--- Log opened Mon May 30 00:00:17 2016 | ||
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has quit [Quit: Leaving.] | 00:11 | |
-!- leagoetz [~leagoetz@host-92-0-162-192.as43234.net] has joined #shogun | 00:12 | |
shogun-buildbot | build #12 of FC23 - libshogun - aarch64 is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/FC23%20-%20libshogun%20-%20aarch64/builds/12 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Ariane Paola Gomes <arianepaola@users.noreply.github.com> | 00:35 |
---|---|---|
shogun-buildbot | build #2884 of bsd1 - libshogun is complete: Failure [failed configure] Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/2884 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Ariane Paola Gomes <arianepaola@users.noreply.github.com> | 00:39 |
shogun-buildbot | build #22 of xenial - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/xenial%20-%20libshogun/builds/22 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Ariane Paola Gomes <arianepaola@users.noreply.github.com> | 00:39 |
shogun-buildbot | build #13 of FC23 - libshogun - aarch64 is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/FC23%20-%20libshogun%20-%20aarch64/builds/13 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Ariane Paola Gomes <arianepaola@users.noreply.github.com> | 01:17 |
-!- besser82 [~besser82@fedora/besser82] has joined #shogun | 01:56 | |
-!- mode/#shogun [+o besser82] by ChanServ | 01:56 | |
-!- besser82 [~besser82@fedora/besser82] has quit [Ping timeout: 246 seconds] | 02:11 | |
-!- leagoetz [~leagoetz@host-92-0-162-192.as43234.net] has quit [Remote host closed the connection] | 02:13 | |
-!- leagoetz [~leagoetz@host-92-0-162-192.as43234.net] has joined #shogun | 02:15 | |
-!- leagoetz [~leagoetz@host-92-0-162-192.as43234.net] has quit [Remote host closed the connection] | 02:27 | |
shogun-buildbot | build #9 of clang - thread analysis is complete: Failure [failed compile] Build details are at http://buildbot.shogun-toolbox.org/builders/clang%20-%20thread%20analysis/builds/9 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Ariane Paola Gomes <arianepaola@users.noreply.github.com>, OXPHOS <engelzora@gmail.com> | 03:44 |
shogun-buildbot | build #8 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/8 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Ariane Paola Gomes <arianepaola@users.noreply.github.com>, OXPHOS <engelzora@gmail.com> | 03:48 |
shogun-buildbot | build #9 of memleak - valgrind is complete: Failure [failed memory check] Build details are at http://buildbot.shogun-toolbox.org/builders/memleak%20-%20valgrind/builds/9 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Ariane Paola Gomes <arianepaola@users.noreply.github.com>, OXPHOS <engelzora@gmail.com> | 06:15 |
shogun-buildbot | build #1011 of nightly_none is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_none/builds/1011 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Ariane Paola Gomes <arianepaola@users.noreply.github.com>, OXPHOS <engelzora@gmail.com> | 06:25 |
shogun-buildbot | build #1140 of nightly_default is complete: Failure [failed test notebooks] Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/1140 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Ariane Paola Gomes <arianepaola@users.noreply.github.com>, OXPHOS <engelzora@gmail.com> | 07:41 |
shogun-buildbot | build #289 of deb1 - libshogun - PR is complete: Failure [failed git] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun%20-%20PR/builds/289 blamelist: OXPHOS | 07:41 |
shogun-buildbot | build #290 of deb1 - libshogun - PR is complete: Success [build successful] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun%20-%20PR/builds/290 | 07:42 |
-!- sanuj [~sanuj@117.204.246.43] has joined #shogun | 08:02 | |
-!- leagoetz [~leagoetz@host-92-0-162-192.as43234.net] has joined #shogun | 09:44 | |
-!- leagoetz [~leagoetz@host-92-0-162-192.as43234.net] has quit [Client Quit] | 09:47 | |
-!- sanuj [~sanuj@117.204.246.43] has quit [Ping timeout: 246 seconds] | 10:37 | |
-!- sanuj [~sanuj@117.204.246.43] has joined #shogun | 10:58 | |
-!- besser82 [~besser82@fedora/besser82] has joined #shogun | 11:43 | |
-!- mode/#shogun [+o besser82] by ChanServ | 11:43 | |
sanuj | lisitsyn, hey | 11:52 |
lisitsyn | sanuj: hi | 11:53 |
sanuj | lisitsyn, getting a weird error with BaseTag | 11:53 |
sanuj | it is std::unordered_map<BaseTag, Any> | 11:54 |
sanuj | and not std::unordered_map<BaseTag*, Any> | 11:54 |
sanuj | ? | 11:54 |
lisitsyn | yes sure, no pointers | 11:54 |
sanuj | i'll share the error with you | 11:54 |
sanuj | lisitsyn, http://pastebin.com/CXG0Ck2U | 11:56 |
lisitsyn | well you've got to implement hash for basetag | 11:58 |
lisitsyn | check out how hashes are implemented for custom types | 11:59 |
sanuj | lisitsyn, okay, i'll see | 12:00 |
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has joined #shogun | 12:08 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 12:08 | |
@HeikoS | Saurabh7: jo | 12:30 |
shogun-buildbot | build #23 of xenial - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/xenial%20-%20libshogun/builds/23 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Saurabh7 <saurabh.mahindre@gmail.com> | 12:32 |
shogun-buildbot | build #2885 of bsd1 - libshogun is complete: Failure [failed configure] Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/2885 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Saurabh7 <saurabh.mahindre@gmail.com> | 12:33 |
sanuj | HeikoS, hi | 12:33 |
@HeikoS | sanuj: hi there | 12:34 |
sanuj | HeikoS, if you are fine with the code example for svr, i can send the integration data PR | 12:34 |
@HeikoS | just looking at yout cookbook | 12:34 |
sanuj | i saw your comments | 12:34 |
@HeikoS | yeah I think this is good | 12:35 |
@HeikoS | so go for it and update | 12:35 |
Saurabh7 | HeikoS: hi! | 12:35 |
@HeikoS | Saurabh7: hi | 12:35 |
@HeikoS | Saurabh7: saw your comment on x-validation | 12:35 |
Saurabh7 | HeikoS: about the get_features ? | 12:36 |
@HeikoS | Saurabh7: yeah for now, just clone the features | 12:36 |
Saurabh7 | ok | 12:36 |
@HeikoS | the labels are stored in CMachine base class, so you can use a getter for them | 12:36 |
Saurabh7 | double clone :) | 12:36 |
@HeikoS | Saurabh7: not nocessarily | 12:36 |
@HeikoS | but we should check that later | 12:36 |
@HeikoS | so add to some todo list | 12:36 |
@HeikoS | But more important to get something working now | 12:37 |
@HeikoS | and then have it tested a bit | 12:37 |
Saurabh7 | oh yeah only if passed via kernel | 12:37 |
@HeikoS | yeah but as I said, lets postpone this for a bit | 12:37 |
@HeikoS | can avoid these cases later | 12:37 |
@HeikoS | for now lets not care about memory too much | 12:38 |
@HeikoS | point is to speed xvalidation up | 12:38 |
Saurabh7 | HeikoS: okay | 12:38 |
Saurabh7 | how would i test ? | 12:38 |
Saurabh7 | imean comapr outputs for simple cases ? | 12:38 |
Saurabh7 | I have checked crudely but if we ahve to add a unit test | 12:39 |
@HeikoS | Saurabh7: first step: | 12:41 |
@HeikoS | write a unit test for one particular model | 12:41 |
@HeikoS | and compare single thread vs multithread results | 12:41 |
@HeikoS | for say LibSVM | 12:41 |
@HeikoS | then valgrind it | 12:41 |
@HeikoS | then put all that cleaned up in the patch | 12:41 |
@HeikoS | then we will go from there | 12:41 |
Saurabh7 | ok got it | 12:42 |
-!- besser82 [~besser82@fedora/besser82] has quit [Ping timeout: 264 seconds] | 12:48 | |
@HeikoS | Saurabh7: let me know how it goes, it should be really fast to get this working btw, not more than a day | 12:50 |
@HeikoS | , let me know if anything blocks you | 12:50 |
@HeikoS | Saurabh7, sanuj, arianepaola keep in mind the weekly email update on Mondays | 12:54 |
Saurabh7 | HeikoS: i will push this today itself | 12:54 |
@HeikoS | Saurabh7: cool! | 12:54 |
Saurabh7 | HeikoS: what should we plan ideally for this week | 12:54 |
Saurabh7 | other than getting this work ofc | 12:54 |
@HeikoS | Saurabh7: yeah good point! | 12:55 |
@HeikoS | the xvalidation wasn't really planned ;) | 12:55 |
Saurabh7 | ye i know | 12:55 |
Saurabh7 | :) | 12:55 |
@HeikoS | so first of all: let us finish this by tomorrow | 12:55 |
@HeikoS | today: send patch and unit test that is memory clean | 12:55 |
@HeikoS | tomorrow: do a few generalisations and more testing | 12:55 |
@HeikoS | preferably all of this today | 12:55 |
Saurabh7 | HeikoS: ok | 12:55 |
@HeikoS | and then, lets have a look at the plan | 12:55 |
@HeikoS | can you share a a link | 12:56 |
@HeikoS | Saurabh7: because we should pick an algo | 12:57 |
@HeikoS | Saurabh7: I am quite open towards things, what would you prefer? | 12:57 |
@HeikoS | Saurabh7: for that algo we do | 12:57 |
@HeikoS | 1) benchmark to ryans platform | 12:57 |
@HeikoS | 2) optimise | 12:58 |
Saurabh7 | ok i will trying to pick one | 12:59 |
@HeikoS | Saurabh7: can you send me the googledoc with the plan | 12:59 |
Saurabh7 | ok sec | 12:59 |
Saurabh7 | HeikoS: https://docs.google.com/document/d/1RtHundU7jVLMKRdVhBDgJzOSzxn1Y_QIDEKlNidxZ6Y/edit | 13:01 |
@HeikoS | Saurabh7: what about LARS? | 13:02 |
@HeikoS | Saurabh7: not much we can do about KRR | 13:02 |
@HeikoS | (there is a guy working on approximate KRR though) | 13:02 |
@HeikoS | so LARS should be a good one, important algorithm | 13:03 |
Saurabh7 | HeikoS: ye I saw | 13:03 |
@HeikoS | start by comparing this to scikit learn? | 13:03 |
@HeikoS | and sending a patch to the benchmark framework? | 13:03 |
@HeikoS | LARS is also not yet eigen3-ised | 13:03 |
Saurabh7 | LARS i already in i believe | 13:03 |
@HeikoS | so thats some potential | 13:03 |
@HeikoS | yeah I took it from your first week :D | 13:03 |
Saurabh7 | oh ok :D | 13:04 |
Saurabh7 | alright I will profiel it havent done yet using perf | 13:04 |
@HeikoS | yeah cool | 13:04 |
@HeikoS | but lets first finish the x-valiudation | 13:04 |
@HeikoS | and only start lars after merge | 13:05 |
Saurabh7 | LARS | 13:05 |
@HeikoS | also, I will try to merge the kmeans stuff soon. | 13:05 |
@HeikoS | needs a rebase though | 13:05 |
Saurabh7 | HeikoS: yes i was resolving those conflicts right now | 13:05 |
@HeikoS | cool | 13:05 |
@HeikoS | good | 13:05 |
Saurabh7 | they are overwritten anyways | 13:05 |
Saurabh7 | beacuse its moved to KMeansBase | 13:05 |
@HeikoS | I see | 13:06 |
@HeikoS | I updated some of the REQUIRES | 13:06 |
@HeikoS | that is really it | 13:06 |
@HeikoS | oh and method name change | 13:06 |
@HeikoS | ruby friendly | 13:06 |
@HeikoS | Saurabh7: btw one more thing | 13:08 |
@HeikoS | Saurabh7: at the end of gsoc, we want to have a summary for this project: what you improved and how things compare in benchmarkls | 13:08 |
@HeikoS | so we should have like a table | 13:08 |
@HeikoS | with before/after gsoc | 13:08 |
Saurabh7 | ah so i need to keep before ones | 13:09 |
@HeikoS | and also all things you improved should be "proven in benchmark", so there should be a link to the new benchmarlk | 13:09 |
Saurabh7 | yes | 13:09 |
@HeikoS | Saurabh7: not really, we can just use ryans benchmark system | 13:09 |
@HeikoS | and execute it with a revision that is pre-gsoc | 13:09 |
@HeikoS | so integrating changes into this system is key | 13:09 |
@HeikoS | then we can automatically generate numbers | 13:09 |
Saurabh7 | ok | 13:09 |
Saurabh7 | will put everything i touch in it | 13:10 |
@HeikoS | but also keep timings you do | 13:10 |
@HeikoS | great | 13:10 |
Saurabh7 | most are already in | 13:10 |
@HeikoS | then we will have a great output of the gsoc :) | 13:10 |
Saurabh7 | except new ones like decision trees | 13:10 |
@HeikoS | kk | 13:10 |
@HeikoS | LARS is a good candidate | 13:10 |
@HeikoS | very very important algorithm | 13:10 |
@HeikoS | ah multi-core liblinear will be cool as well | 13:11 |
shogun-buildbot | build #14 of FC23 - libshogun - aarch64 is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/FC23%20-%20libshogun%20-%20aarch64/builds/14 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Saurabh7 <saurabh.mahindre@gmail.com> | 13:11 |
shogun-buildbot | build #24 of xenial - libshogun is complete: Failure [failed compile] Build details are at http://buildbot.shogun-toolbox.org/builders/xenial%20-%20libshogun/builds/24 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Sanuj <sanuj.sharma.in@gmail.com> | 13:12 |
shogun-buildbot | build #2886 of bsd1 - libshogun is complete: Failure [failed configure] Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/2886 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Sanuj <sanuj.sharma.in@gmail.com> | 13:12 |
@HeikoS | Saurabh7: ok I will check back in a bit | 13:16 |
sanuj | HeikoS, svr initializes the kernel? | 13:18 |
@HeikoS | sanuj: yeah | 13:18 |
sanuj | it's not done in the previous examples | 13:18 |
@HeikoS | I think the train method does that | 13:18 |
sanuj | https://github.com/shogun-toolbox/shogun/pull/3227/files#diff-80ad9ff4477fe8db88875cfa3cb0bf33L32 | 13:18 |
@HeikoS | for testing, is has to be done by hand if a plain "apply()" is called | 13:18 |
@HeikoS | obvioulsy | 13:18 |
@HeikoS | or you call apply(test_features) | 13:18 |
@HeikoS | this does the same | 13:19 |
@HeikoS | for training, I think all kernel methods do this themselves | 13:19 |
@HeikoS | if not, this is a bug | 13:19 |
@HeikoS | so no need to do it from outside | 13:19 |
@HeikoS | if you find that this is not true, feel free to send a mini patch to rectify that | 13:19 |
@HeikoS | but good that you checked :) | 13:20 |
sanuj | HeikoS, but one needs to give a kernel object to libsvr | 13:20 |
@HeikoS | yeah | 13:20 |
@HeikoS | and then svr->train(feats) does | 13:20 |
@HeikoS | kernel->init(feats, feats) | 13:20 |
@HeikoS | so the two are equivalent: | 13:21 |
@HeikoS | kernel->init(feats, feats) | 13:21 |
@HeikoS | svr->train() | 13:21 |
@HeikoS | and | 13:21 |
@HeikoS | svr->train(feats) | 13:21 |
sanuj | HeikoS, so i just pass width to the kernel object | 13:22 |
sanuj | no need to pass training features | 13:22 |
@HeikoS | yes | 13:22 |
sanuj | HeikoS, the constructor requires cache size also | 13:23 |
@HeikoS | yeah I know | 13:23 |
@HeikoS | wanna send a patch that avoids that? :) | 13:23 |
@HeikoS | I hate this | 13:23 |
sanuj | sure | 13:23 |
@HeikoS | cool :) | 13:23 |
@HeikoS | sanuj: ill check back later today, gotta go now | 13:24 |
@HeikoS | see you :) | 13:24 |
@HeikoS | sanuj: btw make sure to send the monday email soon | 13:24 |
sanuj | HeikoS, yes i will | 13:24 |
@HeikoS | great, thanks | 13:24 |
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has quit [Ping timeout: 246 seconds] | 13:31 | |
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has joined #shogun | 13:33 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 13:33 | |
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has quit [Quit: Leaving.] | 13:39 | |
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has joined #shogun | 13:39 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 13:39 | |
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has quit [Client Quit] | 13:40 | |
shogun-buildbot | build #15 of FC23 - libshogun - aarch64 is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/FC23%20-%20libshogun%20-%20aarch64/builds/15 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Sanuj <sanuj.sharma.in@gmail.com> | 13:52 |
arianepaola | hello everyone | 14:44 |
-!- mizari [~mizari@95-174-213-100.nts.su] has joined #shogun | 15:29 | |
-!- sanuj [~sanuj@117.204.246.43] has quit [Ping timeout: 240 seconds] | 15:42 | |
-!- GandalfTheWizard [~Eva@112.10.170.39] has joined #shogun | 15:51 | |
-!- OXPHOS [9d8b131c@gateway/web/freenode/ip.157.139.19.28] has joined #shogun | 16:09 | |
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has joined #shogun | 16:48 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 16:49 | |
@HeikoS | OXPHOS: jo | 16:55 |
@HeikoS | wiking: hihi | 16:55 |
OXPHOS | HeikoS: Hey! | 16:56 |
@HeikoS | OXPHOS: all good? | 16:56 |
OXPHOS | HeikoS: yes. I have a linalg prototype just ready. I'm gonna bring cookbook back to normal now | 16:57 |
@HeikoS | OXPHOS: great about the linalg prototype | 16:57 |
OXPHOS | HeikoS: And I'll try ifdef SWIG later. I think it gave an error during build | 16:57 |
@HeikoS | OXPHOS: this SWIG macro is really just for SWIG | 16:57 |
@HeikoS | not for the compiler, completely ignored by gcc | 16:58 |
@HeikoS | only swig uses it | 16:58 |
@HeikoS | OXPHOS: btw the other issue got resolved with the data :) | 16:58 |
@HeikoS | lucky you, somebody else merged something afterwards and updated ;) | 16:58 |
OXPHOS | HeikoS: saw it. thx. Actually that change often pops up and everytime I just discard it | 16:59 |
@HeikoS | OXPHOS: it is if you change anything in the "data" dir | 16:59 |
@HeikoS | sometimes, running the tests change things | 16:59 |
@HeikoS | but it should say "dirty" | 16:59 |
@HeikoS | if you haven't commited the changes in data | 17:00 |
OXPHOS | I see | 17:00 |
@HeikoS | OXPHOS: so about the cereal | 17:00 |
@HeikoS | I saw your sgvector patch | 17:00 |
@HeikoS | and I dont get why you copied all the content in there, can you explain? | 17:00 |
@HeikoS | No need that we have all methods I think, just the data structure itself is needed for a prototype, no? | 17:01 |
OXPHOS | you mean copied all 'SGVector'? I just didn't take time to screen for unnecessary ones.. | 17:01 |
OXPHOS | Yes I can definitely remove them all @HeikoS | 17:02 |
@HeikoS | I see | 17:02 |
@HeikoS | okok, just was curios what was your thinking | 17:02 |
@HeikoS | thats fine | 17:02 |
@HeikoS | just saving time | 17:02 |
OXPHOS | :) | 17:03 |
@HeikoS | OXPHOS: so that works? | 17:04 |
OXPHOS | HeikoS yep | 17:05 |
@HeikoS | OXPHOS: great news then, curious what wiking will say | 17:08 |
OXPHOS | HeikoS sth went wrong.. | 17:10 |
OXPHOS | for linalg.. | 17:10 |
arianepaola | HeikoS: is there any specific naming convention for the cookbook pages? | 17:10 |
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has quit [Ping timeout: 244 seconds] | 17:13 | |
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has joined #shogun | 17:32 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 17:33 | |
-!- leagoetz [~leagoetz@host-92-0-162-192.as43234.net] has joined #shogun | 17:37 | |
leagoetz | hi all | 17:38 |
leagoetz | OXPHOS:, arianepaola: can you please also send your short reports? | 17:38 |
leagoetz | thanks! | 17:38 |
OXPHOS | leagoetz: hi! sure. working on it. | 17:39 |
leagoetz | great, thanks! | 17:39 |
leagoetz | OXPHOS: all going well? | 17:40 |
OXPHOS | leagoetz: yep. everything starts to work | 17:40 |
leagoetz | nice :) | 17:40 |
arianepaola | leagoetz: ok | 17:42 |
arianepaola | any idea how to convert DNA to meta language? | 17:43 |
arianepaola | for StringCharFeatures | 17:44 |
-!- yorkerlin [63e6385e@gateway/web/freenode/ip.99.230.56.94] has joined #shogun | 17:50 | |
leagoetz | erm ... | 17:51 |
@HeikoS | arianepaola: hi | 17:51 |
@HeikoS | arianepaola: I think the way to go is to load a dataset | 17:52 |
@HeikoS | arianepaola: have a look in the toy data, there are a few examples | 17:52 |
arianepaola | For the meta: StringCharFeatures do I still need DNA then? | 17:52 |
@HeikoS | arianepaola: since you are correct, the meta language has no way to modify data itself | 17:52 |
@HeikoS | not sure i get what you mean | 17:53 |
arianepaola | https://github.com/shogun-toolbox/shogun/blob/develop/examples/undocumented/python_modular/classifier_ssk_modular.py | 17:53 |
arianepaola | CSVFile f_feats_train("../../data/fm_train_dna.dat") | 17:53 |
arianepaola | StringCharFeatures features_train(f_feats_train, DNA) | 17:53 |
-!- yorkerlin [63e6385e@gateway/web/freenode/ip.99.230.56.94] has quit [Client Quit] | 17:54 | |
@HeikoS | ah I see | 17:54 |
@HeikoS | that should be just an enum | 17:54 |
@HeikoS | which is available | 17:54 |
@HeikoS | see the linear svm example | 17:55 |
@HeikoS | is passes an enum for the solver type | 17:55 |
@HeikoS | the DNA should work similarly | 17:55 |
@HeikoS | arianepaola: let me know if that works | 17:55 |
@HeikoS | if not I can help | 17:55 |
arianepaola | ok, thanks | 17:56 |
@HeikoS | arianepaola: found it? | 18:02 |
arianepaola | HeikoS: should we use the abstract in IEEE referenced papers as reference documentation? e.g. for SSK http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=5375806&url=http%3A%2F%2Fieeexplore.ieee.org%2Fiel5%2F5375738%2F5375739%2F05375806.pdf%3Farnumber%3D5375806 | 18:02 |
@HeikoS | arianepaola: is that the one mentioned in the code? | 18:03 |
@HeikoS | which class | 18:03 |
@HeikoS | ? | 18:03 |
OXPHOS | HeikoS: help | 18:04 |
OXPHOS | HeikoS: for the linalg, in my branch, I did "fetch upstream", and "rebase upstream/feature/linalg_refactor" | 18:04 |
OXPHOS | and end up with tons of commits in the history | 18:04 |
arianepaola | HeikoS: it uses libsvm | 18:04 |
arianepaola | https://github.com/shogun-toolbox/shogun/blob/develop/examples/undocumented/python_modular/classifier_ssk_modular.py | 18:04 |
@HeikoS | check the class for SubsequenceStringKernel | 18:05 |
@HeikoS | there is docs which contain a reference | 18:05 |
@HeikoS | OXPHOS: ok lets see | 18:05 |
@HeikoS | so you had a branch which was forked from develop | 18:05 |
@HeikoS | and then did some commits | 18:05 |
OXPHOS | yes yes | 18:06 |
@HeikoS | and then did rebase upstream feature/linalg_refactor? | 18:06 |
OXPHOS | yes | 18:06 |
@HeikoS | and now? | 18:06 |
OXPHOS | All the old commits popped up in my commits history? | 18:06 |
@HeikoS | arianepaola: I think this example can go under kernels, what do you thkink? | 18:06 |
@HeikoS | which old commits? | 18:06 |
@HeikoS | by whim | 18:06 |
@HeikoS | whom | 18:06 |
arianepaola | HeikoS: yes | 18:07 |
OXPHOS | I just closed a PR. https://github.com/shogun-toolbox/shogun/pull/3229 | 18:07 |
@HeikoS | OXPHOS: ok step 1: make a backup copy | 18:07 |
arianepaola | HeikoS: I found it in shogun/kernel/string/SubsequenceStringKernel.h | 18:07 |
arianepaola | so, I can use the class documentation as reference documentation for the meta, right? | 18:08 |
@HeikoS | OXPHOS: step 2: reset your local feature branch to the upstream feature/linalg_refactor, by doing:ggit reset --hard upstream/feature/linalg_refactor | 18:08 |
@HeikoS | step 3: copy in your changes | 18:08 |
@HeikoS | step 3 commit and send PR | 18:08 |
@HeikoS | OXPHOS: but BACKUP first :) | 18:08 |
@HeikoS | arianepaola: no no | 18:08 |
@HeikoS | arianepaola: you should just use the sgclass macro | 18:08 |
@HeikoS | that links to the class docs we have | 18:08 |
@HeikoS | arianepaola: but you should also put in the bibtex for the research paper mentioned | 18:09 |
@HeikoS | into the cookbook page | 18:09 |
@HeikoS | arianepaola: see what I mean? | 18:09 |
arianepaola | HeikoS: yes | 18:09 |
@HeikoS | arianepaola: cool looking forward to finally have a string kernel example in there | 18:10 |
arianepaola | HeikoS: if you check the ssk source, there is a documentation about ssk | 18:10 |
@HeikoS | arianepaola: in the header you mean? | 18:10 |
arianepaola | yes | 18:11 |
@HeikoS | @brief | 18:11 |
@HeikoS | yes | 18:11 |
@HeikoS | you can take the reference from there | 18:11 |
@HeikoS | i.e. convert it into a bibtex | 18:11 |
@HeikoS | (easy via google scholar, search for the paper name, change settings to include bibtex, download bibtex, polish it and put in the references.bib) | 18:11 |
@HeikoS | arianepaola: you can recycle the math | 18:12 |
@HeikoS | in the cookbook page | 18:12 |
arianepaola | there is no way to extract @brief and reuse it in the cookbook? | 18:12 |
arianepaola | with a macro | 18:12 |
OXPHOS | HeikoS: worked. Thanks! curious why though. rebase worked for me last time | 18:12 |
@HeikoS | OXPHOS: maybe had to be git pull --rebase? | 18:13 |
@HeikoS | but if you confuse git, it will strike back and confuse you in turn ;) | 18:13 |
OXPHOS | HeikoS: this is so true.. | 18:13 |
@HeikoS | OXPHOS: I got confused many times in the past ;) | 18:15 |
@HeikoS | arianepaola: it is not really the point | 18:15 |
@HeikoS | arianepaola: @brief is an interface documentation of code | 18:15 |
@HeikoS | arianepaola: cookbook is illustration how to use it | 18:15 |
@HeikoS | so cookbook should be shorter | 18:15 |
arianepaola | HeikoS: I understand that. But @brief has more than just an interface documentation, as it explains the concept. | 18:16 |
arianepaola | HeikoS: just wondered if it was possible, otherwise you have the same documentation snippet in several places. | 18:16 |
@HeikoS | arianepaola: mmh this is a very good point actually | 18:16 |
@HeikoS | I understand now what you mean | 18:16 |
arianepaola | :-) | 18:16 |
@HeikoS | arianepaola: ok lets think | 18:17 |
@HeikoS | there is still a few differences: | 18:17 |
OXPHOS | HeikoS: It was claimed that you don't have to know coding to use github. This is bs. github is a way harder language. | 18:17 |
@HeikoS | some class docs say things like "you can overload this method to do this" | 18:17 |
@HeikoS | arianepaola: that shouldnt be in cookbook | 18:17 |
arianepaola | yes | 18:17 |
@HeikoS | arianepaola: so not sure how we would link them | 18:17 |
@HeikoS | but would be cool if the cookbook and @brief were in sync | 18:17 |
@HeikoS | removes some redundancy | 18:17 |
@HeikoS | arianepaola: at least for intro of cookbook | 18:18 |
@HeikoS | arianepaola: I guess we have to do this at a later point in time | 18:18 |
@HeikoS | wiking, lisitsyn ^ | 18:18 |
@HeikoS | arianepaola: you want to avoid having to copy the math etc right? | 18:18 |
arianepaola | HeikoS: Another thing that I asked earlier was: HeikoS: is there any specific naming convention for the cookbook pages? | 18:19 |
@HeikoS | it is a little nightmare to unify these things, but would be dead cool | 18:19 |
@HeikoS | arianepaola: yeah | 18:19 |
@HeikoS | arianepaola: follow the ones existing | 18:19 |
arianepaola | e.g. short vs long names | 18:19 |
@HeikoS | ah yeah | 18:19 |
@HeikoS | not sure to be honest | 18:19 |
@HeikoS | suggestions? | 18:19 |
@HeikoS | knn - k_nearest_neightboutrs | 18:20 |
@HeikoS | neighbours | 18:20 |
arianepaola | would be better to use long names though | 18:20 |
@HeikoS | ssk - sub_sequence_kernel | 18:20 |
@HeikoS | I agree | 18:20 |
@HeikoS | please do so | 18:20 |
arianepaola | and the modular part? classifier_ssk_modular.py | 18:20 |
@HeikoS | I will update the existing ones | 18:20 |
@HeikoS | no modular | 18:20 |
@HeikoS | yours would be "kernel/sub_sequence_kernel" | 18:20 |
arianepaola | So this one classifier_ssk_modular.py would go in: kernels/classifier_subsequence_kernel | 18:21 |
arianepaola | ah kernel :-) | 18:21 |
arianepaola | thanks! | 18:21 |
@HeikoS | no classifier I think | 18:21 |
@HeikoS | just kernel | 18:21 |
arianepaola | yeah, right :-) | 18:21 |
@HeikoS | think its better to keep this directory structure flat | 18:21 |
@HeikoS | if we have like "classifier" for each folder, then it gets too messy | 18:22 |
@HeikoS | (and the cmake hacks I wrote only support 1 sub-dir for now ;) ) | 18:22 |
arianepaola | agree | 18:22 |
@HeikoS | arianepaola: btw re your suggestion | 18:22 |
@HeikoS | if you use the sgclass macro | 18:22 |
@HeikoS | there will at least be a link to the class docs | 18:22 |
@HeikoS | we can think about following that up a bit later | 18:22 |
@HeikoS | really good idea, I will think a bit more about it | 18:22 |
-!- sanuj [~sanuj@117.203.14.46] has joined #shogun | 18:25 | |
arianepaola | :-) | 18:26 |
sanuj | GandalfTheWizard, !! | 18:43 |
-!- leagoetz [~leagoetz@host-92-0-162-192.as43234.net] has quit [] | 18:49 | |
-!- mizari [~mizari@95-174-213-100.nts.su] has quit [Ping timeout: 246 seconds] | 18:50 | |
@HeikoS | arianepaola: did it work with the enum? | 18:53 |
arianepaola | HeikoS: I am getting there | 18:55 |
arianepaola | HeikoS: have just fixed an error with sphinx in cmake. the sphinx version from apt did not work, missing theme.conf and the version from pypi was not found by default, as it installs in /usr/local/bin | 18:56 |
@HeikoS | arianepaola: cool, send a separate patch if you like | 18:57 |
arianepaola | yes | 18:57 |
-!- OXPHOS [9d8b131c@gateway/web/freenode/ip.157.139.19.28] has quit [Ping timeout: 250 seconds] | 19:02 | |
-!- OXPHOS [9d8b131c@gateway/web/freenode/ip.157.139.19.28] has joined #shogun | 19:05 | |
arianepaola | HeikoS: https://github.com/shogun-toolbox/shogun/pull/3232 | 19:06 |
@HeikoS | arianepaola: does this add the path to be searched | 19:08 |
@HeikoS | or does it only search there? | 19:08 |
arianepaola | it adds the path | 19:08 |
@HeikoS | i.e. does it still work if there is nothing in local/bin | 19:08 |
@HeikoS | ok | 19:08 |
arianepaola | cmake uses NO_DEFAULT_PATH | 19:08 |
arianepaola | https://cmake.org/cmake/help/v3.0/command/find_program.html | 19:08 |
@HeikoS | yep reading that | 19:08 |
arianepaola | k | 19:08 |
@HeikoS | thanks | 19:09 |
arianepaola | np | 19:09 |
@HeikoS | ill leave it to wiking to merge | 19:09 |
arianepaola | ok | 19:09 |
@HeikoS | OXPHOS: btw there is a preview for PRs | 19:13 |
@HeikoS | where you can see the changed files | 19:13 |
@HeikoS | good to check that before sending -- all of us get an email for every PR | 19:13 |
arianepaola | HeikoS: I am looking at meta/generator Python code | 19:15 |
@HeikoS | yep and? | 19:15 |
arianepaola | I ran types/get_type_list.sh to check that there is StringCharFeatures | 19:16 |
arianepaola | the only string reference in the Python code is StringLiteral | 19:16 |
arianepaola | which does not work | 19:16 |
@HeikoS | arianepaola: ah that is more of a hack | 19:16 |
@HeikoS | you can use the exact same class names as in the python example | 19:16 |
@HeikoS | all classes should be available | 19:16 |
OXPHOS | HeikoS: thx I pulled from Desktop. Because I did remove that file?! anyway I'll pull again. | 19:17 |
@HeikoS | OXPHOS: the new PR is fine and clean | 19:17 |
@HeikoS | nice work btw | 19:17 |
@HeikoS | arianepaola: which example in python are you again looking at? | 19:17 |
arianepaola | HeikoS: I get: Exception: Failed to obtain include path for StringCharFeatures or CStringCharFeatures or StringCharFeatures or CStringCharFeatures | 19:17 |
OXPHOS | HeikoS thanks | 19:17 |
@HeikoS | arianepaola: can you paste the gist | 19:18 |
@HeikoS | of the sg | 19:18 |
@HeikoS | I can check then | 19:18 |
arianepaola | https://github.com/shogun-toolbox/shogun/blob/develop/examples/undocumented/python_modular/classifier_ssk_modular.py | 19:18 |
@HeikoS | and your .sg file | 19:19 |
shogun-buildbot | build #2887 of bsd1 - libshogun is complete: Failure [failed configure] Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/2887 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, OXPHOS <engelzora@gmail.com> | 19:19 |
@HeikoS | arianepaola: we need to find a way to load the string features from file using the shogun interface | 19:19 |
@HeikoS | the python example uses this "load_dna" thing | 19:20 |
@HeikoS | cant do that | 19:20 |
@HeikoS | arianepaola: template<class ST> CStringFeatures<ST>::CStringFeatures(CFile* loader, EAlphabet alpha) | 19:20 |
@HeikoS | there is this | 19:21 |
@HeikoS | so that can load directly from file | 19:21 |
@HeikoS | but your error is caused by something else | 19:21 |
@HeikoS | but need to check the .sg listing | 19:21 |
shogun-buildbot | build #25 of xenial - libshogun is complete: Failure [failed compile] Build details are at http://buildbot.shogun-toolbox.org/builders/xenial%20-%20libshogun/builds/25 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, OXPHOS <engelzora@gmail.com> | 19:22 |
@HeikoS | arianepaola: checking it out | 19:26 |
arianepaola | ok thanks | 19:26 |
@HeikoS | arianepaola: ok this is easy to solve | 19:29 |
@HeikoS | Ill send a fix | 19:29 |
@HeikoS | but then you'll run into another problem ;) | 19:29 |
arianepaola | lol | 19:29 |
@HeikoS | arianepaola: pushed to develop | 19:31 |
@HeikoS | see the change | 19:31 |
arianepaola | ok, thanks | 19:31 |
@HeikoS | ill explain what was going on | 19:31 |
@HeikoS | meta examples have two cases: | 19:31 |
@HeikoS | 1.) swig interface (all languages but cpp) | 19:31 |
@HeikoS | 2.) c++ interface | 19:31 |
@HeikoS | in 2.) some class names are different to 1.) | 19:32 |
@HeikoS | as swig renames a few classes | 19:32 |
@HeikoS | see src/interface/modular/Features.i | 19:32 |
arianepaola | so you overwrite the naming and parameter, right? | 19:32 |
@HeikoS | this is why some classes when used for swig are called differently | 19:32 |
arianepaola | instead of default | 19:32 |
@HeikoS | yes, so for the cpp meta language (that doesnt see shogun through swig), we need to use the real names | 19:32 |
@HeikoS | and since there is no (easy) way to tell the c++ class name from the swig class name, we need to keep a list | 19:33 |
@HeikoS | and put them in explicitly | 19:33 |
@HeikoS | but luckily this will be few cases only | 19:33 |
arianepaola | good | 19:33 |
@HeikoS | arianepaola: so now you get some other errors, but they should be possible to fix | 19:33 |
arianepaola | ok | 19:34 |
arianepaola | thank you HeikoS | 19:34 |
-!- sanuj [~sanuj@117.203.14.46] has quit [Quit: Leaving] | 19:35 | |
arianepaola | HeikoS: hmm, I don't get any errors?!? | 19:36 |
@HeikoS | when compiling meta examples? | 19:36 |
arianepaola | yes | 19:36 |
@HeikoS | well then | 19:38 |
@HeikoS | happy days :) | 19:38 |
@HeikoS | send the PR | 19:38 |
arianepaola | sure, just fixing some math formatting errors | 19:38 |
@HeikoS | cool | 19:38 |
@HeikoS | you will also have to send a data patch | 19:38 |
@HeikoS | see the cookbook readme page | 19:39 |
@HeikoS | for integration testing | 19:39 |
@HeikoS | we can merge the page once travis executed the corresponding integration test for your example | 19:39 |
@HeikoS | (and the page looks good) | 19:39 |
@HeikoS | (first meta example takes longest, should be copy/paste aftrer) | 19:39 |
arianepaola | yeah | 19:39 |
arianepaola | I thought the data was already there, the files that I referenced from the example | 19:40 |
@HeikoS | yeah those are | 19:41 |
@HeikoS | that is the training data | 19:41 |
@HeikoS | but the integration testing data, we generate newly atm for every cookbook page | 19:42 |
@HeikoS | it is just the output of the executable | 19:42 |
@HeikoS | and then we compare future commit's output against that reference that you added | 19:42 |
@HeikoS | and the cool thing with the meta examples is: that we can do that in all target languages, so we are increasing test coverage quite a bit and implicitly testing the swig typemaps | 19:42 |
@HeikoS | since cpp is output without swig | 19:43 |
arianepaola | thanks great | 19:43 |
arianepaola | * that's great | 19:48 |
shogun-buildbot | build #2888 of bsd1 - libshogun is complete: Failure [failed configure] Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/2888 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 19:57 |
OXPHOS | Hey HeikoS, so for the name, | 19:59 |
shogun-buildbot | build #26 of xenial - libshogun is complete: Failure [failed compile] Build details are at http://buildbot.shogun-toolbox.org/builders/xenial%20-%20libshogun/builds/26 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 19:59 |
OXPHOS | Quadratic_discriminant_analysis | 19:59 |
OXPHOS | or quadratic_discriminant_analysis | 19:59 |
@HeikoS | OXPHOS: yeah sorry about not mentioning this earlier, but I only just realised with arianepaola | 19:59 |
@HeikoS | lower case | 19:59 |
@HeikoS | the second | 19:59 |
OXPHOS | okay np | 19:59 |
@HeikoS | thanks! | 19:59 |
@HeikoS | unforutnately | 19:59 |
@HeikoS | the data also needs to be changed | 19:59 |
@HeikoS | argh ;) | 20:00 |
@HeikoS | my bad | 20:00 |
OXPHOS | but hierarchical and kmeans are fine ;) | 20:00 |
@HeikoS | OXPHOS: yes they are | 20:02 |
@HeikoS | phew | 20:02 |
shogun-buildbot | build #16 of FC23 - libshogun - aarch64 is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/FC23%20-%20libshogun%20-%20aarch64/builds/16 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, OXPHOS <engelzora@gmail.com> | 20:03 |
@HeikoS | arianepaola: still no error in the cookbook? | 20:13 |
arianepaola | nope | 20:13 |
arianepaola | HeikoS: I think it is useful to add links to the research papers, when they are available online | 20:20 |
arianepaola | In the bibtex I add: url={http://www.jmlr.org/papers/volume2/lodhi02a/lodhi02a.pdf} | 20:20 |
arianepaola | This generates: [LSST+02]H. Lodhi, C. Saunders, J. Shawe-Taylor, C. Cristianini, and C. Wathins. Text classification using string kernels. The Journal of Machine Learning Research, 2:419–444, 2002. URL: http://www.jmlr.org/papers/volume2/lodhi02a/lodhi02a.pdf. | 20:21 |
@HeikoS | arianepaola: I dont like that too much too be honest | 20:21 |
@HeikoS | these links tend to outdate | 20:21 |
@HeikoS | rather one can just use Google | 20:21 |
arianepaola | hmm, ok | 20:21 |
@HeikoS | also it clutters the bibtex | 20:21 |
arianepaola | I'll remove it then | 20:21 |
@HeikoS | cool thanks | 20:21 |
shogun-buildbot | build #17 of FC23 - libshogun - aarch64 is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/FC23%20-%20libshogun%20-%20aarch64/builds/17 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 20:41 |
arianepaola | HeikoS: Do you mean overflow for strlen. It should be >= 0 in https://github.com/shogun-toolbox/shogun/pull/3215 | 20:47 |
@HeikoS | yes | 20:47 |
@HeikoS | so if it is very large | 20:48 |
@HeikoS | and then casted to signed | 20:48 |
@HeikoS | the result might be negative | 20:48 |
@HeikoS | arianepaola: see what I mean? | 20:49 |
arianepaola | yes | 20:50 |
arianepaola | I suggested to change the type of vlen, but wiking and lisitsyn didn't like it | 20:50 |
@HeikoS | no | 20:51 |
@HeikoS | that will break tons of things | 20:51 |
@HeikoS | you can just do what I said | 20:51 |
@HeikoS | i.e. cast vlen to unsigned | 20:51 |
@HeikoS | that is safe | 20:51 |
@HeikoS | but you have to asser >=0 first | 20:52 |
@HeikoS | I mean it is just a small detail really, but since I spotted it, we should do it properly | 20:52 |
arianepaola | yes | 20:53 |
@HeikoS | arianepaola: just saw your cookbook page | 21:06 |
@HeikoS | arianepaola: let me know if my comments make sense to you | 21:07 |
arianepaola | HeikoS: thanks, will check | 21:08 |
@HeikoS | arianepaola: btw how are the debian packages going? | 21:12 |
@HeikoS | wiking: holidays? | 21:23 |
@HeikoS | :) | 21:23 |
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has quit [Quit: Leaving.] | 22:00 | |
-!- OXPHOS [9d8b131c@gateway/web/freenode/ip.157.139.19.28] has quit [Quit: Page closed] | 22:29 | |
--- Log closed Tue May 31 00:00:19 2016 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!