| --- Log opened Tue Feb 25 00:00:13 2014 | ||
| -!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has quit [Quit: Leaving] | 00:02 | |
| -!- new_lido [~walid@41.218.181.141] has joined #shogun | 00:12 | |
| -!- new_lido [~walid@41.218.181.141] has quit [Ping timeout: 272 seconds] | 00:18 | |
| -!- new_lido [~walid@41.218.172.58] has joined #shogun | 00:30 | |
| -!- bluenotebook [~bluenoteb@115.249.18.25] has joined #shogun | 02:16 | |
| -!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 02:26 | |
| -!- siddharth_sharma [~chatzilla@202.78.173.18] has joined #shogun | 02:45 | |
| -!- siddharth_sharma [~chatzilla@202.78.173.18] has quit [Ping timeout: 272 seconds] | 02:50 | |
| -!- siddharth_sharma [~chatzilla@14.139.243.162] has joined #shogun | 02:57 | |
| -!- siddharth_sharma [~chatzilla@14.139.243.162] has quit [Ping timeout: 246 seconds] | 03:01 | |
| -!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 03:54 | |
| shogun-notifier- | shogun-web: Kevin Hughes :master * d8a0099 / pages/views.py: https://github.com/shogun-toolbox/shogun-web/commit/d8a00994f48f314a2fb85d100f14641fb5e75a73 | 03:54 |
|---|---|---|
| shogun-notifier- | shogun-web: make missing pages 404 not raise | 03:54 |
| -!- new_lido [~walid@41.218.172.58] has quit [Ping timeout: 246 seconds] | 04:42 | |
| -!- new_lido [~walid@41.218.175.225] has joined #shogun | 04:56 | |
| -!- bluenotebook [~bluenoteb@115.249.18.25] has quit [Ping timeout: 272 seconds] | 05:22 | |
| -!- new_lido [~walid@41.218.175.225] has quit [Ping timeout: 272 seconds] | 05:32 | |
| -!- siddharth_sharma [~chatzilla@14.139.243.162] has joined #shogun | 05:43 | |
| -!- new_lido [~walid@41.218.178.138] has joined #shogun | 05:44 | |
| -!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 06:54 | |
| -!- bluenotebook [~bluenoteb@115.249.18.25] has joined #shogun | 07:13 | |
| -!- Bluenotebook_ [~Bluenoteb@106.77.242.38] has joined #shogun | 07:16 | |
| -!- Bluenotebook_ [~Bluenoteb@106.77.242.38] has quit [Client Quit] | 07:17 | |
| -!- siddharth_sharma [~chatzilla@14.139.243.162] has quit [Ping timeout: 246 seconds] | 08:26 | |
| -!- ustc [da161503@gateway/web/freenode/ip.218.22.21.3] has joined #shogun | 08:29 | |
| -!- prams [~prams@122.167.224.162] has joined #shogun | 08:31 | |
| ustc | Hello， Sorry to bother you. Recently I tried to use the Shogun toolbox and met some problems.When I used the command 'result=sg('classify');'I can get the predict label of test samples.But in order to see the overall performance, I want to draw an ROC curve,so the score of each sample is needed. Could you please tell me how to get the score and where can I find the instruction of the function 'sg'where I can see the detials of its input | 08:32 |
| -!- lambday [67157f4e@gateway/web/freenode/ip.103.21.127.78] has joined #shogun | 08:58 | |
| -!- lisitsyn [~lisitsyn@80.252.20.67] has quit [Ping timeout: 240 seconds] | 09:04 | |
| sonne|work | ustc: why don't you use a modular interfaces and not the legacy sg('') crap? | 09:10 |
| ustc | could you give me an example? | 09:11 |
| sonne|work | ustc: http://www.shogun-toolbox.org/doc/en/3.0.0/python_modular_examples.html | 09:23 |
| lambday | sonne|work: hi.. | 09:24 |
| lambday | sonne|work: I want to add automagic tests for clone with CSGObjects... | 09:24 |
| lambday | for vector and matrices for which there was a bug | 09:25 |
| sonne|work | lambday: I thought we had that?! | 09:27 |
| @wiking | we had that | 09:28 |
| sonne|work | lambday: look at the jinja2 files | 09:28 |
| sonne|work | wiking: but? | 09:28 |
| @wiking | i mean clone() autogen | 09:28 |
| lambday | sonne|work: we have it for just the objects - https://gist.github.com/lambday/8996223 | 09:28 |
| @wiking | i dont know if there's any but | 09:28 |
| lambday | sonne|work: wiking yeah so this is what I want to add -for all SGObject classes | 09:28 |
| @wiking | {% set ignores = ['Set', 'ParseBuffer', 'TreeMachine'] %} | 09:28 |
| @wiking | only these are in ignore... | 09:29 |
| @wiking | +GUI* | 09:29 |
| lambday | wiking: yeah... | 09:29 |
| @wiking | so i dont understand what you are talking about | 09:29 |
| lambday | wiking: please have a look at the gist.. when we create a vector of SGObjects, we had this bug (which caused any class having a CDynamicObjectArray filled with objs failed in clone) | 09:30 |
| lambday | sonne|work: wiking: the 'calloc' thing that we talked about - after talking to Heiko I changed it | 09:31 |
| lambday | sonne|work: wiking: and I am planning to add unit-test for that... so shall I just modify the jinja file to have this test too? | 09:32 |
| @wiking | lambday: well apparently this has nothing to do with jinja | 09:32 |
| @wiking | as it's not in jinja2 | 09:32 |
| @wiking | sonne|work: SGVector is not in class_list.cpp ? | 09:33 |
| lambday | wiking: I think not - only 'C' classes are there I guess | 09:33 |
| lambday | wiking: and this is about CDynamicObjectArray - filled with objects - when its empty it copies just fine | 09:34 |
| -!- bluenotebook [~bluenoteb@115.249.18.25] has quit [Ping timeout: 244 seconds] | 09:34 | |
| lambday | wiking: like in the gist I tested for an array of CStatistics class, I wanted to make sure it works for all C-classes! | 09:35 |
| @wiking | heh yeah SGVector nor SGMatrix is in not in class_list but it's no wonder because they are inherited fro SGReferencedData and not from SGObject... | 09:35 |
| @wiking | so actually SGVector.clone() should be some custom function in SGVector | 09:35 |
| @wiking | indeed it is | 09:36 |
| @wiking | SGVector.cpp:263 | 09:36 |
| lambday | wiking: ummm.. I am not talking about SGVector clone :( | 09:36 |
| @wiking | lambday: well you cannot have other clone of SGVector... | 09:36 |
| @wiking | it's not a CSGObject... | 09:37 |
| lambday | wiking: I am talking about CDynamicObjectArray.clone when the array *has* some elements already | 09:37 |
| lambday | in fact any array for that matter which has CSGObject type objects inside it | 09:37 |
| @wiking | [ RUN ] SGObject.clone_equals_DynamicObjectArray | 09:38 |
| @wiking | [ OK ] SGObject.clone_equals_DynamicObjectArray (9 ms) | 09:38 |
| @wiking | [ RUN ] SGObject.equals_DynamicObjectArray_equal | 09:38 |
| @wiking | [ OK ] SGObject.equals_DynamicObjectArray_equal (7 ms) | 09:38 |
| @wiking | [ RUN ] SGObject.equals_DynamicObjectArray_equal_after_resize | 09:38 |
| @wiking | [ OK ] SGObject.equals_DynamicObjectArray_equal_after_resize (872 ms) | 09:38 |
| @wiking | [ RUN ] SGObject.equals_DynamicObjectArray_different | 09:38 |
| @wiking | [ OK ] SGObject.equals_DynamicObjectArray_different (6 ms) | 09:38 |
| @wiking | so if these tests are not covering it | 09:38 |
| lambday | wiking: yeah but its for empty arrays | 09:38 |
| @wiking | then you can add your own tests | 09:38 |
| @wiking | but it has nothing to do with jinja2 | 09:38 |
| @wiking | nor automagic testing | 09:38 |
| lambday | wiking: well the thing I had in mind is to check for Parameter::copy() for an array of CSGObject types for *all* CSGObjects there - which easily covers the DynamicObjectArray case here! | 09:42 |
| lambday | wiking: kind of like the one I have here for CStatistics - https://gist.github.com/lambday/8996223 | 09:43 |
| lambday | afk | 09:44 |
| lambday | re | 09:48 |
| -!- new_lido [~walid@41.218.178.138] has quit [Ping timeout: 272 seconds] | 09:58 | |
| -!- ustc [da161503@gateway/web/freenode/ip.218.22.21.3] has quit [Quit: Page closed] | 10:01 | |
| sonne|work | lambday: well maybe you can extend the tests to make them work with non-empty data so they do 'proper' testing | 10:11 |
| lambday | sonne|work: yes.. but all these tests are generated using jinja2 - so, if we want to make sure the array thing works for all 'C' classes, we gotta add test in the jinja2 file, no? | 10:14 |
| lambday | sonne|work: but since it says its auto generated from a template, I am afraid to make any changes to it! I cannot find the template | 10:15 |
| -!- hushell [~hushell@c-50-188-151-130.hsd1.or.comcast.net] has quit [Quit: WeeChat 0.3.7] | 10:26 | |
| -!- vivekrai [uid10305@gateway/web/irccloud.com/x-ubzxykedwjldcxvl] has joined #shogun | 10:36 | |
| -!- lisitsyn [~lisitsin@mxs.kg.ru] has joined #shogun | 10:43 | |
| -!- bluenotebook [~bluenoteb@202.78.175.199] has joined #shogun | 11:10 | |
| @besser82 | sonne|work: do you have any constructive idea about this? ---> http://ur1.ca/gp1mc | 11:33 |
| sonne|work | besser82: no idea - hushell would know | 11:36 |
| @besser82 | sonne|work: unfortunately he isn't here :( | 11:37 |
| @besser82 | sonne|work: we definetively need to run testsuite on iX86 as well... | 11:54 |
| @besser82 | sonne|work: during development... | 11:54 |
| -!- dhruv13 [~dhruv13@84.64.12.242] has joined #shogun | 11:57 | |
| lambday | sonne|work: added automatic test for DynamicObjectArray for all classes.... | 12:09 |
| lambday | sonne|work: could you please check to see if this fits - sending a PR.. | 12:12 |
| -!- makokal [~batman@kgb.informatik.uni-freiburg.de] has joined #shogun | 12:18 | |
| -!- bluenotebook [~bluenoteb@202.78.175.199] has quit [Ping timeout: 272 seconds] | 12:31 | |
| -!- zxtx [~zv@76.91.177.87] has joined #shogun | 12:38 | |
| -!- dhruv13 [~dhruv13@84.64.12.242] has quit [Ping timeout: 244 seconds] | 13:21 | |
| -!- dhruv13 [~dhruv13@84.64.12.242] has joined #shogun | 13:25 | |
| -!- HeikoS1 [~heiko@pat-191-250.internal.eduroam.ucl.ac.uk] has joined #shogun | 13:36 | |
| -!- dhruv13 [~dhruv13@84.64.12.242] has quit [Ping timeout: 242 seconds] | 13:44 | |
| -!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 14:16 | |
| shogun-notifier- | shogun: lambday :develop * eaa2d2d / examples/undocumented/python_modular/kernel_ssk_string_modular.py: https://github.com/shogun-toolbox/shogun/commit/eaa2d2d98cdb43f1569ff08a629ce0ba42095aab | 14:16 |
| shogun-notifier- | shogun: fix mistake in ssk kernel example | 14:16 |
| shogun-notifier- | shogun: lambday :develop * 1d45e03 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/1d45e032d4efef90446a3611ca02a65a9bd191a1 | 14:16 |
| shogun-notifier- | shogun: fix uninitialized memory bug in TParameter::copy | 14:16 |
| shogun-notifier- | shogun: Heiko Strathmann :develop * e171a2b / / (4 files): https://github.com/shogun-toolbox/shogun/commit/e171a2b9582d4e6fea8b32b3687579a570f21ceb | 14:16 |
| shogun-notifier- | shogun: Merge pull request #1892 from lambday/develop | 14:16 |
| shogun-notifier- | shogun: | 14:16 |
| shogun-notifier- | shogun: fix uninitialized memory in TParameter::copy and added unit-test | 14:16 |
| -!- dhruv13 [~dhruv13@84.64.12.242] has joined #shogun | 14:32 | |
| -!- prams [~prams@122.167.224.162] has quit [Ping timeout: 272 seconds] | 14:58 | |
| -!- dhruv13 [~dhruv13@84.64.12.242] has quit [Ping timeout: 246 seconds] | 15:01 | |
| -!- dhruv13 [~dhruv13@84.64.12.242] has joined #shogun | 15:02 | |
| -!- prams [~prams@122.167.224.162] has joined #shogun | 15:10 | |
| -!- batman [~batman@kgb.informatik.uni-freiburg.de] has joined #shogun | 15:20 | |
| -!- makokal [~batman@kgb.informatik.uni-freiburg.de] has quit [Write error: Broken pipe] | 15:20 | |
| -!- pickle27 [~kevin@192-0-136-118.cpe.teksavvy.com] has joined #shogun | 15:26 | |
| pickle27 | sonney2k: can you deploy the site | 15:33 |
| pickle27 | added 2 fixes | 15:33 |
| -!- pickle27 [~kevin@192-0-136-118.cpe.teksavvy.com] has quit [Quit: Leaving] | 15:40 | |
| -!- dhruv13 [~dhruv13@84.64.12.242] has quit [Ping timeout: 244 seconds] | 15:49 | |
| -!- bluenotebook [~bluenoteb@115.249.18.25] has joined #shogun | 16:04 | |
| -!- [paladin] [73f91219@gateway/web/freenode/ip.115.249.18.25] has joined #shogun | 16:04 | |
| prams | lisitsyn: der..? need a little help.. | 16:26 |
| prams | what is the difference between SGMatrix and MatrixXd..? | 16:28 |
| sonne|work | prams: MatrixXd is an eigen3 matrix, SGMatrix is a reference count enabled matrix of shogun | 16:50 |
| -!- adrin [~adrin@guest-90.mpi-inf.mpg.de] has quit [Quit: Leaving] | 16:50 | |
| prams | thanks sonne|work . but i cant find its implementation in eigen3.h . where should i look..? | 16:52 |
| sonne|work | prams well eigen3 documentaion in the web | 16:52 |
| prams | http://www.shogun-toolbox.org/doc/en/current/eigen3_8h.html its not here. am i looking in the wrong place..? | 16:55 |
| -!- lisitsyn1 [~lisitsyn@80.252.20.67] has joined #shogun | 17:14 | |
| -!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 17:16 | |
| -!- lisitsyn1 [~lisitsyn@80.252.20.67] has quit [Client Quit] | 17:18 | |
| -!- bluenotebook [~bluenoteb@115.249.18.25] has quit [Ping timeout: 272 seconds] | 17:29 | |
| -!- dhruv13 [~dhruv13@84.64.15.69] has joined #shogun | 17:36 | |
| -!- [paladin] [73f91219@gateway/web/freenode/ip.115.249.18.25] has quit [Ping timeout: 245 seconds] | 17:44 | |
| -!- [paladin] [73f91219@gateway/web/freenode/ip.115.249.18.25] has joined #shogun | 17:45 | |
| -!- siddharth_sharma [~chatzilla@14.139.243.162] has joined #shogun | 18:04 | |
| -!- dhruv13 [~dhruv13@84.64.15.69] has quit [Ping timeout: 264 seconds] | 18:10 | |
| siddharth_sharma | hello all, i am interested in participation in gsoc in shogun and i have seen problems statements but i am confused on whether to learn about problems in detail first or i should first try to compile , configure and get hang of the code first | 18:15 |
| HeikoS1 | siddharth_sharma: hi there | 18:43 |
| HeikoS1 | siddharth_sharma: welcome here and thanks for your interest | 18:44 |
| HeikoS1 | siddharth_sharma: read the howto that we wrote in order to get started | 18:44 |
| siddharth_sharma | HeikoS1: thank you for your help i will follow the first steps from guide | 18:48 |
| -!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has joined #shogun | 18:58 | |
| -!- mode/#shogun [+o iglesiasg] by ChanServ | 18:58 | |
| @iglesiasg | sonney2k, wiking, besser82, HeikoS1, lisitsyn do we have to fill in this form to request a mentor role? | 19:01 |
| @iglesiasg | http://www.google-melange.com/gsoc/profile/register/org_member/google/gsoc2014 | 19:01 |
| @besser82 | iglesiasg: yes ;) | 19:01 |
| @iglesiasg | I thought that was for the whole organization | 19:01 |
| @iglesiasg | besser82, all right, thank you man! | 19:01 |
| @besser82 | iglesiasg: np, welcome :) | 19:02 |
| -!- lisitsyn1 [~lisitsyn@80.252.20.67] has joined #shogun | 19:17 | |
| lisitsyn1 | getting dense! | 19:22 |
| @iglesiasg | lisitsyn1, haha yeah | 19:23 |
| lisitsyn1 | iglesiasg: how can you dare to ask for mentor connection | 19:25 |
| lisitsyn1 | :D | 19:25 |
| @iglesiasg | lisitsyn1, I like to live to the limit ÑP | 19:25 |
| @iglesiasg | ooops wrong map, :P | 19:25 |
| lisitsyn1 | iglesiasg: you go beyond! | 19:25 |
| lisitsyn1 | iglesiasg: http://www.youtube.com/watch?v=07So_lJQyqw | 19:26 |
| @iglesiasg | lisitsyn, haha wtf | 19:26 |
| naywhayare | HeikoS asked me to make a mentor connection (for the benchmarking project I assume), so I submitted a connection with a Nigerian email scam as text. it seemed like the reasonable thing to do. | 19:34 |
| HeikoS1 | naywhayare: thats exactly what I had in mind ;) | 19:36 |
| lisitsyn1 | naywhayare: you managed to get into the spam folder! yaah! | 19:48 |
| naywhayare | lisitsyn1: did I really? excellent. :) | 19:48 |
| lisitsyn1 | naywhayare: I'll call you wumi from now on | 19:49 |
| naywhayare | fine by me, I'll have to add it to the list of things my IRC client highlights | 19:50 |
| lisitsyn1 | naywhayare: what's up with HASSAN? | 19:50 |
| naywhayare | I have no idea -- I didn't actually read what I pasted before I sent it | 19:51 |
| naywhayare | but I figured there might be a chance that one of you would wire me some large amount of money or something, so hey, worth a shot, right? | 19:51 |
| -!- vivekrai [uid10305@gateway/web/irccloud.com/x-ubzxykedwjldcxvl] has quit [Ping timeout: 245 seconds] | 20:04 | |
| -!- vivekrai [uid10305@gateway/web/irccloud.com/x-aeqxyuzwfnsqtjqx] has joined #shogun | 20:04 | |
| -!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has quit [Quit: Leaving] | 20:10 | |
| HeikoS1 | naywhayare: I almost sent you my parents house | 20:12 |
| naywhayare | damn! so close! | 20:12 |
| lisitsyn1 | naywhayare: yeah it is worth the thing | 20:12 |
| -!- zxtx [~zv@76.91.177.87] has quit [Ping timeout: 246 seconds] | 20:16 | |
| -!- bumblebee [~quassel@202.78.172.162] has joined #shogun | 20:24 | |
| -!- bumblebee is now known as Guest22788 | 20:24 | |
| -!- Guest22788 [~quassel@202.78.172.162] has quit [Client Quit] | 20:24 | |
| -!- bumblebee_ [~quassel@202.78.172.162] has joined #shogun | 20:24 | |
| -!- prams [~prams@122.167.224.162] has left #shogun [] | 20:31 | |
| -!- kaushik94 [~kaushik@14.139.122.114] has joined #shogun | 20:31 | |
| @besser82 | sonney2k: around?!? | 20:42 |
| -!- vijay05_ [0e8b5206@gateway/web/freenode/ip.14.139.82.6] has joined #shogun | 20:57 | |
| vijay05_ | Hi lisitsyn | 21:01 |
| vijay05_ | I'm interested to know more about dictionary learning project in shogun. Could you please elaborate on this? | 21:02 |
| HeikoS1 | lisitsyn, lisitsyn1 it would also be great if you could update the project description and make it more detailed in fact ;) | 21:02 |
| lisitsyn1 | HeikoS1: true | 21:02 |
| lisitsyn1 | vijay05_: you may start with checking what scikit-learn has http://scikit-learn.org/stable/modules/decomposition.html | 21:03 |
| vijay05_ | I'm little bit aware of sparse coding and dictionary learning | 21:04 |
| @besser82 | HeikoS1: how do I add another project for GSoC on the list? | 21:04 |
| lisitsyn1 | vijay05_: oh then you can elaborate on this ;) | 21:04 |
| @besser82 | HeikoS1: I wanted to do some buildsys / cmake action, too :) | 21:04 |
| HeikoS1 | besser82: write a nice proposal on the googledoc, ask people to proofread, and then kindly ask fernando to add it to the site, he know how to do that | 21:05 |
| HeikoS1 | besser82: that would be great. Improve cmake in general, what do you have in mind? | 21:05 |
| @besser82 | HeikoS1: buildspeed, modularization, many more | 21:05 |
| HeikoS1 | besser82: keep in mind some things might be redundant with lisitsyn modularisation project | 21:06 |
| HeikoS1 | besser82: but yeah, good stuff! | 21:06 |
| @besser82 | HeikoS1: thx :D | 21:06 |
| vijay05_ | Lisitsyn1, I basically worked on image denoising and documenrt restoration using dictionary learning algorithms. I improved upon the approach introduced by Julian Mairal and Michal Elad. http://www.cs.technion.ac.il/~elad/publications/journals/2006/KSVD_Color_IEEE_TIP.pdf | 21:07 |
| lisitsyn1 | vijay05_: good to know! | 21:07 |
| vijay05_ | Lisitsyn1, they even have opensource implementation of dictionary learning http://spams-devel.gforge.inria.fr | 21:08 |
| @besser82 | HeikoS1: but i think the modularizaion thing can be somehow go hand in hand with lisitsyn | 21:08 |
| lisitsyn1 | vijay05_: looks good! we can actually think of adding an interface to this think | 21:09 |
| lisitsyn1 | thing* | 21:09 |
| vijay05_ | Cool | 21:09 |
| vijay05_ | Thats what. It does not make much sense to code again when they have open sourced | 21:10 |
| lisitsyn1 | vijay05_: absolutely | 21:10 |
| lisitsyn1 | no intention to recode anything yet | 21:10 |
| lisitsyn1 | only if it is really necessary | 21:10 |
| vijay05_ | So could you elaborate on what is the expectation here? | 21:10 |
| lisitsyn1 | vijay05_: sure - it makes sense to make shogun able to do at least some basic dictionary learning | 21:11 |
| lisitsyn1 | vijay05_: mostly for features - not some reconstruction or whatever | 21:12 |
| vijay05_ | Sure. | 21:12 |
| vijay05_ | I'm interested to work on this. | 21:12 |
| lisitsyn1 | vijay05_: that's good to know | 21:12 |
| vijay05_ | How do I get started on this? | 21:12 |
| lisitsyn1 | vijay05_: lets think about it | 21:12 |
| HeikoS1 | besser82: yeah, push sergey, he is lazy :D | 21:12 |
| @besser82 | HeikoS1: I know :-P | 21:13 |
| @besser82 | HeikoS1: me, too ;) | 21:13 |
| lisitsyn1 | HeikoS1: I wasn't before :D | 21:13 |
| lisitsyn1 | and one day I woke up | 21:13 |
| lisitsyn1 | and hey what the heck | 21:13 |
| -!- besser82 is now known as McLazy | 21:13 | |
| HeikoS1 | lisitsyn1: I know, that why we should kick you ;) | 21:13 |
| HeikoS1 | as in kick your ass | 21:13 |
| lisitsyn1 | kick me out :D | 21:14 |
| -!- McLazy is now known as besser82 | 21:14 | |
| -!- [paladin] [73f91219@gateway/web/freenode/ip.115.249.18.25] has left #shogun [] | 21:14 | |
| lisitsyn1 | vijay05_: okay | 21:16 |
| lisitsyn1 | vijay05_: so what about thinking how do we design basic classes for that | 21:16 |
| -!- kaushik94 [~kaushik@14.139.122.114] has quit [Read error: Connection reset by peer] | 21:16 | |
| lisitsyn1 | so basically how would you like if you try to design some api for this thing | 21:16 |
| lisitsyn1 | it should be quite general I think - not just about dictionary learning but some process of learning features | 21:17 |
| vijay05_ | Ok. | 21:18 |
| vijay05_ | I look at this feature learning basically as a matrix factorization problem Y=AX under some conditions | 21:19 |
| lisitsyn1 | yeah so | 21:19 |
| lisitsyn1 | we get X from Y | 21:19 |
| lisitsyn1 | we have CConverter for that | 21:19 |
| vijay05_ | So this feature learning api basically can perform all kinds of feature learning including PCA, LDA, NMF, dictioanry learning, etc. | 21:20 |
| vijay05_ | end user can select the method he/she wants based on his requirement. | 21:20 |
| vijay05_ | since, all these approaches are inter-related. | 21:21 |
| lisitsyn1 | vijay05_: this matches what shogun 'converter' does | 21:23 |
| lisitsyn1 | vijay05_: you can also forward things and just implement some code interfacing some NMF implementation | 21:24 |
| vijay05_ | Sorry I didnot get you | 21:26 |
| vijay05_ | lisitsyn1: Do you want me to implement the interface for Spams library NMF implementation? | 21:28 |
| lisitsyn1 | vijay05_: I mean you could start with writing some class that fits shogun but delegates real implementation to some library | 21:28 |
| vijay05_ | Ok. Thanks. I will get started on this. I will ask you if I have any doubts. | 21:29 |
| vijay05_ | Thanks a lot | 21:29 |
| lisitsyn1 | vijay05_: we would need to solve the dependency thing somehow though | 21:30 |
| lisitsyn1 | vijay05_: are you versed with cmake things? | 21:30 |
| vijay05_ | not really. have used it for just buliding things. but do not know to play around with it | 21:31 |
| lisitsyn1 | vijay05_: so basically we'd need some detection code | 21:32 |
| lisitsyn1 | which tests whether we have spams somewhere in path | 21:32 |
| lisitsyn1 | vijay05_: could you please also check if it has proper C++ interface | 21:32 |
| lisitsyn1 | I see it has matlab interface but not sure about C++ | 21:32 |
| lisitsyn1 | it could be wrapped up way too much - I have seen crazy things like that before :) so just to be sure | 21:33 |
| vijay05_ | lisitsyn1: You are right. It has a matlab wrapper. | 21:35 |
| vijay05_ | lisitsyn1: But can't we reuse the source code. Sorry if my question sounds stupid. | 21:36 |
| lisitsyn1 | vijay05_: I glanced over code - looks it would be feasible | 21:36 |
| lisitsyn1 | vijay05_: you mean just copy and paste it? | 21:36 |
| lisitsyn1 | yeah sure | 21:36 |
| vijay05_ | ya | 21:36 |
| vijay05_ | cool | 21:36 |
| vijay05_ | where did you find the source code? | 21:38 |
| vijay05_ | I see only matlab files | 21:38 |
| lisitsyn1 | vijay05_: the only issue I see - we wanted to start a process of moving to bsd license | 21:39 |
| lisitsyn1 | it is less restrictive and is appreciated in the community | 21:39 |
| lisitsyn1 | growing codebase if we can avoid it would not be that good | 21:40 |
| lisitsyn1 | vijay05_: if we could avoid putting liblinear to shogun we'd do it I think :D | 21:40 |
| lisitsyn1 | but it is patched | 21:40 |
| -!- lambday [67157f4e@gateway/web/freenode/ip.103.21.127.78] has quit [Ping timeout: 245 seconds] | 21:40 | |
| lisitsyn1 | vijay05_: there is a bunch of .h and .cpp around in various folders | 21:41 |
| @besser82 | lisitsyn1, vijay05_: NOOOOOOOOO!!!! Don't bundle more libs into shogun. \o/ | 21:41 |
| @besser82 | lisitsyn1, vijay05_: that will make the next rpm update / release another horible fight for me | 21:41 |
| lisitsyn1 | besser82: yeah sure | 21:41 |
| @besser82 | lisitsyn1: many thx :D | 21:43 |
| vijay05_ | besser82, lisitsyn1: Ok | 21:51 |
| @sonney2k | guys look at our crappy description! | 21:53 |
| @sonney2k | http://www.google-melange.com/gsoc/org2/google/gsoc2014/shogun | 21:53 |
| @sonney2k | horrible excellente! | 21:53 |
| * sonney2k starts a massacre in this | 21:53 | |
| @besser82 | sonney2k: no response from heise-guys, yet... | 21:54 |
| @besser82 | sonney2k: I'll ping thl tommorow in the early morning ;) | 21:55 |
| @besser82 | sonney2k: perhaps he knows what will be... | 21:55 |
| @besser82 | sonney2k: btw. we need some urgent work on porting the testsuite to different arches | 21:56 |
| @besser82 | sonney2k: some tests fail on iX86, but pass on ARM and vice-versa && all pass on x86_64 :S | 21:56 |
| @sonney2k | besser82, well figure out which tests fail and file bug reports with some link so someone can look at them | 21:58 |
| @besser82 | sonney2k: will do ;) bugreport? you mean gh-issues?!? | 21:59 |
| @besser82 | sonney2k: | 22:00 |
| vijay05_ | lisitsyn1: So, I will install/build and shogun. I will try to do fixes from https://github.com/shogun-toolbox/shogun/issues. Whats the best way to increase my acceptance chances? | 22:02 |
| @sonney2k | somebody please read the blurp here http://www.google-melange.com/gsoc/org2/google/gsoc2014/shogun | 22:03 |
| @sonney2k | vijay05_, the more contributions you do the more likely you are in | 22:03 |
| @sonney2k | vijay05_, it is really as simple as that | 22:04 |
| @sonney2k | vijay05_, when lisitsyn was a GSoC student he was hyper active... | 22:04 |
| vijay05_ | :-) Good to know that. | 22:05 |
| @sonney2k | HeikoS1, please read http://www.google-melange.com/gsoc/org2/google/gsoc2014/shogun and edit this if unhappy | 22:05 |
| @sonney2k | HeikoS1, we need some kind of shogun in education - task for that IMHO and put the baseline ML algorithms in there and again demos and teaching | 22:06 |
| HeikoS1 | sonney2k: ok, will update | 22:06 |
| @besser82 | sonney2k: that reads much better, now :D | 22:06 |
| HeikoS1 | sonney2k: ok I can write that, in fact already started thinking about it | 22:06 |
| HeikoS1 | sonney2k: but I dont really know where its going apart from having a huge list of algorithms to demo | 22:07 |
| HeikoS1 | sonney2k: whats with this guy from zurich? | 22:07 |
| @sonney2k | HeikoS1, I will ping them again tomorrow | 22:07 |
| @sonney2k | HeikoS1, but yes we need that huge list of demos and then we can put structure on top of that later (as some kind of tutorial...) | 22:08 |
| HeikoS1 | sonney2k: its just a bit uncoordinated as a project I mean | 22:08 |
| @sonney2k | HeikoS1, some with notebooks... for me demos / notebooks are pretty similar in that respect. demos are to wet appetite and notebooks to learn the real thing | 22:08 |
| HeikoS1 | ok | 22:09 |
| HeikoS1 | mmh | 22:09 |
| @sonney2k | HeikoS1, true it needs an enumeration of algorithms | 22:09 |
| @sonney2k | but for example hierarchical clustering is missing altogether | 22:09 |
| HeikoS1 | so we want a guy who writes really nice text & examples | 22:09 |
| @sonney2k | GPs don't work | 22:09 |
| HeikoS1 | but this is not about implementing | 22:09 |
| @sonney2k | it is | 22:09 |
| @sonney2k | implement demos | 22:09 |
| HeikoS1 | yeah ok | 22:10 |
| HeikoS1 | but no algorithms | 22:10 |
| HeikoS1 | ok good, I will write that now | 22:10 |
| HeikoS1 | sonney2k: pls feedback the packaging project | 22:10 |
| @sonney2k | no use what shogun has and implement $demo | 22:10 |
| HeikoS1 | ok | 22:10 |
| HeikoS1 | so two parts | 22:10 |
| HeikoS1 | web-demos in a nice way | 22:10 |
| HeikoS1 | and notebooks | 22:10 |
| * sonney2k ZZzzzz | 22:14 | |
| HeikoS1 | sonney2k: GOOOD NIIIGHT | 22:14 |
| @besser82 | sonney2k: sleep tight! :D | 22:14 |
| * besser82 needs some Zzzzzzz, too. | 22:15 | |
| -!- vijay05_ [0e8b5206@gateway/web/freenode/ip.14.139.82.6] has quit [Ping timeout: 245 seconds] | 22:17 | |
| -!- rbrooks [8fd7309a@gateway/web/freenode/ip.143.215.48.154] has joined #shogun | 22:22 | |
| -!- rbrooks [8fd7309a@gateway/web/freenode/ip.143.215.48.154] has quit [Ping timeout: 245 seconds] | 22:26 | |
| HeikoS1 | lisitsyn1: around=? | 22:48 |
| lisitsyn1 | HeikoS1: now yes | 22:49 |
| HeikoS1 | lisitsyn could you please add all the core developers to the http://www.shogun-toolbox.org/page/contact/contacts | 22:49 |
| HeikoS1 | and also all our nicknames? | 22:49 |
| HeikoS1 | in irc | 22:49 |
| HeikoS1 | lisitsyn1: and possible webpages? | 22:49 |
| lisitsyn1 | HeikoS1: oh your fancy tasks again :D | 22:49 |
| HeikoS1 | lisitsyn1: I am writing project descriptions | 22:50 |
| HeikoS1 | lisitsyn1: thanks ! :) | 22:50 |
| @wiking | ooooh yeeeah it's gsoc TIME! we are 10+ on the channel :) | 22:59 |
| lisitsyn1 | wiking: lol | 22:59 |
| -!- zxtx [~zv@static-108-13-230-170.lsanca.fios.verizon.net] has joined #shogun | 23:01 | |
| HeikoS1 | wiking: whooo :) | 23:01 |
| -!- siddharth_sharma [~chatzilla@14.139.243.162] has quit [Ping timeout: 244 seconds] | 23:04 | |
| lisitsyn1 | HeikoS1: dear respected sir I have completed your assignment | 23:09 |
| --- Log closed Wed Feb 26 00:00:15 2014 | ||
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!