--- Log opened Fri Jun 03 00:00:23 2016 | ||
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has quit [Client Quit] | 00:03 | |
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has joined #shogun | 01:26 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 01:26 | |
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has quit [Quit: Leaving.] | 01:42 | |
shogun-buildbot | build #1014 of nightly_none is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_none/builds/1014 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Sanuj <sanuj.sharma.in@gmail.com> | 03:03 |
---|---|---|
-!- GandalfTheWizard [~Eva@112.10.170.39] has quit [Read error: Connection reset by peer] | 03:14 | |
-!- GandalfTheWizard [~Eva@112.10.170.39] has joined #shogun | 03:14 | |
-!- lambday [56a397da@gateway/web/freenode/ip.86.163.151.218] has quit [Ping timeout: 250 seconds] | 03:44 | |
shogun-buildbot | build #12 of clang - thread analysis is complete: Failure [failed compile] Build details are at http://buildbot.shogun-toolbox.org/builders/clang%20-%20thread%20analysis/builds/12 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Sanuj <sanuj.sharma.in@gmail.com> | 03:44 |
shogun-buildbot | build #11 of clang - undefined behaviour analysis is complete: Failure [failed compile] Build details are at http://buildbot.shogun-toolbox.org/builders/clang%20-%20undefined%20behaviour%20analysis/builds/11 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Sanuj <sanuj.sharma.in@gmail.com> | 03:47 |
-!- OXPHOS [9d8b131c@gateway/web/freenode/ip.157.139.19.28] has quit [Quit: Page closed] | 04:30 | |
shogun-buildbot | build #12 of memleak - valgrind is complete: Failure [failed memory check] Build details are at http://buildbot.shogun-toolbox.org/builders/memleak%20-%20valgrind/builds/12 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Sanuj <sanuj.sharma.in@gmail.com> | 05:33 |
shogun-buildbot | build #1143 of nightly_default is complete: Failure [failed test notebooks] Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/1143 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Sanuj <sanuj.sharma.in@gmail.com> | 06:11 |
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has joined #shogun | 09:09 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 09:09 | |
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has quit [Client Quit] | 09:10 | |
-!- besser82 [~besser82@fedora/besser82] has joined #shogun | 09:40 | |
-!- mode/#shogun [+o besser82] by ChanServ | 09:41 | |
-!- besser82 [~besser82@fedora/besser82] has quit [Ping timeout: 246 seconds] | 09:53 | |
-!- deepak_ [0e8bb448@gateway/web/freenode/ip.14.139.180.72] has joined #shogun | 10:30 | |
-!- lambday [8028b10a@gateway/web/freenode/ip.128.40.177.10] has joined #shogun | 11:56 | |
-!- mode/#shogun [+o lambday] by ChanServ | 11:56 | |
-!- lupinix_ [~quassel@v22014041761818086.yourvserver.net] has joined #shogun | 12:01 | |
Saurabh7 | lambday: yo | 12:03 |
-!- lupinix [~quassel@fedora/lupinix] has quit [Quit: No Ping reply in 180 seconds.] | 12:03 | |
Saurabh7 | sry had tucked in early yesterday :) | 12:03 |
@lambday | Saurabh7: hey | 12:03 |
@lambday | Saurabh7: no worries | 12:03 |
Saurabh7 | lambday: i found eigen thing btw | 12:03 |
@lambday | Saurabh7: what is it? | 12:03 |
Saurabh7 | looks something like VectorXd solve = Rmat.transpose().triangularView<Lower>().solve<OnTheLeft>(mcol_k); | 12:03 |
@lambday | aha | 12:04 |
@lambday | and mcol_k is another matrix | 12:04 |
Saurabh7 | yup in my case vector | 12:04 |
@lambday | so your case was to solve Ax=b with A being a lower triangular matrix | 12:05 |
@lambday | b is a vec? | 12:05 |
Saurabh7 | yes b is vec | 12:05 |
@lambday | I see | 12:06 |
@lambday | the blas method is able to do something more I assume | 12:06 |
Saurabh7 | A is cholesky decomposition , we have to update it | 12:07 |
Saurabh7 | yes its similar | 12:07 |
@lambday | Saurabh7: can't you directly use llt solver then? | 12:09 |
Saurabh7 | lambday: no it doesnt do anything more | 12:09 |
Saurabh7 | lambday: i tried llt it was giving some different results | 12:09 |
Saurabh7 | mayb i was doing something wrong | 12:09 |
@lambday | DTRSM solves one of the matrix equations op( A )*X = alpha*B, or X*op( A ) = alpha*B, where alpha is a scalar, X and B are m by n matrices, A is a unit, or non-unit, upper or lower triangular matrix and op( A ) is one of op( A ) = A or op( A ) = A**T. The matrix X is overwritten on B. | 12:10 |
Saurabh7 | lambday: https://eigen.tuxfamily.org/dox/classEigen_1_1TriangularView.html#aa4a11ffc4c8895aaec7ca00668e18c84 | 12:11 |
Saurabh7 | For users coming from BLAS, this function (and more specifically solveInPlace()) offer all the operations supported by the *TRSV and *TRSM BLAS routines. | 12:11 |
Saurabh7 | TRSM so i used this directly | 12:11 |
@lambday | Saurabh7: sorry I was away | 13:02 |
@lambday | Saurabh7: so in order to solve it for matrices X and B, you need to use the eigen equivalent method multiple times | 13:03 |
@lambday | that's what I meant by "able to do more" | 13:03 |
@lambday | but cool that it serves your purpose | 13:03 |
Saurabh7 | lambday: no just using once solves it | 13:03 |
@lambday | Saurabh7: when b is vector, yes | 13:03 |
@lambday | imagine solving it for B (m x n) matrix | 13:04 |
Saurabh7 | ah mayb yes | 13:05 |
@lambday | but for your case, b is vector, so it is perfectly fine | 13:05 |
-!- sanuj [~sanuj@117.203.23.237] has joined #shogun | 13:14 | |
-!- deepak_ [0e8bb448@gateway/web/freenode/ip.14.139.180.72] has quit [Ping timeout: 250 seconds] | 13:26 | |
-!- lupinix_ [~quassel@v22014041761818086.yourvserver.net] has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.] | 13:35 | |
-!- lupinix [~quassel@fedora/lupinix] has joined #shogun | 13:35 | |
Saurabh7 | lambday: there ? | 13:37 |
arianepaola | hello everyone | 13:46 |
-!- deepak_ [0e8bb448@gateway/web/freenode/ip.14.139.180.72] has joined #shogun | 14:01 | |
@lambday | Saurabh7: hey | 14:09 |
@lambday | Saurabh7: actually I'll brb after lunch.. | 14:10 |
* lambday afk | 14:10 | |
sanuj | lisitsyn, hey | 14:51 |
lisitsyn | sanuj: hey | 14:53 |
sanuj | lisitsyn, currently there is only has(tag), shall i also add has(name, type)? | 14:53 |
lisitsyn | sanuj: no I don't think so | 14:54 |
lisitsyn | sanuj: BUT | 14:54 |
lisitsyn | has_type(name, type) | 14:54 |
lisitsyn | :) | 14:54 |
sanuj | lisitsyn, okay, why not has(name, type)? | 14:54 |
lisitsyn | sanuj: well to check whether it has specified type | 14:55 |
lisitsyn | it is not clear for me what has(name, type) checks | 14:56 |
sanuj | lisitsyn, less things for user to remember if we go with has(name,type)... | 14:56 |
lisitsyn | how do you read | 14:57 |
lisitsyn | has("width", Type<float>); | 14:57 |
sanuj | has parameter with name "width" and type = float | 14:58 |
lisitsyn | how's that different from | 14:58 |
lisitsyn | has("width")? | 14:58 |
lisitsyn | as we discussed we can't have float "width" and int "width" | 14:58 |
lisitsyn | to avoid ambiguity | 14:58 |
sanuj | lisitsyn, oh i see :) | 14:59 |
sanuj | lisitsyn, also, if user tries to get a param with correct name but wrong type, it is going to crash | 14:59 |
lisitsyn | yes | 15:00 |
lisitsyn | that's why we'd need | 15:00 |
lisitsyn | some has_type | 15:00 |
lisitsyn | if you want to check | 15:00 |
sanuj | okay | 15:00 |
sanuj | lisitsyn, but do we throw exception in that case? | 15:00 |
sanuj | like SG_ERROR? | 15:01 |
sanuj | lisitsyn, and how to check if a type exists or not? | 15:05 |
sanuj | lisitsyn, by using try and catch? | 15:06 |
sanuj | but which exception shall i catch | 15:07 |
lisitsyn | sanuj: yes you first check has(name) | 15:07 |
lisitsyn | then try to cast | 15:07 |
sanuj | okay | 15:07 |
sanuj | just wanted to confirm this | 15:07 |
lisitsyn | but actually "any" should be able to check *without* exception | 15:07 |
sanuj | cool | 15:07 |
lisitsyn | because exceptions are slower | 15:07 |
sanuj | lisitsyn, how? | 15:07 |
lisitsyn | I think any already has that | 15:08 |
sanuj | lisitsyn, sameType()? | 15:08 |
lisitsyn | sanuj: yes | 15:08 |
lisitsyn | sanuj: sounds like some bad naming of me | 15:09 |
sanuj | lisitsyn, i think it's fine ;) | 15:09 |
lisitsyn | yeah ok easier to keep it | 15:09 |
sanuj | lisitsyn, then i'll also use sameType() for throwing SG_ERROR while getting wrong type | 15:11 |
lisitsyn | where? | 15:12 |
lisitsyn | in get? | 15:12 |
sanuj | yes | 15:12 |
lisitsyn | no need to | 15:12 |
lisitsyn | get of wrong type is already an exceptional case | 15:12 |
lisitsyn | ahh stop I think I got it | 15:12 |
sanuj | lisitsyn, oh, "any" does that | 15:12 |
lisitsyn | let me check your code | 15:12 |
sanuj | i think you are correct | 15:13 |
sanuj | but check the code anyway | 15:13 |
sanuj | :) | 15:13 |
lisitsyn | sanuj: yeah you just recall type | 15:16 |
lisitsyn | and recall type tries to cast it | 15:16 |
lisitsyn | fails otherwise | 15:16 |
lisitsyn | should be ok | 15:16 |
sanuj | lisitsyn, okay | 15:16 |
sanuj | thanks | 15:16 |
* lambday re | 15:21 | |
-!- deepak_ [0e8bb448@gateway/web/freenode/ip.14.139.180.72] has quit [Quit: Page closed] | 15:28 | |
-!- sanuj [~sanuj@117.203.23.237] has quit [Ping timeout: 264 seconds] | 15:30 | |
-!- abc [c133d0aa@gateway/web/freenode/ip.193.51.208.170] has joined #shogun | 15:53 | |
-!- abc is now known as Guest55619 | 15:53 | |
Guest55619 | Hello all | 15:53 |
-!- sanuj [~sanuj@117.203.1.126] has joined #shogun | 16:01 | |
sanuj | lisitsyn, sameType() throws std::logic_error | 16:02 |
lisitsyn | sanuj: yes? | 16:02 |
sanuj | is that fine or do you want it to change it to SG_ERROR? | 16:02 |
lisitsyn | sanuj: I think it's fine although it would make sense to catch it then re-raise as SG_ERROR | 16:05 |
lisitsyn | this way 'any.h' is still generic | 16:05 |
sanuj | lisitsyn, okay | 16:06 |
sanuj | it will be slow | 16:06 |
sanuj | but i think that's fine | 16:06 |
-!- OXPHOS [9d8b131c@gateway/web/freenode/ip.157.139.19.28] has joined #shogun | 16:28 | |
OXPHOS | Hello everyone! | 16:35 |
OXPHOS | log: 6/2/16: linalg, global variable, GPU module; cookbook1. | 16:36 |
arianepaola | hi OXPHOS :-) | 16:37 |
sanuj | arianepaola, OXPHOS hello :) | 16:40 |
arianepaola | hi sanuj | 16:40 |
sanuj | lambday, there? | 16:54 |
-!- travis-ci [~travis-ci@ec2-54-145-106-202.compute-1.amazonaws.com] has joined #shogun | 17:03 | |
travis-ci | it's lambday'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/135044522 | 17:03 |
-!- travis-ci [~travis-ci@ec2-54-145-106-202.compute-1.amazonaws.com] has left #shogun [] | 17:03 | |
sanuj | lisitsyn, there? | 17:06 |
lisitsyn | sanuj: yeas | 17:06 |
sanuj | try{return recall_type<T>(value);}catch(std::logic_error){SG_SERROR("Type is not correct.\n");} | 17:07 |
sanuj | lisitsyn, what's wrong in this ^ | 17:07 |
lisitsyn | well looks correct | 17:08 |
sanuj | getting an error | 17:08 |
sanuj | lisitsyn, error: there are no arguments to ‘SG_SERROR’ that depend on a template parameter, so a declaration of ‘SG_SERROR’ must be available [-fpermissive] | 17:08 |
lisitsyn | well find where SG_SERROR is defined | 17:08 |
lisitsyn | and include it | 17:09 |
sanuj | ohkay | 17:09 |
@lambday | sanuj: yo | 17:23 |
Saurabh7 | how do i make SGVector form std vector anyone | 17:25 |
Saurabh7 | not sharing memory | 17:25 |
@lambday | Saurabh7: you don't want to share memory? | 17:31 |
@lambday | Saurabh7: copy :P | 17:31 |
Saurabh7 | lambday: loop ? | 17:31 |
Saurabh7 | oh | 17:32 |
@lambday | Saurabh7: std::copy(..) | 17:32 |
@lambday | Saurabh7: if sharing memory is fine, then just create a SGVector with refcount=false | 17:32 |
Saurabh7 | still segfault :/ | 17:39 |
@lambday | Saurabh7: can't be if you copy | 17:41 |
Saurabh7 | lambday: std::copy(m_active_set.begin(), m_active_set.end(), active_set.vector); | 17:42 |
Saurabh7 | is this ok ? | 17:42 |
@lambday | Saurabh7: m_active_set is the std::vector? | 17:43 |
Saurabh7 | ye[ | 17:43 |
Saurabh7 | yep | 17:43 |
@lambday | Saurabh7: how much memory did you allocate for "active_set" ? | 17:43 |
@lambday | it has to be "at least" m_active_set.size() | 17:44 |
Saurabh7 | active_set=SGVector<float64_t>(m_num_active); | 17:44 |
@lambday | and m_num_active = m_active_set.size() ? | 17:44 |
Saurabh7 | yes | 17:45 |
@lambday | something is wrong there :) std::copy won't give you segfault unless (a) you don't have sufficient memory in your destination or (b) you're trying to read from uninitialized memory | 17:46 |
Saurabh7 | uninitialized !! | 17:48 |
Saurabh7 | leme check | 17:49 |
-!- travis-ci [~travis-ci@ec2-54-196-75-164.compute-1.amazonaws.com] has joined #shogun | 17:49 | |
travis-ci | it's lambday'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/135057862 | 17:49 |
-!- travis-ci [~travis-ci@ec2-54-196-75-164.compute-1.amazonaws.com] has left #shogun [] | 17:49 | |
@lambday | Saurabh7: in case of any doubt, run valgrind to check what is wrong there | 17:53 |
OXPHOS | lambday: hey. I pushed another commit which works on local | 18:00 |
OXPHOS | lambday: my graphic is Intel integrated so I'll try your way. thx! | 18:02 |
OXPHOS | wiking: there? | 18:02 |
sanuj | lambday, cuda is really hot for deep learning | 18:03 |
sanuj | and cudnn | 18:04 |
sanuj | lisitsyn, this is my get function http://pastebin.com/UqgMSmYJ | 18:06 |
sanuj | lisitsyn, getting: control reaches end of non-void function warning | 18:06 |
sanuj | lisitsyn, i know why is that, what is the best way to fix this | 18:06 |
sanuj | add another catch() without any exception argument with a different error message? | 18:07 |
@wiking | OXPHOS: yes | 18:11 |
@wiking | OXPHOS: wazza/ | 18:11 |
-!- lambday [8028b10a@gateway/web/freenode/ip.128.40.177.10] has quit [Quit: Page closed] | 18:15 | |
-!- lambday [8028b10a@gateway/web/freenode/ip.128.40.177.10] has joined #shogun | 18:17 | |
-!- mode/#shogun [+o lambday] by ChanServ | 18:17 | |
-!- travis-ci [~travis-ci@ec2-54-196-75-164.compute-1.amazonaws.com] has joined #shogun | 18:35 | |
travis-ci | it's lambday'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/135067683 | 18:35 |
-!- travis-ci [~travis-ci@ec2-54-196-75-164.compute-1.amazonaws.com] has left #shogun [] | 18:35 | |
OXPHOS | wiking: if I add cereal save/load method to SGVector, I got errors: | 18:39 |
OXPHOS | wiking: https://gist.github.com/OXPHOS/b9c363c2de6129029428e270f3bc815c | 18:39 |
@wiking | cool | 18:40 |
@wiking | can you send me the patch | 18:40 |
@wiking | plz | 18:40 |
@wiking | via email | 18:40 |
@wiking | i'lll try it locallyh | 18:40 |
OXPHOS | wiking: okay | 18:48 |
sanuj | lambday, there? | 18:55 |
sanuj | lisitsyn, there | 18:55 |
sanuj | ? | 18:55 |
sanuj | wiking, there? | 18:56 |
-!- lambday [8028b10a@gateway/web/freenode/ip.128.40.177.10] has quit [Quit: Page closed] | 18:58 | |
@wiking | got the patch OXPHOS checking... | 18:59 |
@wiking | sanuj: yes | 18:59 |
@wiking | wazza | 18:59 |
sanuj | hey | 18:59 |
arianepaola | hi wiking | 19:00 |
sanuj | wiking, can i use "for ( auto it : map)" to iterate over an unordered_map in sgobject.cpp? | 19:00 |
sanuj | i mean is it slow? | 19:00 |
arianepaola | wiking: I am working on the Python package and will be working also on cookbooks today | 19:00 |
OXPHOS | wiking: thx! | 19:03 |
@wiking | arianepaola: cool thnx! if there's anything you need lemme know | 19:03 |
arianepaola | thanks wiking | 19:04 |
arianepaola | wiking: btw you do not have a written setuptools, distutils documentation? (joking) | 19:04 |
arianepaola | there is hardly any | 19:04 |
-!- sanuj [~sanuj@117.203.1.126] has quit [Quit: Leaving] | 19:07 | |
@wiking | sanuj: arianepaola as there's none | 19:07 |
@wiking | i mean there was never any effore before | 19:07 |
@wiking | to do this | 19:07 |
@wiking | but actually it's a good thing to have | 19:07 |
arianepaola | wiking: :-) | 19:08 |
-!- Guest55619 [c133d0aa@gateway/web/freenode/ip.193.51.208.170] has quit [Quit: Page closed] | 19:09 | |
-!- HeikoS [~heiko@82-132-247-219.dab.02.net] has joined #shogun | 19:41 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 19:41 | |
-!- HeikoS [~heiko@82-132-247-219.dab.02.net] has quit [Remote host closed the connection] | 19:43 | |
-!- HeikoS [~heiko@82-132-247-219.dab.02.net] has joined #shogun | 19:44 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 19:44 | |
-!- HeikoS [~heiko@82-132-247-219.dab.02.net] has quit [Quit: Leaving.] | 19:55 | |
-!- OXPHOS_ [8ca3fe9e@gateway/web/freenode/ip.140.163.254.158] has joined #shogun | 20:29 | |
-!- GandalfTheWizard [~Eva@112.10.170.39] has quit [Read error: Connection reset by peer] | 21:34 | |
-!- GandalfTheWizard [~Eva@112.10.170.39] has joined #shogun | 21:35 | |
-!- OXPHOS_ [8ca3fe9e@gateway/web/freenode/ip.140.163.254.158] has quit [Quit: Page closed] | 21:59 | |
-!- HeikoS [~heiko@82-132-246-136.dab.02.net] has joined #shogun | 22:47 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 22:47 | |
-!- lambday [56a397da@gateway/web/freenode/ip.86.163.151.218] has joined #shogun | 22:50 | |
-!- mode/#shogun [+o lambday] by ChanServ | 22:50 | |
@HeikoS | test | 22:51 |
@HeikoS | wiking: lambday jo | 22:51 |
@HeikoS | Saurabh7: jo | 22:51 |
-!- OXPHOS_ [8ca3fe9e@gateway/web/freenode/ip.140.163.254.158] has joined #shogun | 22:52 | |
@lambday | HeikoS: hola | 22:52 |
@lambday | HeikoS: isle of man? | 22:52 |
@HeikoS | OXPHOS hihi | 22:52 |
lisitsyn | hallo | 22:52 |
lisitsyn | wie gehts | 22:53 |
@HeikoS | lambday: all good over there in London? | 22:53 |
@lambday | HeikoS: same old, same old.. it has just been one day, you see :P | 22:54 |
@HeikoS | lisitsyn: es geht gut und dir=? | 22:54 |
@HeikoS | lambday: portland | 22:54 |
@HeikoS | (close to portland) | 22:54 |
@HeikoS | wanted to check and reply to some emails | 22:54 |
@lambday | HeikoS: BTW I took the bike to gatsby today :D | 22:54 |
@HeikoS | but my phone internet is way too slow | 22:54 |
@HeikoS | cant connect to github | 22:54 |
@HeikoS | lambday: haha indeed | 22:54 |
@HeikoS | hah great! and you even survived it | 22:54 |
@HeikoS | lisitsyn: see, you should come to london to cycle with us | 22:54 |
lisitsyn | oh I actually love to cycle | 22:54 |
lisitsyn | not like hundreds of kms | 22:54 |
lisitsyn | not very used to that :P | 22:55 |
@HeikoS | next too hundreds of buses? | 22:55 |
@lambday | lisitsyn: I cycled 30kms today | 22:55 |
@lambday | :D | 22:55 |
lisitsyn | haha is it common there? | 22:55 |
@lambday | lisitsyn: seems like | 22:55 |
@lambday | HeikoS himself cycles 25+km a day | 22:56 |
lisitsyn | just go there you would have *a lot* of roads to cycle haha | 22:56 |
@HeikoS | yeah thats my commute | 22:56 |
lisitsyn | I live like 12 km from my job | 22:56 |
@HeikoS | ah this internet sucks | 22:57 |
@HeikoS | cannoit write github comments | 22:57 |
lisitsyn | where are you? | 22:57 |
lisitsyn | why portland? :P | 22:57 |
@HeikoS | lambday: can you inform OXPHOS, Saurabh7 that I cannot reply to things today | 22:57 |
@HeikoS | will do tomorrow from some cafe | 22:57 |
lisitsyn | it sounds like the us | 22:57 |
@HeikoS | lisitsyn: portland | 22:57 |
lisitsyn | but you don't like this country ;) | 22:57 |
@HeikoS | good climbing here | 22:58 |
@HeikoS | nono | 22:58 |
@HeikoS | island in south england | 22:58 |
lisitsyn | ohh okay | 22:58 |
@HeikoS | nice cliffs to climb | 22:58 |
lisitsyn | other portland | 22:58 |
@lambday | HeikoS: yeah don't worry, things are good h | 22:59 |
@HeikoS | lambday: feel free to comment under Saurabh7 s PRs as well bte | 22:59 |
@lambday | HeikoS: okay | 22:59 |
@HeikoS | I think we can merge the xvalidation if one of you guys is OK with it | 22:59 |
@lambday | Saurabh7: any luck with the segfault? | 23:00 |
lisitsyn | can one point me | 23:00 |
lisitsyn | to this xval? | 23:00 |
lisitsyn | I am procrastinating my job so I can check | 23:00 |
@HeikoS | lambday: he was also checking low rank updates of cholesky factors | 23:00 |
@lambday | HeikoS: yeah he found a eigen-ized way to do this | 23:01 |
@HeikoS | lisitsyn: https://github.com/shogun-toolbox/shogun/pull/3224 | 23:01 |
lisitsyn | tanks | 23:01 |
lisitsyn | thanks | 23:01 |
@lambday | lisitsyn: you're still @office? | 23:01 |
@HeikoS | lisitsyn: great | 23:01 |
@HeikoS | lisitsyn: so what this is doing: cloning all objects | 23:01 |
@HeikoS | for the multiple threads | 23:01 |
lisitsyn | lambday: no | 23:01 |
@HeikoS | so while memory usage increases quite a bit, runtime goes down a bit | 23:02 |
@HeikoS | not very scalable | 23:02 |
@HeikoS | BUT i think it is a good first step | 23:02 |
@HeikoS | lambday: yeah that what I told him | 23:02 |
@HeikoS | lisitsyn: next step is to use shared memory for the feature data | 23:02 |
@HeikoS | not the object though as we need to assign sugsets | 23:02 |
@HeikoS | lisitsyn: need to sort out the const method mess we have for t hat | 23:02 |
@HeikoS | lisitsyn: merge it btw | 23:03 |
lisitsyn | ok it seems I get the idea | 23:03 |
@HeikoS | the PR is done | 23:03 |
@HeikoS | lambday: can you merge the patch by esben | 23:03 |
lisitsyn | HeikoS: merge what? | 23:03 |
@HeikoS | it should help with meta examples | 23:03 |
@HeikoS | oh and some of the guys had troubles with the meta examples not parsing, even after my patch | 23:03 |
lisitsyn | 3224? | 23:03 |
@HeikoS | that comes from old parser files. they need to remove the output manually (git status reveals it) | 23:03 |
@HeikoS | lisitsyn: yes | 23:04 |
lisitsyn | HeikoS: is your comment on sg_unref already addressed? | 23:04 |
@HeikoS | lisitsyn: Can't tell, ask Saurabh7 | 23:04 |
@lambday | HeikoS: okay I'll check.. I thought it was merged | 23:04 |
@HeikoS | the thing was that clone() gives an object with refcount =1 | 23:04 |
lisitsyn | HeikoS: then we shouldn't merge I geuss | 23:04 |
@HeikoS | lisitsyn: not sure, Saurabh7 told me it leaked a while ago, but I can't check | 23:05 |
lisitsyn | ok HeikoS put a comment | 23:05 |
lisitsyn | (I did I mean) | 23:06 |
@HeikoS | basically, nothing that is cloned needs to be SG_REFed | 23:06 |
@HeikoS | only unrefed | 23:06 |
@HeikoS | kk | 23:06 |
@HeikoS | thats it from me, see you tomorrow | 23:06 |
lisitsyn | cu | 23:06 |
@HeikoS | lisitsyn: cool | 23:06 |
@HeikoS | lisitsyn: Saurabh7 knows what to do there | 23:06 |
@HeikoS | lisitsyn: it also helps time-wise btw | 23:06 |
@HeikoS | lisitsyn: next step: do it for locked training (precomputed kernel matrix) | 23:07 |
@HeikoS | bye | 23:07 |
@HeikoS | lambday: bye | 23:07 |
@lambday | HeikoS: have fun! :) | 23:07 |
@lambday | HeikoS: see you on Monday :) | 23:07 |
@HeikoS | yeah see you Monday :) | 23:07 |
lisitsyn | lambday: https://github.com/shogun-toolbox/shogun/pull/3243/files#r65775409 | 23:07 |
@HeikoS | ill have lots of marks on my hands then hehe | 23:07 |
@lambday | HeikoS: haha | 23:08 |
@lambday | HeikoS: careful :) | 23:08 |
-!- HeikoS [~heiko@82-132-246-136.dab.02.net] has quit [Quit: Leaving.] | 23:08 | |
lisitsyn | ok RED MACHINE going to CRITIQUE your pull requests now | 23:08 |
lisitsyn | :D | 23:08 |
@lambday | lisitsyn: | 23:08 |
@lambday | RED MACHINE :/ | 23:08 |
@lambday | I recall you as blackburn | 23:08 |
@lambday | at least I vagurely remember | 23:09 |
@lambday | vaguely* | 23:09 |
lisitsyn | it seems I used to be that guy! | 23:09 |
@lambday | ah good old days | 23:09 |
lisitsyn | I am teaching some ML now | 23:09 |
@lambday | lisitsyn: class with only one static method? | 23:09 |
@lambday | lisitsyn: teach me TEACH ME!! | 23:09 |
lisitsyn | curious if I can use shogun | 23:10 |
lisitsyn | you know a lot more already | 23:10 |
@lambday | lisitsyn: zero! | 23:10 |
lisitsyn | and you have Heiko by your side ;) | 23:10 |
@lambday | lisitsyn: gotta learn machine learning and c++ before I die | 23:10 |
@lambday | lisitsyn: yeah he's a great teacher.. | 23:11 |
@lambday | super clear knowledge about everything | 23:11 |
@lambday | lisitsyn: but if you're teaching, I won't miss a single class either :D | 23:11 |
@lambday | lisitsyn: about the PR you showed me, what exactly am I suppose to check? | 23:12 |
@lambday | or check it in general? | 23:12 |
lisitsyn | lambday: well just checking if you agree | 23:12 |
lisitsyn | that single class with the single method is a function | 23:12 |
lisitsyn | :D | 23:12 |
@lambday | lisitsyn: haha yeah class with one static method doesn't make any sense.. unless he plans to add more methods to that class in the future | 23:13 |
lisitsyn | class with static methods = namespace | 23:13 |
lisitsyn | as long as we don't use java | 23:13 |
@lambday | lisitsyn: true that! | 23:14 |
lisitsyn | ok I start to like putting some asshole-style comments | 23:14 |
lisitsyn | :D | 23:14 |
@lambday | lisitsyn: haha yeah class with one static method doesn't make any sense.. unless he plans to add more methods to that class in the future | 23:14 |
lisitsyn | https://github.com/shogun-toolbox/shogun/pull/3240/files#r65776397 | 23:15 |
lisitsyn | lambday: this time I realize I talk to a neural network | 23:15 |
lisitsyn | you're busted, the LSTM bot | 23:15 |
@lambday | lisitsyn: errr, what? | 23:16 |
@lambday | ahahaha asshole style coments | 23:16 |
lisitsyn | lambday: passive aggresive! | 23:16 |
lisitsyn | :D | 23:16 |
@lambday | hahahaha | 23:16 |
lisitsyn | wiking is my teacher | 23:16 |
@lambday | people in India are sleepin' :( | 23:18 |
@lambday | hahahaha | 23:18 |
lisitsyn | lambday: but moscow never sleeps (c) | 23:19 |
@lambday | lisitsyn: you're not in samara anymore? | 23:19 |
lisitsyn | for 2 years | 23:19 |
lisitsyn | :D | 23:19 |
@lambday | oh yeah, ever since you took the job at yandex you moved to moscow | 23:19 |
@lambday | lisitsyn: hah when I was in India I never slept by this time :D | 23:20 |
lisitsyn | lambday: I am quite disappointed by the fact I have to get up in 7 hrs from now | 23:21 |
lisitsyn | and no chance I will not stay a few hours more | 23:21 |
-!- sonne|osx [~sonne@x4db3e15e.dyn.telefonica.de] has joined #shogun | 23:21 | |
lisitsyn | and here sonne|osx comes | 23:21 |
@lambday | lisitsyn: and why is that? tomorrow is saturday, remember? | 23:21 |
@lambday | sonne|osx: hello! :D | 23:21 |
lisitsyn | lambday: yes but students want to learn some vowpal wabbit I believe | 23:21 |
lisitsyn | ;) | 23:21 |
@lambday | lisitsyn: wait.. so you do a job mon-fri, and in the weekends you teach | 23:22 |
@lambday | :/ | 23:22 |
lisitsyn | sure! | 23:22 |
@lambday | lisitsyn: teach as in, in some uni or... ? | 23:23 |
lisitsyn | just a sec | 23:23 |
sonne|osx | hey | 23:23 |
lisitsyn | lambday: you've got a link | 23:24 |
@lambday | sonne|osx: how have you been? :D | 23:24 |
OXPHOS_ | so surprised the channel is active at this time | 23:24 |
lisitsyn | OXPHOS_: tell us a good story then ;) | 23:25 |
@lambday | OXPHOS_: hah it happens when me and lisitsyn both are wasted :D | 23:25 |
@lambday | btw OXPHOS_ and OXPHOS do you guys know each other? | 23:25 |
@lambday | xD | 23:25 |
lisitsyn | you must have met each other before | 23:25 |
OXPHOS_ | lisitsyn: I only listen to stories XD | 23:25 |
lisitsyn | you look quite similar | 23:25 |
OXPHOS_ | lambday: wait whos that guy?! | 23:25 |
--- Log closed Sat Jun 04 00:00:24 2016 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!