--- Log opened Sun May 15 00:00:56 2016 | ||
-!- Saurabh7 [Saurabh7@gateway/shell/panicbnc/x-fyndgtihyyznoojx] has quit [Ping timeout: 260 seconds] | 00:42 | |
-!- Saurabh7 [Saurabh7@gateway/shell/panicbnc/x-sdfzxrbvlkiwkahs] has joined #shogun | 00:43 | |
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has quit [Quit: Leaving.] | 01:23 | |
shogun-buildbot | build #662 of debian wheezy - memcheck is complete: Failure [failed memory check] Build details are at http://buildbot.shogun-toolbox.org/builders/debian%20wheezy%20-%20memcheck/builds/662 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Leon Kuchenbecker <lkuchenb@inf.fu-berlin.de> | 03:24 |
---|---|---|
-!- mizari [~mizari@95-174-213-100.nts.su] has joined #shogun | 04:10 | |
shogun-buildbot | build #1131 of nightly_default is complete: Failure [failed notebooks] Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/1131 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Leon Kuchenbecker <lkuchenb@inf.fu-berlin.de> | 04:54 |
-!- sanuj [75cb1298@gateway/web/freenode/ip.117.203.18.152] has joined #shogun | 05:56 | |
sanuj | wiking: there? | 06:30 |
sanuj | lisitsyn: there? | 06:31 |
-!- sanuj [75cb1298@gateway/web/freenode/ip.117.203.18.152] has quit [Ping timeout: 250 seconds] | 06:39 | |
-!- sanuj [75cb08d8@gateway/web/freenode/ip.117.203.8.216] has joined #shogun | 06:56 | |
-!- sanuj [75cb08d8@gateway/web/freenode/ip.117.203.8.216] has quit [Quit: Page closed] | 07:09 | |
-!- sanuj [~sanuj@117.203.8.216] has joined #shogun | 07:13 | |
-!- sanuj [~sanuj@117.203.8.216] has quit [Client Quit] | 07:16 | |
-!- sanuj [75cb08d8@gateway/web/freenode/ip.117.203.8.216] has joined #shogun | 07:18 | |
sanuj | Saurabh7: hey | 07:21 |
sanuj | what's the timezone for the doodle meeting thing | 07:21 |
Saurabh7 | sanuj: yo | 07:21 |
Saurabh7 | utc i think | 07:21 |
sanuj | okay | 07:22 |
sanuj | wiking: ping me when you are available :) | 07:39 |
@wiking | bping | 07:45 |
@wiking | ah doodle | 07:45 |
@wiking | yes | 07:45 |
@wiking | lemme see didn't fill out | 07:45 |
sanuj | wiking: i grepped "()=0" or "() = 0" | 07:48 |
sanuj | found some files | 07:48 |
sanuj | one is MulticlassMachine.h | 07:49 |
sanuj | the class hierarchy goes like CMulticlassMachine --> CBaseMulticlassMachine --> CMachine --> CSGObject | 07:50 |
sanuj | then shall i add all these classes to the base-shogun list? | 07:50 |
sanuj | wiking: ^^\ | 07:51 |
sanuj | Saurabh7: did you look at these warnings? | 07:58 |
sanuj | http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/2853/steps/libshogun/logs/warnings%20(10)/text | 07:58 |
Saurabh7 | sanuj: no, whats that ? | 08:00 |
Saurabh7 | oh right, dunno seem to be caused by cdflib | 08:01 |
sanuj | yeah | 08:04 |
sanuj | Saurabh7: can we replicate these warnings on our pc? | 08:07 |
sanuj | i was thinking for fixing them right now | 08:07 |
sanuj | i'm building it....let's see if i get the warnings | 08:11 |
Saurabh7 | sanuj: those might be compiler specific | 08:13 |
sanuj | i see | 08:13 |
Saurabh7 | anyways first few look straightforward if we can add parentheses | 08:14 |
Saurabh7 | in expressions arnd && and others | 08:14 |
sanuj | yeah | 08:15 |
sanuj | should be easy to fix | 08:15 |
sanuj | i just want to replicate the warnings on my pc first so i no for sure that it's fixed before i send a PR | 08:16 |
sanuj | know* | 08:17 |
Saurabh7 | adding parentheses is good anyways :) | 08:17 |
sanuj | aight | 08:19 |
-!- sanuj [75cb08d8@gateway/web/freenode/ip.117.203.8.216] has quit [Quit: Page closed] | 08:19 | |
-!- sanuj [~sanuj@117.203.8.216] has joined #shogun | 08:21 | |
sanuj | today is the day to experiment with multiple irc clients :P | 08:21 |
Saurabh7 | irssi is pretty cool i ditched others | 08:24 |
sanuj | too much commands | 08:24 |
sanuj | Saurabh7, does it show logs? | 08:25 |
sanuj | irssi | 08:25 |
Saurabh7 | sanuj: you mean logs from when you were last around ? | 08:26 |
sanuj | yeah | 08:26 |
Saurabh7 | I havent set it up why would you need that tho | 08:27 |
sanuj | oh i was just asking | 08:27 |
sanuj | i'm getting these warnings on my local also | 08:36 |
sanuj | fixing them now | 08:36 |
sanuj | lunch break :D | 09:47 |
-!- sanuj [~sanuj@117.203.8.216] has quit [Ping timeout: 276 seconds] | 09:52 | |
-!- sanuj [~sanuj@59.91.129.165] has joined #shogun | 10:45 | |
sanuj | how to generate documentation | 11:11 |
sanuj | what's the make flag for it | 11:11 |
sanuj | i tried make doc | 11:12 |
sanuj | i think it worked | 11:12 |
-!- mizari [~mizari@95-174-213-100.nts.su] has quit [Read error: Connection reset by peer] | 12:07 | |
-!- mizari [~mizari@95-174-213-100.nts.su] has joined #shogun | 12:09 | |
sanuj | enough warning fixing for one day | 12:45 |
Saurabh7 | :D | 12:54 |
Saurabh7 | ah didn't realise kernels had subsets too | 13:23 |
Saurabh7 | zz | 13:23 |
sanuj | Saurabh7, do you know SWIG? | 13:23 |
Saurabh7 | have to do all over again :) | 13:24 |
Saurabh7 | sanuj: a bit | 13:24 |
sanuj | hmm | 13:24 |
Saurabh7 | what are you trying? | 13:24 |
sanuj | i'm trying to write swig interface for my TAG prototype | 13:25 |
sanuj | let me give you the link | 13:25 |
sanuj | Saurabh7, https://github.com/sanuj/shogun-plugin-arch/tree/master/tags-string | 13:29 |
sanuj | i haven't started writing the interface for this | 13:29 |
sanuj | did it for much easier code and generated python bindings | 13:29 |
sanuj | which had no classes or templates | 13:30 |
sanuj | now i need to write it for my tags-string prototype | 13:30 |
sanuj | i tried to decipher some of the swig code in shogun | 13:31 |
sanuj | but it's huge | 13:31 |
sanuj | don't know where to begin | 13:31 |
Saurabh7 | sanuj, have a look at some of shogun's templated classes then | 13:43 |
sanuj | yes | 13:52 |
sanuj | Saurabh7, which cookbook page are you writing? | 14:17 |
sanuj | i was thinking of writing it on GMMs | 14:18 |
sanuj | need to confirm that with Heiko | 14:18 |
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has joined #shogun | 15:05 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 15:05 | |
sanuj | HeikoS, hi | 15:09 |
sanuj | can i write a cookbook page on GMMs? | 15:09 |
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has quit [Quit: Leaving.] | 15:12 | |
lisitsyn | sanuj: feel free ;) | 15:53 |
sanuj | lisitsyn, okay thanks :) | 15:54 |
sanuj | also i wanted to ask | 15:54 |
lisitsyn | yes? | 15:54 |
sanuj | how to decipher shogun's swig code | 15:54 |
lisitsyn | haha decipher | 15:55 |
sanuj | reading swig docs gets boring...examples will help | 15:55 |
lisitsyn | ok | 15:55 |
lisitsyn | sanuj: did you manage to create some simple example yet? | 15:55 |
sanuj | yes i have done that for python | 15:55 |
sanuj | only functions | 15:55 |
lisitsyn | ok | 15:56 |
sanuj | no classes or templates | 15:56 |
lisitsyn | what about classes? | 15:56 |
sanuj | i need to do that | 15:56 |
sanuj | was reading this for that | 15:56 |
sanuj | http://www.swig.org/Doc3.0/SWIGPlus.html#SWIGPlus_nn38 | 15:56 |
sanuj | i was hoping to find similar implementation in shogun code | 15:56 |
sanuj | to get a better idea | 15:57 |
lisitsyn | sanuj: well we don't do anything but just rename our classes | 15:57 |
lisitsyn | to remove C from CClass | 15:57 |
lisitsyn | CSVM -> SVM | 15:57 |
sanuj | yeah | 15:57 |
lisitsyn | this is why we have %rename | 15:57 |
lisitsyn | but we don't have anything else on classes I think | 15:57 |
-!- mizari [~mizari@95-174-213-100.nts.su] has quit [Read error: Connection reset by peer] | 15:58 | |
sanuj | what are all those .i files in src/interfaces/python_modular | 15:58 |
lisitsyn | sanuj: just the same thing as you write for your example | 15:58 |
lisitsyn | but splitted by modules | 15:58 |
lisitsyn | they are all included so it's just to split things | 15:59 |
sanuj | for my simple swig example | 15:59 |
-!- mizari [~mizari@95-174-213-100.nts.su] has joined #shogun | 15:59 | |
sanuj | i just had to include headers | 16:00 |
sanuj | and declare all the functions | 16:00 |
lisitsyn | yes | 16:00 |
lisitsyn | we have other things but they are unrelated for now | 16:00 |
lisitsyn | just add some class and you should get a wrapper | 16:00 |
lisitsyn | templates are easy as well | 16:00 |
sanuj | i know that templates are instantiated with all types in swig | 16:01 |
lisitsyn | you just have to specialize them with a declaration | 16:01 |
lisitsyn | I think it is better to be explicit | 16:01 |
lisitsyn | say you have template<typename T> func(T); | 16:01 |
lisitsyn | you just put there two lines | 16:01 |
lisitsyn | template<> func(int); | 16:01 |
lisitsyn | template<> func(double); | 16:02 |
sanuj | yes | 16:02 |
lisitsyn | this means you just 'export' two of them | 16:02 |
sanuj | you were anticipating a problem with swig for string based parameters right? | 16:02 |
lisitsyn | well | 16:02 |
lisitsyn | you can't infer the type by string | 16:02 |
lisitsyn | so you would have to return something generic | 16:03 |
lisitsyn | I mean get(string) can't return correct type | 16:03 |
sanuj | okay | 16:03 |
lisitsyn | it can return just some generic holder like any | 16:03 |
lisitsyn | then you have to cast it | 16:03 |
sanuj | so lets go step by step | 16:04 |
lisitsyn | yes | 16:04 |
lisitsyn | 1) generate a wrapper for a class | 16:04 |
lisitsyn | 2) make it a template | 16:04 |
lisitsyn | 3) declare a few specialization of this template | 16:04 |
sanuj | okay, first i'll get this done with my already existing string-tags prototype | 16:05 |
sanuj | then i'll look further into things | 16:06 |
lisitsyn | yeah sure | 16:06 |
lisitsyn | sanuj: to work with string | 16:06 |
lisitsyn | create a class | 16:06 |
lisitsyn | basically you can use any | 16:06 |
lisitsyn | 'any' | 16:06 |
lisitsyn | but add a few *special* impls of cast | 16:06 |
lisitsyn | asInt = as<int> | 16:06 |
lisitsyn | asDouble = as<double> | 16:06 |
lisitsyn | see what I mean? | 16:06 |
sanuj | okay | 16:07 |
lisitsyn | sanuj: do you see why we have to specialize these casts? | 16:07 |
sanuj | no | 16:08 |
sanuj | i don't remember your any implementation | 16:08 |
lisitsyn | sanuj: then ask ;) | 16:08 |
lisitsyn | ok | 16:08 |
lisitsyn | which one did you use? | 16:08 |
lisitsyn | in this repo I see this thing from aer | 16:09 |
sanuj | i think i got it from aer | 16:09 |
lisitsyn | template <typename T> | 16:10 |
lisitsyn | T& as() const | 16:10 |
lisitsyn | but as I said before you have to specialize for swig | 16:10 |
lisitsyn | this would be ambiguous for swig | 16:10 |
lisitsyn | so you have to specialize them with other name | 16:10 |
sanuj | ohkay | 16:11 |
sanuj | i see now | 16:11 |
sanuj | like templates you need to specialize for swig | 16:11 |
lisitsyn | yes | 16:11 |
lisitsyn | but as<int> and as<float> would be just the same for swig | 16:11 |
lisitsyn | so you have to add asInt, asFloat, etc | 16:11 |
sanuj | we need special implementations | 16:11 |
lisitsyn | then it would look like | 16:12 |
lisitsyn | get("my parameter").asInt() | 16:12 |
lisitsyn | this would return int or fail | 16:12 |
sanuj | okay | 16:12 |
sanuj | get("my parameter") would not work with swig | 16:12 |
sanuj | so we need asInt() as well | 16:13 |
lisitsyn | no it would work but | 16:13 |
lisitsyn | you can't do anything with that generic value | 16:13 |
lisitsyn | maybe you can print it | 16:13 |
lisitsyn | but you have to cast it if you need its typed value | 16:13 |
sanuj | okay :) | 16:13 |
lisitsyn | sanuj: ok so the plan is to make things like (10 + object.get("x").asInt()) work | 16:15 |
sanuj | yes | 16:15 |
sanuj | i'll make it work in python | 16:15 |
sanuj | python is fine right? | 16:16 |
lisitsyn | yeah sure | 16:17 |
lisitsyn | sanuj: and lets speed up a bit, please let me know if you're stuck | 16:18 |
sanuj | lisitsyn, yes, that's what i'm trying to do | 16:18 |
sanuj | i got back home yesterday | 16:18 |
lisitsyn | cool | 16:18 |
sanuj | now will be able to give more time to shogun | 16:18 |
sanuj | i had exams and submissions before that | 16:19 |
lisitsyn | I see | 16:19 |
sanuj | lisitsyn, do you also code in c++ for yandex? | 16:29 |
lisitsyn | sanuj: yeah sometimes | 16:29 |
sanuj | okay | 16:29 |
lisitsyn | sanuj: mostly in python though | 16:29 |
lisitsyn | because it is faster to write :) | 16:29 |
sanuj | oh yes | 16:29 |
sanuj | going for dinner | 16:41 |
sanuj | will be back soon | 16:41 |
-!- sanuj [~sanuj@59.91.129.165] has quit [Ping timeout: 276 seconds] | 16:46 | |
-!- sanuj [~sanuj@59.91.129.165] has joined #shogun | 17:32 | |
-!- sanuj [~sanuj@59.91.129.165] has quit [Quit: Leaving] | 19:20 | |
-!- Saurabh7 [Saurabh7@gateway/shell/panicbnc/x-sdfzxrbvlkiwkahs] has quit [Ping timeout: 260 seconds] | 19:22 | |
-!- Saurabh7 [Saurabh7@gateway/shell/panicbnc/x-kxbypopcevntqryc] has joined #shogun | 19:23 | |
-!- leagoetz [~leagoetz@host-92-0-162-192.as43234.net] has joined #shogun | 19:23 | |
-!- leagoetz [~leagoetz@host-92-0-162-192.as43234.net] has quit [Remote host closed the connection] | 19:34 | |
-!- leagoetz [~leagoetz@host-92-0-162-192.as43234.net] has joined #shogun | 19:36 | |
-!- leagoetz [~leagoetz@host-92-0-162-192.as43234.net] has quit [Remote host closed the connection] | 19:39 | |
-!- mizari [~mizari@95-174-213-100.nts.su] has quit [Ping timeout: 260 seconds] | 20:05 | |
-!- leagoetz [~leagoetz@host-92-0-162-192.as43234.net] has joined #shogun | 20:43 | |
-!- leagoetz [~leagoetz@host-92-0-162-192.as43234.net] has quit [Remote host closed the connection] | 21:16 | |
-!- leagoetz [~leagoetz@host-92-0-162-192.as43234.net] has joined #shogun | 21:18 | |
-!- leagoetz [~leagoetz@host-92-0-162-192.as43234.net] has quit [Ping timeout: 260 seconds] | 21:23 | |
-!- Saurabh7 [Saurabh7@gateway/shell/panicbnc/x-kxbypopcevntqryc] has quit [Ping timeout: 260 seconds] | 21:31 | |
-!- Saurabh7 [Saurabh7@gateway/shell/panicbnc/x-udvywipalculvqrs] has joined #shogun | 21:32 | |
--- Log closed Mon May 16 00:00:58 2016 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!