsd_licensy'14:33
@HeikoSshogun-buildbot: force build nightly_bsd_license14:34
@HeikoSmicmn:14:35
@HeikoSthis compares signed and unsigned14:36
@HeikoShttps://github.com/shogun-toolbox/shogun/blob/develop/tests/unit/preprocessor/KernelPCA_unittest.cc#L2314:36
@HeikoSprobably should change something as it is not really consistent14:36
@wikingmuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu14:37
@wikingTrixis, mmmsure thing14:37
@wikingTrixis, as said all PRs are most welcome14:38
@wikingwe can give creditz and fame :D14:38
@wikingmmm i haven't seen olivier around lately14:39
@HeikoSwiking: he is away14:40
@wikinglast week?14:41
@HeikoSalmost two weeks14:41
@wikingkk14:41
micmnHeikoS: what do you mean by not consistent? I'll add a cast to index_t to get rid of the warning but other than that should be ok14:41
@wikingthen that's why14:41
@wikingHeikoS, have u seen this14:41
@HeikoSmicmn: the cast is not safe ;)14:41
@wikingHeikoS, https://github.com/ericniebler/range-v314:41
@HeikoSso in shogun, we should either use unsigned for positive quantities such as size14:42
@wikingHeikoS, especially note view::14:42
@HeikoSor signed (seems better sometimes)14:42
@HeikoSbut not both14:42
@wikingHeikoS, https://github.com/ericniebler/range-v3/blob/master/doc/index.md#views14:42
micmnso that's a problem with index_t14:42
@HeikoSmicmn: yeah not your fault14:43
@HeikoSI think the return type of size() should be index_t14:43
@HeikoSor int64_t14:43
@HeikoSwiking: uh that is good stuff14:43
@HeikoSIll check14:43
micmnbut why have index_t signed?14:43
@wikingmicmn, ok so there's  huge discission about this14:44
@wikingmicmn, google et.al. :)14:44
@HeikoSproblem with signed is that STL is not14:44
@HeikoSso we cannot nicely nix SG* and std::14:44
@HeikoSthat is annoying14:44
@HeikoSI'd rather follow STL14:45
@HeikoSbut no hard feelings14:45
micmnok thx :)14:45
@wikingmicmn, You should not use the unsigned integer types such as uint32_t, unless there is a valid reason such as representing a bit pattern rather than a number, or you need defined overflow modulo 2^N. In particular, do not use unsigned types to say a number will never be negative. Instead, use assertions for this.14:45
@HeikoSwiking: how to talk to STL? Want to put assertions everywhere in code?14:46
@wikingmicmn, since for example... what if you wanna represent relative index ;)14:46
micmnyeah...14:46
@wikingmicmn, that could and should be possible to be negative14:46
@wikingso that's the idea about why index actually should be a signed value14:46
@wikinginstead of an unsigned14:46
@wikingHeikoS, well size cannnot be ever negative14:47
@wikingright? :)14:47
@wikingi mean if we are in non-loco mode14:47
@wikingHeikoS, i presume you are talking about size_t container::size() const;14:47
@wikingright?14:47
@HeikoSyes14:47
@wikingcast-o machine assertation14:47
@HeikoSso if you iterate over a std::vector to touch SG* then you get these warnings14:47
@wikinglife's a bitch14:47
@wikingHeikoS, of course!14:47
@wikingi mean14:48
@wikingof course, nobody14:48
@wikingshould ever use14:48
@wikingindex for iterating over a data14:48
@wiking:)14:48
@wikingonly for random accessing it14:48
@wiking:P14:48
@HeikoSthat is true, BUT14:48
@HeikoSthat is what shogun does14:48
@wikingyeye of ocurse14:48
@HeikoSand currently if you start using std:: that sucks14:48
@wikingit's a bad practice14:48
@wikingnot only in shogun14:48
@wikingbut many other places14:48
@wikingbut at least14:48
@wikingstl stuff14:48
@HeikoSso what to do right now?14:48
@wikinghas .begin() and .end()14:48
@wikingso you could actually iterate through STLs14:49
@HeikoSI guess we should give SG* a begin() and end() as well14:49
@wikingwith iterator14:49
micmnSGMatrix::size() is not related to STL14:49
@wikingHeikoS, yeah but we know this :)14:49
@HeikoSstill need count if you want to put stuff in SG*14:49
@wikingHeikoS, count?14:49
@wikingHeikoS, like size()?14:49
@HeikoSvector[i]14:49
@HeikoSthe i14:49
@wikingi = index_t14:49
@HeikoSwhatever14:49
@HeikoSanyways14:49
@wikingbut size = can be unsgined :D14:49
@wikingit's just shitty14:49
@HeikoSI think we should change return type of SGVector::size14:50
@HeikoSto index_t14:50
@HeikoScurrently it is unsigned14:50
@HeikoSor change to int64_t14:50
@HeikoSwiking: complaints on that?14:50
@wikingHeikoS, index_t should actually be14:50
@wikingint64_t :D14:50
@HeikoSwhatever14:50
@wikingbecause i's a bit of a shitty stuff atm14:51
@wikingthat we can only index14:51
@HeikoSbut size() method should return index_t14:51
@wiking2^3114:51
@HeikoSmicmn: so that return type can be changed and all good with that warning14:51
@HeikoSwiking: and yes let's change it to 6414:51
@HeikoSi mean index_t14:51
@wikingHeikoS, feature/BigInteger14:51
@wiking;)14:51
@HeikoSah yes I forgot14:52
@wikingit's crazy fucking shit14:52
@HeikoSin that case14:52
@wikingmany many places14:52
@HeikoSmicmn: it needs to be changed to int64_t for now :D14:52
@HeikoSas otherwise it might well overflow14:52
@HeikoSfor big matrices14:52
@HeikoSSG_MALLOC(8*1024*size);14:53
@HeikoSwiking: I think I once saw that somewhere14:53
@wikingHeikoS, what ?14:53
@HeikoSwiking: hard-coded sizeof14:53
@wikinghehehe14:53
@wikingwhen i started that branch14:54
@wikingand turned on overflow detection as error14:54
@wikinglaaaaaaaaaaalalala land14:54
@wikingbut yeah needs to be done one day14:54
@wikingas it is stupid that14:54
@wikingnowadays you could have a memory of 256G14:54
@HeikoSlala land welcomes you14:55
@wikingand then you cannot index14:55
@wiking200G of doubles14:55
@wikingso yeah :) i mean of course14:55
@wikingwe know what we know :P14:55
@wiking(never admit in public)14:55
@wikingbut yeah theoretically if we would see and SGMatrix of 200g14:56
@wikingwe could index it :)14:56
@wiking*we couldn't index it14:56
@wikingbuut yeah stuff like this is crazy nice14:57
@wiking    auto rng = vi | view::remove_if([](int i){return i % 2 == 1;})14:57
@wiking                  | view::transform([](int i){return std::to_string(i);});14:57
@wikingboom14:57
-!- mikeling [uid89706@gateway/web/irccloud.com/x-jfouepcimcfulsss] has joined #shogun15:12
-!- geektoni [c1cdd252@gateway/web/freenode/ip.193.205.210.82] has joined #shogun15:20
mikelingHeikoS: ping15:26
@HeikoSmikeling: pong15:26
mikelingHi , after I add set_global_seed(1) to calls.sg, the set_global_seed(1) been translated to set_global_seed()15:26
mikelingwhich missing the parameter '1'15:26
@HeikoSmikeling: oh15:26
@HeikoSmikeling: I might have screwed something up15:26
@HeikoSchecking!15:27
-!- geektoni_ [c1cdd221@gateway/web/freenode/ip.193.205.210.33] has joined #shogun15:53
-!- geektoni [c1cdd252@gateway/web/freenode/ip.193.205.210.82] has quit [Ping timeout: 260 seconds]15:54
@wikinggeektoni_, wadaaaaaaaaaaaaaapppppppppp15:58
geektoni_wiking: it still takes too much time15:58
@wikingnooooooooooooooooouuuuuuuu15:59
geektoni_and it also seg fault now15:59
@wikingah buuutt15:59
@wikingbuuut15:59
geektoni_but it is a minor problem :P15:59
@wikinggeektoni_, cc5628d262fb09eac56b6bd8fa5d82df3776886f15:59
@wikinginterractive rebase and drop this15:59
@wikingi told15:59
@wikingu15:59
@wiking:)15:59
@wikingas this removes -j2 from make16:00
@wikingso that's for sure one reason why it's slower than usual16:00
@sukey[https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3919 synchronized by geektoni16:00
geektoni_boom16:01
@wikingzuuuuuuuuuuupzuuuuuuuuuup16:01
@wikingbtw honsly16:06
@wikinginstead of making insane amount of traffic for our loved github/aws16:06
@wikingmaybe we should cache gtest source :))))16:07
@wikinggeektoni_, but yead there's some weird memory error16:10
geektoni_wiking: yep. I'm working on it.16:14
@sukey[https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3859 synchronized by MikeLing16:15
@sukey[https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3859 synchronized by MikeLing16:17
@wikingmmm16:21
@wikingHeikoS, can we drop all linalg/16:22
@wikingand gpuvector/matrix16:22
@wikingor at least on the cleanup16:22
@wikingto drop it16:22
@wiking?16:22
-!- iglesiasg [~iglesiasg@217.119.234.214] has quit [Quit: leaving]16:37
@wikinggeektoni_, we almost managed to do it :)))16:51
@wikingbut ccache is populated so i guess next time it'll be faster16:52
@wikingHeikoS, can we drop serialization format tests?16:52
@wikinglike json/xml/ascii?16:52
geektoni_wiking: I've managed to fix the memory error16:52
@wikingcool16:53
@wikingpush16:53
geektoni_the next push will (probably) make travis green16:53
@sukey[https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3919 synchronized by geektoni17:06
@sukey[https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3919 synchronized by geektoni17:07
@HeikoSwiking: so dropping linalg, what do you mean exactly?17:24
@HeikoSwiking: and drop serialization format tests, yes do it, though we havent got the new stuff in yet, so maybe wait a bit more17:25
@sukey[https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3919 synchronized by geektoni17:54
@sukey[https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3919 synchronized by geektoni17:56
lisitsyndrop linalg? :D17:57
-!- geektoni_ [c1cdd221@gateway/web/freenode/ip.193.205.210.33] has quit [Ping timeout: 260 seconds]18:14
-!- HeikoS [~heiko@untrust-out.swc.ucl.ac.uk] has quit [Ping timeout: 240 seconds]18:45
-!- mikeling [uid89706@gateway/web/irccloud.com/x-jfouepcimcfulsss] has quit [Quit: Connection closed for inactivity]19:08
@sukey[https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3919 merged by vigsterkr22:48
@sukey[https://github.com/shogun-toolbox/shogun] vigsterkr pushed 5 commits:22:48
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/e3dde552813cb04bd1fc79b76c6415ac0ae778cc22:48
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/8ac2b028f00560346d282cc4b7bd1fe5861ee81522:48
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/6f1603a08822d224528e43f09a5d0c27fbcdb10722:48
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/4ad2814cdac331de6fbbdda15eda950e019dded522:48
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/9be1025f596e1325fbc478a93647a93fe75474ee22:48
--- Log closed Wed Jul 12 00:00:51 2017
« Monday, 2017-07-10 Index Wednesday, 2017-07-12 »

Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!