| --- Log opened Tue Apr 23 00:00:28 2013 | ||
| -!- sijin [~smuxi@144.214.222.109] has quit [Excess Flood] | 00:18 | |
| -!- sijin [~smuxi@144.214.222.109] has joined #shogun | 00:20 | |
| -!- utsavjain [~tiripS@14.139.82.6] has left #shogun [] | 00:44 | |
| -!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 01:05 | |
| -!- iglesiasg [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 02:15 | |
| -!- sijin__ [~smuxi@144.214.222.109] has joined #shogun | 02:32 | |
| -!- sijin_ [~smuxi@144.214.222.109] has quit [Remote host closed the connection] | 02:34 | |
| -!- sijin__ [~smuxi@144.214.222.109] has quit [Remote host closed the connection] | 02:35 | |
| -!- sijin [~smuxi@144.214.222.109] has quit [Remote host closed the connection] | 02:40 | |
| -!- sijin [~smuxi@144.214.222.109] has joined #shogun | 02:49 | |
| sijin | @lisitsyn morning, are you there? | 02:55 |
|---|---|---|
| -!- rgirish28 [~rgirish28@61.17.226.220] has joined #shogun | 02:56 | |
| sijin | @lisitsyn I have modified takpee_shogun.cpp/tapkee_shogun.hpp for support of KPCA and retriving projecting function, could you please check it? If it is ok, i can move to writing the warper | 03:20 |
| shogun-buildbot | build #370 of nightly_default is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/370 | 03:53 |
| -!- kevin [820f1060@gateway/web/freenode/ip.130.15.16.96] has joined #shogun | 04:21 | |
| -!- kevin is now known as Guest98041 | 04:22 | |
| Guest98041 | hello is anyone on here right now? I have a question about blas | 04:22 |
| -!- votjakovr [~chatzilla@host-46-241-3-209.bbcustomer.zsttk.net] has joined #shogun | 04:27 | |
| Guest98041 | hey I have a question about blas | 04:29 |
| -!- Guest98041 [820f1060@gateway/web/freenode/ip.130.15.16.96] has quit [Quit: Page closed] | 04:35 | |
| -!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has joined #shogun | 05:43 | |
| -!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has quit [Quit: Page closed] | 06:03 | |
| -!- pickle27 [~kevin@rcv3-lab-pc.ee.queensu.ca] has joined #shogun | 06:13 | |
| -!- pickle27 [~kevin@rcv3-lab-pc.ee.queensu.ca] has left #shogun [] | 06:13 | |
| -!- pickle27 [~kevin@rcv3-lab-pc.ee.queensu.ca] has joined #shogun | 06:13 | |
| -!- pickle27 [~kevin@rcv3-lab-pc.ee.queensu.ca] has quit [Client Quit] | 06:14 | |
| -!- poojits [75d35d4a@gateway/web/freenode/ip.117.211.93.74] has joined #shogun | 06:15 | |
| -!- poojits [75d35d4a@gateway/web/freenode/ip.117.211.93.74] has quit [Client Quit] | 06:20 | |
| -!- saket [0e8b5206@gateway/web/freenode/ip.14.139.82.6] has joined #shogun | 06:43 | |
| -!- rgirish28 [~rgirish28@61.17.226.220] has quit [Ping timeout: 240 seconds] | 06:46 | |
| -!- AlphaCentauri [73fe6352@gateway/web/freenode/ip.115.254.99.82] has joined #shogun | 06:46 | |
| -!- rgirish28 [~rgirish28@111.92.71.249] has joined #shogun | 06:50 | |
| -!- rgirish28 [~rgirish28@111.92.71.249] has quit [Ping timeout: 256 seconds] | 07:20 | |
| -!- AlphaCentauri [73fe6352@gateway/web/freenode/ip.115.254.99.82] has left #shogun [] | 07:21 | |
| -!- rgirish28 [~rgirish28@111.92.71.249] has joined #shogun | 07:23 | |
| saket | hi | 07:26 |
| saket | how do we run examples in the folder examples/documented/libshogun ? | 07:27 |
| saket | Compiling them with " g++ classifier_nearest_centroid.cpp -lshogun " | 07:27 |
| saket | give this error classifier_nearest_centroid.cpp:1:36: fatal error: shogun/features/Labels.h: No such file or directory | 07:28 |
| saket | do i need to give some parameter to ./configure | 07:28 |
| votjakovr | saket: Hi, Have you installed shogun by running 'make install'? | 07:32 |
| saket | votjakovr: yes i did | 07:34 |
| saket | other examples do compiel | 07:34 |
| saket | *compile | 07:34 |
| saket | but on running i get error while loading shared libraries: libshogun.so.14: cannot open shared object file: No such file or directory | 07:35 |
| saket | I think i need to append path to libshogun.so.14 to some env variable..right? any idea ? | 07:36 |
| saket | got it ..needed to add the path to /etc/ld.so.conf | 07:44 |
| votjakovr | saket: great :) | 07:45 |
| votjakovr | saket: I think you can't compile classifier_nearest_centroid.cpp, because Labels.h in shogun/labels/ now | 07:46 |
| saket | oh | 07:48 |
| saket | shall i move it ? | 07:48 |
| votjakovr | saket: No, I don't think so... | 07:49 |
| saket | votjakovr: I will be grateful to you if can help me out with the v_array issue .. i mail yesterday to the mailing list | 07:50 |
| saket | votjakovr: subject of the mail = "v_array issue && ball_tree update." | 07:51 |
| -!- vikram360 [~vikram360@117.216.147.142] has joined #shogun | 08:22 | |
| votjakovr | saket: I think it's better to ask lisitsyn about this | 08:31 |
| -!- user215806 [~Thunderbi@119.157.235.83] has joined #shogun | 08:32 | |
| user215806 | Is anyone around? | 08:33 |
| saket | votjakovr: okay | 08:35 |
| -!- rgirish28 [~rgirish28@111.92.71.249] has quit [Quit: rgirish28] | 08:48 | |
| -!- rgirish28 [~rgirish28@111.92.71.249] has joined #shogun | 09:07 | |
| -!- sumit_ [73f91219@gateway/web/freenode/ip.115.249.18.25] has joined #shogun | 09:11 | |
| -!- sumit_ [73f91219@gateway/web/freenode/ip.115.249.18.25] has quit [Ping timeout: 245 seconds] | 09:32 | |
| -!- rgirish28 [~rgirish28@111.92.71.249] has quit [Quit: rgirish28] | 09:32 | |
| -!- lisitsyn [~blackburn@37.61.179.23] has quit [Quit: Leaving.] | 09:46 | |
| -!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has joined #shogun | 09:47 | |
| -!- sids_aquarius [~sidi_@li400-124.members.linode.com] has quit [Remote host closed the connection] | 09:50 | |
| -!- sids_aquarius [~sidi_@li400-124.members.linode.com] has joined #shogun | 10:09 | |
| -!- lisitsyn [~lisitsin@mxs.kg.ru] has joined #shogun | 10:12 | |
| -!- mode/#shogun [+o lisitsyn] by ChanServ | 10:12 | |
| -!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has quit [Ping timeout: 245 seconds] | 10:16 | |
| -!- vikram360 [~vikram360@117.216.147.142] has quit [Ping timeout: 256 seconds] | 10:22 | |
| -!- iglesiasg [~nando@n162-p214.kthopen.kth.se] has joined #shogun | 10:24 | |
| -!- mode/#shogun [+o iglesiasg] by ChanServ | 10:24 | |
| -!- user215807 [~Thunderbi@39.48.53.85] has joined #shogun | 10:28 | |
| -!- user215806 [~Thunderbi@119.157.235.83] has quit [Read error: Connection reset by peer] | 10:29 | |
| sijin | @lisitsyn hello, do you have time now? | 10:30 |
| @lisitsyn | sijin: yes a bit | 10:30 |
| sijin | I have modified takpee_shogun.cpp/tapkee_shogun.hpp for support of KPCA and retriving projecting function, could you please check it? If it is ok, i can move to writing the warper | 10:31 |
| sijin | Since projecting_func->implemtation is NULL currently, I can't test the program by myself | 10:32 |
| @lisitsyn | sijin: I am doing some changes in that code last days | 10:32 |
| @lisitsyn | yes | 10:32 |
| @lisitsyn | that's a blocker for changing kpca to use tapkee | 10:32 |
| @lisitsyn | I will test things tonight and commit | 10:32 |
| sijin | Also, I notice that kpca has get_transformation_matrix(), | 10:33 |
| @lisitsyn | it is kind of bad naming for me | 10:34 |
| sijin | can projecting function provide that? or I would like to delete this member function from that class | 10:35 |
| @lisitsyn | lets get back to that once I prepare some things | 10:36 |
| sijin | ok. | 10:36 |
| wiking | rer | 10:41 |
| @lisitsyn | vgorbati: bah that test fails | 10:53 |
| @lisitsyn | 0.425 | 10:53 |
| -!- HeikoS [~heiko@nat-180-153.internal.eduroam.ucl.ac.uk] has joined #shogun | 10:59 | |
| -!- mode/#shogun [+o HeikoS] by ChanServ | 10:59 | |
| wiking | i want op | 11:10 |
| wiking | i want oooop | 11:10 |
| wiking | :) | 11:10 |
| @iglesiasg | wiking: feeling well? :) | 11:10 |
| wiking | votjakovr around? | 11:10 |
| votjakovr | wiking: yep :) | 11:10 |
| wiking | iglesiasg: gimme op :D | 11:10 |
| wiking | votjakovr: check my comments | 11:11 |
| wiking | votjakovr: and apply changes accordingly | 11:11 |
| @iglesiasg | wiking: op? | 11:11 |
| wiking | iglesiasg: @ | 11:11 |
| wiking | operator status man :> | 11:11 |
| wiking | when have u started ircing :) | 11:11 |
| @iglesiasg | aaah ok | 11:11 |
| @iglesiasg | last year when I started appeared around here :) | 11:11 |
| wiking | iglesiasg: lol | 11:11 |
| wiking | young :P | 11:11 |
| @iglesiasg | haha | 11:12 |
| @iglesiasg | I am still a toddler you know | 11:12 |
| wiking | yeah realised :) | 11:12 |
| wiking | just do /op wiking ;) | 11:12 |
| -!- mode/#shogun [+o wiking] by lisitsyn | 11:12 | |
| -!- mode/#shogun [-o wiking] by ChanServ | 11:12 | |
| wiking | thereugo | 11:13 |
| @iglesiasg | only the captatin may do that | 11:13 |
| wiking | :) | 11:13 |
| wiking | hahahahah | 11:13 |
| @lisitsyn | wiking: register your nick first | 11:13 |
| -!- erlenda [~erlenda@fw-oslo.intelcom.no] has joined #shogun | 11:13 | |
| wiking | lisitsyn: /whois wiking | 11:13 |
| wiking | lisitsyn: u think my nick is not registered with nickserv? :) | 11:13 |
| @lisitsyn | wiking: oh then you are not in the list of channel ops | 11:13 |
| votjakovr | wiking: thanks for your comments | 11:13 |
| wiking | votjakovr: nw | 11:13 |
| @iglesiasg | wiking: oh you live in US? | 11:13 |
| wiking | iglesiasg: atm no | 11:14 |
| wiking | iglesiasg: now i'm asturias ;P | 11:14 |
| @lisitsyn | sonney2k: ha! someone needs our domain | 11:14 |
| @iglesiasg | haha Spain! | 11:14 |
| wiking | iglesiasg: yes | 11:14 |
| @sonney2k | lisitsyn, no additional domain in china! | 11:14 |
| @iglesiasg | wiking: holidays or living there? | 11:14 |
| @sonney2k | if that was not spam | 11:14 |
| wiking | iglesiasg: 1 month living... came for a research visit | 11:14 |
| @iglesiasg | sonney2k: what is this domain name issue? haha | 11:14 |
| @iglesiasg | it seems fishy idd | 11:14 |
| @iglesiasg | wiking: nice - aprendiendo espa?ol? :) | 11:15 |
| wiking | iglesiasg: ye | 11:15 |
| @lisitsyn | wiking: so you are in spain? | 11:15 |
| @sonney2k | wiking, for 9 years you are hanging around on freenode? | 11:15 |
| wiking | sonney2k: yeah since the beginning | 11:16 |
| @lisitsyn | crazy | 11:16 |
| @iglesiasg | how do you get that info? | 11:16 |
| wiking | sonney2k: i've even chated on the phone with the guy who started freenode, who died recently | 11:16 |
| @sonney2k | wiking, did you create that network or what? | 11:16 |
| @iglesiasg | haha | 11:16 |
| wiking | sonney2k: back in those days you had to chat with him on the phone to get the funky 'domains' for your registered user | 11:16 |
| @sonney2k | heh | 11:17 |
| wiking | i don't know what's the current proceedure.... but we should get /shogun :) | 11:17 |
| @sonney2k | wiking, heh | 11:18 |
| -!- mode/#shogun [+o wiking] by ChanServ | 11:19 | |
| @wiking | cheers | 11:20 |
| @wiking | lol | 11:20 |
| @wiking | there's like 6 integration tests error on my machine :) | 11:21 |
| @wiking | no it's a lot | 11:21 |
| @wiking | :) | 11:21 |
| @HeikoS | sonney2k, wiking just received an application | 11:22 |
| @HeikoS | per mail | 11:22 |
| @wiking | serialization_complex_example.py setting 1/2 OK | 11:22 |
| @wiking | Warning: program compiled against libxml 209 using older 207 | 11:22 |
| @wiking | woah | 11:22 |
| @HeikoS | ah it was cced on the list | 11:22 |
| @wiking | wtf is this? :D | 11:22 |
| @wiking | but yeah | 11:22 |
| @wiking | i managed to reproduce the error \o/ | 11:23 |
| @wiking | and even more of it! :D | 11:23 |
| @HeikoS | man, those indian/chinese CVs are always packed ;) | 11:24 |
| @HeikoS | Lead Vocal & Cofounder of a Rock Band in Zhejiang University (the Rover) | 11:24 |
| @HeikoS | ? Cover punk and metal bands like Sex Pistols and Judas Priest | 11:24 |
| @HeikoS | yeah :D | 11:24 |
| @lisitsyn | HeikoS: 15 years of advanced C++ metaprogramming experience | 11:25 |
| @HeikoS | lisitsyn: hehe | 11:25 |
| @HeikoS | I wasnt born 15 years ago ;) | 11:25 |
| @HeikoS | lisitsyn: has the guy sent any PR? I haven't seen any but just to be sure | 11:26 |
| @lisitsyn | HeikoS: haven't seen | 11:27 |
| @iglesiasg | wiking, HeikoS : btw guys, are you coming to the workshop too? | 11:28 |
| @HeikoS | iglesiasg: workshop? | 11:28 |
| @iglesiasg | omg | 11:28 |
| @HeikoS | iglesiasg: ours? | 11:28 |
| @iglesiasg | yeah :) | 11:28 |
| @HeikoS | I see :) | 11:28 |
| @HeikoS | yes of course :) | 11:28 |
| @iglesiasg | cool | 11:28 |
| @HeikoS | iglesiasg: so you too I guess! | 11:29 |
| @wiking | iglesiasg: yes sure | 11:29 |
| @HeikoS | aweseom, really looking forward to meet all you guys! | 11:29 |
| @iglesiasg | HeikoS: sure :) | 11:29 |
| @iglesiasg | wiking: nice | 11:29 |
| @iglesiasg | HeikoS: yeah, it is going to be nice | 11:29 |
| @wiking | sonney2k: this seems like a stochastic machine for me.... classifier_averaged_perceptron_modular.py setting 1/2 passes 50% of the times and 50% of the times fails... ran 10 times now | 11:33 |
| @lisitsyn | iglesiasg: HeikoS: wiking: I am very probably in berlin starting from late evening of 10th, July | 11:36 |
| @lisitsyn | oops | 11:36 |
| @lisitsyn | 9th | 11:36 |
| @iglesiasg | lisitsyn: I will come on the 10th or 11th | 11:36 |
| @iglesiasg | I think | 11:36 |
| @HeikoS | no idea yet when I will come, but will book flights soon | 11:37 |
| @HeikoS | Berlin is always worth a week :) | 11:37 |
| @lisitsyn | HeikoS: is there a lot of sightseeings? | 11:37 |
| @HeikoS | yes | 11:37 |
| @lisitsyn | HeikoS: remind me how to say that in german | 11:37 |
| @lisitsyn | some funny long word | 11:38 |
| @HeikoS | mostly underground clubs/bars/cool places :D | 11:38 |
| @lisitsyn | (sightseeing) | 11:38 |
| @wiking | i think i fly in around 10 or 11 | 11:38 |
| @HeikoS | Sehensw?rdigkeit | 11:38 |
| @wiking | haven't decided yet | 11:38 |
| @lisitsyn | wurdig or wurdigskeit? | 11:38 |
| @HeikoS | lisitsyn: but also some sightseeing which is nice | 11:38 |
| @lisitsyn | HeikoS: I will spend 5 days there | 11:38 |
| @lisitsyn | hope it would be enough | 11:38 |
| @HeikoS | Sehensw?rdigkeit | 11:38 |
| @lisitsyn | alright | 11:39 |
| @wiking | ok | 11:39 |
| @wiking | who knows anything about this script: classifier_averaged_perceptron_modular.py | 11:39 |
| @wiking | ? | 11:39 |
| @wiking | as in the integration test for it | 11:40 |
| @lisitsyn | wiking: does it fail? | 11:40 |
| @wiking | w SGVector<float64> 2 ({0.5}{0.5}) | 11:40 |
| @wiking | vs | 11:40 |
| @wiking | w SGVector<float64> 2 ({0.49}{1.160182291326443e+75}) | 11:40 |
| @HeikoS | wiking: I thought I fixed the seed for that? | 11:41 |
| @HeikoS | since when does it fail? | 11:41 |
| @wiking | a is the last one b is the [D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[C[C[C[C[C[C[C[C[C[C[C[C[C[C[C [C is in a and the first one in [C[C[C[C[C[C[C[C[C | 11:41 |
| @HeikoS | maybe its the labelss that I changed? | 11:41 |
| @HeikoS | wiking: ehm? | 11:41 |
| @wiking | mmm shit | 11:41 |
| @wiking | sorry something went wrong | 11:41 |
| @wiking | anyhow | 11:42 |
| @wiking | b is this one data/testsuite/tests/classifier_averaged_perceptron_modular0.txt | 11:42 |
| @wiking | i.e. the first one with w SGVector<float64> 2 ({0.5}{0.5}) | 11:43 |
| @wiking | and the other one is generated on my machine | 11:43 |
| @wiking | and of course since they are not the same | 11:44 |
| @wiking | the integration test fails | 11:44 |
| @HeikoS | wiking: are the seeds fixed? | 11:44 |
| @wiking | HeikoS: dunno | 11:44 |
| @wiking | the only thing i know is that i compiled the code | 11:45 |
| @wiking | from the HEAD | 11:45 |
| @wiking | develop branch | 11:45 |
| @HeikoS | check the code pls | 11:45 |
| @wiking | HeikoS: where what? | 11:45 |
| @HeikoS | classifier_averaged_perceptron_modular.py | 11:46 |
| @HeikoS | python modular | 11:46 |
| @HeikoS | wiking: tests work fine here | 11:46 |
| @wiking | HeikoS: ok what should be there? | 11:46 |
| @HeikoS | last compiled yesterday | 11:46 |
| @HeikoS | wiking: fixed seeds for python numpy and shogun | 11:46 |
| @wiking | random.seed(seed) | 11:46 |
| @wiking | seed=1 | 11:46 |
| @wiking | if u mean that | 11:46 |
| @HeikoS | yes | 11:47 |
| @wiking | well it's there | 11:47 |
| @HeikoS | ok | 11:47 |
| @HeikoS | mmh | 11:47 |
| @HeikoS | no idea then | 11:47 |
| @wiking | but still it fails | 11:47 |
| @HeikoS | just pulled and compiling | 11:47 |
| @wiking | and not only this | 11:47 |
| @HeikoS | will try agian | 11:47 |
| @wiking | but a lot other | 11:47 |
| @HeikoS | I also have some tests locally failing | 11:47 |
| @HeikoS | since I have not all libs installed | 11:47 |
| @wiking | ok so here's a list of failing integration tests on osx: http://pastebin.com/sAbq05kN | 11:49 |
| @HeikoS | wiking: there must be something wrong | 11:50 |
| @HeikoS | I recently broke all tests | 11:50 |
| @HeikoS | but corrected afterwards | 11:50 |
| @HeikoS | so maybe non uptpdate code? | 11:50 |
| @wiking | HeikoS: i've tried | 11:50 |
| @wiking | but no help | 11:50 |
| @HeikoS | Warning: program compiled against libxml 209 using older 207 | 11:50 |
| @HeikoS | weird | 11:50 |
| @wiking | git pull && make clean && ./configure && make install | 11:50 |
| @HeikoS | git pull --rebase upstream develop | 11:51 |
| @HeikoS | thats what I do | 11:51 |
| @HeikoS | while being in dev branch | 11:51 |
| @wiking | f2dae9dbaa0b8fbf34b01c5bb53a3108986e4184 | 11:51 |
| @wiking | this is my last commit in dev branch | 11:51 |
| @wiking | and that's the last one in github as well | 11:51 |
| @HeikoS | f2dae9dbaa0b8fbf34b01c5bb53a3108986e4184 | 11:52 |
| @HeikoS | yes | 11:52 |
| @HeikoS | mine too | 11:52 |
| @HeikoS | strange! | 11:52 |
| @HeikoS | but no idea, probably mac stuff | 11:52 |
| @wiking | i guess so | 11:52 |
| @HeikoS | wiking: sorry, I have to do some work now, see you! | 11:53 |
| @wiking | but still then i guess this could be a background for why one integration test fails on travis | 11:53 |
| @wiking | okok cya | 11:53 |
| @HeikoS | wiking: yes good point! | 11:53 |
| @lisitsyn | vgorbati: ping | 11:55 |
| vgorbati | isitsyn: yes, I saw the test failure - guess it was that 1e-5 probability. Make the number smaller? | 11:56 |
| vgorbati | lisitsyn:or another way - the things are much better, when target_dim==initial. I can make it so.. | 11:57 |
| @lisitsyn | vgorbati: yes please | 11:58 |
| vgorbati | lisitsyn: yes for the second option? | 11:58 |
| @lisitsyn | vgorbati: yes | 11:59 |
| @lisitsyn | and we could change the data actually | 11:59 |
| @lisitsyn | vgorbati: you may try some displaced grid in 2d | 12:00 |
| vgorbati | lisitsyn: to predefined? I guess I can generate one matrix randomly, see the minimum intersection rate, save the matrix to file, and the read it during the test. Then there wont be any randomness (is ther such a word) | 12:00 |
| @lisitsyn | vgorbati: heh, well we are of nearly the same native language so I understand your words (about your wondering if there is such a word) | 12:01 |
| @lisitsyn | vgorbati: err not in 2d but 3d | 12:01 |
| @lisitsyn | x = [0,0.001,....,1] | 12:02 |
| @lisitsyn | y = [0,0.001,...,1] | 12:02 |
| @lisitsyn | z = some noise | 12:02 |
| @lisitsyn | this should preserve neighborhoods much more | 12:02 |
| vgorbati | lisitsyn: and project that to 2D? | 12:02 |
| vgorbati | lisitsyn: i mean embed | 12:03 |
| @lisitsyn | yes | 12:03 |
| @lisitsyn | vgorbati: do you feel the difference btw? | 12:03 |
| @lisitsyn | it is actually pretty the same but it looks better to say 'project' about PCA | 12:03 |
| vgorbati | lisitsyn: in data? there's a linear dependenc | 12:03 |
| @lisitsyn | and 'embed' for some non-linear thing like LLE | 12:04 |
| @lisitsyn | vgorbati: anyway please take a look | 12:04 |
| vgorbati | lisitsyn: yes, I did feel that it is wrong to say 'project' for non-linear embedding methods | 12:04 |
| vgorbati | lisitsyn: ok, I will take a look after an hour I guess | 12:05 |
| @lisitsyn | alrrgiht | 12:05 |
| -!- vgorbati [c3ee5cb1@gateway/web/freenode/ip.195.238.92.177] has quit [Quit: Page closed] | 12:09 | |
| -!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 12:09 | |
| -!- user215807 [~Thunderbi@39.48.53.85] has quit [Ping timeout: 256 seconds] | 12:10 | |
| -!- user215806 [~Thunderbi@39.48.53.85] has joined #shogun | 12:11 | |
| -!- user215806 [~Thunderbi@39.48.53.85] has quit [Ping timeout: 252 seconds] | 12:34 | |
| -!- user215806 [~Thunderbi@39.48.53.85] has joined #shogun | 12:35 | |
| -!- user215806 [~Thunderbi@39.48.53.85] has quit [Ping timeout: 255 seconds] | 12:39 | |
| -!- user215806 [~Thunderbi@39.48.53.85] has joined #shogun | 12:40 | |
| -!- user215807 [~Thunderbi@39.48.180.228] has joined #shogun | 12:43 | |
| -!- user215806 [~Thunderbi@39.48.53.85] has quit [Ping timeout: 256 seconds] | 12:44 | |
| -!- user215807 [~Thunderbi@39.48.180.228] has quit [Ping timeout: 256 seconds] | 12:47 | |
| -!- user215806 [~Thunderbi@39.48.180.228] has joined #shogun | 12:48 | |
| -!- user215806 [~Thunderbi@39.48.180.228] has quit [Ping timeout: 264 seconds] | 12:53 | |
| -!- user215806 [~Thunderbi@39.48.180.228] has joined #shogun | 12:54 | |
| -!- user215806 [~Thunderbi@39.48.180.228] has quit [Ping timeout: 264 seconds] | 12:59 | |
| -!- user215806 [~Thunderbi@39.48.180.228] has joined #shogun | 12:59 | |
| -!- user215806 [~Thunderbi@39.48.180.228] has quit [Ping timeout: 264 seconds] | 13:03 | |
| -!- user215806 [~Thunderbi@119.157.240.239] has joined #shogun | 13:05 | |
| -!- HeikoS [~heiko@nat-180-153.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.] | 13:12 | |
| -!- erlenda [~erlenda@fw-oslo.intelcom.no] has quit [Read error: Connection reset by peer] | 13:20 | |
| -!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 13:20 | |
| -!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 13:20 | |
| -!- rgirish28 [~rgirish28@111.92.71.249] has joined #shogun | 13:34 | |
| @iglesiasg | I did a bad PR sorry, I forgot to re-compile after renaming a couple of variables and it seems I introduced a typo :D | 13:36 |
| @iglesiasg | I am closing that PR and will issue a new clean one | 13:37 |
| -!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 13:53 | |
| shogun-notifier- | shogun: Roman Votyakov :develop * d3a4d29 / src/shogun/regression/gp/ (5 files): https://github.com/shogun-toolbox/shogun/commit/d3a4d29e5c8a7aa438b917f2f6734a47736c1d0f | 13:53 |
| shogun-notifier- | shogun: Degrees of freedom (m_df) field moved from CLikelihoodModel to CStudentsTLikelihood | 13:53 |
| shogun-notifier- | shogun: Roman Votyakov :develop * a5b629f / src/shogun/regression/gp/LaplacianInferenceMethod.cpp: https://github.com/shogun-toolbox/shogun/commit/a5b629ff0d8777178959db4dd0d776585ae04e04 | 13:53 |
| shogun-notifier- | shogun: Make more efficient usage of cholesky | 13:53 |
| shogun-notifier- | shogun: Roman Votyakov :develop * 77b2d0d / src/shogun/regression/gp/ (3 files): https://github.com/shogun-toolbox/shogun/commit/77b2d0d5e572ceae619ecf5f6a0502704ba42c7e | 13:53 |
| shogun-notifier- | shogun: Some fixes in GP framework | 13:53 |
| shogun-notifier- | shogun: Roman Votyakov :develop * 77e00f8 / src/shogun/regression/gp/LaplacianInferenceMethod.cpp: https://github.com/shogun-toolbox/shogun/commit/77e00f87078b160379f8ecf29a85e940c7ddbb40 | 13:53 |
| shogun-notifier- | shogun: SG_UNREF added to prevent memory leak | 13:53 |
| shogun-notifier- | shogun: Heiko Strathmann :develop * 395dba3 / src/shogun/regression/gp/ (5 files): https://github.com/shogun-toolbox/shogun/commit/395dba39729c41a9b7b574db9bac654a6ea0b0a6 | 13:53 |
| shogun-notifier- | shogun: Merge pull request #1004 from votjakovr/develop | 13:54 |
| shogun-notifier- | shogun: | 13:54 |
| shogun-notifier- | shogun: Some fixes in GP framework | 13:54 |
| -!- HeikoS [~heiko@nat-180-153.internal.eduroam.ucl.ac.uk] has joined #shogun | 13:54 | |
| -!- mode/#shogun [+o HeikoS] by ChanServ | 13:54 | |
| -!- gsomix [~gsomix@37.61.179.23] has joined #shogun | 13:57 | |
| gsomix | hi | 13:57 |
| -!- iglesiasg [~nando@n162-p214.kthopen.kth.se] has quit [Read error: Operation timed out] | 14:01 | |
| -!- rgirish28 [~rgirish28@111.92.71.249] has quit [Quit: rgirish28] | 14:04 | |
| shogun-notifier- | shogun: Patrick Pletscher :develop * 077b7ce / .gitignore: https://github.com/shogun-toolbox/shogun/commit/077b7ce747b74eb2e99deac21df1155a643b3ad2 | 14:04 |
| shogun-notifier- | shogun: ignore temporary files created by csharp and octave interfaces | 14:04 |
| shogun-notifier- | shogun: Sergey Lisitsyn :develop * 07ca07b / .gitignore: https://github.com/shogun-toolbox/shogun/commit/07ca07bb9a3b6503d782ba5af94f1220c4647703 | 14:04 |
| shogun-notifier- | shogun: Merge pull request #1006 from ppletscher/feature/gitignore | 14:04 |
| shogun-notifier- | shogun: | 14:04 |
| shogun-notifier- | shogun: ignore temporary files created by csharp and octave interfaces | 14:04 |
| @lisitsyn | HeikoS: mentors are doing pull requests too hah | 14:05 |
| @lisitsyn | nice | 14:05 |
| -!- rgirish28 [~rgirish28@111.92.71.249] has joined #shogun | 14:06 | |
| @HeikoS | lisitsyn: ? | 14:06 |
| vgorbati | lisitsyn: so, I generated data as you have said, but the average similarity ratio stays the same | 14:07 |
| @lisitsyn | HeikoS: check that pr I merged | 14:08 |
| @HeikoS | lisitsyn: I see -- cool! | 14:08 |
| @lisitsyn | vgorbati: I am just curious | 14:10 |
| @lisitsyn | if you use local tangent space alignment | 14:10 |
| @lisitsyn | is that getting better? | 14:10 |
| vgorbati | shogun/converter/LocalTangentSpaceAlignment.h? | 14:11 |
| @lisitsyn | vgorbati: yes | 14:11 |
| vgorbati | checking | 14:12 |
| -!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has joined #shogun | 14:12 | |
| -!- travis-ci [~travis-ci@ec2-50-17-100-41.compute-1.amazonaws.com] has joined #shogun | 14:12 | |
| travis-ci | [travis-ci] it's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/6564742 | 14:12 |
| -!- travis-ci [~travis-ci@ec2-50-17-100-41.compute-1.amazonaws.com] has left #shogun [] | 14:12 | |
| vgorbati | lisitsyn: even worse( | 14:14 |
| -!- rgirish28 [~rgirish28@111.92.71.249] has quit [Quit: rgirish28] | 14:14 | |
| @lisitsyn | vgorbati: could you please save both original and embedded data and scatter plot it? | 14:16 |
| vgorbati | lisitsyn: is there an easy way to do it?) | 14:18 |
| vgorbati | lisitsyn: should I use CDenseFeatures::save method? | 14:18 |
| @lisitsyn | vgorbati: won't help you much | 14:19 |
| -!- gsomix_ [~gsomix@188.168.3.116] has joined #shogun | 14:20 | |
| vgorbati | lisitsyn: ok, then will do it some other way | 14:20 |
| -!- gsomix [~gsomix@37.61.179.23] has quit [Ping timeout: 256 seconds] | 14:24 | |
| deerishi | @HeikoS: hello | 14:24 |
| @HeikoS | deerishi: hi | 14:25 |
| deerishi | i am just gonna give the pull request. | 14:25 |
| @HeikoS | deerishi: did you put the data generator into tools? | 14:25 |
| @HeikoS | and import it from both the modular and the graphical ex? | 14:25 |
| deerishi | nope.by data generator do you mean the circular data. | 14:25 |
| deerishi | ? | 14:25 |
| deerishi | in tools do i jyst need to write import preprocessor_kernel.py and import preprocessor_kernel_graphical.py? | 14:26 |
| @HeikoS | deerishi: no | 14:26 |
| @HeikoS | you import the circle data from the examples | 14:27 |
| @HeikoS | not the other way round | 14:27 |
| @HeikoS | and please rename the files to preprocessor_kpca ... | 14:27 |
| @HeikoS | kernel doesnt mean much :) | 14:27 |
| deerishi | m still a bit unsure abt what i need to do in the tools folder . i have done shogun / examples / undocumented / python_modular ---->The KPCA CODE: preprocessor_kpca.py shogun / examples / undocumented / python_modular / graphical ----->The circular input data: preprocessor_kpca_graphical.py | 14:28 |
| @HeikoS | you generate your data with some function in a file right? | 14:29 |
| @HeikoS | that has to go to the tools folder | 14:29 |
| @HeikoS | then, both examples import this file to generate data | 14:29 |
| @HeikoS | one example plots, the other one doesnt | 14:30 |
| @HeikoS | makes three files | 14:30 |
| @HeikoS | 1) circle_data.py - located in python_modular/tools | 14:31 |
| @HeikoS | 2) modular example - located in python_modular, imports 1 | 14:31 |
| @HeikoS | 3) graphicla example - located in python_modular/graphical, imports 1, plots | 14:31 |
| deerishi | k.i think i got it.it was a slight misunderstanding on my part. the data generation file goes to the tools folder.. | 14:31 |
| deerishi | thanks @HeikoS | 14:32 |
| deerishi | :) | 14:32 |
| @HeikoS | welcome :) | 14:32 |
| -!- iglesiasg [~nando@n162-p214.kthopen.kth.se] has joined #shogun | 14:32 | |
| -!- mode/#shogun [+o iglesiasg] by ChanServ | 14:32 | |
| -!- Yanglittle [deb20af1@gateway/web/freenode/ip.222.178.10.241] has joined #shogun | 14:42 | |
| Yanglittle | hey, can i make crossvalidate in mkl_multiclass for all the parameters? | 14:45 |
| vgorbati | lisitsyn:here? | 14:45 |
| @lisitsyn | vgorbati: yes | 14:45 |
| Yanglittle | hey, can i make crossvalidate in mkl_multiclass for all the parameters? | 14:46 |
| @lisitsyn | Yanglittle: not sure what you mean | 14:46 |
| vgorbati | lisitsyn: http://www.youtube.com/watch?v=SpbeUMYwr7g&feature=youtu.be, original - on the left, target - right (its actually 2D, i just used 3D so that I could rotate it). Looks perfect, so I guess the problem is somewhere in the test( | 14:46 |
| @lisitsyn | hmm | 14:47 |
| @lisitsyn | looks ok indeed | 14:47 |
| @lisitsyn | vgorbati: could you please check it then? | 14:47 |
| -!- travis-ci [~travis-ci@ec2-50-19-58-94.compute-1.amazonaws.com] has joined #shogun | 14:49 | |
| travis-ci | [travis-ci] it's Sergey Lisitsyn's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/6565036 | 14:49 |
| -!- travis-ci [~travis-ci@ec2-50-19-58-94.compute-1.amazonaws.com] has left #shogun [] | 14:49 | |
| vgorbati | lisitsyn: sure, but debugging using cout<< is soo painful) | 14:50 |
| @lisitsyn | vgorbati: well you may try gdb or anything like that :) | 14:50 |
| Yanglittle | in the grid search, assuming the gaussian kernel, there are two parameters, delta and C. this is no problem. But in mkl_multiclass, assuming five gaussian kernels, and it needs to find five different delta and a C, does grid_serach can support this? | 14:53 |
| @lisitsyn | ahh | 14:53 |
| @lisitsyn | HeikoS: ^ could you please comment that? | 14:54 |
| @HeikoS | Yanglittle: in MKL you usually do not optimize over kernel parameters but just put in a set of kernels and then look at the weights | 14:55 |
| @HeikoS | MKL solves for the weights for you | 14:55 |
| @HeikoS | then you only do a gridsearch over C | 14:55 |
| @HeikoS | 5 different kernel parameters also is completely infeasible | 14:55 |
| @HeikoS | params_per_kernel^5 * C_possiilities | 14:55 |
| @lisitsyn | actually it sounds exp yeah | 14:57 |
| Yanglittle | yes i know this but when calculate the single kernel, we search a good pair of delta and C, but the good delta may not be the best for the mkl_multiclass. what do you think? | 14:59 |
| vgorbati | lisitsyn: i am having troubles on how to use gdb. Ok, I added -g option in makefile, so that now 'shogun-unit-test' executable contains debug info. Ok, I can run 'gdb shogun-unit-test'. But then, how do I break the execution in needed test? | 15:05 |
| Yanglittle | gdb.set_trace()? | 15:08 |
| vgorbati | I know about pdb.set_trace() in python, but does it exist in C/C++? | 15:09 |
| -!- saket [0e8b5206@gateway/web/freenode/ip.14.139.82.6] has quit [Quit: Page closed] | 15:09 | |
| @lisitsyn | vgorbati: break that? | 15:09 |
| @lisitsyn | like? | 15:09 |
| @lisitsyn | you mean set a breakpoint? | 15:10 |
| @lisitsyn | http://www.delorie.com/gnu/docs/gdb/gdb_29.html | 15:10 |
| vgorbati | lisitsyn: yes) | 15:10 |
| @HeikoS | Yanglittle: what is delta for you? | 15:14 |
| @HeikoS | the kernel width? | 15:14 |
| @HeikoS | Yanglittle: sorry I did not get that, could you explain again? | 15:15 |
| Yanglittle | yeah, the width, the kernel parameter. | 15:15 |
| -!- rgirish28 [~rgirish28@111.92.71.249] has joined #shogun | 15:19 | |
| @iglesiasg | lisitsyn: any idea why travis build started failing suddenly? | 15:19 |
| @iglesiasg | I see now error in serialization_complex_example.py setting 1/2 | 15:19 |
| @lisitsyn | iglesiasg: http://www.youtube.com/watch?feature=player_detailpage&v=__DrJI7mTHQ | 15:22 |
| @HeikoS | Yanglittle: could you explain your question again? | 15:23 |
| @HeikoS | why do you want to select the bandwidths of all kernels in MKL? | 15:23 |
| @HeikoS | the bad kernels will receive low weight anyway | 15:23 |
| -!- gsomix_ [~gsomix@188.168.3.116] has quit [Ping timeout: 260 seconds] | 15:24 | |
| @iglesiasg | lisitsyn: haha ok | 15:26 |
| @iglesiasg | lisitsyn: I think it is why this appears as build failed https://github.com/shogun-toolbox/shogun/pull/1007 | 15:27 |
| -!- rgirish28 [~rgirish28@111.92.71.249] has quit [Quit: rgirish28] | 15:27 | |
| @iglesiasg | aham I see it is for one interface | 15:28 |
| Yanglittle | i want to make each subkernel get a good accuracy on the data. and so , is it beneficial for mkl? in mkl, each subkernel needs the width, and how to determine the width? | 15:30 |
| @HeikoS | Yanglittle: you don't select the width but rather throw some kernels at the MKL which then discards the unimportant ones via the weights. | 15:43 |
| @HeikoS | Yanglittle: also if one single kernel is best for your problem, then all MKL kernels will have the same width | 15:47 |
| -!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has quit [Ping timeout: 245 seconds] | 15:49 | |
| Yanglittle | i input many kernel with different width , and then let mkl choose, how about this? | 15:49 |
| Yanglittle | hehe | 15:49 |
| -!- rgirish28 [~rgirish28@111.92.71.249] has joined #shogun | 15:51 | |
| -!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has joined #shogun | 16:02 | |
| -!- rgirish28 [~rgirish28@111.92.71.249] has quit [Quit: rgirish28] | 16:04 | |
| vgorbati | lisitsyn: here? | 16:05 |
| -!- gsomix [~Miranda@46.20.65.94] has joined #shogun | 16:10 | |
| -!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has quit [Ping timeout: 245 seconds] | 16:12 | |
| @lisitsyn | vgorbati: yes | 16:14 |
| vgorbati | lisitsyn: is it possible that embedding shuffles columns a bit? Then the picture will look right, but neighbor indices not | 16:14 |
| @lisitsyn | vgorbati: columns like vectors? | 16:15 |
| vgorbati | lisitsyn: yes, feature vectors | 16:16 |
| @lisitsyn | no, the order is preserved | 16:16 |
| @HeikoS | Yanglittle: yes, thats the idea :) | 16:21 |
| @lisitsyn | HeikoS: are you reading that iranian madness in the mentors list? | 16:24 |
| @lisitsyn | :D | 16:24 |
| @HeikoS | lisitsyn: I can understand the discussion though | 16:25 |
| @lisitsyn | HeikoS: I am quite opposite to that - citizens are responsible for their country | 16:27 |
| @HeikoS | lisitsyn: still the reaction is understandable - it counters the open-source principle | 16:27 |
| @HeikoS | while google has a point | 16:27 |
| @HeikoS | and they cannot do anything anyway | 16:28 |
| @HeikoS | so .... | 16:28 |
| @lisitsyn | HeikoS: do you think it does counter that? | 16:28 |
| @lisitsyn | I mean open-source is not about gsoc | 16:28 |
| @HeikoS | yes | 16:28 |
| @lisitsyn | nobody can ever deny open-source | 16:28 |
| @HeikoS | but gsoc is about open-source | 16:28 |
| @HeikoS | anyway, doesnt matter | 16:28 |
| @HeikoS | google cannot chang eit | 16:29 |
| @lisitsyn | it is about funding which comes from google | 16:29 |
| @lisitsyn | if google decided not to fund something it is just their choice | 16:29 |
| @HeikoS | its not even their choice in this case | 16:29 |
| @lisitsyn | yeah | 16:29 |
| @lisitsyn | HeikoS: we actually have one iranian phd student arya | 16:30 |
| @HeikoS | where? | 16:30 |
| @lisitsyn | which has been around | 16:30 |
| @HeikoS | ah yes I remember | 16:31 |
| @lisitsyn | well he posts some things to the mailing list sometimes | 16:31 |
| @lisitsyn | and we can't take him as gsoc student | 16:31 |
| @lisitsyn | he wanted too IIRC | 16:31 |
| @lisitsyn | to* | 16:31 |
| @HeikoS | yes, remember | 16:31 |
| -!- iglesiasg [~nando@n162-p214.kthopen.kth.se] has quit [Quit: leaving] | 16:48 | |
| Yanglittle | in the grid_search, how can i get the array of the build_values? It seems no the method. | 16:50 |
| @HeikoS | Yanglittle: no there is no method, you can print the parameter tree though | 16:53 |
| vgorbati | lisitsyn: so I checked really carefully, and it looks like the test is also ok. The thing is, that for every point there are sometimes a lot of points nearly at the same distance to it - and when we embed them, this distances change slightly, that can't be seen visually, but actually changes nearest neighbors | 16:55 |
| @lisitsyn | vgorbati: I see, thanks! | 17:02 |
| -!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 17:04 | |
| -!- Yanglittle [deb20af1@gateway/web/freenode/ip.222.178.10.241] has quit [Quit: Page closed] | 17:05 | |
| -!- saket [0e8b5206@gateway/web/freenode/ip.14.139.82.6] has joined #shogun | 17:08 | |
| vgorbati | lisitsyn: the bad news is that it is still unclear what to do with the test) | 17:08 |
| vgorbati | lisitsyn: what about giving the data that does not have such problem (I mean a lot of points at nearly the same distance)? | 17:10 |
| vgorbati | lisitsyn: so that small changes in distance between points, caused by embedding, should not influence the nearest neighbors? | 17:13 |
| @lisitsyn | vgorbati: if you have such data :) | 17:18 |
| -!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has joined #shogun | 17:33 | |
| -!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has quit [Ping timeout: 245 seconds] | 17:39 | |
| -!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 17:44 | |
| -!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 17:45 | |
| -!- user215807 [~Thunderbi@39.48.192.124] has joined #shogun | 17:45 | |
| -!- user215806 [~Thunderbi@119.157.240.239] has quit [Ping timeout: 240 seconds] | 17:47 | |
| -!- vikram360 [~vikram360@117.221.124.81] has joined #shogun | 17:47 | |
| -!- user215806 [~Thunderbi@119.157.145.53] has joined #shogun | 17:48 | |
| -!- user215807 [~Thunderbi@39.48.192.124] has quit [Ping timeout: 264 seconds] | 17:50 | |
| -!- lisitsyn [~lisitsin@mxs.kg.ru] has quit [Quit: Leaving.] | 17:50 | |
| -!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 17:51 | |
| votjakovr | HeikoS: Hi, I sent PR, please look at it :) | 17:58 |
| @HeikoS | looking | 17:59 |
| @HeikoS | votjakovr: so you are running the matlab gpml locally to create comparison results? | 18:00 |
| votjakovr | HeikoS: Yep. Not matlab, but octave... But I think it doesn't metter. | 18:02 |
| @HeikoS | votjakovr: no it doesnt | 18:02 |
| @HeikoS | very cool that you simplified things | 18:03 |
| @HeikoS | were there bugs? | 18:03 |
| @HeikoS | what kind of? | 18:03 |
| -!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has joined #shogun | 18:10 | |
| votjakovr | HeikoS: So there was some mistakes... It was hard to debug old code and actually rewrote much of update_cholesky | 18:10 |
| @HeikoS | votjakovr: I see, nice work! | 18:11 |
| votjakovr | HeikoS: Thanks :) | 18:11 |
| @HeikoS | It would be nice if you created examples for the methods that you unit-tested | 18:11 |
| @HeikoS | graphical would be even better :) | 18:11 |
| @HeikoS | but you can also continue on other things | 18:12 |
| @HeikoS | as you seem to have a good clue of what to do | 18:12 |
| votjakovr | HeikoS: On Octave? | 18:12 |
| @HeikoS | votjakovr: python would be best | 18:12 |
| @HeikoS | and then once these work translate to octave | 18:12 |
| @HeikoS | votjakovr: in the style of the existing regression_gaussian_process_modular.py | 18:14 |
| votjakovr | HeikoS: Ok, I will think about this :) | 18:14 |
| -!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 18:14 | |
| shogun-notifier- | shogun: Roman Votyakov :develop * 9d697fa / tests/unit/regression/gp/FITCInferenceMethod_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/9d697fa4facefc0bd1286eeef8983f12cee108ea | 18:14 |
| shogun-notifier- | shogun: Added unit-test for Fully Independent Conditional Training inference method | 18:14 |
| shogun-notifier- | shogun: Roman Votyakov :develop * 0472313 / src/shogun/regression/gp/FITCInferenceMethod.cpp: https://github.com/shogun-toolbox/shogun/commit/04723130dd62ff251cfc556bbb07a2a76c27bc2e | 18:14 |
| shogun-notifier- | shogun: Fixed various bugs in CFITCInferenceMethod class | 18:14 |
| shogun-notifier- | shogun: Heiko Strathmann :develop * 97a14c3 / / (2 files): https://github.com/shogun-toolbox/shogun/commit/97a14c39ed13463380bbd6b6b28249641faa626b | 18:14 |
| @HeikoS | merged :) | 18:14 |
| shogun-notifier- | shogun: Merge pull request #1008 from votjakovr/develop | 18:14 |
| shogun-notifier- | shogun: | 18:14 |
| shogun-notifier- | shogun: Added unit-test for FITC inference method and fixed bugs | 18:14 |
| deerishi | @HeikoS: u der> | 18:15 |
| deerishi | ? | 18:15 |
| @HeikoS | deerishi: ? | 18:15 |
| deerishi | i have a doubt ,how do i import the preprocessor_kpca_graphical.py from the graphical folder | 18:16 |
| deerishi | i could import it from the tools folder in python_modular | 18:16 |
| deerishi | but how do i set the path? | 18:17 |
| deerishi | ? | 18:18 |
| @HeikoS | deerishi: see the other python example | 18:18 |
| @HeikoS | most of them import a Matrix loader | 18:18 |
| -!- sumit [73f91219@gateway/web/freenode/ip.115.249.18.25] has joined #shogun | 18:19 | |
| -!- rgirish28 [~rgirish28@121.246.65.77] has joined #shogun | 18:24 | |
| -!- cwidmer [8ca3fe9d@gateway/web/freenode/ip.140.163.254.157] has joined #shogun | 18:29 | |
| deerishi | @HeikoS : I got it | 18:29 |
| cwidmer | hi all | 18:29 |
| cwidmer | HeikoS, around? | 18:30 |
| deerishi | i used the os.path module | 18:30 |
| -!- rgirish28 [~rgirish28@121.246.65.77] has quit [Ping timeout: 264 seconds] | 18:33 | |
| -!- cwidmer [8ca3fe9d@gateway/web/freenode/ip.140.163.254.157] has quit [Quit: Page closed] | 18:36 | |
| -!- vikram360 [~vikram360@117.221.124.81] has quit [Ping timeout: 258 seconds] | 18:37 | |
| deerishi | @HeikoS : I sent the pull request https://github.com/shogun-toolbox/shogun/pull/1009 | 18:37 |
| deerishi | please get back to me about it | 18:42 |
| -!- arinkverma [b4953e34@gateway/web/freenode/ip.180.149.62.52] has joined #shogun | 18:42 | |
| -!- user215807 [~Thunderbi@39.48.231.69] has joined #shogun | 18:43 | |
| -!- lisitsyn [~blackburn@46.20.65.94] has joined #shogun | 18:43 | |
| -!- mode/#shogun [+o lisitsyn] by ChanServ | 18:43 | |
| -!- travis-ci [~travis-ci@ec2-50-19-58-94.compute-1.amazonaws.com] has joined #shogun | 18:44 | |
| travis-ci | [travis-ci] it's Heiko Strathmann's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/6573069 | 18:44 |
| -!- travis-ci [~travis-ci@ec2-50-19-58-94.compute-1.amazonaws.com] has left #shogun [] | 18:44 | |
| -!- user215806 [~Thunderbi@119.157.145.53] has quit [Ping timeout: 264 seconds] | 18:46 | |
| arinkverma | Hi! I m interested in developing interactive machine learning demos using django. But I not very much aware of shogun framework, can someone help me with python_modular examples for classification of digits? | 18:48 |
| -!- user215807 [~Thunderbi@39.48.231.69] has quit [Ping timeout: 264 seconds] | 18:50 | |
| -!- user215806 [~Thunderbi@39.48.231.69] has joined #shogun | 18:52 | |
| -!- arinkverma_ [b4953e34@gateway/web/freenode/ip.180.149.62.52] has joined #shogun | 18:52 | |
| -!- arinkverma [b4953e34@gateway/web/freenode/ip.180.149.62.52] has quit [Ping timeout: 245 seconds] | 18:54 | |
| -!- sumit [73f91219@gateway/web/freenode/ip.115.249.18.25] has quit [Ping timeout: 245 seconds] | 18:54 | |
| -!- user215806 [~Thunderbi@39.48.231.69] has quit [Ping timeout: 248 seconds] | 18:56 | |
| -!- user215806 [~Thunderbi@39.48.231.69] has joined #shogun | 18:57 | |
| -!- arinkverma_ [b4953e34@gateway/web/freenode/ip.180.149.62.52] has quit [Quit: Page closed] | 19:00 | |
| -!- arink [~arink@180.149.62.52] has joined #shogun | 19:00 | |
| arink | from where I can subscribe to shogun mail-list? | 19:01 |
| -!- user215806 [~Thunderbi@39.48.231.69] has quit [Ping timeout: 252 seconds] | 19:02 | |
| -!- rgirish28 [~rgirish28@61.17.226.66] has joined #shogun | 19:04 | |
| -!- arink [~arink@180.149.62.52] has quit [Ping timeout: 252 seconds] | 19:13 | |
| -!- arink [~arink@180.149.62.52] has joined #shogun | 19:23 | |
| -!- arink [~arink@180.149.62.52] has quit [Read error: Connection reset by peer] | 19:53 | |
| -!- arink [~arink@180.149.62.52] has joined #shogun | 19:54 | |
| @HeikoS | arink: see the webpage | 20:02 |
| -!- mikhailBelous [~towelenee@109.205.249.161] has joined #shogun | 20:07 | |
| mikhailBelous | hello is your project need contribution in decision trees implimentation? | 20:08 |
| @HeikoS | mikhailBelous: best ask on the mailing list | 20:15 |
| @HeikoS | I currently dont know :) | 20:15 |
| -!- van51 [~van51@athedsl-407661.home.otenet.gr] has joined #shogun | 20:15 | |
| mikhailBelous | can somebody add me to mailing list? | 20:16 |
| mikhailBelous | And Georg have already send about it to mailing list | 20:17 |
| @lisitsyn | mikhailBelous: any concrete suggestion? | 20:18 |
| -!- HeikoS [~heiko@nat-180-153.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.] | 20:20 | |
| mikhailBelous | Georg said that there is no decision trees in your project | 20:22 |
| mikhailBelous | So maybe it will be usefull to write some | 20:23 |
| -!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 20:30 | |
| -!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has quit [Ping timeout: 245 seconds] | 20:30 | |
| @lisitsyn | yes we don't have anything on that | 20:32 |
| mikhailBelous | And I'm writing proposal in improving decision trees, so I can write it for your in GSOC format | 20:35 |
| @lisitsyn | you mean you want to develop decision trees as your gsoc project? | 20:36 |
| mikhailBelous | As an option | 20:39 |
| @lisitsyn | mikhailBelous: we'd need a mentor for that | 20:40 |
| @lisitsyn | Georg probably won't take it | 20:41 |
| mikhailBelous | So I can do lmnn as an option | 20:41 |
| -!- Vikalp [~AndChat49@223.185.83.111] has joined #shogun | 20:41 | |
| -!- Vikalp [~AndChat49@223.185.83.111] has quit [Client Quit] | 20:42 | |
| -!- saket [0e8b5206@gateway/web/freenode/ip.14.139.82.6] has quit [Quit: Page closed] | 20:53 | |
| -!- votjakovr [~chatzilla@host-46-241-3-209.bbcustomer.zsttk.net] has quit [Quit: ChatZilla 0.9.90 [Firefox 20.0.1/20130409194949]] | 20:57 | |
| van51 | hello all | 21:01 |
| van51 | I believe I have a fix for broken json bug | 21:02 |
| van51 | can someone walk me through the further testing needed and how to submit a patch? | 21:02 |
| -!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 21:14 | |
| -!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 21:25 | |
| shogun-notifier- | shogun: Sergey Lisitsyn :develop * 607c2ee / src/shogun/lib/tapkee/ (51 files): https://github.com/shogun-toolbox/shogun/commit/607c2ee1c1e12693186d039f3148cba0ea9a5958 | 21:25 |
| shogun-notifier- | shogun: Updated tapkee library | 21:25 |
| @lisitsyn | van51: submit it using a pull request | 21:26 |
| @lisitsyn | as for testing - well just describe what have you fixed we will take a look | 21:26 |
| -!- iglesiasg [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 21:35 | |
| -!- mode/#shogun [+o iglesiasg] by ChanServ | 21:35 | |
| -!- mikhailBelous [~towelenee@109.205.249.161] has quit [Ping timeout: 264 seconds] | 21:35 | |
| -!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has joined #shogun | 21:47 | |
| -!- travis-ci [~travis-ci@ec2-50-19-58-94.compute-1.amazonaws.com] has joined #shogun | 21:50 | |
| travis-ci | [travis-ci] it's Sergey Lisitsyn's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/6578855 | 21:50 |
| -!- travis-ci [~travis-ci@ec2-50-19-58-94.compute-1.amazonaws.com] has left #shogun [] | 21:50 | |
| -!- gsomix [~Miranda@46.20.65.94] has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org] | 21:50 | |
| @iglesiasg | Patrick mentioned that ./configure --enable-swig-directors crashes | 21:54 |
| @iglesiasg | error in csharp_modular interface when doing that | 21:54 |
| @iglesiasg | has someone seen that before? | 21:54 |
| -!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has quit [Ping timeout: 245 seconds] | 21:56 | |
| @wiking | lisitsyn: there's mp for clang! :) | 21:58 |
| @iglesiasg | wiking: macports? | 21:58 |
| @wiking | iglesiasg: ye | 21:59 |
| -!- deerishi_ [73f88294@gateway/web/freenode/ip.115.248.130.148] has joined #shogun | 22:05 | |
| -!- deerishi_ [73f88294@gateway/web/freenode/ip.115.248.130.148] has quit [Ping timeout: 245 seconds] | 22:12 | |
| -!- arink [~arink@180.149.62.52] has quit [Ping timeout: 255 seconds] | 22:12 | |
| -!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has joined #shogun | 22:24 | |
| deerishi | ? | 22:28 |
| deerishi | @iglesiasg : der? | 22:28 |
| @iglesiasg | yeah | 22:28 |
| deerishi | i am working on global alignment kernels. i am having some doubts understanding them.could you direct me to some material apart from the link you mentioned in the issue? | 22:30 |
| @iglesiasg | deerishi: mm no, not really | 22:31 |
| @iglesiasg | I don't know much about that apart from what I wrote in the issue | 22:31 |
| @iglesiasg | I think the references were ok though | 22:31 |
| deerishi | k.will try working on that itself. :) | 22:32 |
| @iglesiasg | deerishi: sure, IIRC there is some code there as well | 22:33 |
| @iglesiasg | deerishi: I am confident you can get it to work :) | 22:33 |
| deerishi | @iglesiasg : thanks.will try my level best. | 22:34 |
| deerishi | @iglesiasg : just an idea i had. i am very interested in speech recognition . is there any possibility of incorporating some speech recognition algorithims in shogun | 22:42 |
| @iglesiasg | deerishi: indeed | 22:46 |
| -!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has quit [Ping timeout: 245 seconds] | 22:46 | |
| @iglesiasg | I have done something in speech recognition too | 22:46 |
| @iglesiasg | we could discuss about it | 22:46 |
| -!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 22:49 | |
| -!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has joined #shogun | 22:49 | |
| deerishi | @iglesiasg : just an idea i had. i am very interested in speech recognition . is there any possibility of incorporating some speech recognition algorithims in shogun | 22:50 |
| @iglesiasg | deerishi: yeah, why not | 22:50 |
| @iglesiasg | deerishi: we could put something in the applications directory | 22:50 |
| @iglesiasg | deerishi: I have done a bith of speech recognition (just isolated word recognition) we could discuss about it | 22:51 |
| deerishi | sure .that 'll be great . :) | 22:52 |
| deerishi | i'll working on some things and i'l get back to you tommorow | 22:56 |
| deerishi | i'll start working on some things and i'l get back to you tommorow | 22:56 |
| @iglesiasg | ok | 22:56 |
| -!- vikram360 [~vikram360@117.221.124.81] has joined #shogun | 22:56 | |
| -!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has quit [Quit: Page closed] | 23:07 | |
| --- Log closed Wed Apr 24 00:00:28 2013 | ||
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!