IRC logs of #shogun for Tuesday, 2017-06-06

« Monday, 2017-06-05 Index Wednesday, 2017-06-07 »
--- Log opened Tue Jun 06 00:00:01 2017
@sukeyPull Request #3831 "[ProgressBar] Refactor the progress bar code and add documentation."  merged by vigsterkr - https://github.com/shogun-toolbox/shogun/pull/383103:11
@sukeyNew Commit "Merge pull request #3831 from geektoni/feature/progress_refactor03:11
@sukey[ProgressBar] Refactor the progress bar code and add documentation." to shogun-toolbox/shogun by vigsterkr: https://github.com/shogun-toolbox/shogun/commit/4312e6e11bd5690f96372d0957197f4eceab9f8a03:11
@sukeyPull Request #3830 "Add to linalg trace, zero, identity, add_col_vec (CPU only)"  merged by vigsterkr - https://github.com/shogun-toolbox/shogun/pull/383003:14
@sukeyNew Commit "Merge pull request #3830 from micmn/feature/linalg-ops03:14
@sukeyAdd to linalg trace, zero, identity, add_col_vec (CPU only)" to shogun-toolbox/shogun by vigsterkr: https://github.com/shogun-toolbox/shogun/commit/87a9ccb2c457d4a25b17f7ee30949d45f2caa70003:14
-!- olinguyen [81615ad9@gateway/web/freenode/ip.129.97.90.217] has quit [Quit: Page closed]03:24
shogun-buildbotbuild #196 of yakkety - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/yakkety%20-%20libshogun/builds/196  blamelist: Michele Mazzoni <mic.mazzoni@gmail.com>, Viktor Gal <vigsterkr@gmail.com>03:25
shogun-buildbotbuild #31 of osx1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/31  blamelist: Michele Mazzoni <mic.mazzoni@gmail.com>, Viktor Gal <vigsterkr@gmail.com>03:25
@wikingmicmn, ^ need to address these when you have the time ASAP03:26
shogun-buildbotbuild #180 of xenial - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/xenial%20-%20libshogun/builds/180  blamelist: Michele Mazzoni <mic.mazzoni@gmail.com>, Viktor Gal <vigsterkr@gmail.com>03:28
@sukeyIssue #3834 "add CoreML model serialization format" opened by vigsterkr - https://github.com/shogun-toolbox/shogun/issues/383403:32
@sukeyIssue #3834 "add CoreML model serialization format" vigsterkr added label: "development tasks" - https://github.com/shogun-toolbox/shogun/issues/383403:32
-!- travis-ci [~travis-ci@ec2-54-161-159-95.compute-1.amazonaws.com] has joined #shogun04:03
travis-ciit's Viktor Gal'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/23979507004:03
-!- travis-ci [~travis-ci@ec2-54-161-159-95.compute-1.amazonaws.com] has left #shogun []04:03
-!- mikeling [uid89706@gateway/web/irccloud.com/x-rzcgahdfjctgovtj] has joined #shogun04:30
-!- travis-ci [~travis-ci@ec2-54-158-26-86.compute-1.amazonaws.com] has joined #shogun04:52
travis-ciit's Viktor Gal'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/23979546504:52
-!- travis-ci [~travis-ci@ec2-54-158-26-86.compute-1.amazonaws.com] has left #shogun []04:52
mikelingwiking: ping05:01
@wikingpong05:01
mikelingmmmm05:03
mikelingwiking: so here is my gdb trace https://pastebin.mozilla.org/902360105:03
mikelingwhat's that Exception for?05:03
@wiking?05:05
@wikingi'm not so sure05:05
@wikingwhat you05:05
@wikingm ean05:05
@wikingbut here i cannot help you anymore05:06
@wikingbecause this is really a simple task of writing05:06
mikelingok :)05:06
@wikingunit tests for a set of apis :)05:06
@wikingwrite it for the original one05:07
@wikingand make sure all of it's tested and used05:07
@wikingonce that's done then you should try your patch05:07
@wikingif i need to help in this more05:07
@wikingit'd be easier for me to actually write the tests myself05:07
@wikinglisitsyn, hahaha in may we've burned 8,072.89 USD credit :))))06:58
@sukeyPull Request #3767 "LinalgRefactor -  specialpurpose"  synchronized by OXPHOS - https://github.com/shogun-toolbox/shogun/pull/376707:01
@sukeyPull Request #3767 "LinalgRefactor -  specialpurpose"  synchronized by OXPHOS - https://github.com/shogun-toolbox/shogun/pull/376707:17
@sukeyIssue #3835 "Installation configuration problem" opened by xxnyufengzhe - https://github.com/shogun-toolbox/shogun/issues/383508:41
@sukeyIssue #3835 "Installation configuration problem" closed by vigsterkr - https://github.com/shogun-toolbox/shogun/issues/383508:45
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Ping timeout: 260 seconds]08:52
-!- shogun-toolbox [~shogun@7nn.de] has quit [Ping timeout: 260 seconds]08:52
--- Log closed Tue Jun 06 08:52:12 2017
--- Log opened Tue Jun 06 08:52:20 2017
-!- shogun-toolbox [~shogun@7nn.de] has joined #shogun08:52
-!- Irssi: #shogun: Total of 20 nicks [4 ops, 0 halfops, 0 voices, 16 normal]08:52
-!- Irssi: Join to #shogun was synced in 7 secs08:52
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun08:52
micmnwiking: checking09:02
-!- geektoni [~geektoni@93-34-234-212.ip52.fastwebnet.it] has joined #shogun09:02
-!- iglesiasg [~iglesiasg@217.119.234.214] has joined #shogun09:04
-!- mode/#shogun [+o iglesiasg] by ChanServ09:04
@wikingmicmn, thnx... it's about eigen 3.309:09
@wikingimo09:09
@wikingas those buildbots failed which has eigen 3.3+09:09
@wikingsee here http://buildbot.shogun-toolbox.org/waterfall09:09
micmnthx09:11
@sukeyPull Request #3836 "[ProgressBar] Replace SG_PROGRESS with the new progress bar."  opened by geektoni - https://github.com/shogun-toolbox/shogun/pull/383609:47
@sukeyPull Request #3836 "[ProgressBar] Replace SG_PROGRESS with the new progress bar."  synchronized by geektoni - https://github.com/shogun-toolbox/shogun/pull/383609:53
@sukeyPull Request #3836 "[ProgressBar] Replace SG_PROGRESS with the new progress bar." - https://github.com/shogun-toolbox/shogun/pull/383609:54
micmnwiking: eigen3.3 fails with T=bool10:05
micmnif I change DEFINE_FOR_ALL_PTYPE to DEFINE_FOR_NUMERIC_PTYPE10:05
micmnit works10:06
micmnadd is not defined for bool too10:06
mikelingwiking: I use CDynamicArray.resize_array(10), and get_array_size() tell me the array size is 15???10:10
mikelingwhy?10:10
@wikingmikeling, bug?10:11
mikelingI'm not sure, just want to know if that we expected result10:12
mikelinglet me update the commit first10:12
mikelingand we could discuss in there10:12
@wikingmikeling, good idea10:13
@wikingmikeling, look what i would do is basically git grep resize_array and see if in code we rely on this in a way that it's 15 and not 10 :P10:13
@wikingi think10:14
@wikingthat it's a bug10:14
@wikingone way to do this is basically10:14
@wiking'fix' it and see if anything breaks :)10:14
@sukeyPull Request #3833 "unit test for DynamicArray"  synchronized by MikeLing - https://github.com/shogun-toolbox/shogun/pull/383310:14
micmnwiking: https://github.com/shogun-toolbox/shogun/issues/364410:15
@wikingand this supposed to be the fix?10:17
@wikinghttps://github.com/shogun-toolbox/shogun/pull/3656/files10:17
micmnyep10:18
micmnchange DEFINE_FOR_ALL_PTYPE to DEFINE_FOR_NUMERIC_PTYPE10:18
@sukeyPull Request #3833 "unit test for DynamicArray"  synchronized by MikeLing - https://github.com/shogun-toolbox/shogun/pull/383310:18
@wikingmicmn, good :)10:18
@wikingmicmn, any questions? :)10:18
mikelingwiking: https://github.com/shogun-toolbox/shogun/pull/3833/files#diff-8129456668aa9135d6340bd89fdad5bcR6410:19
micmnshould i open a new pr?10:19
@wikingmicmn, yeah since we merged the other one10:20
@wikingjust create a fixer PR :)10:20
@wikingmikeling, yes i commented some stuff10:21
mikelingwiking: thank you, I will address the pr as commented :)10:22
@wikingbut about the resize10:22
@wikingi have no clue10:22
@wikingthat needs to be tested10:22
@wikingwhether fixing it has an effect on anything in the library10:22
@wikingmikeling, but based on the code10:24
@wikingthis is actually a possible bug in DynArray10:24
@wikingright?10:24
mikelingyep10:24
@wikingaz this just calls DynArray.resize10:25
@wikingi mean DynArray.resize_array10:25
mikelingwiking: https://github.com/shogun-toolbox/shogun/blob/46f91d517c877ade80f990a77fbaf29d034076d5/src/shogun/base/DynArray.h#L34610:25
mikelingis that because we add one in here?10:26
@wikingresize_granularity = 128 by default right?10:26
mikelingyep10:26
@sukeyPull Request #3837 "Fix add_col_vec compatibility with eigen 3.3.x"  opened by micmn - https://github.com/shogun-toolbox/shogun/pull/383710:27
@wikingmikeling, i mean in that case new_num_elements = 128 ?10:27
@wikingah no ok10:28
@wikingit starts with resize_granularity = 5 irght?10:29
@wikingand then10:29
@wikingbecause (10/5 +1)*5 = 15 :)10:29
@wikingmicmn, i'm gonna merge this before acually waiting for the CIs as those will not catch this (none of them uses 3.3 eigen)10:30
@sukeyPull Request #3837 "Fix add_col_vec compatibility with eigen 3.3.x"  merged by vigsterkr - https://github.com/shogun-toolbox/shogun/pull/383710:30
@sukeyNew Commit "Merge pull request #3837 from micmn/feature/add_col_vec-fix10:30
@sukeyFix add_col_vec compatibility with eigen 3.3.x" to shogun-toolbox/shogun by vigsterkr: https://github.com/shogun-toolbox/shogun/commit/e3b31a74ada03bd9adc59b104bc74e7c4bf07d7510:30
@wikinglet's see what buildbot does10:30
micmn:)10:31
mikelingwiking: so, the resize_array() is about to10:33
@wikingbut honestly10:33
mikelingorign_size + size10:33
@wikingthis shouldn't be a problem10:33
@wikingif you dont follow this resize strategy imo10:34
@wikingbecause this all depends on the starting size10:34
@wikingi.e. granularity10:34
@wikingmikeling, and there's 0 resize_array call in https://github.com/shogun-toolbox/shogun/blob/develop/src/shogun/mathematics/Integration.cpp10:35
@wikingand that's where things go bad10:35
mikelingwiking: mmm, but in std::vector or std::deque,  resize() means the size you want to have rather than to add10:36
mikelinglike resize(10)10:36
@wikingyes i know10:36
mikelingok10:36
mikelingso, we want to keep that move? like resize_array(10) will change the size depend on original size10:37
mikelingrather than change it to 10?10:37
@wikingi wouldn't do this10:37
@wikingagain10:37
@wikingthe bug is not here10:37
@wikingi'm 100%10:37
@wikingthere's a serious problem10:37
@wikingwith the refactor of the set_array10:37
@wikingin your deque10:37
@wikingpr10:37
@wikingthat's why things happens10:37
@wikinghave you checked what i told10:38
@wikingintegrate_quadgk10:38
@wikinghttps://github.com/shogun-toolbox/shogun/blob/develop/src/shogun/mathematics/Integration.cpp#L34310:38
@wikingif you add here10:38
@wikingsubs->display_array(); q_subs->display_array(); err_subs->display_array();10:39
@wikingand you run on both of the branches10:39
@wikingand run10:39
@wiking./bin/shogun-unit-test --gtest_filter=GaussianProcessClassification.apply_preprocessor_and_binary10:40
@wikingyou'll see what i mean10:40
@wikingthe arrays are different10:40
@wikingmikeling, see what i mean?10:43
mikelingyeah, let me do it right now.10:44
@wikingk10:44
shogun-buildbotbuild #32 of osx1 - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx1%20-%20libshogun/builds/3210:46
@wikingmicmn, ^10:46
@wikingso good10:47
@wikingthnx10:47
@sukeyPull Request #3836 "[ProgressBar] Replace SG_PROGRESS with the new progress bar."  synchronized by geektoni - https://github.com/shogun-toolbox/shogun/pull/383610:47
micmnnice10:47
micmnso yesterday while i was working on fisherlda10:48
micmnand digging through old prs i had a plot twist10:48
micmnthere is a third lda class10:48
micmnmclda (multiclass)10:49
micmnwith the third implementation of the canonvar solver10:50
@wikingwooooooohooooooooo10:51
@wiking:D10:51
micmnand actually it's the best to benchmark since in the benchmark system there are multiclass datasets10:51
@wikingrefactor & kill10:51
shogun-buildbotbuild #181 of xenial - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/xenial%20-%20libshogun/builds/18110:51
micmn2/310:51
shogun-buildbotbuild #197 of yakkety - libshogun is complete: Success [build successful]  Build details are at http://buildbot.shogun-toolbox.org/builders/yakkety%20-%20libshogun/builds/19710:52
micmn3/3 woohoo10:53
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]11:09
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun11:09
@sukeyNew branch feature/smart-pointers created on shogun-toolbox/shogun11:21
@sukeyNew Commit "Merge pull request #3837 from micmn/feature/add_col_vec-fix11:21
@sukeyFix add_col_vec compatibility with eigen 3.3.x" to shogun-toolbox/shogun by vigsterkr: https://github.com/shogun-toolbox/shogun/commit/e3b31a74ada03bd9adc59b104bc74e7c4bf07d7511:21
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting]11:23
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun11:23
-!- geektoni [~geektoni@93-34-234-212.ip52.fastwebnet.it] has quit [Quit: Leaving.]12:46
-!- geektoni [~geektoni@93-34-234-212.ip52.fastwebnet.it] has joined #shogun12:46
-!- geektoni [~geektoni@93-34-234-212.ip52.fastwebnet.it] has quit [Ping timeout: 240 seconds]12:50
-!- travis-ci [~travis-ci@ec2-54-158-26-86.compute-1.amazonaws.com] has joined #shogun12:55
travis-ciit's Viktor Gal'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/23987306412:55
-!- travis-ci [~travis-ci@ec2-54-158-26-86.compute-1.amazonaws.com] has left #shogun []12:55
-!- geektoni [~geektoni@93-34-234-212.ip52.fastwebnet.it] has joined #shogun13:06
micmnwiking: check this https://github.com/shogun-toolbox/shogun/blob/c7640fa97b44936770a2b10762ee2431ab505e94/src/shogun/multiclass/MCLDA.cpp#L30113:27
micmnthe mean is calculated on the centered data13:27
@wikingeh13:29
@wikinglemme check sorry i'm just in a middle of eveyrthin13:30
micmnsure no hurry13:31
@wikingok i'm here13:42
@wikingso what's the problem ? :)13:42
@wikingmicmn, // subtract the mean of class k from each sample of class k and store the centered data in Xc13:43
@wikingXc or rather X no?13:44
micmnthis line: Em_xbar = X.colwise().sum();13:52
@wikingyeah13:53
@wikingi was jsut trying to figure out the content of X13:53
@wikingbut yeah i see13:53
@wikingthe X basically standardised  right?13:54
micmnyeah13:54
micmnthe result is that Em_xbar is ~eps13:55
@wikingah yeah13:56
micmni was comparing mclda with scikit which is almost identical13:56
micmnbut the results where different13:57
micmn*were13:57
@wikingmmm you mean identical in sense of?13:57
@wikinglogic?13:57
micmnyes13:57
@wikingoh i wonder13:57
@wikingwhich is the scikit learn implementation?13:58
@wiking(do you have github for it?)13:58
micmnhttps://github.com/scikit-learn/scikit-learn/blob/14031f6/sklearn/discriminant_analysis.py#L34713:58
micmnxbar is computed on L37013:58
@wikingmmm13:59
@wikingit looks like a copy paste :)14:00
@wikingi wonder which direction14:01
-!- travis-ci [~travis-ci@ec2-54-158-26-86.compute-1.amazonaws.com] has joined #shogun14:02
travis-ciit's Viktor Gal'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/23988702014:02
-!- travis-ci [~travis-ci@ec2-54-158-26-86.compute-1.amazonaws.com] has left #shogun []14:02
micmnthe unit test just checks the output labels and not the model params...14:04
@wiking:)14:08
micmnOr at least the raw output like other tests14:09
-!- iglesiasg [~iglesiasg@217.119.234.214] has quit [Quit: leaving]14:12
@sukeyPull Request #3836 "[ProgressBar] Replace SG_PROGRESS with the new progress bar."  synchronized by geektoni - https://github.com/shogun-toolbox/shogun/pull/383615:01
@wikinggeektoni, ping16:06
geektonipong16:06
geektoniwiking: what's up?16:08
@wikingerrors16:08
@wikingin the pr :)16:08
geektoniyeah, I know16:09
geektoniwrong #include positions mostly16:09
@sukeyPull Request #3836 "[ProgressBar] Replace SG_PROGRESS with the new progress bar."  synchronized by geektoni - https://github.com/shogun-toolbox/shogun/pull/383616:24
@wikingmmm i saw other errors16:27
@wikingas well16:27
geektoniargh16:28
geektonilet me see16:28
@wikinggeektoni, so lemme show16:50
@wikinggeektoni, have u seen this https://travis-ci.org/shogun-toolbox/shogun/jobs/239953964#L87916:51
@wikingah i guess this is what cause by wrong include? :)16:51
geektoniexactly16:52
geektonisome include were placed inside some #ifdef guards16:53
geektoninever trust IDEs auto-complete features16:56
@wiking:P17:00
@wikinggeektoni, btw https://ci.appveyor.com/project/vigsterkr/shogun/build/4.2.829/job/bcnhpg9b28acbd3g#L188218:01
@wikinggood luck with windows :)18:02
geektoni:(18:03
@sukeyPull Request #3836 "[ProgressBar] Replace SG_PROGRESS with the new progress bar."  synchronized by geektoni - https://github.com/shogun-toolbox/shogun/pull/383618:08
geektoniwiking: AppVeyor failure was probably  caused by a #if directive which should be #ifdef instead.18:12
@wikinggeektoni, imo18:12
@wiking#ifdef should be fine18:12
@wikingbtw18:12
@wikingwhy do u need windows.h there?18:12
geektonibecause I need to know the terminal size to set the progress bar18:13
geektoniand there is a different procedure on windows to get the terminal height and width.18:13
@sukeyPull Request #3832 "use std::deque instead of DynArray(on going)"  synchronized by MikeLing - https://github.com/shogun-toolbox/shogun/pull/383218:35
@sukeyPull Request #3833 "unit test for DynamicArray"  synchronized by MikeLing - https://github.com/shogun-toolbox/shogun/pull/383318:41
mikelingwiking: ping18:42
@wikingpong18:42
mikelingwiking: new error comes out after old one been fixed :/18:42
@wikingwhere?18:43
@wikingbut i mean i see your push18:43
@wikingbut there are like 3 more set_array apis18:43
@wikingright?18:43
@wikinganyhow18:43
@wikingmanged to see18:43
@wikingthe error18:43
@wikingi was talking about?18:43
@wikingwith integration18:44
@wiking?18:44
mikelingwiking: yeah, maybe places have  double free or corruption  issues, let me past the gdb trace in here18:45
@wikingbut i mean18:45
@wikingthe problem is not this18:45
@wikingi mean in integration you will see18:46
@wikingthat totally different sizes18:46
@wikingof arrays are being created18:46
@wikingbut anyways18:46
@wikingi need to crash now18:46
mikelingcrash?18:47
mikelingyou mean go to sleep?18:48
mikelingok later :)18:48
@wikingyes18:52
-!- OXPHOS [92bd305b@gateway/web/freenode/ip.146.189.48.91] has quit [Quit: Page closed]19:22
@sukeyPull Request #3767 "LinalgRefactor -  specialpurpose"  synchronized by OXPHOS - https://github.com/shogun-toolbox/shogun/pull/376719:40
-!- olinguyen [81615ad9@gateway/web/freenode/ip.129.97.90.217] has joined #shogun20:02
@sukeyPull Request #3836 "[ProgressBar] Replace SG_PROGRESS with the new progress bar."  synchronized by geektoni - https://github.com/shogun-toolbox/shogun/pull/383620:17
-!- geektoni [~geektoni@93-34-234-212.ip52.fastwebnet.it] has left #shogun []20:17
-!- mikeling [uid89706@gateway/web/irccloud.com/x-rzcgahdfjctgovtj] has quit [Quit: Connection closed for inactivity]22:50
-!- olinguyen [81615ad9@gateway/web/freenode/ip.129.97.90.217] has quit [Quit: Page closed]23:57
--- Log closed Wed Jun 07 00:00:02 2017
« Monday, 2017-06-05 Index Wednesday, 2017-06-07 »

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