--- Log opened Fri Mar 29 00:00:17 2013 | ||
@sonney2k | it is 24:00 :) | 00:00 |
---|---|---|
@sonney2k | thought my phone crashed when it displayed 00:00 | 00:01 |
@sonney2k | blackburn, how shall we do it about the proposal? | 00:04 |
blackburn | sonney2k: do whatever you want to do with it | 00:04 |
@sonney2k | shall I directly edit the proposal gdoc document? | 00:04 |
@sonney2k | and you fill in the stuff tomorrow? | 00:04 |
blackburn | sonney2k: if you want to do it not solely - yes | 00:04 |
@sonney2k | not sure what you mean | 00:05 |
@sonney2k | It would be best to have another iteration | 00:05 |
blackburn | sonney2k: if you want help from fernando or heiko side | 00:05 |
blackburn | you'd have to put it in gdocs | 00:05 |
blackburn | as they can't see the real application | 00:05 |
blackburn | sonney2k: sent a mail | 00:11 |
@sonney2k | blackburn, I did too :) | 00:13 |
blackburn | sonney2k: what for? | 00:14 |
blackburn | okay whatever I am off for today | 00:15 |
n4nd0 | blackburn: ook, good night! and good job :) | 00:17 |
blackburn | n4nd0: good night | 00:17 |
-!- zhuoran [80eefe5d@gateway/web/freenode/ip.128.238.254.93] has joined #shogun | 00:32 | |
zhuoran | Hi, is anyone online? | 00:32 |
n4nd0 | zhuoran: Yep | 00:32 |
zhuoran | I am a newbie here, trying to do a program solving a graph problem | 00:32 |
n4nd0 | zhuoran: It sounds fun | 00:33 |
zhuoran | Is there any general common graph class or structure in shogun? | 00:33 |
n4nd0 | zhuoran: not explicitly | 00:33 |
n4nd0 | but you could use SGMatrix for dense representation or also something for sparse | 00:34 |
n4nd0 | SGSparseMatrix | 00:34 |
zhuoran | Or I should write my own adjacency list in program | 00:34 |
zhuoran | SGSparseMatrix | 00:35 |
n4nd0 | zhuoran: as you wish :) | 00:35 |
zhuoran | That seems what I want. | 00:35 |
zhuoran | Let me have a glance. | 00:35 |
zhuoran | Thanks!!! | 00:35 |
n4nd0 | zhuoran: You are welcome | 00:35 |
n4nd0 | zhuoran: What are you working on btw? | 00:35 |
zhuoran | https://github.com/shogun-toolbox/shogun/issues/873 | 00:39 |
zhuoran | Issue 873 | 00:39 |
zhuoran | I have already wrote my own program to solve this problem. | 00:40 |
n4nd0 | aham cool! | 00:40 |
zhuoran | However, I really not familiar with SHOGUN. I am trying to merge it in. | 00:40 |
n4nd0 | zhuoran: Ok, just ask away stuck | 00:41 |
n4nd0 | if* stuck | 00:41 |
zhuoran | Thanks :D | 00:41 |
@sonney2k | n4nd0, fighting w/ the application | 00:46 |
n4nd0 | sonney2k: do you think it needs that much improvement?? | 00:46 |
@sonney2k | some parts yes | 00:46 |
n4nd0 | I thought it was ok.. | 00:46 |
n4nd0 | mm I see | 00:46 |
n4nd0 | I am going to open the gdocs | 00:47 |
n4nd0 | sonney2k: are you modifying it there? | 00:47 |
@sonney2k | for example first part was asking about the organizaton | 00:47 |
@sonney2k | n4nd0, yes | 00:47 |
@sonney2k | but we told only about shogun | 00:47 |
@sonney2k | not so much about us | 00:47 |
n4nd0 | aham | 00:47 |
n4nd0 | I see a much longer description now there | 00:48 |
@sonney2k | but it is not much more work | 00:48 |
@sonney2k | yes mostly from previous year with adjustments | 00:48 |
@sonney2k | what you are marking is what I am working on currenlty | 00:48 |
n4nd0 | hehe ok :) | 00:48 |
n4nd0 | I tend to highlight things as I read them | 00:49 |
@sonney2k | what also will need improvemnt is 6) | 00:49 |
@sonney2k | we have to explicitly say why we choose which mentor | 00:49 |
@sonney2k | doing that will leave no room for confusion | 00:49 |
n4nd0 | I understand | 00:49 |
@sonney2k | e.g. vojtech -> libqp expert, shogun mentor me shogun expert etc | 00:50 |
n4nd0 | let me write a list with the mentors in there | 00:50 |
@sonney2k | yes please I will continue with 2 | 00:51 |
@sonney2k | the rest likely needs not much polishing | 00:51 |
n4nd0 | sonney2k: I just realized - Christian is not mentoring this year? Nor Alex Binder? | 01:01 |
@sonney2k | no one asked them I guess? | 01:01 |
n4nd0 | too bad | 01:02 |
@sonney2k | well we can still add to the ideas list next week | 01:02 |
-!- FSCV [~FSCV@206.225.135.24] has quit [Quit: Leaving] | 01:05 | |
n4nd0 | sonney2k: there are some people I don't really know about | 01:13 |
-!- zhuoran [80eefe5d@gateway/web/freenode/ip.128.238.254.93] has quit [Ping timeout: 245 seconds] | 01:15 | |
@sonney2k | n4nd0, yeah well | 01:15 |
@sonney2k | jsut put their names there and I will write more details | 01:15 |
n4nd0 | ok | 01:16 |
n4nd0 | there are some TODOs together with their names | 01:16 |
@sonney2k | ok | 01:18 |
@sonney2k | n4nd0, somehow the questions in the gdoc don't match the one in melange | 01:18 |
n4nd0 | ouch | 01:19 |
@sonney2k | anyway please read 1 & 2 | 01:19 |
n4nd0 | ok | 01:19 |
n4nd0 | shit idk how to make proper comments | 01:21 |
n4nd0 | got it now | 01:22 |
n4nd0 | you established the JMLR software track? wow | 01:25 |
@sonney2k | yeah, mloss.org, mldata.org too | 01:26 |
@sonney2k | n4nd0, please read 3-6 now too | 01:27 |
n4nd0 | ok | 01:29 |
n4nd0 | I wrote a couple of minor comments | 01:29 |
n4nd0 | if you agree I can modify | 01:30 |
@sonney2k | arg I am tired | 01:31 |
@sonney2k | so what is missing | 01:31 |
n4nd0 | write sth about this 4 guys | 01:32 |
@sonney2k | 7-11 | 01:32 |
@sonney2k | let me start from the end | 01:32 |
n4nd0 | ok | 01:32 |
@sonney2k | please read 11 | 01:38 |
n4nd0 | ok | 01:38 |
@sonney2k | please read 10 | 01:41 |
@sonney2k | n4nd0, ^ | 01:41 |
n4nd0 | parsing... | 01:41 |
n4nd0 | sonney2k: should we put in 10) that students are also encourage to "idle" at IRC and the communication between students and some mentors is daily there? | 01:43 |
@sonney2k | n4nd0, yes please! | 01:44 |
@sonney2k | [;ease do | 01:45 |
@sonney2k | pl | 01:45 |
n4nd0 | ok | 01:45 |
@sonney2k | n4nd0, do you know what blackburn meant in 8 with Second, we do coordinate plans and schedule before GSoC as much as possible and chances are low for busy students.? | 01:48 |
n4nd0 | sonney2k: since it is required to write an schedule for the application, students who will be busy during the summer with other occupations are detected?? | 01:49 |
n4nd0 | I didn't express myself well either TBH... | 01:50 |
@sonney2k | so this is about overloaded students! | 01:50 |
n4nd0 | I think so | 01:50 |
@sonney2k | you can read 9 and 8 | 01:55 |
@sonney2k | n4nd0, I hope 8 is better now | 01:56 |
n4nd0 | yes, I think it is good | 01:58 |
n4nd0 | 7 and we are done then | 01:58 |
@sonney2k | I am dying | 02:03 |
@sonney2k | please read the first paragraph of 7 | 02:03 |
n4nd0 | hehe tomorrow is holiday :) | 02:03 |
n4nd0 | yeah I am reading on the way | 02:03 |
n4nd0 | hehe hardcore machine learners? :) | 02:04 |
@sonney2k | kids wake up in a few hours they won't understand why I waste time on randomly hitting keystrokes at night | 02:04 |
n4nd0 | in a few hours?! what time do they normally wake up? | 02:05 |
n4nd0 | my gf will wake me up latest at 9h saying, I am hungry, frukost! :D | 02:05 |
n4nd0 | Fr?hst?ck | 02:06 |
@sonney2k | well I have 4 hours max before being woken up | 02:06 |
n4nd0 | wow they really wake up early then | 02:06 |
@sonney2k | n4nd0, ok done | 02:12 |
@sonney2k | please check | 02:12 |
@sonney2k | 7 | 02:12 |
n4nd0 | we are good then I think | 02:14 |
n4nd0 | sonney2k: are you going to put it in melange right away? | 02:15 |
@sonney2k | n4nd0, let me try | 02:18 |
@sonney2k | alright submitted | 02:30 |
n4nd0 | yaay! | 02:30 |
@sonney2k | bed time!!!!! | 02:31 |
n4nd0 | yep | 02:32 |
n4nd0 | good night! | 02:32 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 02:39 | |
shogun-buildbot | build #337 of nightly_default is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/337 | 03:46 |
-!- naywhayare [~ryan@spoon.lugatgt.org] has joined #shogun | 03:51 | |
-!- blackburn1 [~blackburn@185.6.239.210] has joined #shogun | 06:12 | |
-!- blackburn [~blackburn@185.6.239.176] has quit [Ping timeout: 256 seconds] | 06:15 | |
-!- sumit_ [73f91219@gateway/web/freenode/ip.115.249.18.25] has joined #shogun | 08:21 | |
blackburn1 | http://i.imgur.com/Q3cUg29.gif | 08:37 |
-!- blackburn1 [~blackburn@185.6.239.210] has quit [Quit: Leaving.] | 09:02 | |
-!- blackburn [~lisitsin@mxs.kg.ru] has joined #shogun | 09:38 | |
-!- sumit_ [73f91219@gateway/web/freenode/ip.115.249.18.25] has quit [Ping timeout: 245 seconds] | 09:47 | |
-!- hoijui [~hoijui@dslb-088-075-046-171.pools.arcor-ip.net] has joined #shogun | 10:19 | |
@sonney2k | blackburn, yes indeed :D | 11:29 |
@sonney2k | blackburn, please check whether you are OK with the proposal now | 11:30 |
blackburn | sonney2k: yeah why not | 11:31 |
blackburn | I am ok with everything | 11:31 |
blackburn | I see you filled in a lot of stuff there | 11:32 |
@sonney2k | blackburn, ok | 12:14 |
blackburn | sonney2k: I am thinking about removing dimension reduction thing as it is a separate developed module - would be a bad thing | 12:20 |
blackburn | moving code back and forth is an idiot thing | 12:21 |
@sonney2k | remove from where? | 12:22 |
blackburn | sonney2k: from the ideas list | 12:23 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 12:23 | |
@sonney2k | blackburn, I thought you do that in tapkee only anyways and only write wrappers to make it work in shogun? | 12:25 |
blackburn | sonney2k: yes and it sounds like a bad idea for *shogun* | 12:26 |
@sonney2k | why? | 12:26 |
blackburn | because student would work on tapkee directly and can tend to avoid learning shogun | 12:26 |
@sonney2k | we did the same with some other libs and have libqp similarly in the proposal | 12:26 |
@sonney2k | blackburn, well you mentor him right? | 12:26 |
blackburn | yes that's why vojtech student didn't know about shogun at all | 12:26 |
@sonney2k | so you can make sure he gets it right | 12:27 |
@sonney2k | blackburn, yeah but he did integrate the stuff | 12:27 |
blackburn | ok whatever lets keep it then | 12:27 |
@sonney2k | there can be synergies | 12:27 |
blackburn | yeah if I didn't develop a standalone library I wouldn't manage to learn a lot things | 12:28 |
@sonney2k | i.e. the external lib profits, shogun profits from the external lib (less code directly maintained in shogun, interfaceses to all $LANG) | 12:28 |
blackburn | yes that was my point too but you didn't like that before? | 12:28 |
@sonney2k | blackburn, so what is missing now is to get shogun's build back to green | 12:29 |
blackburn | I have no idea what is missing | 12:29 |
@sonney2k | blackburn, of course not in your case | 12:29 |
@sonney2k | it is like 1 shogun developer less | 12:29 |
@sonney2k | the stuff with vojtech scales as in he continues to develop libqp | 12:30 |
@sonney2k | same with libvw | 12:30 |
blackburn | I see nothing different | 12:30 |
blackburn | if you'd rather keep my skills low - yes | 12:31 |
@sonney2k | blackburn, ohh you did remove VW ? | 12:31 |
blackburn | sonney2k: yes I didn't yet write an description | 12:31 |
@sonney2k | I thought you wanted to have it in the ideas list? | 12:31 |
-!- lambday [3ba2174d@gateway/web/freenode/ip.59.162.23.77] has joined #shogun | 12:32 | |
blackburn | I have no plan about it yet | 12:32 |
@sonney2k | blackburn, the difference is that john langford / vojtech will never work on shogun but their library | 12:32 |
blackburn | may be I'll will write it this weekend | 12:32 |
blackburn | sonney2k: and? | 12:32 |
@sonney2k | blackburn, you are very active w/ shogun. when develop a separate library your shogun activity decreases | 12:33 |
@sonney2k | that is the difference | 12:33 |
blackburn | but we are not striving for producing a noise but quality right? | 12:33 |
@sonney2k | sure why? | 12:34 |
blackburn | what do you mean by decreasing activity then? | 12:34 |
@sonney2k | blackburn, one core shogun developer less | 12:35 |
@sonney2k | in contrast to an indirect win of one shogun developer when libvw and libqp are improved | 12:35 |
@sonney2k | n4nd0, good morning! I will go to bed now to get some more sleep. If you have time please fix the dead link on the front page (middle more button). | 12:37 |
n4nd0 | sonney2k: good afternoon :P | 12:37 |
n4nd0 | let me see what link is wrong | 12:37 |
blackburn | sonney2k: well people lose interest that's normal | 12:38 |
-!- heiko [~heiko@027fc803.bb.sky.com] has joined #shogun | 12:39 | |
-!- heiko [~heiko@027fc803.bb.sky.com] has quit [Client Quit] | 12:39 | |
blackburn | I also learned that have not the best BDFL | 12:39 |
n4nd0 | sonney2k: the digit recognition application shows completely wrong in my browse :S | 12:39 |
n4nd0 | blackburn, sonney2k : is it fine there? http://shogun-toolbox.org/page/about/examples | 12:39 |
blackburn | n4nd0: it has kinda 2x of normal width | 12:40 |
blackburn | n4nd0: do you mean that? | 12:40 |
n4nd0 | blackburn: https://dl.dropbox.com/u/11020840/2013-03-29-124249_1600x900_scrot.png | 12:41 |
blackburn | n4nd0: oh even worse | 12:42 |
n4nd0 | sonney2k: I think that PR should fix it, merge please if you have a second | 12:48 |
n4nd0 | blackburn: do you have shogun-web rights as well? :) | 12:49 |
blackburn | n4nd0: yes | 12:53 |
n4nd0 | thanks! | 12:53 |
n4nd0 | I don't really know how it works for shogun-web when one merges things | 12:54 |
n4nd0 | if they go directly to the webpage or one should do something else | 12:54 |
n4nd0 | it seems it didn't change yet | 12:55 |
blackburn | n4nd0: probably that doesn't change anything | 13:05 |
n4nd0 | il maestro must probably update something too | 13:06 |
-!- heiko [~heiko@nat-180-131.internal.eduroam.ucl.ac.uk] has joined #shogun | 14:26 | |
lambday | hi heiko | 14:41 |
lambday | just saw your comments.. | 14:41 |
lambday | fixing the unittest.. | 14:41 |
heiko | lambday: these are just minor comments | 14:41 |
heiko | good work otherwise | 14:41 |
lambday | heiko: thanks :) | 14:43 |
lambday | and you mean fixed matrix of 100x100? | 14:44 |
heiko | lambday: yes | 14:44 |
heiko | and do not use log(det(M)) | 14:44 |
heiko | but use matlab or something to compute the number to compare to | 14:44 |
lambday | heiko: okay.. | 14:45 |
heiko | use the log(det(M)) for the small one | 14:45 |
heiko | should be safe | 14:46 |
lambday | okay.. | 14:46 |
lambday | hmmm.. | 14:46 |
heiko | in facts, its better to use the number | 14:46 |
heiko | to leave it :) | 14:46 |
heiko | no need for randomness | 14:46 |
heiko | fixed tests are better | 14:46 |
lambday | okay.. will be updating asap.. | 14:47 |
heiko | if you want more than one test, what you could do is to introduce some "difficult" matrices with a large condition number, but I think thats not really needed | 14:47 |
lambday | one more thing.. I was working on the sparse case... | 14:47 |
heiko | lambday: yes I saw the mail, let me read it :) | 14:49 |
lambday | heiko: okay.. so I have to manually enter 100,000 numbers, in the test? will be using matlab rand for generating one though.. :( | 14:49 |
lambday | alright :) | 14:49 |
heiko | lambday: | 14:49 |
heiko | no | 14:49 |
heiko | create some matrix with structure and then use that one | 14:50 |
heiko | let me think of an example ... | 14:51 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 14:51 | |
heiko | lambday: just do A=1:100 | 14:54 |
heiko | (as col vector) | 14:54 |
heiko | and then A=A*A' | 14:54 |
heiko | and then add some number to the diagonal | 14:54 |
lambday | oh | 14:54 |
lambday | got it | 14:54 |
heiko | to make prevent problems | 14:54 |
lambday | hmmm... | 14:54 |
heiko | you have to add something otherwise smalles Eigenvalue will be zero | 14:55 |
heiko | sum(log(diag(chol(A*A'+eye(size(A,1)))))) = 6.365919548592109 | 14:56 |
heiko | on my machine | 14:56 |
heiko | with 1:100 | 14:56 |
lambday | testing... | 14:57 |
lambday | yes, same result | 15:01 |
heiko | nice | 15:01 |
lambday | okay.. using that in the test then | 15:01 |
heiko | maybe for the small test, use the log(det(M)) thing and the number | 15:02 |
heiko | lambday: btw you can assert high accuracy floating point results with EXPECT_EQ(CMath::abs(number - expected), 1E-15) | 15:03 |
heiko | ehm | 15:03 |
heiko | EXPECT_LE (less or equal) | 15:03 |
heiko | thats how I usually do it | 15:03 |
lambday | heiko: okay.. | 15:06 |
lambday | with floating points this really sounds the best way to check I guess.. I had to fight with it for quite a long time that day.. | 15:07 |
heiko | yes the google test one only checks a few digits, which is not enough for us | 15:07 |
lambday | yes.. | 15:08 |
blackburn | heiko: ??? | 15:08 |
blackburn | heiko: EXPECT_NEAR | 15:08 |
lambday | i tried with expect double eq but that failed | 15:08 |
heiko | blackburn: same thing | 15:09 |
heiko | but yes, actually less to write :) | 15:09 |
heiko | lambday: so check EXPECT_NEAR | 15:09 |
heiko | EXPECT_NEAR(val1, val2, abs_error); the difference between val1 and val2 doesn't exceed the given absolute error | 15:09 |
lambday | heiko: alright... | 15:10 |
blackburn | heiko: but if there is a bug in cmath::abs? | 15:10 |
heiko | sometimes one has to play with the error tolerances a bit ... | 15:10 |
heiko | blackburn: well we should have a test :) | 15:10 |
blackburn | heiko: and if you don't cover a bug? | 15:11 |
heiko | blackburn: I think its possible to cover all cases for CMath::abs | 15:12 |
heiko | in fact, for this one, I can even prove that it works correctly with some hoare calculus or whatever ;) | 15:12 |
blackburn | heiko: you can't prove that | 15:14 |
heiko | blackburn: I can | 15:14 |
heiko | I learned that at university | 15:15 |
blackburn | heiko: I don't think you can prove that code will work on any compiler/machine | 15:15 |
blackburn | that's why people stick with other tools rather than writing their own | 15:15 |
heiko | blackburn, well wrt to some conditions | 15:15 |
heiko | blackburn: but the same holds for gtest right? | 15:16 |
blackburn | heiko: of course but it is a duty of gtest to write that | 15:16 |
blackburn | heiko: it is sonney2k-like approach to rewrite all the shit and pray | 15:16 |
blackburn | I don't like that | 15:16 |
blackburn | I don't believe my code more than any other code | 15:16 |
heiko | blackburn: so it is the duty of any compiler to check if then else right? | 15:17 |
heiko | from our perspective, it doesnt matter whether we use CMath::abs or gtest since for both to be correct, we have to assume that either compiler or gtest work correct | 15:17 |
heiko | but under the assumption that compiler works correct, the abs code is fine | 15:17 |
blackburn | yes but you have to stick to testing framework when checking invariants | 15:18 |
heiko | blackburn: well what happens if there is a bug in abs? | 15:18 |
heiko | test fails | 15:18 |
blackburn | it is not about that abs thing but in general | 15:18 |
heiko | so with my test I also test abs | 15:18 |
heiko | blackburn: yes | 15:18 |
blackburn | testing through untested | 15:18 |
heiko | agreed | 15:18 |
blackburn | it cascades bugs | 15:18 |
heiko | yes | 15:18 |
-!- sumit [73f91219@gateway/web/freenode/ip.115.249.18.25] has joined #shogun | 15:19 | |
blackburn | heiko: btw the same principle applies to the test coverage thing | 15:20 |
heiko | blackburn: what do you mean? | 15:20 |
blackburn | if we use a lot of CMath etc | 15:20 |
blackburn | it is marked as covered | 15:20 |
blackburn | giving a false feeling of safety | 15:20 |
heiko | test coverage is always a fals feeling of safety right? | 15:21 |
heiko | test coverage can only be used to create feelings of non-safety | 15:21 |
heiko | namely if its low | 15:21 |
blackburn | heiko: yes but the more things you avoid the better that estimate is | 15:21 |
heiko | true | 15:21 |
heiko | blackburn: but this is also a hard problem right? | 15:21 |
blackburn | of course | 15:21 |
heiko | you cannot always check everything from scratch in your tests | 15:22 |
blackburn | heiko: that's why you should not use code you don't test there if possible | 15:23 |
blackburn | and that's why I don't like using any of abs or etc there | 15:23 |
heiko | blackburn: well I agree, but I would say for abs it is not important due to the simplicity of it | 15:23 |
heiko | blackburn: but will avoid in the future | 15:23 |
blackburn | feel free heiko I just warned | 15:24 |
blackburn | heiko: about the proposal - have you checked? | 15:27 |
blackburn | it now contains a lot of stuff | 15:28 |
blackburn | sonney2k rewrote the whole thing or so | 15:28 |
heiko | blackburn: no thats why I am here in fact | 15:30 |
heiko | is it still in the google thingi? | 15:31 |
blackburn | heiko: yes | 15:32 |
heiko | blackburn: btw I read your paper draft. | 15:52 |
heiko | I dont have anything to complain except for a few typos /grammar things | 15:52 |
heiko | will send you a scan of my notes soon | 15:52 |
blackburn | heiko: cool! thanks for reading it | 15:52 |
blackburn | heiko: I owe you a beer :) | 15:53 |
heiko | also like it! | 15:53 |
heiko | beer or bear? :) | 15:53 |
blackburn | heiko: bear hmm I am not sure they live there | 15:54 |
heiko | haha :) | 15:54 |
blackburn | heiko: if you go with me a little eastern | 15:54 |
blackburn | I could find you some bear | 15:54 |
heiko | hehe :) | 15:54 |
heiko | so when is deadline? | 15:54 |
blackburn | deadline for? | 15:55 |
blackburn | heiko: gsoc? | 15:55 |
heiko | jlmr | 15:55 |
heiko | jmlr | 15:55 |
blackburn | heiko: I don't know - is there any deadline? :) | 15:55 |
blackburn | I mean it is not even submitted yet | 15:56 |
heiko | blackburn: ok I see, so no stress in sending you the feedback | 15:56 |
blackburn | yeah a month is ok | 15:56 |
blackburn | :D | 15:56 |
heiko | my parents are here thats why I ask , so I do this afterwards :D | 15:57 |
lambday | heiko: I made the changes, it seems that 1E-15 is not working... I've kept it 1E-10 | 15:58 |
heiko | lambday, thats fine - go as accurate as you can | 15:58 |
blackburn | heiko: yeah you should rather get to your parents :) | 15:59 |
blackburn | lambday: 1e-15 is too tight | 15:59 |
lambday | hmmm... | 15:59 |
heiko | blackburn: I sent them to the british museum which is 5 mins from here :) | 15:59 |
heiko | blackburn: I sometimes use 1E-15 | 15:59 |
heiko | but 10 should also be fine | 15:59 |
lambday | I'm committing then | 15:59 |
heiko | lambday: nice | 15:59 |
blackburn | heiko: it should be ok for small numbers | 15:59 |
heiko | I have to leave soon, so if you hurry, I will merge before :) | 16:00 |
blackburn | heiko: but 1 != 1 with 1e-15 I think? | 16:00 |
blackburn | may be 1==1 but could be that 3!=3 :) | 16:01 |
heiko | blackburn, lambday, just go maximum that gets accepted :) | 16:02 |
heiko | just more than 6 that was my point | 16:02 |
lambday | heiko: okay.. :) | 16:02 |
lambday | and I need a good tutorial for git I guess :( 80% of the time I am wondering about git.. | 16:02 |
lambday | I git a git add Statistics_unittest.cc and then git commit... | 16:03 |
heiko | lambday: haha, I totally know that and have that too from time to time :) | 16:03 |
heiko | best is ask blackburn | 16:03 |
heiko | lambday: thats fine --- add/commit for new files | 16:03 |
lambday | heiko: :( :( ... | 16:04 |
blackburn | http://www.youtube.com/watch?feature=player_detailpage&v=zRIbf6JqkNc | 16:04 |
heiko | blackburn: the application is very good I think | 16:04 |
heiko | blackburn: so not changing anything - you should commit :) | 16:04 |
heiko | I have to leave in 2 mins | 16:05 |
blackburn | heiko: boss made the deal already I thnik | 16:05 |
lambday | blackburn: :D | 16:05 |
heiko | blackburn: I see, nice then, should be no problems | 16:05 |
lambday | how to do the git pull --rebase thing? git pull --rebase upstream master... here shall I put upstream as "https://github.com/shogun-toolbox/shogun" or my forked repo? | 16:05 |
heiko | lambday, from your master branch you do | 16:06 |
heiko | git fetch upstream | 16:06 |
heiko | git pull --rebase upstream master | 16:06 |
heiko | the you change to your feature branch and do | 16:06 |
blackburn | fetch is redundant here | 16:06 |
heiko | git rebase master | 16:06 |
heiko | then you go to master again and do | 16:06 |
lambday | heiko: alright.. | 16:06 |
heiko | git merge feature_branch_name | 16:06 |
heiko | then you do git push | 16:07 |
lambday | okay.. | 16:07 |
heiko | lambday: and if you have changes you have to stash or commit them first | 16:07 |
heiko | and a good tip is: | 16:07 |
heiko | never modify the master branch | 16:07 |
heiko | only for pushing :) | 16:07 |
heiko | otherwise you can create a mess easily | 16:07 |
heiko | had that often | 16:07 |
heiko | but might be my stupidity :) | 16:08 |
heiko | lambday: ok sorry, gotta go now, will check tonight | 16:08 |
heiko | blackburn,lambday bye! | 16:09 |
lambday | heiko: okay... thanks :) ciao | 16:09 |
-!- heiko [~heiko@nat-180-131.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.] | 16:09 | |
-!- blackburn [~lisitsin@mxs.kg.ru] has quit [Quit: Leaving.] | 16:40 | |
-!- lambday [3ba2174d@gateway/web/freenode/ip.59.162.23.77] has quit [] | 16:42 | |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 18:25 | |
shogun-notifier- | shogun: rahul :master * afb7f2b / src/ (4 files): https://github.com/shogun-toolbox/shogun/commit/afb7f2bd3f88500eace69c8beea69cab81037d2b | 18:25 |
shogun-notifier- | shogun: added a new kernel SSKStringKernel which implements the subsequence kernel (by lodhi et. al.). Also added to the modular interface | 18:25 |
shogun-notifier- | shogun: lambday :master * c5a7cac / src/shogun/mathematics/Statistics.cpp,src/shogun/mathematics/Statistics.h: https://github.com/shogun-toolbox/shogun/commit/c5a7cacef649510b0e929a3a9e4ed2da3bd7851e | 18:25 |
shogun-notifier- | shogun: Added log_det method in CStatistics class | 18:25 |
shogun-notifier- | shogun: lambday :master * 0891f5a / src/shogun/mathematics/Statistics.h: https://github.com/shogun-toolbox/shogun/commit/0891f5a38bcb927d3c2349f1d3a006b975893d11 | 18:25 |
shogun-notifier- | shogun: indentation fixed in Statistics.h | 18:25 |
shogun-notifier- | shogun: lambday :master * ac917f2 / / (7 files): https://github.com/shogun-toolbox/shogun/commit/ac917f29ab1ad60b4933f130d352e7aeab487b90 | 18:25 |
shogun-notifier- | shogun: Fixed the previous messy commits | 18:25 |
shogun-notifier- | shogun: lambday :master * 6cf55d3 / src/shogun/mathematics/Statistics.cpp: https://github.com/shogun-toolbox/shogun/commit/6cf55d34f52bdeb253262b391d610bcdd0134c80 | 18:25 |
shogun-notifier- | shogun: Unit test for log_det added. log_det updated. | 18:25 |
shogun-notifier- | shogun: lambday :master * 6b28e87 / src/shogun/mathematics/Statistics.h: https://github.com/shogun-toolbox/shogun/commit/6b28e8717efe557d399351ca75684dc7acbb101d | 18:25 |
shogun-notifier- | shogun: Fixed doxygen info in log_det | 18:25 |
-!- n4nd0 [~fircuser@c83-251-227-64.bredband.comhem.se] has joined #shogun | 18:28 | |
shogun-buildbot | build #843 of deb2 - static_interfaces is complete: Failure [failed test python_static] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/843 blamelist: lambday <heavensdevil6909@gmail.com> | 18:34 |
shogun-buildbot | build #844 of deb2 - static_interfaces is complete: Failure [failed test python_static] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/844 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, lambday <heavensdevil6909@gmail.com> | 18:46 |
-!- n4nd0 [~fircuser@c83-251-227-64.bredband.comhem.se] has quit [Remote host closed the connection] | 18:48 | |
-!- n4nd0 [~fircuser@c83-251-227-64.bredband.comhem.se] has joined #shogun | 18:48 | |
shogun-buildbot | build #845 of deb2 - static_interfaces is complete: Failure [failed test python_static] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/845 blamelist: rahul <rahul@localhost.localdomain> | 18:56 |
wiking | mmm we are still committing to master branch ? :S | 19:05 |
@sonney2k | wiking, yes sure until we have the build fixed | 19:12 |
@sonney2k | hmmhh but not features | 19:12 |
@sonney2k | ohh well | 19:12 |
@sonney2k | we are good in producing features and suck in fixing bugs | 19:12 |
shogun-buildbot | build #971 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/971 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, lambday <heavensdevil6909@gmail.com> | 19:15 |
-!- lambday [0e8b614d@gateway/web/freenode/ip.14.139.97.77] has joined #shogun | 19:28 | |
shogun-buildbot | build #972 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/972 blamelist: rahul <rahul@localhost.localdomain> | 19:34 |
@sonney2k | lambday, I just realize that your PR did contain all the previous commits - please for next time use git commit --amend or git reset <version> and commit a single clean version | 19:39 |
@sonney2k | lambday, this way the buildbot will not fail for the old broken commits (it builds every commit...). | 19:39 |
lambday | sonney2k: yes.. sorry for messing it up this time.. | 19:39 |
@sonney2k | n4nd0, thanks for fixing the link | 19:40 |
@sonney2k | lambday, don't worry we are no better in all that | 19:40 |
lambday | sonney2k: next time onwards I'll be more careful before sending a PR.. | 19:40 |
@sonney2k | lambday, btw please also set a proper user name for your git stuff | 19:40 |
@sonney2k | lambday, don't worry it is our mistake not to tell you upfront | 19:41 |
n4nd0 | is it fixed now? | 19:41 |
@sonney2k | we do / did the same stuff | 19:41 |
@sonney2k | n4nd0, I didn't put it to the website yet | 19:41 |
n4nd0 | I tried but it goes still to the old link! | 19:41 |
lambday | sonney2k: I have been struggling with git... hope to get better asap :( | 19:41 |
@sonney2k | I was thinking that maybe we should put the workshop instead of the demos to the front page | 19:41 |
n4nd0 | all right | 19:41 |
@sonney2k | lambday, we still struggle with git too so welcome to the team :) | 19:42 |
@sonney2k | n4nd0, would you agree? | 19:42 |
n4nd0 | writing on the phone here - 5 keystrokes per minute :) | 19:42 |
@sonney2k | I mean it does help a lot | 19:42 |
@sonney2k | n4nd0, pretty fast you are then :D | 19:42 |
lambday | sonney2k: everyone around here is so helpful.. :) thanks for being so patient | 19:43 |
n4nd0 | git yeah... our eternal archienemy/best friend :D | 19:43 |
-!- sumit [73f91219@gateway/web/freenode/ip.115.249.18.25] has quit [Ping timeout: 245 seconds] | 19:44 | |
n4nd0 | I agree with the workshop idea, probably lot of people will visit the page soon | 19:45 |
n4nd0 | the more visible the better I think | 19:45 |
@sonney2k | n4nd0, ok then we should do it rather now. If you have time please write up sth. I am trying to find /ask for a nice c-base figure to put on the front page | 19:47 |
n4nd0 | I am not at home now unfortunately | 19:48 |
n4nd0 | no computer around to work, maybe tonight when I get back | 19:48 |
@sonney2k | ahh true | 19:48 |
@sonney2k | phone :D | 19:49 |
n4nd0 | yeah | 19:49 |
* sonney2k is asking for a pic at #c-base | 19:49 | |
n4nd0 | what kind of pic you have in mind? | 19:49 |
n4nd0 | the spaceship one is fancy hehe | 19:50 |
@sonney2k | http://gallery.c-base.org/ | 19:50 |
@sonney2k | the logo on top would be cool | 19:50 |
@sonney2k | or the starshop | 19:51 |
@sonney2k | looks cool for sure | 19:51 |
n4nd0 | so you have to ask them for the permission to use it I understand | 19:52 |
@sonney2k | or this one http://vimeo.com/cbase | 19:52 |
@sonney2k | I think so | 19:52 |
n4nd0 | I like these ones better than the spaceship | 19:58 |
n4nd0 | but any is fine I think | 19:59 |
lambday | sonney2k: one stupid question - why did you named this project shogun? I mean, what does "sho gun" mean? | 20:03 |
n4nd0 | soeren sonnenburg and gunnar raetsch I think - main original authors | 20:08 |
n4nd0 | sort of acronym of their names | 20:08 |
@sonney2k | lambday, yes indeed what n4nd0 said | 20:13 |
@sonney2k | n4nd0, I will just fix the website as is - please when you have time write some text for the middle block | 20:14 |
lambday | n4nd0: hmmm.. :) | 20:14 |
@sonney2k | it seems it takes some time to get permission to use a c-base logo | 20:14 |
@sonney2k | lambday, we had other ideas like Reason/ Raeson / but in some survey people did prefer shogun :D | 20:15 |
lambday | sonney2k: shogun sounds cool.. I put that in google translator from japanese to english and it showed "military treatment" :D | 20:17 |
n4nd0 | ok will do tonight probably | 20:17 |
n4nd0 | tomorrow afternoon I am going to travel and I will be until Tuesday eveningn completely off so I will do it before going | 20:18 |
@sonney2k | n4nd0, thanks. I will fix the website now and then we still have this burning issue with tests not being all passing :/ | 20:19 |
@sonney2k | I could use some help for that... | 20:20 |
n4nd0 | any clue what is wrong? | 20:20 |
n4nd0 | I can spend some time tonight | 20:20 |
n4nd0 | sonney2k ^ | 20:25 |
-!- blackburn [~blackburn@109.226.90.103] has joined #shogun | 20:27 | |
@sonney2k | n4nd0, thanks | 20:35 |
@sonney2k | blackburn, do you have time to help fixing the buildbot errors? | 20:57 |
blackburn | sonney2k: yes | 20:58 |
wiking | sonney2k: what's the bug that holds back switching to dev branch? | 20:58 |
@sonney2k | wiking, I've made tests / example failures critical | 20:58 |
blackburn | wiking: something is wrong | 20:58 |
@sonney2k | i.e. when one fails -> boom | 20:58 |
@sonney2k | buildbot fails | 20:59 |
@sonney2k | before that we didn't enable this stuff | 20:59 |
wiking | ah ok | 20:59 |
@sonney2k | and because of that we saw lots of breakage recently | 20:59 |
wiking | yeah i know that examples were not critical eariler | 20:59 |
@sonney2k | so we didn't detect the issues | 20:59 |
wiking | sonney2k: some of the code always core dumped on the fbsd bot i saw that in dmesg always | 21:00 |
@sonney2k | wiking, I am still not sure I really enabled all of them | 21:00 |
blackburn | sonney2k: we've got to ignore these 3 kernels | 21:00 |
blackburn | and hmm | 21:00 |
@sonney2k | wiking, yeah some examples generate random data | 21:00 |
blackburn | that's not possible to solve | 21:00 |
@sonney2k | and so they fail only sometimes (well did fail - I've set the seed now) | 21:00 |
wiking | ok, where's the list of failures? :) | 21:01 |
@sonney2k | wiking, buildbot | 21:02 |
wiking | blackburn: u have your visa for july workshop? :) | 21:02 |
@sonney2k | but caution | 21:02 |
wiking | sonney2k: just checking | 21:02 |
blackburn | wiking: no | 21:02 |
@sonney2k | we only see python failures for now | 21:02 |
wiking | blackburn: :( get it! | 21:02 |
blackburn | wiking: yeah I should | 21:03 |
@sonney2k | the other stuff will only become visible once python failures are fixed | 21:03 |
@sonney2k | wiking, and once we fixed that all I still don't know what we do with buildbots | 21:03 |
wiking | blackburn: is the tsn or what the hell in tapkee now? | 21:04 |
@sonney2k | I've checked config and we explicitly run the buildbot for master branch only | 21:04 |
blackburn | wiking: yes t-sne is in | 21:04 |
wiking | blackburn: mmm i ought to test that | 21:04 |
wiking | blackburn: in shogun as well? or only tapkee for now? | 21:04 |
blackburn | wiking: no I didn't manage to write a wrapper class yet | 21:04 |
wiking | blackburn: so tapkee only cool i'll hae to check that tonight then | 21:04 |
blackburn | wiking: if your data is dense you may use that simple app that comes with tapkee | 21:05 |
@sonney2k | wiking, you know there was more important stuff to do gsoc application! | 21:05 |
wiking | blackburn: yeah it is so then i can test it out of box | 21:05 |
-!- lambday [0e8b614d@gateway/web/freenode/ip.14.139.97.77] has quit [Ping timeout: 245 seconds] | 21:05 | |
wiking | sonney2k: :S | 21:05 |
blackburn | wiking: it reads ascii and outputs ascii | 21:05 |
blackburn | wiking: one row = one feature | 21:05 |
wiking | blackburn: cool | 21:05 |
wiking | blackburn: csv? | 21:05 |
blackburn | wiking: no I am lazy | 21:06 |
blackburn | just ascii | 21:06 |
wiking | blackburn: space? | 21:06 |
blackburn | yes | 21:06 |
wiking | ok | 21:06 |
wiking | blackburn: i'll test it then soonish and let you know what did it do with my stuff | 21:06 |
blackburn | wiking: ./bin/tapkee -m t-sne -i input.dat -o output.dat | 21:06 |
blackburn | wiking: ./bin/tapkee -m t-sne -i input.dat -o output.dat --sne-perplexity 50 --sne-theta 0.5 | 21:07 |
blackburn | wiking: two params | 21:07 |
wiking | blackburn: where's more info about those two params | 21:08 |
wiking | blackburn: and what's the output btw? clustered vectors? | 21:08 |
blackburn | wiking: I have to admit I don't know the exact meaning | 21:08 |
blackburn | wiking: clustered?? | 21:08 |
blackburn | wiking: ./bin/tapkee -m t-sne -i input.dat -o output.dat -td 2 --sne-perplexity 50 --sne-theta 0.5 | 21:08 |
blackburn | -td is target dimension | 21:08 |
blackburn | 2 by default | 21:08 |
blackburn | http://arxiv.org/abs/1301.3342 that's the paper on barnes-hut-sne | 21:09 |
wiking | well okeeey | 21:09 |
wiking | i guess i'll need a bit bigger -td :P | 21:09 |
blackburn | wiking: why? | 21:09 |
wiking | i have a feeling | 21:10 |
wiking | but then again no idea | 21:10 |
wiking | so i'll just test it | 21:10 |
blackburn | wiking: 3? | 21:10 |
blackburn | do you see through multidimensional spaces? | 21:10 |
blackburn | :D | 21:10 |
* wiking haz magic eyez | 21:10 | |
wiking | blackburn: http://animalsbeingdicks.com/ | 21:10 |
wiking | blackburn: just for those hours when u give a fuck | 21:11 |
wiking | ;P | 21:11 |
blackburn | wiking: t-SNE is purposed for visualization (d=2,3) | 21:12 |
blackburn | that's why I am surprised | 21:12 |
blackburn | wiking: furthermore barnes hut is *specialized* to support only 2 IIRC | 21:12 |
wiking | mmmmhm | 21:13 |
blackburn | wiking: yeah you can't set it to 3 even :D | 21:13 |
wiking | ahhaha ok | 21:13 |
blackburn | wiking: I can check if I can extend it | 21:13 |
blackburn | wiking: alrightt should work with other valuesto | 21:14 |
blackburn | too* | 21:14 |
blackburn | wiking: PARAMETER(ScalarType, perplexity, SNE_PERPLEXITY, IN_RANGE(perplexity,0,(N-1)/3)); | 21:14 |
blackburn | wiking: last parameter tells you about possible range of perplexity :) | 21:15 |
blackburn | wiking: and theta=0 means t-SNE while theta>0 means barnes-hut-sne | 21:15 |
-!- n4nd0 [~fircuser@c83-251-227-64.bredband.comhem.se] has quit [Remote host closed the connection] | 21:20 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 21:25 | |
blackburn | wiking: I can't say it works with td!=2 | 21:29 |
blackburn | :D | 21:29 |
blackburn | lets check basic t-sne | 21:30 |
blackburn | without barnes-hut | 21:30 |
blackburn | wiking: https://github.com/shogun-toolbox/shogun/issues/959 what do you think? | 21:32 |
@sonney2k | blackburn, looks nice? gsoc idea? | 21:33 |
blackburn | sonney2k: I think a week of work | 21:34 |
@sonney2k | blackburn, I don't remember whether you responded to my comment to this: one can do this all in c++ | 21:35 |
@sonney2k | with some kind of builder pattern and then it should immediately work from all *modular interfaces like that | 21:36 |
blackburn | sonney2k: it is not a pure builder pattern | 21:36 |
blackburn | it is a builder without need to call build() | 21:36 |
@sonney2k | but you are right thinking about it it should be very easy. would even say just 1 day of work | 21:36 |
blackburn | sonney2k: not a day - we do not support list of sgobjects | 21:37 |
@sonney2k | true woudl need a new typemap | 21:38 |
blackburn | sonney2k: with proper design it could be heavily advertised | 21:38 |
@sonney2k | what I don't understand with this syntax though is how one would do nested things | 21:38 |
@sonney2k | like in your example try out gaussian_kernel with certain params and then linear kernel with other params etc | 21:39 |
blackburn | sonney2k: how does it work now? | 21:39 |
@sonney2k | I don't recall exactly - heiko would know | 21:40 |
blackburn | sonney2k: actually I see it like that | 21:40 |
@sonney2k | I had this bracket based solution that supports that | 21:40 |
@sonney2k | but it would require typemaps all over | 21:40 |
blackburn | gaussian_kernel.parameter("width").as(ExponentialRange).from(-10).to(10) | 21:41 |
blackburn | my_kernel.parameter("happiness").as(LinearRange).from(0).to(10) | 21:42 |
blackburn | svm.parameter("kernels").as(ValuesSet).values([my_kernel,gaussian_kernel]) | 21:43 |
blackburn | that's all | 21:43 |
blackburn | when you are done | 21:44 |
blackburn | svm.deactivate_parameters() | 21:44 |
blackburn | no tree messing at all | 21:44 |
blackburn | sonney2k: ^ | 21:44 |
@sonney2k | blackburn, I totally swapped out how that works internally | 21:45 |
blackburn | sonney2k: step by step | 21:45 |
@sonney2k | might work I am not sure | 21:45 |
blackburn | gaussian_kernel.parameter("width").as(ExponentialRange).from(-10).to(10) | 21:45 |
@sonney2k | I mean this syntax is certainly powerful enough | 21:46 |
blackburn | 1) gaussian_kernel.parameter("width") returns an instance of some class parameter (not parameter we have) | 21:46 |
@sonney2k | I just don't know if we can easily map it to the structure we have | 21:46 |
blackburn | 2) it sets default parameters to that parameter | 21:46 |
blackburn | 3) as(ExponentialRange) pushes the parameter to the exponential state | 21:46 |
blackburn | 4) from(-10) sets lower bound | 21:47 |
blackburn | 5) to(10) sets upper bound | 21:47 |
@sonney2k | blackburn, I've committed the submitted proposal to publicity & updated the website | 21:47 |
blackburn | okay a few issues more | 21:47 |
blackburn | 1.1) each parameter("width") call marks parameter inside a class as *to-be-modelselected* | 21:47 |
blackburn | 3.1) as(ExponentialRange) has some default thing too | 21:48 |
blackburn | sonney2k: the only flaw | 21:48 |
blackburn | if we call wrong order or wrong methods | 21:48 |
blackburn | it is runtime error | 21:48 |
blackburn | not a compile-time error | 21:48 |
blackburn | but that's default tradeoff | 21:48 |
blackburn | I am totally sure people would get excited about that | 21:49 |
@sonney2k | blackburn, I just checked modelselection_grid_search_kernel.cpp ? | 21:54 |
@sonney2k | and I think your idea would work if one includes a | 21:54 |
@sonney2k | CModelSelectionParameters* in SGObject | 21:55 |
@sonney2k | only issue is that the complete tree has to be constructed somehow | 21:55 |
blackburn | sonney2k: I do not know exactly but I see it constructing each time parameter() method is called | 21:55 |
@sonney2k | ModelSelectionParameters can handle nested stuff | 21:57 |
@sonney2k | so I guess simplest soluton would be to have non-nested stuff inside each class | 21:57 |
blackburn | sonney2k: there should be major changes in the framework to support that | 21:57 |
@sonney2k | I don't think so | 21:57 |
@sonney2k | then when the whole object to be model selected is ready | 21:58 |
@sonney2k | one creates the tree by going top->down | 21:58 |
@sonney2k | creating the nested CModelSelectionParameters | 21:58 |
@sonney2k | I think that's all | 21:58 |
blackburn | sonney2k: no it should not be top to down in final | 21:58 |
blackburn | I'd vote for magic | 21:59 |
@sonney2k | what magic? | 21:59 |
blackburn | each time parameter() of some sgobject is activated | 21:59 |
blackburn | the tree is reconstructed etc | 21:59 |
blackburn | or may be it is constructed only when doing model selection | 22:00 |
blackburn | that makes sense to me as well | 22:00 |
@sonney2k | that is what I meant | 22:00 |
@sonney2k | (the latter idea) | 22:00 |
blackburn | I see | 22:00 |
blackburn | I have two major ideas now | 22:00 |
blackburn | client server and that | 22:00 |
@sonney2k | but no time to do either one | 22:01 |
blackburn | haha | 22:01 |
@sonney2k | that is why I am saying -> gsoc idea! | 22:01 |
blackburn | that's a tough idea | 22:02 |
blackburn | basically one has to be a shogun developer to understand what to change | 22:03 |
blackburn | I imagine if I would be asked to do that 2 years ago | 22:03 |
blackburn | :D | 22:03 |
blackburn | sonney2k: ahh btw yesterday was the day I joined #shogun | 22:04 |
blackburn | 2 years ago | 22:04 |
blackburn | I feel youth is slipping away haha | 22:04 |
@sonney2k | habby 2nd b'day | 22:08 |
@sonney2k | blackburn, well heiko did the whole model selection framework in his first gsoc | 22:09 |
@sonney2k | so what is more tough? | 22:09 |
blackburn | sonney2k: his current idea | 22:09 |
@sonney2k | yes indeed *lol* | 22:09 |
blackburn | develop a new machine learning paradigm | 22:10 |
blackburn | abstract | 22:10 |
blackburn | structured risk minimization was developed by vapnik in the end of 20th century. we need a new framework | 22:11 |
@sonney2k | structured output on general graphs & large scale is already sufficient | 22:12 |
blackburn | sonney2k: but it is not heiko thing? | 22:14 |
@sonney2k | blackburn, no but if you have some idea how to even do that - there is a workshop at this years icml about it :D | 22:19 |
blackburn | I don't know about machine learning | 22:19 |
blackburn | so no | 22:19 |
@sonney2k | yeah and you don't drink vodka all day | 22:20 |
* sonney2k is in css hell | 22:20 | |
blackburn | both are true | 22:20 |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 22:21 | |
shogun-notifier- | shogun: Sergey Lisitsyn :master * 99135b4 / tests/integration/python_modular/regression.py: https://github.com/shogun-toolbox/shogun/commit/99135b47342c390f900fb0841d79dd27775dc0f3 | 22:21 |
shogun-notifier- | shogun: A splat for KRR test | 22:21 |
blackburn | sonney2k: svmsgd parameters were changed you say? | 22:22 |
@sonney2k | blackburn, I hardly remember | 22:24 |
@sonney2k | there was something | 22:24 |
blackburn | sonney2k: trying to understand what is the difference between modular and static sgd calls | 22:24 |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: leaving] | 22:26 | |
shogun-buildbot | build #846 of deb2 - static_interfaces is complete: Failure [failed test python_static] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/846 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 22:27 |
blackburn | alright will take a look at these examples tomorrow | 22:42 |
@sonney2k | blackburn, look http://shogun-toolbox.org/page/Events/gsoc2013_ideas | 22:53 |
@sonney2k | I managed to get the border back | 22:53 |
@sonney2k | some 'trivial' css | 22:53 |
@sonney2k | .ideas tr:not(:first-child) td {border:1px solid #8cacbb;} | 22:53 |
blackburn | sonney2k: right border is not showed here | 22:54 |
@sonney2k | blackburn, relaod! | 22:54 |
@sonney2k | reload! | 22:54 |
blackburn | ohh | 22:54 |
blackburn | good you remind me | 22:54 |
blackburn | I forgot to put benoit rostykys | 22:54 |
@sonney2k | who? | 22:54 |
blackburn | sonney2k: rostykus | 22:55 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 22:55 | |
blackburn | guy with olivier | 22:55 |
@sonney2k | ohh yes | 22:55 |
blackburn | n4nd0: sonney2k: did you fix anything in content of that gsoc2013 html? | 22:55 |
@sonney2k | was there any bug? | 22:56 |
n4nd0 | blackburn: mmmm I don't remember right now | 22:56 |
@sonney2k | blackburn, ohh btw you had this nice header telling students that they should select at most 3 ideas | 22:56 |
@sonney2k | that should also be in there again | 22:56 |
@sonney2k | and the patch requirement | 22:56 |
n4nd0 | why? | 22:56 |
blackburn | sonney2k: alright | 22:57 |
blackburn | n4nd0: I want to sync with repo stuff | 22:57 |
@sonney2k | n4nd0, some students don't read what we write in the form... | 22:57 |
@sonney2k | blackburn, what? | 22:57 |
@sonney2k | which repo? | 22:57 |
@sonney2k | ahh the html | 22:57 |
blackburn | sonney2k: I want to copy from publicity to the website | 22:57 |
n4nd0 | blackburn: yes take the version in the page then | 22:57 |
blackburn | yes | 22:57 |
@sonney2k | I see | 22:57 |
@sonney2k | blackburn, plesae do the other way round | 22:58 |
@sonney2k | copy from website | 22:58 |
n4nd0 | blackburn: I think there was sth minor I changed | 22:58 |
blackburn | aok | 22:58 |
@sonney2k | yeah | 22:58 |
@sonney2k | better be safe | 22:58 |
n4nd0 | oh yeah some carriage returns | 22:58 |
@sonney2k | we have no history or so | 22:58 |
@sonney2k | on the website | 22:58 |
blackburn | tabs are ruined | 22:58 |
blackburn | argh | 22:58 |
n4nd0 | sonney2k: why? | 22:58 |
n4nd0 | that they don't read forms | 22:58 |
@sonney2k | n4nd0, bern4rd didn't program any | 22:59 |
@sonney2k | it is some effort... | 22:59 |
blackburn | that's a problem | 22:59 |
blackburn | sonney2k: we have history | 22:59 |
@sonney2k | blackburn, where? | 22:59 |
blackburn | sonney2k: but it doesn't store old | 22:59 |
@sonney2k | yes | 22:59 |
blackburn | just mentions changes | 22:59 |
blackburn | ohh it has eaten all spaces | 23:00 |
@sonney2k | n4nd0, well you could program some simple solution easily - just add a revision to each object | 23:00 |
@sonney2k | and increase that revision all the time | 23:00 |
@sonney2k | and show only the largest revision | 23:00 |
n4nd0 | we are talking about storing history of the articles right? | 23:01 |
@sonney2k | n4nd0, yes | 23:01 |
@sonney2k | I think subpages/pages is not so important | 23:01 |
@sonney2k | that's more layout | 23:01 |
blackburn | ohh I can't stop laughing about heiko idea | 23:01 |
blackburn | there are three mentors now | 23:01 |
n4nd0 | "easily", I think there would be some things to handle like what if two people are modifying at the same time? | 23:01 |
n4nd0 | there must be for sure something done already somewhere to draw inspiration from though | 23:02 |
@sonney2k | n4nd0, ahh no we don't do that | 23:02 |
shogun-buildbot | build #973 of deb3 - modular_interfaces is complete: Failure [failed test python_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/973 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 23:02 |
@sonney2k | n4nd0, we usually coordinate stuff no? | 23:03 |
blackburn | sonney2k: are they using shogun at that 'criteo'? | 23:03 |
n4nd0 | sonney2k: yes | 23:03 |
n4nd0 | simple idea would be if someone is editing one article then another one cannot have the chance to do it | 23:03 |
@sonney2k | blackburn, no idea | 23:03 |
@sonney2k | n4nd0, yes but this is (not yet) a problem | 23:04 |
blackburn | sonney2k: I think heiko need a co-mentor to increase # of mentors to 4 | 23:04 |
n4nd0 | sonney2k: btw, there is no place yet in the homepage for the workshop thing or am I missing sth? | 23:04 |
@sonney2k | when we are 100 people then 2 might edit the website at the same time | 23:04 |
@sonney2k | n4nd0, I was hoping you write something? | 23:04 |
n4nd0 | sonney2k: aham got it | 23:05 |
n4nd0 | I misunderstood | 23:05 |
n4nd0 | so the idea is to write a few words about the workshop, put it in the homepage together with a link to the main workshop page under events | 23:05 |
@sonney2k | n4nd0, exactly | 23:05 |
n4nd0 | sonney2k: ok, I am on it then | 23:06 |
@sonney2k | n4nd0, btw I managed to get the table border back | 23:07 |
@sonney2k | in the ideas list | 23:07 |
n4nd0 | wow I see | 23:08 |
n4nd0 | sonney2k: it is looking very nice! | 23:08 |
@sonney2k | `just' .ideas tr:not(:first-child) td {border:1px solid #8cacbb;} | 23:08 |
n4nd0 | css I guess | 23:08 |
@sonney2k | anyway I have to sleep | 23:08 |
@sonney2k | yeah | 23:08 |
@sonney2k | hard w/o a css editor | 23:09 |
n4nd0 | sonney2k: I will mail you these short text for the homepage soon | 23:09 |
n4nd0 | this* | 23:09 |
@sonney2k | thanks | 23:09 |
@sonney2k | & nite | 23:09 |
n4nd0 | good night | 23:09 |
blackburn | sonney2k: that css family guy pic fits the case | 23:09 |
n4nd0 | hehe indeed | 23:09 |
--- Log closed Sat Mar 30 00:00:18 2013 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!