--- Log opened Sun Jun 25 00:00:28 2017 | ||
-!- TingMiao [uid229534@gateway/web/irccloud.com/x-zmqavftcpyclebqv] has quit [Quit: Connection closed for inactivity] | 02:06 | |
-!- mikeling [uid89706@gateway/web/irccloud.com/x-izuvhokhnizxraux] has joined #shogun | 04:56 | |
mikeling | wiking: ping | 05:04 |
---|---|---|
@wiking | pong | 05:04 |
mikeling | Hi~ Had you read the those messages in personal channel? | 05:04 |
@wiking | yes | 05:04 |
mikeling | About the GMM notebook | 05:04 |
@wiking | but those errors are due to some matplotlib things | 05:05 |
@wiking | mikeling, have you tried running GMM on cloud? | 05:05 |
mikeling | yes, but it doesn't work | 05:05 |
@wiking | lemme check | 05:05 |
@wiking | but i mean can't you debug why? | 05:06 |
@wiking | based on the error maybe some matplotlib thing is wrong | 05:06 |
mikeling | nono, I can do that of course. I just want to make sure if you can get same output like mine | 05:07 |
mikeling | on the main branch | 05:07 |
@wiking | if it is the same on | 05:08 |
@wiking | cloud | 05:08 |
@wiking | then yes | 05:08 |
@wiking | then there's a problem with the functions | 05:09 |
@wiking | http://shogun.ml/notebook/latest/GMM.html | 05:09 |
@wiking | this is not what you get in cloud | 05:09 |
@wiking | so yeah i guess the matplotlib functions | 05:09 |
@wiking | needs to be fixed :( | 05:09 |
mikeling | ok | 05:10 |
mikeling | i will do it | 05:10 |
mikeling | the cloud reset my request yesterday | 05:10 |
@wiking | what do you mean reset your requet? | 05:10 |
@wiking | *request | 05:10 |
mikeling | like: 404.... I guess it's because the firewall in china | 05:11 |
mikeling | :( | 05:11 |
@wiking | could be | 05:11 |
@wiking | i'm not so sure :( | 05:11 |
shogun-buildbot | build #84 of nightly_all is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_all/builds/84 blamelist: Michele Mazzoni <mic.mazzoni@gmail.com>, Viktor Gal <vigsterkr@gmail.com>, Viktor Gal <viktor.gal@maeth.com> | 07:30 |
mikeling | wiking: How to get eigenvectors by DirectEigenSolver? | 10:03 |
mikeling | I only found we can get max and min eigenvalue by our EigenSolver | 10:04 |
micmn | mikeling: check my PR https://github.com/shogun-toolbox/shogun/pull/3843 | 10:07 |
micmn | includes a couple of eigensolvers | 10:07 |
micmn | in linalg | 10:07 |
micmn | maybe useful | 10:07 |
mikeling | micmn: oh, great! | 10:07 |
micmn | can be extended as needed | 10:08 |
micmn | wiking, mikeling: it seems that the SVMOcas test failing has to do with running tests in parallel (-j2) | 10:58 |
micmn | so it should be fixed with mikeling's work on rng... | 10:59 |
mikeling | micmn: could you tell me more about that? why it's related to rng? | 11:02 |
mikeling | thank you | 11:02 |
mikeling | BTW, the eigensolver you write is awesome :) | 11:03 |
micmn | thx but it's just a simple wrapper XD | 11:05 |
micmn | let me try to reproduce it and see if my guess is correct | 11:07 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting] | 11:24 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 11:25 | |
@wiking | shogun-buildbot: force build --branch=develop 'xenial - libshogun - aarch64' | 11:26 |
shogun-buildbot | build #0 forced | 11:26 |
shogun-buildbot | I'll give a shout when the build finishes | 11:26 |
micmn | mikeling: now I was wrong, managed to reproduce it repeating the test | 11:27 |
micmn | but train_feats are different between runs | 11:27 |
shogun-buildbot | build #182 of FC23 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/FC23%20-%20libshogun/builds/182 blamelist: Viktor Gal <viktor.gal@maeth.com> | 11:29 |
@wiking | mmm dman | 11:30 |
@wiking | we need to fix the ocas test asap | 11:30 |
@wiking | otherwise we are in trouble with all the PRs | 11:30 |
@wiking | ok i guess what we can do is | 11:30 |
@wiking | basically fix the PRNG | 11:30 |
@wiking | when the fixtures are created | 11:31 |
@wiking | mikeling, could u do that plz | 11:31 |
@wiking | ? | 11:31 |
@wiking | just fix the seed | 11:31 |
@wiking | before any fixture is generated | 11:31 |
mikeling | wiking: sure | 11:31 |
@wiking | thnx | 11:31 |
mikeling | BTW, I just fix the broken GMM | 11:32 |
mikeling | will commit a pr | 11:32 |
@wiking | k | 11:32 |
shogun-buildbot | Hey! build xenial - libshogun - aarch64 #0 is complete: Failure [failed compile] | 12:11 |
shogun-buildbot | Build details are at http://buildbot.shogun-toolbox.org/builders/xenial%20-%20libshogun%20-%20aarch64/builds/0 | 12:11 |
@wiking | argh | 12:13 |
@wiking | --benchmark_list_tests | 12:13 |
@wiking | virtual memory exhausted: Cannot allocate memory | 12:13 |
@wiking | sukey, flip | 12:13 |
@sukey | (╯°□°)╯︵ ┻━┻ | 12:13 |
@wiking | shogun-buildbot: force build --branch=develop 'xenial - libshogun - aarch64' | 12:15 |
shogun-buildbot | build #1 forced | 12:15 |
shogun-buildbot | I'll give a shout when the build finishes | 12:15 |
@wiking | arg | 12:21 |
@wiking | 1.651g | 12:21 |
@wiking | RAM for LinalgEigenBackend | 12:21 |
@wiking | :( | 12:21 |
@wiking | micmn OXPHOS we need to slice up the linalgbackend file more | 12:22 |
@wiking | it is taking insane amount of RAM to compile that one object | 12:22 |
micmn | k | 12:22 |
@wiking | if you guys have ideas let's dig into it | 12:23 |
@wiking | otherwise it's just | 12:23 |
@wiking | go with the flow :) | 12:23 |
@wiking | micmn, i couldn't make MSVC compiling :( | 12:23 |
@wiking | same story there | 12:23 |
micmn | there was a comment by OXPHOS in my pr | 12:24 |
@wiking | the problem is that i'm trying an aarch64 vps | 12:24 |
@wiking | and it's running out of ram as well | 12:24 |
micmn | https://github.com/shogun-toolbox/shogun/pull/3843#issuecomment-308319590 | 12:24 |
@wiking | micmn, yeah that is reasonable | 12:25 |
micmn | ok I can do that | 12:25 |
@wiking | one can even structure it with namespaces :P | 12:25 |
@wiking | micmn, yeah as i'd love to merge all that stuff you've PRed like a week ago :D | 12:25 |
micmn | more like two XD | 12:26 |
@wiking | but appveyor keeps dying on us | 12:26 |
@wiking | -rw-r--r-- 1 buildslave buildslave 4.8M Jun 25 10:15 /home/buildslave/.ccache/tmp/LinalgBack.stdout.shogun-arm-xenial.31557.m2pHS4.ii | 12:27 |
@wiking | the src itself is 5M | 12:27 |
@wiking | this is LinalgEigenBackend | 12:27 |
micmn | so quick fix: just split that one | 12:28 |
@wiking | hehe | 12:29 |
micmn | find a partition of this set: {add, add_col_vec, dot, element_prod, identity, logistic, matrix_prod, max, mean, range_fill, scale, set_const, sum, sum_symmetric, row/colwise_sum, trace, zero} | 12:40 |
micmn | my guess: | 12:43 |
micmn | {add, add_col_vec, dot, element_prod, matrix_prod} | 12:43 |
micmn | {identity, range_fill, scale, set_const, zero} | 12:43 |
micmn | {logistic, max, mean, sum..., trace} | 12:43 |
micmn | names? :p | 12:45 |
@sukey | [https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3858 synchronized by geektoni | 13:28 |
shogun-buildbot | Hey! build xenial - libshogun - aarch64 #1 is complete: Failure [failed test] | 14:22 |
shogun-buildbot | Build details are at http://buildbot.shogun-toolbox.org/builders/xenial%20-%20libshogun%20-%20aarch64/builds/1 | 14:22 |
-!- TingMiao [uid229534@gateway/web/irccloud.com/x-dhurootcfzsrcxiz] has joined #shogun | 15:06 | |
@sukey | [https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3865 opened by micmn | 15:21 |
@sukey | [https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3858 merged by vigsterkr | 16:02 |
@sukey | [https://github.com/shogun-toolbox/shogun] vigsterkr pushed 7 commits: | 16:02 |
@sukey | https://github.com/shogun-toolbox/shogun/commit/960597095423f16c6e981f9de0044e614891614c | 16:03 |
@sukey | https://github.com/shogun-toolbox/shogun/commit/1b13ce45c0f7931e28d14d0ca207b485e9bfd6fb | 16:03 |
@sukey | https://github.com/shogun-toolbox/shogun/commit/7b343f956dc3045f0af022e052715039aa873851 | 16:03 |
@sukey | https://github.com/shogun-toolbox/shogun/commit/9360256a984a6aee89cff29edad2bfee03b00324 | 16:03 |
@sukey | https://github.com/shogun-toolbox/shogun/commit/308206b41d923720a397743cd61054fc0b8e4525 | 16:03 |
@sukey | https://github.com/shogun-toolbox/shogun/commit/6b78f3273d295c625d019067149f53597a5fc083 | 16:03 |
@sukey | https://github.com/shogun-toolbox/shogun/commit/10bb3bac044dd7e071ccfff91a411abc76f63087 | 16:03 |
@sukey | [https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3866 opened by MikeLing | 16:39 |
@sukey | [https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3866 synchronized by MikeLing | 16:41 |
-!- OXPHOS [92bd305b@gateway/web/freenode/ip.146.189.48.91] has quit [Ping timeout: 260 seconds] | 19:04 | |
-!- mikeling [uid89706@gateway/web/irccloud.com/x-izuvhokhnizxraux] has quit [Quit: Connection closed for inactivity] | 21:43 | |
@sukey | [https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3867 opened by OXPHOS | 23:27 |
--- Log closed Mon Jun 26 00:00:06 2017 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!