--- Log opened Fri Apr 19 00:00:28 2013 | ||
wiking | trying | 00:36 |
---|---|---|
-!- utsavjain [~tiripS@14.139.82.6] has left #shogun [] | 01:27 | |
-!- iglesiasg [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 04:16 | |
-!- vikram360 [~vikram360@117.192.184.107] has joined #shogun | 04:41 | |
-!- gsomix [~Miranda@46.20.65.215] has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org] | 05:59 | |
-!- Shell_ [80c1080c@gateway/web/freenode/ip.128.193.8.12] has joined #shogun | 06:07 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has joined #shogun | 07:26 | |
foulwall | sonney2k: sonne|work: which version of python do we use? I use a importlib in the PR, and need python version>=2.7 | 07:28 |
-!- Shell_ [80c1080c@gateway/web/freenode/ip.128.193.8.12] has quit [Ping timeout: 245 seconds] | 07:41 | |
-!- hoijui [~hoijui@dslb-088-074-124-102.pools.arcor-ip.net] has joined #shogun | 08:12 | |
foulwall | sonney2k: I came up with an idea if I'm at leisure I can do some Chinese translation for shogun-web and make docs a Chinese version | 08:14 |
-!- travis-ci [~travis-ci@ec2-23-22-22-174.compute-1.amazonaws.com] has joined #shogun | 08:42 | |
travis-ci | [travis-ci] it's Viktor Gal'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/6410096 | 08:42 |
-!- travis-ci [~travis-ci@ec2-23-22-22-174.compute-1.amazonaws.com] has left #shogun [] | 08:42 | |
-!- travis-ci [~travis-ci@ec2-23-22-22-174.compute-1.amazonaws.com] has joined #shogun | 08:43 | |
travis-ci | [travis-ci] it's Viktor Gal'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/6415457 | 08:43 |
-!- travis-ci [~travis-ci@ec2-23-22-22-174.compute-1.amazonaws.com] has left #shogun [] | 08:43 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has quit [Read error: Connection reset by peer] | 09:17 | |
-!- gzhd79 [~zeller@embln.embl.de] has joined #shogun | 09:21 | |
-!- vikram360 [~vikram360@117.192.184.107] has quit [Ping timeout: 256 seconds] | 09:23 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has joined #shogun | 09:29 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has quit [Ping timeout: 272 seconds] | 09:36 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has joined #shogun | 09:38 | |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 09:41 | |
-!- foulwall` [~user@li379-21.members.linode.com] has joined #shogun | 09:42 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has quit [Remote host closed the connection] | 09:43 | |
-!- foulwall` [~user@li379-21.members.linode.com] has left #shogun [] | 09:43 | |
-!- foulwall [~user@li379-21.members.linode.com] has joined #shogun | 09:47 | |
-!- foulwall [~user@li379-21.members.linode.com] has quit [Remote host closed the connection] | 09:48 | |
-!- foulwall [~user@li379-21.members.linode.com] has joined #shogun | 09:51 | |
foulwall | sonney2k: Hey there? | 09:52 |
-!- vgorbati_ [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 10:21 | |
sonne|work | foulwall: hey! | 10:23 |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 10:24 | |
foulwall | sonne|work: I've mentioned in the email. we need a module named importlib, just pip it on the server. | 10:29 |
sonne|work | foulwall: why do we need it? | 10:29 |
foulwall | sonne|work: since we need to import the function at runtime, and the old __import__ is messy. | 10:31 |
foulwall | sonne|work: importlib is just a wrapper of __import__ | 10:31 |
sonne|work | foulwall: I don't understand why you need it though? Couldn't you just import everything statically? | 10:32 |
foulwall | sonne|work: for the extensibility. when a new demo is written, there's no need to modify the views.py, just put them in the demos' directory and they'll work. | 10:38 |
sonne|work | foulwall: ok I understand but for that you need to add checks that it doesn't import some random stuff | 10:39 |
sonne|work | otherwise this introduces security issues | 10:39 |
foulwall | it just import the modules with the prefix 'demos.' | 10:40 |
foulwall | when import something that is not exist, it respond with a 404 | 10:42 |
-!- iglesiasg [~nando@dyna218-117.csc.kth.se] has joined #shogun | 10:44 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 10:44 | |
sonne|work | ok that should work | 10:44 |
foulwall | Wait I'll pull another request | 10:46 |
-!- vgorbati_ [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 10:47 | |
foulwall | hide the funtion in the modules that is not permitted to invoke by user. | 10:47 |
sonne|work | foulwall: you basically have to make sure that no user input can harm the system | 10:49 |
sonne|work | so try to avoid imports based on user import | 10:49 |
sonne|work | function calls based on user input | 10:50 |
sonne|work | always only call hardcoded stuff when you see certain input | 10:50 |
sonne|work | iglesiasg: please announce the workshop on twitter | 10:50 |
@iglesiasg | sonne|work: ok | 10:52 |
* wiking is writing the unit tester for rocevaluation | 10:55 | |
foulwall | sonne|work: ok | 11:00 |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 11:01 | |
@iglesiasg | sonne|work: done! | 11:01 |
@iglesiasg | retweet it people! | 11:01 |
vgorbati | hi guys, I am trying to write a unit test according to https://github.com/shogun-toolbox/shogun/issues/882, but when I run 'make' in shogun/tests/unit it runs 40 tests, without my newly added unit test (some other tests are missing too). However, the obj file is generated. Can anyone help me? | 11:03 |
foulwall | iglesiasg: retweeted and followed you | 11:03 |
wiking | wtf: mathematics/Statistics_unittest.cc:85:29: error: variable length array of non-POD element type 'SGSparseVector<float64_t>' | 11:04 |
wiking | who wrote this?! | 11:05 |
wiking | and who merged it?! | 11:05 |
@iglesiasg | foulwall: nice, thank you! | 11:05 |
wiking | ok i know who wrote it.... | 11:05 |
wiking | heiko i will kill u | 11:07 |
wiking | :) | 11:07 |
@iglesiasg | haha | 11:07 |
-!- vgorbati_ [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 11:08 | |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 11:09 | |
wiking | oh motherfucker | 11:12 |
wiking | why is this not possible | 11:12 |
wiking | to not to fucking commit broken code into the repo | 11:12 |
-!- gzhd79 [~zeller@embln.embl.de] has quit [Quit: gzhd79] | 11:12 | |
@iglesiasg | wiking: you ok man? :D | 11:12 |
wiking | no because apparently whatever i try to make shogun work | 11:13 |
wiking | it's fucking broken | 11:13 |
wiking | ok here's a very funny unit test that fails | 11:16 |
wiking | oh even better | 11:17 |
wiking | i have a segfault | 11:17 |
wiking | ok | 11:19 |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 11:19 | |
shogun-notifier- | shogun: Viktor Gal :develop * ee8ab82 / tests/unit/evaluation/ROCEvaluation_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/ee8ab823011b611244c5ced763befa1f8eeb89fb | 11:19 |
shogun-notifier- | shogun: Add unit test for ROCEvaluation | 11:19 |
shogun-notifier- | shogun: as currently this little example fails | 11:19 |
wiking | can anybody tell me why this unittest fails? | 11:19 |
@iglesiasg | wiking: what does it say when it crashes? | 11:20 |
wiking | iglesiasg: [ERROR] assertion m_current_values.vector && idx<get_num_labels() failed in file labels/Labels.cpp line 57 | 11:21 |
wiking | unknown file: Failure | 11:21 |
wiking | Unknown C++ exception thrown in the test body. | 11:21 |
wiking | if i do the set_labels(vector) | 11:21 |
wiking | it's even worse | 11:21 |
wiking | it just segfaults | 11:21 |
@iglesiasg | wiking: mm I don't see what is wrong either | 11:26 |
-!- abinash [75ef5e6b@gateway/web/freenode/ip.117.239.94.107] has joined #shogun | 11:27 | |
@iglesiasg | wiking: I think it will work if you creat an SGVector<int32_t> with the labels and you give this to the CBinaryLabels constructor | 11:29 |
wiking | iglesiasg: i've tried almost the same | 11:29 |
@iglesiasg | but what you are doing you work too idd | 11:29 |
wiking | instead of giving it to the constructor | 11:29 |
wiking | use label->set_labels(vector) | 11:29 |
wiking | that segfaults | 11:29 |
wiking | :) | 11:29 |
@iglesiasg | yeah.. but try using it in the constructor | 11:29 |
@iglesiasg | I bet that will work | 11:30 |
wiking | but what is the difference?! | 11:30 |
wiking | i mean come on! | 11:30 |
@iglesiasg | I am checking Labels code | 11:30 |
wiking | that *must* work | 11:30 |
@iglesiasg | yeah, but I see something that might be weird | 11:30 |
wiking | yes for sure | 11:30 |
-!- vgorbati_ [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 11:30 | |
@iglesiasg | wiking: does it work fine using the SGVector in the constructor? | 11:31 |
sonne|work | wiking: labels are copied with one | 11:31 |
sonne|work | (in constructor) | 11:31 |
foulwall | sonne|work: hi sonne. | 11:31 |
sonne|work | not with the set_labels() | 11:31 |
sonne|work | foulwall: hi! | 11:31 |
wiking | sonne|work: should we remove set_labels? | 11:32 |
abinash | Hello I want to submit a code for the issue Probabilities for SVMs. What is the workflow for code submission | 11:32 |
sonne|work | wiking: no | 11:32 |
sonne|work | wiking: it should work | 11:32 |
foulwall | sonne|work: Excuse me. and I've changed the demoHandler into a more security one. | 11:32 |
wiking | ctor version works | 11:32 |
@iglesiasg | sonne|work, wiking : I think it might be something related to m_current_values and m_labels | 11:32 |
wiking | sonne|work iglesiasg [ERROR] assertion m_current_values.vector && idx<get_num_labels() failed in file labels/Labels.cpp line 57 | 11:33 |
wiking | with set_labels(vector) | 11:33 |
@iglesiasg | set_labels and what wiking was using does not set m_current_values I think | 11:33 |
foulwall | sonne|work: but it's a bit messy | 11:33 |
sonne|work | foulwall: btw please dont' do things like import * but just import what you need | 11:33 |
foulwall | sonne|work: I've changed that. | 11:33 |
foulwall | sonne|work: and i'll never do that. | 11:33 |
sonne|work | iglesiasg: ahh I see | 11:33 |
@iglesiasg | sonne|work: maybe it is a bug, I am not sure though | 11:34 |
wiking | iglesiasg: well for sure it's a bug :) | 11:34 |
sonne|work | wiking: no | 11:34 |
sonne|work | we have 2 functions | 11:34 |
sonne|work | set_labels | 11:34 |
sonne|work | and set_values | 11:34 |
sonne|work | they don't set both | 11:34 |
wiking | sonne|work: set_label(idx, value) | 11:34 |
wiking | ? | 11:34 |
sonne|work | only constructor does | 11:34 |
wiking | sonne|work: using set_labels(vector) and set_label(idx, value) | 11:35 |
sonne|work | that's not nice though - but no idea (currenlty) how to do it better | 11:35 |
wiking | will not work with ROC | 11:35 |
wiking | and this is a bug | 11:35 |
wiking | as we let this happen | 11:35 |
sonne|work | wiking: so do we ever want to be able to set these arrays *separately* | 11:36 |
sonne|work | iglesiasg: could you please help abinash? | 11:36 |
wiking | sonne|work: currently i have a usecase where yes i would like to | 11:36 |
sonne|work | iglesiasg: thanks for twittering... | 11:36 |
@iglesiasg | sonne|work: but what do m_labels and m_current_values represent? There must be a difference | 11:36 |
@iglesiasg | sonne|work: no problem | 11:36 |
@iglesiasg | hey abinash | 11:36 |
@iglesiasg | abinash: we have recently started using git flow in our development workflow | 11:37 |
sonne|work | wiking, iglesiasg - yes sure there is a difference: consider e.g. binary labels. m_labels would then just be -1/+1 values | 11:37 |
wiking | and what the hell is set_values anyways? | 11:37 |
-!- travis-ci [~travis-ci@ec2-23-23-3-13.compute-1.amazonaws.com] has joined #shogun | 11:38 | |
travis-ci | [travis-ci] it's Viktor Gal'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/6467279 | 11:38 |
-!- travis-ci [~travis-ci@ec2-23-23-3-13.compute-1.amazonaws.com] has left #shogun [] | 11:38 | |
@iglesiasg | abinash: see this mail on the mailing list for more info: http://article.gmane.org/gmane.comp.ai.machine-learning.shogun/3342 | 11:38 |
sonne|work | wiking: the confidence values | 11:39 |
wiking | sonne|work: and then set_labels ? | 11:39 |
@iglesiasg | sonne|work: so m_labels would be the -1/+1 got it, but m_current_values? | 11:40 |
sonne|work | iglesiasg: indeed no idea | 11:40 |
sonne|work | I guess some subset stuff? | 11:40 |
@iglesiasg | I thought so | 11:41 |
@iglesiasg | some of Heiko's magic for x-val | 11:41 |
wiking | mmm we need to enable unit testing on travis | 11:41 |
@iglesiasg | wiking, sonne|work: even if it is not a bug, I think it is a bit confusing | 11:42 |
sonne|work | it is totally | 11:42 |
sonne|work | I don't understand it | 11:42 |
@iglesiasg | I wouldn't understand why the test didn't work without digging into Labels stuff | 11:42 |
wiking | sonne|work: so now my answer for yesterday: shogun auROC: 0.900061 vs weka auroc: 0.955 | 11:42 |
sonne|work | we have to ask hypergenious heiko to parse that I guess | 11:42 |
@iglesiasg | maybe we should rename so stuff | 11:42 |
@iglesiasg | sonne|work: sure, let's see what he tells | 11:42 |
@iglesiasg | us | 11:43 |
wiking | sonne|work: so we are still off by a lot | 11:44 |
-!- abinash [75ef5e6b@gateway/web/freenode/ip.117.239.94.107] has quit [Ping timeout: 245 seconds] | 11:45 | |
sonne|work | wiking: yeah so now tell us what you do | 11:46 |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 11:46 | |
wiking | sonne|work: so now what's happening is that i use PolyKernel with 2.0 exponent | 11:46 |
wiking | sonne|work: + normalize the input with p=2.0 | 11:47 |
wiking | i.e. euclidean distance | 11:47 |
wiking | and use libsvm with C=1.0 | 11:47 |
wiking | bias enabled | 11:47 |
wiking | epsilon set to 1E-12 | 11:47 |
wiking | same settings in weka, using their smo | 11:48 |
wiking | and of course run it on the same train/test set | 11:48 |
wiking | i can right away try different kernels if u want me to, as i suspected that it should be better with other kernels | 11:50 |
wiking | and using kernel normalization does not change the auroc | 11:50 |
sonne|work | wiking: so with weka you use poly kernel | 11:52 |
wiking | sonne|work: yes | 11:52 |
sonne|work | in-homogeneous ? | 11:53 |
wiking | sonne|work: the non-normalized polykernel | 11:53 |
sonne|work | then dont' normalize data in shogun either | 11:53 |
wiking | yeah i've just tried both | 11:53 |
sonne|work | by default the poly kernel does | 11:53 |
sonne|work | set the identity kernel normalizer | 11:53 |
wiking | ok | 11:55 |
-!- HeikoS [~heiko@nat-172-14.internal.eduroam.ucl.ac.uk] has joined #shogun | 11:55 | |
wiking | HeikoS: ARGH | 11:55 |
wiking | HeikoS: get to work ! :) | 11:55 |
HeikoS | wiking: ? :) | 11:55 |
wiking | HeikoS: today you have a lot of task :))) | 11:55 |
HeikoS | what | 11:55 |
wiking | HeikoS: well first of all u merged code that breaks unit testing | 11:56 |
HeikoS | just arrived at werk :D | 11:56 |
HeikoS | work | 11:56 |
HeikoS | oh | 11:56 |
HeikoS | thats good, which ones? | 11:56 |
wiking | HeikoS: second check the rocevaluator unittest | 11:56 |
HeikoS | already did | 11:56 |
wiking | HeikoS: as i have nooooo clue why that fails | 11:56 |
HeikoS | it fails? | 11:56 |
wiking | HeikoS: yeps big time :) | 11:56 |
* sonne|work ~~~ooo~~~~~who~~let~~~the~~~dogs~~~out~~~ooo~~~~ | 11:56 | |
wiking | sonne|work: with identitynormalizer: 0.910194 | 11:57 |
HeikoS | sonne|work: ha you are alright? :) | 11:57 |
HeikoS | wiking: what dies it return? | 11:57 |
wiking | HeikoS: well ASSERT error | 11:57 |
sonne|work | HeikoS: wiking is biting you already :) | 11:57 |
wiking | HeikoS: cd tests/unit && make | 11:57 |
wiking | and you'll see | 11:57 |
sonne|work | wiking: then don't normalize in shogun - like you do in weka | 11:57 |
wiking | sonne|work: i mean i'm normalizing the input itself | 11:58 |
wiking | sonne|work: i shouldn't? | 11:58 |
HeikoS | wiking: does the buildbot fail? | 11:58 |
wiking | HeikoS: afaik buildbot doesn't run on develop branch | 11:58 |
wiking | or? | 11:58 |
HeikoS | I see | 11:58 |
HeikoS | oh thats not good | 11:58 |
HeikoS | I am confused | 11:59 |
HeikoS | the buildbot runs on master? | 11:59 |
HeikoS | and we commit to develop? | 11:59 |
HeikoS | why is that? | 11:59 |
wiking | HeikoS: yes | 11:59 |
wiking | HeikoS: because travis should do tit :) | 11:59 |
wiking | HeikoS: and it actually does most of it :P | 11:59 |
HeikoS | I see | 11:59 |
sonne|work | wiking: ohh man I guess you have to start from the beginning | 11:59 |
HeikoS | ehm? | 11:59 |
sonne|work | wiking: you got data - you normalize than save that data | 11:59 |
HeikoS | yes pls :) | 11:59 |
HeikoS | wiking so when does code migrate from develop to master? | 12:00 |
sonne|work | then use exactly that data in weka and shogun? | 12:00 |
wiking | sonne|work: L2Norm(input) -> PolyKernel(Exp=2.0) -> LIBSVM | 12:00 |
wiking | HeikoS: release | 12:00 |
sonne|work | wiking: nope | 12:00 |
wiking | sonne|work: that's the current pipeline for the data | 12:00 |
wiking | HeikoS: hahahah you grasped the most important part of the untittest :D | 12:01 |
HeikoS | wiking: I see | 12:02 |
HeikoS | what I dont really get is, why dont we run the buildbot on develop? | 12:02 |
wiking | HeikoS: i mean with your comment... the problem is that currently that little snippet fails | 12:02 |
wiking | HeikoS: hahahahahahahah | 12:02 |
wiking | HeikoS: travis not enough? | 12:02 |
HeikoS | wiking: so what does it return | 12:02 |
HeikoS | wiking: not all architectures | 12:03 |
wiking | HeikoS: [ERROR] assertion m_current_values.vector && idx<get_num_labels() failed in file labels/Labels.cpp line 57 | 12:03 |
HeikoS | so if we have errors that would not be detected there we will only notice when migrating to master | 12:03 |
wiking | HeikoS: ok i guess then we have to discuss this with sonne|work | 12:03 |
HeikoS | wiking: I see, so this seems to be a different problem | 12:03 |
HeikoS | wiking: also this warning stuff, its so nice to see them in the buildbot | 12:03 |
HeikoS | wiking: wait | 12:04 |
HeikoS | I actually saw commits by lambday compiling on the buildbot | 12:04 |
HeikoS | why is that? | 12:04 |
wiking | HeikoS: PR | 12:04 |
HeikoS | did he commit to master and I did not realise? | 12:04 |
wiking | or no | 12:04 |
HeikoS | explain for dumb people please ;) | 12:04 |
wiking | pr is not buildboted | 12:04 |
wiking | i have no idea | 12:04 |
wiking | HeikoS: explain me why that code fails | 12:04 |
wiking | ;) | 12:04 |
wiking | i'm really dumb as well in that case | 12:04 |
HeikoS | wiking: ok will do | 12:04 |
HeikoS | wiking: I notices some warnings in lambdays code in the build bot log | 12:05 |
HeikoS | and then he fixed them and the warnings went away | 12:05 |
wiking | HeikoS: https://github.com/shogun-toolbox/shogun/commits/master | 12:05 |
wiking | because it was merged into master for some reason | 12:05 |
wiking | ;P | 12:05 |
wiking | which is NONO NOOOO good | 12:06 |
wiking | but i guess you did not realise | 12:06 |
wiking | that he sent the PR | 12:06 |
HeikoS | no I did not | 12:06 |
wiking | for master branch | 12:06 |
wiking | https://github.com/shogun-toolbox/shogun/commits/master | 12:06 |
wiking | Merge pull request #961 from lambday/master ... | 12:06 |
wiking | dba5cb777c Browse code | 12:06 |
wiking | karlnapf authored 15 days ago | 12:06 |
wiking | ;) | 12:06 |
HeikoS | I see | 12:06 |
HeikoS | okay | 12:06 |
wiking | dont lie my friend, github catches u | 12:06 |
wiking | ;) | 12:06 |
HeikoS | haha | 12:06 |
HeikoS | so I have to make sure to merge into develop | 12:06 |
wiking | well just check the PR | 12:07 |
wiking | whether it's sent for master or develop branch | 12:07 |
HeikoS | ok | 12:07 |
HeikoS | good | 12:07 |
HeikoS | to know | 12:07 |
wiking | if it's coming for master branch | 12:07 |
HeikoS | and then travis catches things | 12:07 |
wiking | just drop it straight away | 12:07 |
wiking | we dont accept PR for master | 12:07 |
HeikoS | wiking, and then when we release we merge develop into master? | 12:07 |
-!- vgorbati_ [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 12:08 | |
HeikoS | wiking: okay! I knew this but forgot the check | 12:08 |
HeikoS | good that you tell me these things :) | 12:08 |
wiking | HeikoS: indeed we do the merging | 12:08 |
wiking | HeikoS: now it's going to be very interesting | 12:08 |
HeikoS | wiking: so could you add something to travis that collects warnings? | 12:08 |
wiking | HeikoS: becuase you got some stuff into master that is not in develop | 12:08 |
HeikoS | also for libshogun I would even vote to allow warnings as errors | 12:08 |
HeikoS | as for python | 12:08 |
HeikoS | but you have to ask sonne|work and lisitsyn on this | 12:09 |
wiking | HeikoS: so now we'll have to do some funky shit merging | 12:09 |
HeikoS | wiking: shit, sorry for that | 12:09 |
HeikoS | well, we can rebase right? | 12:09 |
wiking | HeikoS: heheh that's going to be a merge | 12:09 |
wiking | and not a simple rebase | 12:09 |
wiking | HeikoS: anyhow | 12:10 |
wiking | HeikoS: why is that unit test fails | 12:10 |
wiking | now it's your turn | 12:10 |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 12:10 | |
HeikoS | wiking: I will check | 12:10 |
HeikoS | lisitsyn: could you have a thought on this out-of-synch master and develop? I fucked it up :( | 12:10 |
HeikoS | how to restore that all in master is also in develop | 12:11 |
HeikoS | wiking: running tests | 12:11 |
-!- iglesiasg [~nando@dyna218-117.csc.kth.se] has quit [Ping timeout: 264 seconds] | 12:11 | |
wiking | HeikoS: cool | 12:11 |
HeikoS | wiking: so which tests did I make to fail? | 12:12 |
HeikoS | since only the ROC fail | 12:12 |
HeikoS | s | 12:12 |
wiking | HeikoS: no other fails | 12:12 |
wiking | HeikoS: but that is already being taken care of ;) | 12:12 |
wiking | HeikoS: dont worry about that | 12:12 |
wiking | HeikoS: now it's ROC time | 12:12 |
wiking | ok? :) | 12:12 |
wiking | oooohooooo | 12:12 |
wiking | sonne|work: another problem :))) | 12:13 |
-!- iglesiasg [~nando@n175-p42.kthopen.kth.se] has joined #shogun | 12:13 | |
wiking | sonne|work: my input is DenseFeatures<float32_t> | 12:13 |
wiking | sonne|work: and lets say i want to use Chi2Kernel | 12:13 |
wiking | sonne|work: Chi2Kernel.cpp:61 | 12:13 |
wiking | float64_t* avec= ((CDenseFeatures<float64_t>*) lhs)->get_feature_vector(idx_a, alen, afree); | 12:13 |
wiking | this will lead to a segfault | 12:13 |
wiking | :( | 12:13 |
wiking | i mean not this line | 12:13 |
wiking | but because of this | 12:14 |
wiking | soooooo | 12:14 |
wiking | how can we assure that no such case every can happen? | 12:14 |
wiking | i.e. if the input is not CDenseFeatures<float64_t> | 12:14 |
wiking | then at least it fails with an assert | 12:14 |
wiking | and not a segfault | 12:14 |
HeikoS | wiking: easy explanation | 12:14 |
HeikoS | ROC works on the scores | 12:15 |
wiking | HeikoS: okok listening, what's the fix, is it already in? | 12:15 |
HeikoS | not the integer labels | 12:15 |
HeikoS | I suggest the following: | 12:15 |
HeikoS | add a proper error message instead of the assert | 12:15 |
HeikoS | then, change the test a bit so that value vector is set | 12:15 |
HeikoS | I will do that | 12:15 |
wiking | HeikoS: ok just you know that this is confusing a lot | 12:15 |
wiking | right? | 12:15 |
HeikoS | what? | 12:16 |
HeikoS | wiking: the test is weird anyway | 12:16 |
wiking | HeikoS: well this that we have set_label and set_value | 12:16 |
HeikoS | ROC on binary integers doesnt make any sense | 12:16 |
HeikoS | set_value doenst work | 12:17 |
HeikoS | if the value vector is not set | 12:17 |
wiking | HeikoS: even if it doesn't make sense | 12:17 |
HeikoS | ok | 12:17 |
wiking | we should not die like this imo | 12:17 |
HeikoS | wiking: true | 12:19 |
HeikoS | we should have some thoughts on the label vs values vs scores design | 12:19 |
HeikoS | lisitsyn: ^ | 12:19 |
wiking | holaaa! | 12:21 |
wiking | sonne|work: ok now i've beaten weka | 12:21 |
wiking | sonne|work: as i suspected ;) | 12:21 |
wiking | sonne|work: but i had the segfaults because of the prev error i was mentioning | 12:21 |
HeikoS | wiking, sonne|work, lisitsyn we should have better error messages: I think it would be good to include the name of the method one is in class::name(): Error message | 12:22 |
HeikoS | I never know where the asserts are coming from | 12:22 |
wiking | HeikoS: lisitsyn sonne|work we should have a discussion about this: we have DenseFeatures<T> but then for example in some preprocessors and kernels have DenseFeatures<float64_t> hardcoded... this will cause segfaults... :( | 12:22 |
wiking | somehow it'd be great if we could do this part automatically | 12:22 |
wiking | e.g. add a preprocessor to the DenseFeature and that will use the right 'version' (template paramter) of the preprocessor | 12:23 |
wiking | any thoughts on this one? | 12:23 |
HeikoS | wiking: the test might also fail if you have only labels from one class (might happen since random is used) | 12:26 |
HeikoS | better to use odd/even for labels | 12:26 |
wiking | HeikoS: heheh ok ok | 12:26 |
wiking | HeikoS: i admit | 12:26 |
wiking | not the most sofisticated test | 12:26 |
wiking | i just wanted to push something to show that something is wrong atm | 12:27 |
HeikoS | wiking: I see, thats good | 12:28 |
-!- vgorbati__ [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 12:29 | |
@lisitsyn | re | 12:30 |
@lisitsyn | vgorbati__: still having the same issue? | 12:30 |
-!- vgorbati_ [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 12:30 | |
@lisitsyn | wiking: yes that's why casts are let me cite you NOO ONOONON | 12:30 |
-!- vgorbati_ [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 12:31 | |
wiking | lisitsyn: :D | 12:31 |
wiking | lisitsyn: "you dont love me and i know NOW!" :D | 12:31 |
-!- abinash [75ef5e6b@gateway/web/freenode/ip.117.239.94.107] has joined #shogun | 12:31 | |
wiking | lisitsyn: you mean that song? :P | 12:31 |
@lisitsyn | wiking: no, your "NONONO" | 12:32 |
@lisitsyn | :) | 12:32 |
-!- iglesiasg [~nando@n175-p42.kthopen.kth.se] has quit [Read error: Connection reset by peer] | 12:32 | |
wiking | lisitsyn: 'fuckit!' | 12:32 |
wiking | tableflipp | 12:32 |
wiking | :) | 12:32 |
-!- vgorbati__ [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 12:33 | |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 12:34 | |
abinash | Ok I got the workflow, but could someone explain how to submit code ? | 12:35 |
wiking | abinash: PR | 12:36 |
wiking | abinash: github pull request. make sure that it's against the develop branch | 12:36 |
-!- vgorbati_ [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 12:37 | |
-!- lambday [3ba2174d@gateway/web/freenode/ip.59.162.23.77] has joined #shogun | 12:38 | |
lambday | HeikoS: hi | 12:38 |
HeikoS | lambday: hi! | 12:39 |
lambday | should we always use SG_MALLOC while initializing SG-classes? | 12:39 |
HeikoS | lambday: just a quick note: if you are doing PR's please always do them on the develop branch | 12:40 |
HeikoS | lambday: the SGVector SGMatrix etc are treated differently | 12:40 |
HeikoS | we have automatic reference counting for them | 12:40 |
HeikoS | you *always* create them on the stack | 12:40 |
HeikoS | rest is done internally | 12:40 |
HeikoS | so always SGVector=SGVector(10) | 12:40 |
HeikoS | never SGVector*=new SGVector(10) | 12:40 |
@lisitsyn | do not do SGVector a = SGVector(10); | 12:41 |
@lisitsyn | do SGVector a(10); | 12:41 |
@lisitsyn | :) | 12:41 |
wiking | hahah :) | 12:41 |
HeikoS | lambday: you heard the man :D | 12:41 |
wiking | lambday: now u know :) | 12:41 |
@lisitsyn | HeikoS: it creates an object two times | 12:41 |
HeikoS | lisitsyn: yes you are in fact right | 12:42 |
HeikoS | lisitsyn: could you help in rebasing the develop branch against master? | 12:42 |
@lisitsyn | HeikoS: alright let me check | 12:42 |
HeikoS | wiking: how can I see the travis for my PR? | 12:42 |
HeikoS | ah found it! | 12:43 |
HeikoS | lisitsyn: thanks | 12:44 |
-!- lambday [3ba2174d@gateway/web/freenode/ip.59.162.23.77] has quit [Ping timeout: 245 seconds] | 12:44 | |
HeikoS | lisitsyn: what do you think of warnings as errors for libshogun and python_modular on travis? | 12:44 |
-!- lambday_ [3ba2174d@gateway/web/freenode/ip.59.162.23.77] has joined #shogun | 12:44 | |
HeikoS | we have to remove those anyway | 12:44 |
lambday_ | sorry I got disconnected | 12:44 |
@lisitsyn | HeikoS: doc warnings? | 12:44 |
HeikoS | in particular I like them for python since doxygen comments cannot be forgotten | 12:44 |
lambday_ | in one of my previous PR, I did something like SGSparseVector<float64_t> v[size-1];.... viktor and sonne|work pointed out that its not right | 12:44 |
@lisitsyn | I do not like that | 12:44 |
HeikoS | lisitsyn: yes, and also libshogun gcc | 12:44 |
HeikoS | why? | 12:44 |
@lisitsyn | I'd rather get rid of all the dummy doc we have | 12:44 |
@lisitsyn | it is useless | 12:45 |
HeikoS | lisitsyn: yes, do not merge dummy doc | 12:45 |
HeikoS | but the warning thing then helps you remembering to check | 12:45 |
HeikoS | since I tend to forget | 12:45 |
HeikoS | and others do to | 12:45 |
@lisitsyn | HeikoS: yes but it is not an error | 12:45 |
HeikoS | to get rid of dummy is responsibility of PR merger | 12:45 |
HeikoS | lisitsyn: okay agreed | 12:45 |
HeikoS | what about gcc warnings? | 12:45 |
HeikoS | my point is just: we tend to collect them | 12:46 |
@lisitsyn | HeikoS: I do not mind treating them as errors | 12:46 |
HeikoS | since its easy to oversee | 12:46 |
@lisitsyn | HeikoS: we should enable -Weffc++ :D :D | 12:46 |
HeikoS | whats that? | 12:46 |
@lisitsyn | warnings from the effective C++ book by meyers | 12:46 |
HeikoS | what do they tell you? | 12:47 |
@lisitsyn | I expect thousands in shogun | 12:47 |
@lisitsyn | virtual destructor, operator=, blabla | 12:47 |
HeikoS | I dont wont to be mean to people, just want to avoid piling up warnings | 12:47 |
HeikoS | but I guess one can also just check the warnings | 12:47 |
HeikoS | wiking: could we collect them in travis and output at the end? or make it yellow as in the buildbot? | 12:48 |
wiking | HeikoS: :))) google? | 12:48 |
@lisitsyn | HeikoS: so we want to rebase develop on top of last commit in master/ | 12:48 |
HeikoS | vgorbati: hi! | 12:48 |
@lisitsyn | ? | 12:48 |
-!- lambday_ [3ba2174d@gateway/web/freenode/ip.59.162.23.77] has quit [Ping timeout: 245 seconds] | 12:49 | |
@lisitsyn | wiking: ^ ? | 12:49 |
HeikoS | vgorbati: please use main channel for aksing questions, they might help others | 12:49 |
wiking | lisitsyn: kindof | 12:49 |
HeikoS | vgorbati: in order to make your unit test being executed, just use similar naming to the existing files | 12:49 |
HeikoS | lisitsyn: exactly | 12:49 |
@lisitsyn | I am curious if it would work without force push | 12:49 |
wiking | lisitsyn: merge | 12:50 |
@lisitsyn | wiking: yeah merge is the only possible thing I see | 12:50 |
@lisitsyn | wait wait | 12:51 |
wiking | i didnt touch anything | 12:51 |
@lisitsyn | wiking: but we have it already | 12:51 |
@lisitsyn | wiking: develop is on top of master d10e022a23a26b51e73a217c5014f3cf8d879ee6 | 12:52 |
@lisitsyn | there is nothing to merge | 12:52 |
HeikoS | sonne|work: could you explain me the problem with the sparse vector | 12:57 |
HeikoS | why is SGSparseVector v[size-1]; wrong? | 12:57 |
@lisitsyn | HeikoS: []? | 12:57 |
HeikoS | lisitsyn: ?? | 12:57 |
@lisitsyn | SGSparseVector v[size-1]; | 12:58 |
@lisitsyn | what does it mean? | 12:58 |
foulwall | sonne|work: Hi there? | 12:58 |
HeikoS | Ah | 12:58 |
HeikoS | I see! | 12:58 |
HeikoS | I guess he meant (size) | 12:58 |
HeikoS | with round brackets? | 12:58 |
HeikoS | thats how I read it | 12:58 |
@lisitsyn | HeikoS: yes | 13:00 |
@lisitsyn | HeikoS: with [] it probably means create vector v and get its size-1 element | 13:01 |
@lisitsyn | I'd expect a segfault there | 13:01 |
HeikoS | lisitsyn: yeah | 13:02 |
HeikoS | I was confused | 13:02 |
HeikoS | what am I saying | 13:03 |
HeikoS | I am confused ;) | 13:03 |
-!- iglesiasg [~nando@dyna218-117.csc.kth.se] has joined #shogun | 13:04 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 13:04 | |
HeikoS | wiking: travis fails complaining about not being able to read tem pfiles | 13:04 |
wiking | HeikoS: link | 13:04 |
HeikoS | https://travis-ci.org/shogun-toolbox/shogun/jobs/6468861 | 13:04 |
HeikoS | check the tests on the bottom | 13:05 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 966dd6b / src/shogun/evaluation/ROCEvaluation.cpp: https://github.com/shogun-toolbox/shogun/commit/966dd6bc8d386fab2fd998641e97976fd5eaf866 | 13:05 |
shogun-notifier- | shogun: replaced assert by proper error message | 13:05 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 65da3fd / src/shogun/labels/BinaryLabels.cpp: https://github.com/shogun-toolbox/shogun/commit/65da3fd2646dc13e2d36baaca314c5f7f2e36411 | 13:05 |
shogun-notifier- | shogun: added method name to error messages and re-formated | 13:05 |
shogun-notifier- | shogun: Heiko Strathmann :develop * ce363f3 / src/shogun/labels/DenseLabels.cpp: https://github.com/shogun-toolbox/shogun/commit/ce363f3ca42b8fc5729e3b00eeee9f1048290170 | 13:05 |
shogun-notifier- | shogun: made set_cont to affect values/scores | 13:05 |
shogun-notifier- | shogun: allocate value memory | 13:05 |
shogun-notifier- | shogun: add values to parameters | 13:05 |
shogun-notifier- | shogun: Heiko Strathmann :develop * a19166f / tests/unit/evaluation/ROCEvaluation_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/a19166f93719665d21705f032d50b3d8f2245156 | 13:05 |
shogun-notifier- | shogun: made unit-test work | 13:05 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 53378a5 / / (4 files): https://github.com/shogun-toolbox/shogun/commit/53378a5092d298c6e9cd30d373ae11efcb23f69f | 13:05 |
shogun-notifier- | shogun: Merge pull request #985 from karlnapf/develop | 13:05 |
shogun-notifier- | shogun: | 13:05 |
shogun-notifier- | shogun: Fixed ROC unit test | 13:05 |
HeikoS | I still feel a bit uncomfortable that the buildbot doesn't run on develop ... like being in the dark. But should be fine with travis you are right | 13:06 |
@iglesiasg | HeikoS, wiking, sonne|work : so it was bug in Labels at the end? :) | 13:07 |
wiking | HeikoS: who creates those tmp files? | 13:07 |
HeikoS | wiking: ehm, the tests? | 13:07 |
wiking | HeikoS: no idea that's why i'm asking | 13:07 |
sonne|work | wiking please take raw data for both | 13:08 |
sonne|work | foulwall: yes | 13:08 |
sonne|work | HeikoS: SGSparseVector v[size-1] creates | 13:08 |
sonne|work | size-1 sparse vectors | 13:08 |
sonne|work | (empty ones) | 13:08 |
sonne|work | but on the stack | 13:08 |
sonne|work | very dangerous thing - can make the system run out of stack | 13:09 |
foulwall | sonne|work: What could I do next? | 13:09 |
sonne|work | and with a dynamic parameter like size it is even a non-standard gnu extension | 13:09 |
@lisitsyn | sonne|work: bah I forgot about arrays, true | 13:10 |
sonne|work | foulwall: eye candy? | 13:11 |
sonne|work | I mean some animation | 13:11 |
sonne|work | grid etc? | 13:11 |
foulwall | sonne|work: ok | 13:12 |
-!- vgorbati_ [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 13:12 | |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 13:14 | |
-!- foulwall [~user@li379-21.members.linode.com] has left #shogun ["ERC Version 5.3 (IRC client for Emacs)"] | 13:18 | |
-!- abinash [75ef5e6b@gateway/web/freenode/ip.117.239.94.107] has quit [Ping timeout: 245 seconds] | 13:19 | |
-!- travis-ci [~travis-ci@ec2-23-22-22-174.compute-1.amazonaws.com] has joined #shogun | 13:24 | |
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/6469168 | 13:24 |
-!- travis-ci [~travis-ci@ec2-23-22-22-174.compute-1.amazonaws.com] has left #shogun [] | 13:24 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has joined #shogun | 13:26 | |
-!- Marty28 [~marty@5.28.86.117] has joined #shogun | 14:06 | |
-!- foulwall_ [~foulwall@li379-21.members.linode.com] has joined #shogun | 14:13 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has quit [Read error: Connection reset by peer] | 14:13 | |
-!- Marty28 [~marty@5.28.86.117] has quit [Quit: Colloquy for iPad - http://colloquy.mobi] | 14:15 | |
HeikoS | wiking: could you have a look at one of the travis ones? This is happeneing all the time | 14:17 |
HeikoS | https://travis-ci.org/shogun-toolbox/shogun/jobs/6469528 | 14:17 |
shogun-notifier- | shogun: lambday :develop * a28c730 / tests/unit/mathematics/Statistics_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/a28c730a47c97b4ee8a5ba79885557e927d033e2 | 14:18 |
shogun-notifier- | shogun: replaced variable-length array with SG_MALLOC in log_det_test_3 | 14:18 |
shogun-notifier- | shogun: Heiko Strathmann :develop * a343111 / tests/unit/mathematics/Statistics_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/a343111f0b8c5512c2c0892bd88c324ff4aadbf0 | 14:18 |
shogun-notifier- | shogun: Merge pull request #986 from lambday/develop | 14:18 |
shogun-notifier- | shogun: | 14:18 |
shogun-notifier- | shogun: replaced variable-length array with SG_MALLOC in log_det_test_3 | 14:18 |
-!- vgorbati_ [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 14:20 | |
wiking | HeikoS: the /tmp stuff? | 14:26 |
HeikoS | wiking yes | 14:26 |
wiking | HeikoS: no idea man | 14:27 |
wiking | HeikoS: i'll try to google it but... | 14:27 |
wiking | i'm not a travis expert ;) | 14:27 |
HeikoS | wiking: ok | 14:28 |
HeikoS | mmh | 14:28 |
HeikoS | then lets ask somebody else :) | 14:28 |
HeikoS | why does it only fail on this particular build? | 14:28 |
shogun-notifier- | shogun: Heiko Strathmann :develop * ff28dd0 / examples/undocumented/libshogun/labels_binary_fit_sigmoid.cpp: https://github.com/shogun-toolbox/shogun/commit/ff28dd0306d1b59f1fd71252a9169a3372ce6811 | 14:32 |
shogun-notifier- | shogun: removed assertions and print probabilities | 14:32 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 7f31c48 / tests/unit/labels/BinaryLabels_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/7f31c48370ed1781aabea83e951c3069206051e7 | 14:32 |
shogun-notifier- | shogun: added unit test for sigmoid fitting | 14:32 |
shogun-notifier- | shogun: Heiko Strathmann :develop * d30f205 / / (2 files): https://github.com/shogun-toolbox/shogun/commit/d30f20557e0282e8bd8580608b8583aa2586b1a7 | 14:32 |
shogun-notifier- | shogun: Merge pull request #988 from karlnapf/develop | 14:32 |
shogun-notifier- | shogun: | 14:32 |
shogun-notifier- | shogun: unit test for binary label sigmoid fitting | 14:32 |
-!- rishabh [~rishabh@202.78.172.162] has joined #shogun | 14:35 | |
-!- travis-ci [~travis-ci@ec2-23-22-22-174.compute-1.amazonaws.com] has joined #shogun | 14:37 | |
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/6470306 | 14:37 |
-!- travis-ci [~travis-ci@ec2-23-22-22-174.compute-1.amazonaws.com] has left #shogun [] | 14:37 | |
@lisitsyn | argh come on will they finally catch this asshole? | 14:49 |
HeikoS | ? | 14:51 |
@lisitsyn | HeikoS: watertown, usa | 14:52 |
@iglesiasg | lisitsyn: oh so the police killed the guy | 14:53 |
@lisitsyn | iglesiasg: yes one moron killed | 14:53 |
HeikoS | I see | 14:54 |
@lisitsyn | iglesiasg: as I tweeted - these nations are troublesome for anyone - so usa should think of sheltering such animals | 14:54 |
@lisitsyn | they usually do the same here | 14:54 |
@iglesiasg | lisitsyn: to what do you refer to with these nations? | 14:54 |
@iglesiasg | where these guys that did it come from? | 14:54 |
@lisitsyn | iglesiasg: chechens, dagestanians | 14:55 |
@iglesiasg | I see | 14:55 |
@lisitsyn | iglesiasg: they came from dagestan but they are chechens | 14:55 |
-!- travis-ci [~travis-ci@ec2-23-22-22-174.compute-1.amazonaws.com] has joined #shogun | 14:55 | |
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/6470583 | 14:55 |
-!- travis-ci [~travis-ci@ec2-23-22-22-174.compute-1.amazonaws.com] has left #shogun [] | 14:55 | |
@lisitsyn | iglesiasg: well it is adjacent regions | 14:55 |
@iglesiasg | I see | 14:56 |
-!- foulwall_ [~foulwall@li379-21.members.linode.com] has quit [Read error: Connection reset by peer] | 15:07 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has joined #shogun | 15:08 | |
foulwall | sonne|work: some eye candy | 15:09 |
foulwall | sonne|work: http://foulwall.com:8000/demo/svr/entrance | 15:09 |
@iglesiasg | foulwall: you are an artist! :) | 15:10 |
sonne|work | foulwall: nice. it could move a bit faster | 15:10 |
foulwall | lisitsyn: sonne|work thankyou | 15:10 |
-foulwall:#shogun- I'll add some animation on 'clear' | 15:11 | |
foulwall | sonne|work: I'm wondering what kind of animation can be added to the cluster demo. | 15:12 |
sonne|work | foulwall: yeah would be cool | 15:12 |
sonne|work | foulwall: make the diameter of the circle from 0 -> r | 15:12 |
@iglesiasg | foulwall: if you are using k-means, you can do a nice one | 15:12 |
@iglesiasg | foulwall: since this method is iterative | 15:13 |
sonne|work | iglesiasg: he only has the end result | 15:13 |
sonne|work | it's the shogun output | 15:13 |
@iglesiasg | foulwall: mmm however that would require intermediate results | 15:13 |
@iglesiasg | sonne|work: idd I noted that | 15:13 |
foulwall | sonne|work: ok | 15:13 |
@iglesiasg | sonne|work: I think it is worth trying | 15:13 |
sonne|work | iglesiasg: the same would work for an svm | 15:13 |
sonne|work | or svr | 15:13 |
sonne|work | think of it just training one iteration | 15:14 |
@iglesiasg | aham true | 15:14 |
@iglesiasg | I didn't come up with it for the svm and svr | 15:14 |
HeikoS | foulwall, could you connect the lines between the points in the regression function? | 15:15 |
HeikoS | when you make the kernel width very small you see the individual points | 15:16 |
foulwall | HeikoS: I note that, wait me and I'll fix that | 15:16 |
HeikoS | foulwall, this is really cool! | 15:17 |
foulwall | HeikoS: thank you | 15:17 |
HeikoS | foulwall: you could do this for the GP classifier and regression | 15:17 |
HeikoS | foulwall, you know what would be even cooler? | 15:17 |
HeikoS | if you could plot the underlying distribution | 15:17 |
HeikoS | like a heatmap | 15:17 |
HeikoS | at least for classifier | 15:17 |
foulwall | HeikoS: that's what I want to do next | 15:17 |
HeikoS | wow | 15:17 |
HeikoS | I am amazed :D | 15:17 |
HeikoS | keep on the good work! | 15:18 |
sonne|work | HeikoS, foulwall that is sth daniel has done already | 15:18 |
HeikoS | sonne|work, ok, I didnt see that | 15:18 |
sonne|work | HeikoS: https://github.com/dvalcarce/shogun-gsoc | 15:18 |
sonne|work | it is also pretty cool | 15:18 |
foulwall | sonne|work: I'll go for class. later i'll fix them. | 15:18 |
sonne|work | foulwall: thanks for your work | 15:19 |
sonne|work | cu! | 15:19 |
HeikoS | sonne|work seeing, wow that also very cool | 15:20 |
thoralf | sonne|work: about text features again: i think it still makes sense to keep the preprocessing part outside shogun, but: | 15:25 |
-!- foulwall [~foulwall@li379-21.members.linode.com] has quit [Remote host closed the connection] | 15:25 | |
thoralf | allowing files "feat1 feat2 ..." would be the way | 15:25 |
thoralf | n-grams etc. could be created outside like feat1_feat2_... feat2_feat3_... etc. | 15:26 |
thoralf | All you have to to is to create TextFeature classes - and eventually something that maps them into a feature space. | 15:27 |
-!- sonne|work [~sonnenbu@sams-office-nat.tomtomgroup.com] has quit [Quit: Leaving.] | 15:28 | |
-!- Asix3 [452f9490@gateway/web/freenode/ip.69.47.148.144] has joined #shogun | 15:32 | |
-!- lambday [0e8b614d@gateway/web/freenode/ip.14.139.97.77] has joined #shogun | 15:37 | |
lambday | HeikoS: hi | 15:37 |
-!- rishabh [~rishabh@202.78.172.162] has quit [Ping timeout: 264 seconds] | 15:38 | |
lambday | HeikoS: I added the sample_from_gaussian for sparse cov matrices and a unit test.. | 15:38 |
lambday | HeikoS: shall I send the PR and you could take a look? | 15:38 |
lambday | HeikoS: also changed the array of pointers that in the log_det unit test to double pointer and used SG_MALLOC instead | 15:39 |
HeikoS | lambday: I saw that, nice! already merged | 15:41 |
HeikoS | send the PR! | 15:42 |
HeikoS | looking forward to see it :) | 15:42 |
lambday | HeikoS: okay sending :) | 15:42 |
lambday | HeikoS: as of now, I used a constant diagonal matrix for sparse unit-test.. in the next unit test will add a random one | 15:44 |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 15:46 | |
-!- Asix3_ [452f9490@gateway/web/freenode/ip.69.47.148.144] has joined #shogun | 15:52 | |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 15:53 | |
-!- Asix3 [452f9490@gateway/web/freenode/ip.69.47.148.144] has quit [Ping timeout: 245 seconds] | 15:53 | |
-!- trtr3434 [~trtr3434@slb-96-33.tm.net.my] has joined #shogun | 16:03 | |
-!- trtr3434 [~trtr3434@slb-96-33.tm.net.my] has left #shogun [] | 16:03 | |
-!- iglesiasg [~nando@dyna218-117.csc.kth.se] has quit [Quit: leaving] | 16:08 | |
wiking | dudeeez | 16:13 |
wiking | HeikoS: mathematics/Statistics.cpp: In constructor 'shogun::CStatistics::log_det(shogun::SGSparseMatrix<double>)::EigenSimplicialLLT::EigenSimplicialLLT()': | 16:14 |
wiking | mathematics/Statistics.cpp:2030:12: error: 'SimplicialCholeskyLLt' was not declared in this scope | 16:14 |
wiking | make[1]: [mathematics/Statistics.cpp.o] Error 1 | 16:14 |
wiking | make[1]: Leaving directory `/home/niki/Dropbox/work/workspace/shogun/src/shogun' | 16:14 |
wiking | fails on ubuntu | 16:14 |
wiking | make: *** [libshogun.so.14.0] Error 2 | 16:14 |
HeikoS | oh, then we should add this PR recently | 16:14 |
HeikoS | and a buildbot on develop :) | 16:14 |
-!- lambday [0e8b614d@gateway/web/freenode/ip.14.139.97.77] has quit [Ping timeout: 245 seconds] | 16:14 | |
HeikoS | wiking: gsomix sent this PR recently to solve this problem | 16:15 |
-!- lambday [0e8b614d@gateway/web/freenode/ip.14.139.97.77] has joined #shogun | 16:15 | |
lambday | HeikoS: I think I made a mistake.. checking again | 16:15 |
-!- Asix3_ [452f9490@gateway/web/freenode/ip.69.47.148.144] has quit [Ping timeout: 245 seconds] | 16:16 | |
wiking | HeikoS: this is *master* branch | 16:18 |
HeikoS | wiking: argh | 16:18 |
wiking | so there should be definitely done something about this | 16:18 |
HeikoS | my fault again I guess :( | 16:18 |
-!- rishabh [~rishabh@202.78.172.162] has joined #shogun | 16:18 | |
HeikoS | wiking, link | 16:18 |
HeikoS | ? | 16:18 |
HeikoS | wiking: good chance to try the hotfix branch ;) | 16:18 |
wiking | yes | 16:20 |
wiking | it is a good chance for that | 16:20 |
HeikoS | lambday: no worries, not merged yet | 16:20 |
HeikoS | wiking: have we already rebased cev against master? | 16:20 |
HeikoS | dev | 16:20 |
HeikoS | since I suggest to do this before | 16:21 |
HeikoS | it might get messy otherwise | 16:21 |
-!- juanmanpr [c2fe7d88@gateway/web/freenode/ip.194.254.125.136] has joined #shogun | 16:21 | |
juanmanpr | hello guys, about gsoc, what kine of profiles are you searching on the students? | 16:22 |
juanmanpr | I am studying master in computer vision, and im really interested in machine learning | 16:23 |
juanmanpr | i have experience in C, C++, Matlab, and a little python | 16:23 |
HeikoS | juanmanpr: we are looking for motivated people than can code and have some background in the topic their project is about | 16:24 |
HeikoS | juanmanpr: the best way to get involved is to start sending patches and discuss project ideas | 16:24 |
HeikoS | with the mentors | 16:24 |
juanmanpr | But, is there any special requirement that you are looking, for example to Interface and develop the general quadratic solver library libqp | 16:24 |
juanmanpr | ok, where can I finf a list of menthors for shogun? | 16:25 |
HeikoS | on our ideas page | 16:25 |
juanmanpr | thanks, i'll take a look | 16:25 |
HeikoS | there is no "hard" requirement, you just have to convince us of you :) | 16:25 |
HeikoS | as a computer vision student, you should have the right background | 16:25 |
juanmanpr | You also would like to see some code, right? | 16:26 |
HeikoS | so just show us that you can code and are motivated by sending patches :) see our issues list on github for initial tasks you could do | 16:26 |
HeikoS | yes, it is a requirement for your application to be considered | 16:26 |
HeikoS | not only show | 16:26 |
HeikoS | but actually merged to shogun codebase | 16:26 |
-!- lambday [0e8b614d@gateway/web/freenode/ip.14.139.97.77] has quit [Ping timeout: 245 seconds] | 16:27 | |
juanmanpr | I see, well, I better do something now! | 16:27 |
juanmanpr | Thanks | 16:28 |
HeikoS | juanmanpr: looking forward to see a pull request! | 16:28 |
HeikoS | I gotta go now, bye! | 16:28 |
juanmanpr | bye | 16:28 |
-!- juanmanpr [c2fe7d88@gateway/web/freenode/ip.194.254.125.136] has quit [Quit: Page closed] | 16:32 | |
-!- dvalcarce [~dvalcarce@86.117.117.91.dynamic.mundo-r.com] has joined #shogun | 16:37 | |
-!- rishabh [~rishabh@202.78.172.162] has quit [Ping timeout: 264 seconds] | 16:40 | |
dvalcarce | hi all | 16:40 |
wiking | anybody seen this error before: /usr/local/include/shogun/lib/Signal.h:102:14: error: expected unqualified-id before 'protected' | 16:51 |
wiking | ok we have a problem | 16:59 |
wiking | i mean it's not our problem | 16:59 |
wiking | but it's qt's fault | 17:00 |
wiking | basically qmake has a #define signals | 17:00 |
wiking | -> we get that error with shogun if u use a code in qmake that uses shogun (headers) | 17:00 |
wiking | ok | 17:18 |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 17:32 | |
-!- foulwall [~foulwall@2001:da8:215:c252:edf7:a7fe:6f7f:ef7f] has joined #shogun | 17:40 | |
foulwall | Hi, dvalcarce | 17:40 |
dvalcarce | foulwall: hi! | 17:41 |
foulwall | dvalcarce: I am also applying for GSoC the interactive demos. | 17:41 |
foulwall | dvalcarce: we can work together and make better demos. | 17:41 |
dvalcarce | yeah, that's a good idea | 17:42 |
dvalcarce | i was thinking the same thing | 17:42 |
foulwall | dvalcarce: :-) | 17:42 |
dvalcarce | i see your regression demo, it's cool | 17:43 |
dvalcarce | maybe we can split server and client sides work | 17:43 |
dvalcarce | i don't know what sonney2k will think about it | 17:43 |
foulwall | dvalcarce: good idea. | 17:44 |
foulwall | dvalcarce: and dont' know either. | 17:45 |
dvalcarce | let's ask him then :P | 17:46 |
dvalcarce | which part do you prefer? | 17:46 |
dvalcarce | because right now i think we are doing duplicated work on django side | 17:47 |
dvalcarce | i think we should define some basic interfaces | 17:48 |
foulwall | both are ok, and the front end needs more creativity. | 17:49 |
dvalcarce | i agree, bootstrap is beautiful but it isn't enough | 17:50 |
foulwall | I've considered building a interface, but for the security reasons... | 17:50 |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 17:51 | |
dvalcarce | i understand | 17:52 |
dvalcarce | it would be convenient to have a list of demos so we can plan the website, too | 17:52 |
thoralf | Bye. See you monday. | 17:53 |
-!- thoralf [~thoralf@enki.zib.de] has left #shogun ["Konversation terminated!"] | 17:53 | |
foulwall | sonney2k told me it's better to limit the access. | 17:53 |
vgorbati | lisitsyn, hi are you there? | 17:55 |
dvalcarce | we need server communication, but we don't want a ddos | 17:56 |
foulwall | dvalcarce: yes. | 17:57 |
dvalcarce | some kind of authentication? | 17:57 |
dvalcarce | or maybe it is too overhead because of some demos? | 17:57 |
foulwall | dvalcarce: auth is easy to implement but we need somekind of elastic way to make the demo code extensibility | 17:59 |
foulwall | dvalcarce: code reuse and general interface | 17:59 |
foulwall | but a general interface may cause some function exposed to the bad user. | 18:02 |
dvalcarce | we can implement data size or computing time limitations | 18:04 |
-!- rishabh [~rishabh@202.78.172.162] has joined #shogun | 18:11 | |
-!- hoijui [~hoijui@dslb-088-074-124-102.pools.arcor-ip.net] has quit [Ping timeout: 246 seconds] | 18:12 | |
-!- se7en7 [~babu@2001:6b0:1:1430:21e:4fff:fec0:f4e2] has quit [Remote host closed the connection] | 18:13 | |
-!- vikram360 [~vikram360@117.216.145.131] has joined #shogun | 18:14 | |
-!- lambday [0e8b614d@gateway/web/freenode/ip.14.139.97.77] has joined #shogun | 18:14 | |
lambday | HeikoS: hi.. there? | 18:14 |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 18:16 | |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 18:29 | |
-!- lambday [0e8b614d@gateway/web/freenode/ip.14.139.97.77] has quit [] | 18:33 | |
-!- foulwall [~foulwall@2001:da8:215:c252:edf7:a7fe:6f7f:ef7f] has quit [Ping timeout: 245 seconds] | 18:41 | |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 18:47 | |
-!- dvalcarce [~dvalcarce@86.117.117.91.dynamic.mundo-r.com] has quit [Quit: dvalcarce] | 18:54 | |
-!- dvalcarce [~dvalcarce@86.117.117.91.dynamic.mundo-r.com] has joined #shogun | 18:54 | |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 18:57 | |
-!- rishabh [~rishabh@202.78.172.162] has quit [Ping timeout: 256 seconds] | 19:01 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has joined #shogun | 19:03 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has quit [Ping timeout: 258 seconds] | 19:16 | |
-!- rishabh [~rishabh@202.78.172.162] has joined #shogun | 19:19 | |
-!- HeikoS [~heiko@nat-172-14.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.] | 19:25 | |
-!- vgorbati_ [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 19:26 | |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 19:30 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has joined #shogun | 19:37 | |
foulwall | lisitsyn: dvalcarce sonney2k: I've add another eyecandy for the clustering / svr demo. | 19:39 |
foulwall | you can see that on http://foulwall.com:8000/demo/clustering/entrance http://foulwall.com:8000/demo/svr/entrance | 19:39 |
-!- hoijui [~hoijui@wlan-141-23-66-81.tubit.tu-berlin.de] has joined #shogun | 19:40 | |
foulwall | it's late in beijing, i'll go sleeping. goodnight. | 19:40 |
dvalcarce | nice :) | 19:41 |
dvalcarce | good night | 19:41 |
@lisitsyn | nite | 19:41 |
@lisitsyn | foulwall: and nice | 19:42 |
foulwall | HeikoS: I've connected the dots into line so svr demo can work correctly when the width is small. | 19:43 |
-!- foulwall [~foulwall@li379-21.members.linode.com] has quit [Ping timeout: 258 seconds] | 19:52 | |
-!- rishabh [~rishabh@202.78.172.162] has quit [Ping timeout: 256 seconds] | 20:02 | |
-!- rishabh [~rishabh@202.78.172.162] has joined #shogun | 20:13 | |
-!- dvalcarce [~dvalcarce@86.117.117.91.dynamic.mundo-r.com] has quit [Quit: dvalcarce] | 20:25 | |
-!- sidbee [~sidbee@202.78.172.162] has joined #shogun | 20:40 | |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 20:45 | |
-!- vgorbati_ [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 20:46 | |
-!- hoijui [~hoijui@wlan-141-23-66-81.tubit.tu-berlin.de] has quit [Quit: Leaving] | 20:48 | |
@lisitsyn | vgorbati: what you may do now is to replace this 100% check with something weaker | 20:54 |
-!- sidbee [~sidbee@202.78.172.162] has quit [] | 20:55 | |
vgorbati | lisitsyn: experimentally find some percentage, that will pass? | 20:55 |
@lisitsyn | vgorbati: kind of | 20:55 |
@lisitsyn | vgorbati: you may check the other thing though | 20:56 |
vgorbati | lisitsyn: ? | 20:56 |
@lisitsyn | whether weights are preserved | 20:56 |
@lisitsyn | like for any vector x in initial space | 20:56 |
@lisitsyn | the linear combination weights are the same | 20:57 |
@lisitsyn | for initial space and embedded space | 20:57 |
vgorbati | well, that will complicate things a little bit) | 20:58 |
vgorbati | it's like doing the big part of LLE itself | 20:59 |
vgorbati | lisitsyn: I guess I can't receive the neighbors weights from CLocallyLinearEmbedding somehow, can I/ | 21:01 |
-!- n4nd0 [~fircuser@c83-251-227-64.bredband.comhem.se] has joined #shogun | 21:14 | |
-!- n4nd0 is now known as iglesiasg | 21:15 | |
@lisitsyn | vgorbati: yes I don't like that way too | 21:17 |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 252 seconds] | 21:21 | |
vgorbati | lisitsyn: I think I will implement the "weakening" part on Monday - is it ok? | 21:23 |
@lisitsyn | vgorbati: totally | 21:26 |
-!- vgorbati_ [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 21:30 | |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 21:32 | |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun | 21:36 | |
-!- qpalzm [~ck@78.96.110.81] has joined #shogun | 21:45 | |
-!- lisitsyn [~blackburn@109-226-120-160.clients.tlt.100megabit.ru] has quit [Quit: Leaving.] | 21:51 | |
-!- lisitsyn [~blackburn@109-226-120-160.clients.tlt.100megabit.ru] has joined #shogun | 21:52 | |
-!- mode/#shogun [+o lisitsyn] by ChanServ | 21:52 | |
-!- vgorbati_ [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 21:54 | |
-!- vikram360 [~vikram360@117.216.145.131] has quit [Ping timeout: 248 seconds] | 22:04 | |
-!- gsomix [~Miranda@37.61.180.100] has joined #shogun | 22:10 | |
gsomix | good evening | 22:10 |
iglesiasg | hello gsomix | 22:10 |
iglesiasg | how is gsoc preparation going? | 22:11 |
gsomix | iglesiasg: bit by bit. | 22:15 |
gsomix | lisitsyn: https://dl.dropboxusercontent.com/u/19029407/uzkyIYZDWzo.jpg | 22:35 |
@lisitsyn | the dream city | 22:35 |
gsomix | lisitsyn: I have one more https://dl.dropboxusercontent.com/u/19029407/khKJokVI0ZY.jpg | 22:36 |
gsomix | huh | 22:36 |
@lisitsyn | anyone watching for the 18th hour of that madness? | 22:44 |
@lisitsyn | :D | 22:44 |
-!- mdragan [6d66bb92@gateway/web/freenode/ip.109.102.187.146] has joined #shogun | 22:48 | |
-!- gsomix [~Miranda@37.61.180.100] has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org] | 22:52 | |
-!- iglesiasg [~fircuser@c83-251-227-64.bredband.comhem.se] has quit [Ping timeout: 264 seconds] | 22:54 | |
@lisitsyn | suspect in custody | 23:17 |
@lisitsyn | ? | 23:17 |
@lisitsyn | uh | 23:17 |
-!- iglesiasg [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 23:40 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 23:40 | |
@lisitsyn | iglesiasg: join me listening to scanner :D | 23:41 |
@lisitsyn | iglesiasg: I admit I am feeling bad already have been listening to it a hour or so | 23:41 |
@iglesiasg | lisitsyn: let me check | 23:43 |
@iglesiasg | I have never listened to it | 23:43 |
@lisitsyn | iglesiasg: http://www.ustream.tv/channel/ma-rt-9-window-cam?utm_campaign=t.co&utm_source=6807252&utm_medium=social | 23:43 |
@iglesiasg | lisitsyn: btw, have you seen? https://github.com/shogun-toolbox/shogun/pull/990 | 23:43 |
@iglesiasg | nice code I think | 23:44 |
@lisitsyn | iglesiasg: yes sure | 23:44 |
@iglesiasg | lisitsyn: what is this man? | 23:45 |
@iglesiasg | is it like the police radio | 23:45 |
@lisitsyn | iglesiasg: yes | 23:45 |
@iglesiasg | that wouldn't make sense | 23:45 |
@lisitsyn | exactly | 23:45 |
@lisitsyn | why not? | 23:45 |
@iglesiasg | how is it possible we can listen to it? | 23:45 |
@iglesiasg | shouldn't it be secret? hehe | 23:46 |
-!- kaushalb09 [~kaushalb0@dhcp205-011-002.wireless.buffalo.edu] has joined #shogun | 23:46 | |
@lisitsyn | iglesiasg: boston PD asked to not share addresses/etc but you can't control the internet you know | 23:46 |
@lisitsyn | iglesiasg: they are deadly tired | 23:47 |
@lisitsyn | 19th or 20th hour already | 23:47 |
@iglesiasg | someone must be helping him IMHO | 23:47 |
@lisitsyn | iglesiasg: helping like? | 23:49 |
@lisitsyn | 2 guys | 23:50 |
@iglesiasg | lisitsyn: helping like keeping him at a place | 23:50 |
@lisitsyn | ah they are going to fail this time | 23:50 |
@iglesiasg | I bet he is at somenone's place | 23:50 |
@lisitsyn | ahh | 23:50 |
@lisitsyn | yes may be | 23:50 |
@iglesiasg | just gut feeling | 23:50 |
@lisitsyn | they have scanned basically all the city already I guess | 23:51 |
@lisitsyn | iglesiasg: they are going to have a trouble with sunsetting | 23:53 |
@iglesiasg | lisitsyn: why? | 23:54 |
@lisitsyn | iglesiasg: if they haven't found it with daylight.. | 23:54 |
@lisitsyn | iglesiasg: 6 streets left | 23:56 |
@iglesiasg | lisitsyn: are they actually going house by house? | 23:58 |
@iglesiasg | checking inside, I guess not | 23:58 |
@lisitsyn | iglesiasg: yes | 23:58 |
@lisitsyn | actually | 23:58 |
@lisitsyn | iglesiasg: not the detailed inside I guess by yes - door-by-door | 23:58 |
@iglesiasg | lisitsyn: but this radio thing is fucked up | 23:59 |
@iglesiasg | I mean I don't understand | 23:59 |
--- Log closed Sat Apr 20 00:00:28 2013 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!