--- Log opened Fri Mar 30 00:00:19 2012 | ||
blackburn | btw should labels created in apply be SG_REFed? | 00:01 |
---|---|---|
@sonney2k | these are returned as new objects and so are not used internally | 00:02 |
@sonney2k | so no | 00:02 |
blackburn | aha I see | 00:02 |
shogun-buildbot | build #408 of java_modular is complete: Failure [failed git] Build details are at http://www.shogun-toolbox.org/buildbot/builders/java_modular/builds/408 blamelist: blackburn91@gmail.com | 00:03 |
@sonney2k | blackburn, you killed the build :) | 00:05 |
blackburn | sonney2k: lol | 00:05 |
@sonney2k | dammed github issue | 00:06 |
@sonney2k | blackburn, btw we should all start to use git amend | 00:06 |
blackburn | for what? | 00:06 |
@sonney2k | blackburn, to not have stupid/broken git commits | 00:07 |
@sonney2k | amend enables merging multiple commits into one big one | 00:07 |
blackburn | probably | 00:08 |
@sonney2k | so if sth was forgotten in one commit (missing file, other error) | 00:08 |
@sonney2k | buildbot doesn't need to build it | 00:08 |
@sonney2k | and fail once | 00:08 |
@sonney2k | and then succeed | 00:08 |
@sonney2k | and we don't have a crap commit :) | 00:08 |
blackburn | we could also consider bigger commits | 00:12 |
blackburn | actually I try to do all my commits non-breaking anything | 00:12 |
@sonney2k | blackburn, bah philipps' opencl test is slower on my GPU | 00:51 |
@sonney2k | by a lot ... | 00:51 |
blackburn | sonney2k: than CPU? | 00:51 |
@sonney2k | yeah | 00:51 |
blackburn | haha | 00:51 |
@sonney2k | like 10 times | 00:51 |
@sonney2k | on his it is like 5 times faster | 00:51 |
blackburn | sonney2k: see you | 00:59 |
-!- blackburn [~qdrgsm@62.106.114.183] has quit [Quit: Leaving.] | 01:00 | |
-!- vikram360 [~vikram360@117.192.161.171] has quit [Read error: Connection reset by peer] | 01:00 | |
-!- vikram360 [~vikram360@117.192.161.171] has joined #shogun | 01:00 | |
-!- flxb_ [~cronor@g231217050.adsl.alicedsl.de] has joined #shogun | 01:29 | |
-!- flxb [~cronor@g231240112.adsl.alicedsl.de] has quit [Ping timeout: 246 seconds] | 01:30 | |
-!- flxb_ is now known as flxb | 01:31 | |
-!- flxb_ [~cronor@g230122054.adsl.alicedsl.de] has joined #shogun | 01:32 | |
-!- flxb [~cronor@g231217050.adsl.alicedsl.de] has quit [Ping timeout: 246 seconds] | 01:36 | |
-!- flxb_ is now known as flxb | 01:36 | |
-!- vikram360 [~vikram360@117.192.161.171] has quit [Ping timeout: 246 seconds] | 03:11 | |
-!- romi__ [~mizobe@189-18-40-233.dsl.telesp.net.br] has quit [Ping timeout: 246 seconds] | 03:17 | |
-!- romi_ [~mizobe@187.57.6.229] has joined #shogun | 03:45 | |
-!- harshit_ [~harshit@182.68.160.94] has joined #shogun | 04:39 | |
-!- harshit_ [~harshit@182.68.160.94] has quit [Ping timeout: 246 seconds] | 05:37 | |
-!- romi_ [~mizobe@187.57.6.229] has quit [Ping timeout: 246 seconds] | 05:37 | |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 05:39 | |
-!- vikram360 [~vikram360@117.192.161.171] has joined #shogun | 05:56 | |
-!- Miggy [~piggy@14.139.82.6] has quit [Ping timeout: 248 seconds] | 06:43 | |
-!- av3ngr [av3ngr@nat/redhat/x-lgbneiugylvjahzy] has joined #shogun | 06:48 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 07:34 | |
-!- novice [b4953264@gateway/web/freenode/ip.180.149.50.100] has joined #shogun | 08:00 | |
-!- vikram360 [~vikram360@117.192.161.171] has quit [Ping timeout: 265 seconds] | 08:01 | |
-!- vikram360 [~vikram360@117.192.191.6] has joined #shogun | 08:08 | |
novice | Hi, can someone help me with github installation ? | 08:15 |
novice | I am trying to set it up in a virtual machine running ubuntu | 08:16 |
novice | added the ssh key, but im unable to test it | 08:16 |
n4nd0 | hi novice | 08:17 |
novice | hello :) | 08:17 |
n4nd0 | so github is not something you really "install", you need to create an account | 08:18 |
novice | yup created account | 08:18 |
n4nd0 | for what you have just said, I understand you have already done this step | 08:18 |
n4nd0 | ok | 08:18 |
n4nd0 | so now you should be able to use git | 08:18 |
n4nd0 | git is something you do install | 08:19 |
novice | installed git via sudo apt-get | 08:19 |
n4nd0 | ok cool | 08:19 |
novice | I am trying to follow the steps mentioned in the home page after we create an account for the first time | 08:19 |
n4nd0 | send me the link please | 08:20 |
novice | it says to genreate ssh key and add it | 08:20 |
novice | k jus a second | 08:20 |
n4nd0 | ok | 08:20 |
novice | http://help.github.com/linux-set-up-git/ | 08:20 |
n4nd0 | so have you already generated the keys? | 08:20 |
novice | yup | 08:20 |
novice | in the 5th step whihc says test it out | 08:20 |
novice | i get an error which says domain name not registered | 08:21 |
n4nd0 | what command are you executing? | 08:21 |
novice | sry the exact error message is ssh: Could not resolve hostname | 08:21 |
n4nd0 | ssh -T git@github.com | 08:21 |
novice | http://help.github.com/linux-set-up-git/ | 08:21 |
n4nd0 | ?? | 08:21 |
novice | yup ssh -T git@github.com | 08:21 |
n4nd0 | mmm I gind that weird | 08:22 |
n4nd0 | find* | 08:22 |
n4nd0 | can you try to do: ping github.com | 08:22 |
novice | I am in my college which uses NAT, u think its anythin to do with the firewall of my college? | 08:22 |
-!- Netsplit *.net <-> *.split quits: flxb, av3ngr | 08:23 | |
n4nd0 | well yes it might be they filtered ssh connections, port 22 basically | 08:23 |
novice | ping : unkown host github.com | 08:24 |
novice | ohh | 08:24 |
n4nd0 | but that is something they normally do in incoming connections | 08:24 |
novice | ping : unkown host github.com, this si the message i get | 08:24 |
novice | ok | 08:24 |
n4nd0 | can you browse github.com?? | 08:24 |
novice | yup | 08:24 |
n4nd0 | eeeh that is actually weird | 08:25 |
novice | i am able to see the issues of shogun and all | 08:25 |
n4nd0 | the shell tells you that it doesn't know where that is | 08:25 |
-!- av3ngr [av3ngr@nat/redhat/session] has joined #shogun | 08:25 | |
n4nd0 | but the browser can | 08:25 |
novice | im using virtual box | 08:25 |
novice | and trying to set ii up in virtual box | 08:25 |
novice | it* | 08:25 |
n4nd0 | might be something related | 08:25 |
av3ngr | hey novice what network set up did you use in virtual box? | 08:26 |
n4nd0 | try to ping 207.97.227.239 | 08:26 |
novice | i tried pinging from my windows cmd, it says similar thing | 08:26 |
novice | @av3ngr : i selected teh default option in it, nat | 08:26 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: Lost terminal] | 08:27 | |
av3ngr | novice, does ssh work for you normally? | 08:27 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 08:27 | |
novice | this is my first time using ssh :| | 08:27 |
n4nd0 | novice: what do you with ping to the IP address | 08:27 |
n4nd0 | what result do you get* | 08:28 |
av3ngr | n4nd0, it might be because 22 is blocked by the college. | 08:28 |
novice | linux the terminal doenst say anyhting, from windoes host i get 100% loss | 08:28 |
novice | Request timed out | 08:28 |
n4nd0 | av3ngr: could be, but now we are just trying to ping | 08:28 |
n4nd0 | av3ngr: it is not normal that you can browse a page and the ping doesn't work ... | 08:29 |
av3ngr | yea just realized that - this is odd | 08:29 |
n4nd0 | novice: do you browse from windows? | 08:29 |
av3ngr | unless college is blocking ping requests as well | 08:29 |
novice | naa browsing from virtual machine now | 08:29 |
av3ngr | (which is annoying - but often done by college admins_ | 08:29 |
n4nd0 | I have never seen that, but as you say, may be possible | 08:31 |
novice | lemme just check from my friends room, brb | 08:31 |
n4nd0 | ok | 08:31 |
av3ngr | http://returnbooleantrue.blogspot.com.au/2009/06/using-github-through-draconian-proxies.html << might help | 08:32 |
av3ngr | also http://stackoverflow.com/questions/5093558/using-github-behind-a-firewall-without-ssh-access | 08:33 |
-!- vikram360 [~vikram360@117.192.191.6] has quit [Ping timeout: 246 seconds] | 08:33 | |
novice | request timed out for him as well | 08:34 |
n4nd0 | av3ngr: yeah! those are good references, probably novice can fix the problem like that | 08:34 |
-!- vikram360 [~vikram360@117.192.167.113] has joined #shogun | 08:34 | |
novice | @av3ngr : checking the links, ty :) | 08:34 |
av3ngr | no worries | 08:34 |
av3ngr | also you could just email the IT admins to enable ssh for github :) | 08:36 |
-!- Netsplit over, joins: flxb | 08:41 | |
-!- PSmitAalto [82e9b263@gateway/web/freenode/session] has joined #shogun | 08:41 | |
-!- blackburn [~qdrgsm@62.106.114.183] has joined #shogun | 08:42 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 08:52 | |
-!- novice [b4953264@gateway/web/freenode/ip.180.149.50.100] has quit [Ping timeout: 245 seconds] | 08:55 | |
-!- novice [b4953264@gateway/web/freenode/session] has joined #shogun | 08:58 | |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 09:13 | |
-!- vikram360 [~vikram360@117.192.167.113] has quit [Read error: Connection reset by peer] | 09:16 | |
-!- vikram360 [~vikram360@117.192.187.187] has joined #shogun | 09:17 | |
-!- novice [b4953264@gateway/web/freenode/session] has quit [Quit: Page closed] | 09:22 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 09:25 | |
-!- av3ngr [av3ngr@nat/redhat/session] has quit [Changing host] | 09:57 | |
-!- av3ngr [av3ngr@nat/redhat/x-awjigafhwklslbst] has joined #shogun | 09:57 | |
-!- PSmitAalto [82e9b263@gateway/web/freenode/session] has quit [Changing host] | 09:57 | |
-!- PSmitAalto [82e9b263@gateway/web/freenode/ip.130.233.178.99] has joined #shogun | 09:57 | |
-!- av3ngr [av3ngr@nat/redhat/x-awjigafhwklslbst] has quit [Quit: That's all folks!] | 10:04 | |
-!- ada [~Adium@vera2g14-20.wi-fi.upv.es] has joined #shogun | 10:08 | |
-!- Miggy [~piggy@14.139.82.6] has joined #shogun | 10:26 | |
n4nd0 | blackburn: I think it really sucks for students in Iran and others not to be able to participate :S | 10:27 |
blackburn | pretty unfair probably yes | 10:29 |
blackburn | n4nd0: do you know anyone from iran wanting to participate? | 10:30 |
blackburn | ah | 10:31 |
blackburn | mailing list | 10:31 |
n4nd0 | yeah | 10:31 |
-!- jckrz [~jacek@89-69-164-5.dynamic.chello.pl] has joined #shogun | 10:43 | |
blackburn | sonney2k: I give up, you was right | 10:49 |
blackburn | n4nd0: damn this Soeren is wise | 10:50 |
blackburn | :D | 10:50 |
n4nd0 | blackburn: what is it now? :) | 10:51 |
blackburn | what do you mean? | 10:51 |
n4nd0 | you said that he was right | 10:51 |
blackburn | ah | 10:52 |
n4nd0 | what did you discuss? | 10:52 |
n4nd0 | something about STL or another thing? | 10:52 |
blackburn | there is no need for this libedrt | 10:52 |
blackburn | I spent a lot of time for useless shit | 10:52 |
n4nd0 | aha, why do you think so? | 10:52 |
blackburn | I got stucked with it | 10:52 |
n4nd0 | yeah man ... I checked your pull request, there is lot of stuff there, can't we get something profitable from it? | 10:52 |
blackburn | no need to rewrite things again | 10:52 |
blackburn | well it could be done in an easier way | 10:53 |
blackburn | oh fuck it I'll do this later | 10:54 |
n4nd0 | ok | 10:55 |
blackburn | n4nd0: didn't you send a pull request with kernel multiclass stuff? | 10:55 |
n4nd0 | I am on it right now | 10:55 |
blackburn | aha I just get hallucinated probably :D | 10:57 |
n4nd0 | blackburn: check this a moment please | 10:58 |
n4nd0 | blackburn: http://snipt.org/ughie7 | 10:58 |
blackburn | oh shit | 10:58 |
blackburn | :D | 10:58 |
blackburn | wtf | 10:58 |
n4nd0 | blackburn: is there a way to make presentingPolygon use the method of myGayPolygon?? | 10:58 |
n4nd0 | it would be useful for the kernel stuff | 10:58 |
blackburn | hahhaah | 10:59 |
n4nd0 | I want to have set_subset and remove_subset in Kernel and rewrite them in CustomKernel | 10:59 |
blackburn | it should work well already | 10:59 |
-!- jckrz [~jacek@89-69-164-5.dynamic.chello.pl] has quit [Quit: Ex-Chat] | 10:59 | |
n4nd0 | but the trick is that KernelMulticlassMachine would call the one in CustomKernel when it is one | 10:59 |
n4nd0 | mmm | 10:59 |
blackburn | doesn't it? | 10:59 |
n4nd0 | no gay polygon on the screen | 11:00 |
n4nd0 | just polygons | 11:00 |
blackburn | try to add constructors | 11:00 |
blackburn | and GayPolygon(): Polygon() | 11:00 |
blackburn | oh damn man that's funny | 11:01 |
blackburn | gay polygon | 11:01 |
n4nd0 | the same | 11:01 |
blackburn | what if you use pointers? | 11:02 |
n4nd0 | aham | 11:02 |
n4nd0 | that's a good idea actually | 11:02 |
n4nd0 | blackburn: you're wise too :) | 11:03 |
blackburn | so pointers make things work? | 11:04 |
n4nd0 | yeah | 11:04 |
blackburn | that's interesting | 11:04 |
n4nd0 | I can say a Gay polygon now | 11:04 |
n4nd0 | http://snipt.org/ughig7 | 11:04 |
blackburn | probably your gaypolygon was casted to polygon somehow | 11:04 |
n4nd0 | yeah, something like that | 11:04 |
blackburn | actually for methods like that | 11:05 |
n4nd0 | I have studied all this stuff at school with java so I know that it can be done | 11:05 |
blackburn | you would rather use const Polygon& poly | 11:05 |
n4nd0 | but sometimes I screw when I say ok, how do I do this in C++ | 11:05 |
blackburn | but not Polygon poly | 11:05 |
blackburn | try with const Polygon& | 11:05 |
n4nd0 | let me see | 11:05 |
n4nd0 | yeah | 11:06 |
n4nd0 | works as well with references | 11:06 |
blackburn | one more puzzle for interview LOL | 11:06 |
blackburn | did you get what is going on? | 11:07 |
n4nd0 | I guess that when the function is called | 11:07 |
n4nd0 | since the argument was not a reference | 11:08 |
n4nd0 | but the type directly | 11:08 |
n4nd0 | it was up-casting it | 11:08 |
blackburn | no, it copies | 11:08 |
n4nd0 | oh | 11:08 |
blackburn | actually it should copy this instance | 11:09 |
n4nd0 | then whatever I give to the function | 11:09 |
blackburn | if it is not a const | 11:09 |
n4nd0 | if it is a Polygon | 11:09 |
blackburn | one more idea | 11:09 |
n4nd0 | it will be a Polygon | 11:09 |
blackburn | what if you add some field to gay polygon | 11:09 |
blackburn | will it avoid gayness still? | 11:09 |
blackburn | they are almost equal in definition I mean | 11:10 |
n4nd0 | I would guess that yes | 11:10 |
n4nd0 | but let's check | 11:10 |
blackburn | but actually good lesson there - use const & | 11:10 |
blackburn | :D | 11:10 |
n4nd0 | still not gay if function gets Polygon | 11:11 |
blackburn | aha | 11:11 |
blackburn | in fact I didn't get how can you use it for kernels? | 11:12 |
n4nd0 | Kernel.h defines set_subset and remove_subset | 11:13 |
n4nd0 | or CKernel better | 11:13 |
n4nd0 | and CCustomKernel overrides it | 11:13 |
blackburn | yeah right | 11:14 |
Miggy | hi blackburn | 11:16 |
blackburn | hi | 11:16 |
Miggy | do u know if anyone is working on "Built generic multiclass learning framework" project? | 11:16 |
blackburn | working? | 11:17 |
blackburn | not really | 11:17 |
Miggy | i mean has anyone given any proposal for that project? | 11:18 |
Miggy | i m very much interested in doing that project | 11:19 |
blackburn | oh I don't know | 11:19 |
Miggy | ok | 11:20 |
Miggy | blackburn: when does sonney2k come...he is the mentor and i would like to discuss abt the project | 11:21 |
n4nd0 | well guys, gtg now | 11:22 |
n4nd0 | blackburn: I will push the multiclass kernel stuff a bit later | 11:22 |
n4nd0 | bye | 11:22 |
blackburn | n4nd0: see you | 11:23 |
Miggy | n4nd0: u working on multiclass learning framework? | 11:23 |
n4nd0 | Miggy: oh sorry, I didn't read | 11:24 |
blackburn | Miggy: just some patch | 11:24 |
blackburn | Miggy: yeah try to catch him this evening | 11:24 |
n4nd0 | Miggy: so I am planning to apply for it, but it is not my first option, my first is SO | 11:24 |
Miggy | ok | 11:25 |
n4nd0 | bye people! | 11:25 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 11:25 | |
Miggy | blackburn: really finding it hard to do patch...suggest me something please | 11:26 |
Miggy | any small algo or lib functions will do | 11:27 |
blackburn | sorry have to go now | 11:28 |
blackburn | will be back in a hour or so | 11:28 |
Miggy | ok but please come with something good :) | 11:29 |
-!- flxb [~cronor@g230122054.adsl.alicedsl.de] has quit [Quit: flxb] | 11:35 | |
-!- flxb [~cronor@141.23.80.206] has joined #shogun | 12:25 | |
-!- flxb [~cronor@141.23.80.206] has quit [Quit: flxb] | 13:00 | |
-!- flxb [~cronor@fb.ml.tu-berlin.de] has joined #shogun | 13:06 | |
blackburn | Miggy: ok still here/ | 13:37 |
blackburn | ? | 13:37 |
Miggy | yes blackburn | 14:02 |
blackburn | Miggy: anything you would prefer to do? | 14:12 |
-!- n4nd0 [~nando@n179-p53.kthopen.kth.se] has joined #shogun | 14:18 | |
Miggy | yes smthing which is not too long..dont want to start with big code | 14:27 |
blackburn | one patch could be to make use of covertree in knn classifier | 14:31 |
blackburn | simpler one is to implement some graphical regression or classification example | 14:33 |
blackburn | with some popular dataset | 14:33 |
-!- vikram360 [~vikram360@117.192.187.187] has quit [Ping timeout: 250 seconds] | 14:39 | |
-!- PhilTillet [~Philippe@tillet-p42154.maisel.int-evry.fr] has joined #shogun | 14:51 | |
-!- harshit_ [~harshit@182.68.160.94] has joined #shogun | 14:51 | |
PhilTillet | Hi | 14:53 |
harshit_ | hey PhilTillet | 14:54 |
harshit_ | and hi n4nd0 | 14:54 |
PhilTillet | :) | 14:54 |
-!- pluskid [~chatzilla@173.254.214.60] has joined #shogun | 14:54 | |
n4nd0 | harshit_: hey man | 14:54 |
harshit_ | whats going on .. what are you working on, these days ? | 14:55 |
n4nd0 | harshit_: so I pushed yesterday some multiclass stuff, one-vs-one for generic | 14:56 |
n4nd0 | harshit_: now I am preparing an example with multiclass kernel using that stuff | 14:57 |
n4nd0 | harshit_: what about you? | 14:57 |
harshit_ | ohh great .. | 14:57 |
harshit_ | actually i didnt get time, last 2 days were really hectic for me.. now next 2 days are free | 14:58 |
harshit_ | so planning to do something big | 14:58 |
harshit_ | hey i had a problem | 14:58 |
harshit_ | in my data folder in shogun directory | 14:58 |
-!- vikram360 [~vikram360@117.192.186.163] has joined #shogun | 14:58 | |
harshit_ | i dont have any files | 14:58 |
n4nd0 | yeah | 14:59 |
harshit_ | i dont knw why its like that, coz of that many octave examples are not running | 14:59 |
n4nd0 | data is in another repo | 14:59 |
n4nd0 | check it on github | 14:59 |
harshit_ | ohh | 14:59 |
n4nd0 | go to shogun user and check its repos | 14:59 |
blackburn | git submodule init --update | 14:59 |
n4nd0 | there is another one called shogun-data | 14:59 |
n4nd0 | ok ... that's probably a better option :) | 14:59 |
n4nd0 | blackburn: thank you! I had the other repo in another place and soft links all around the place :O | 15:00 |
blackburn | hah | 15:00 |
-!- vikram360 [~vikram360@117.192.186.163] has quit [Ping timeout: 260 seconds] | 15:03 | |
-!- vikram360 [~vikram360@117.192.170.195] has joined #shogun | 15:03 | |
-!- pluskid [~chatzilla@173.254.214.60] has quit [Ping timeout: 260 seconds] | 15:13 | |
harshit_ | dam, this data directory is really big !! | 15:19 |
blackburn | yes it is | 15:22 |
-!- PhilTillet [~Philippe@tillet-p42154.maisel.int-evry.fr] has quit [Ping timeout: 244 seconds] | 15:35 | |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Read error: Connection reset by peer] | 15:43 | |
-!- ada [~Adium@vera2g14-20.wi-fi.upv.es] has quit [Quit: Leaving.] | 15:45 | |
n4nd0 | see you guys! | 16:03 |
-!- n4nd0 [~nando@n179-p53.kthopen.kth.se] has quit [Quit: leaving] | 16:04 | |
harshit_ | blackburn : do you knw any multiclass svm classifier that doesnt have its example in octave_modular ? | 16:08 |
blackburn | harshit_: IIRC examples of octave_modular are pretty outdated | 16:14 |
harshit_ | so i have created an example where 20NewsGroup dataset is being classified (Actually there is a lil bit of work left ) but i am using cgmnpsvm, whose example exists beforehand | 16:18 |
harshit_ | and i am not using any context, its just BOW representation of 20NewsGroup that i am using | 16:19 |
blackburn | harshit_: consider using multiclassliblinear then | 16:21 |
harshit_ | okay, thanks blackburn | 16:22 |
blackburn | harshit_: gmnp is kernelized one, no need to use it there | 16:22 |
harshit_ | but kernelized features might give better result . no ? | 16:23 |
harshit_ | never mind multiclassliblinear example also doesnt exist so i think doing that would be fruitful .. | 16:25 |
blackburn | harshit_: yes if you use good kernel for that | 16:32 |
blackburn | however even linear would work | 16:32 |
harshit_ | blackburn : also i was thinking to include some nlp stuff in shogun like porter's algorithm ,or any other stemming algo | 16:34 |
blackburn | feel free! | 16:35 |
harshit_ | do you think shogun requires it | 16:35 |
harshit_ | :) gr | 16:35 |
harshit_ | gr8* | 16:35 |
-!- gsomix [~gsomix@188.168.128.120] has joined #shogun | 16:37 | |
gsomix | hi | 16:40 |
blackburn | harshit_: not really but it shows how can you use it, it is rather application | 16:43 |
-!- ada [~Adium@veraw142-195.wi-fi.upv.es] has joined #shogun | 16:48 | |
-!- jckrz [~jacek@89-69-164-5.dynamic.chello.pl] has joined #shogun | 16:55 | |
harshit_ | blackburn: So I think i should ask sonney2k before doing it . | 16:55 |
-!- ada [~Adium@veraw142-195.wi-fi.upv.es] has quit [Quit: Leaving.] | 17:06 | |
-!- harshit_ [~harshit@182.68.160.94] has quit [Ping timeout: 252 seconds] | 17:44 | |
-!- PhilTillet [~Philippe@tillet-p42154.maisel.int-evry.fr] has joined #shogun | 18:47 | |
-!- flxb [~cronor@fb.ml.tu-berlin.de] has quit [Quit: flxb] | 18:48 | |
-!- harshit_ [~harshit@182.68.160.94] has joined #shogun | 19:17 | |
-!- PhilTillet_ [~Philippe@tillet-p42154.maisel.int-evry.fr] has joined #shogun | 19:36 | |
-!- PhilTillet [~Philippe@tillet-p42154.maisel.int-evry.fr] has quit [Read error: Connection reset by peer] | 19:36 | |
PhilTillet_ | Hi again | 19:43 |
blackburn | hi | 19:47 |
PhilTillet_ | I hate jet lag. | 19:47 |
blackburn | heh | 19:48 |
PhilTillet_ | I arrived at 2pm, slept more or less until now | 19:48 |
PhilTillet_ | Now I am working on my GSoC proposal. | 19:49 |
PhilTillet_ | I pretty much got some results with SVM on the GPU | 19:54 |
blackburn | PhilTillet_: actually sonney2k said he has got worse results than you | 20:04 |
PhilTillet_ | I know, i saw his mail, the results are weird but I have seen on the internet some complaints about OpenCL on the 9600GT M | 20:05 |
PhilTillet_ | so I tested it on a dedicated benchmark platform and got even better results | 20:05 |
PhilTillet_ | that is a factor 8 | 20:05 |
PhilTillet_ | (i7 950 vs GTX 470) | 20:06 |
blackburn | I see | 20:12 |
PhilTillet_ | I'll try to test it on more hardware though, and for more problem sizes | 20:14 |
PhilTillet_ | Essentially, the results should be about the same then for Matrix-Matrix product | 20:17 |
PhilTillet_ | as I am using somewhat the same layout of parallelization | 20:17 |
PhilTillet_ | as for* XD (still tired :( ) | 20:19 |
blackburn | :) | 20:25 |
-!- prabhakarr [0e8b5206@gateway/web/freenode/ip.14.139.82.6] has joined #shogun | 20:33 | |
prabhakarr | Am trying to install shogun for using the "converter" tools. Everything else seems to be working except "converter". Any help is appreciated ! | 20:34 |
blackburn | prabhakarr: what is the exact problem? | 20:38 |
prabhakarr | "from shogun.Converter import Isomap" results in "ImportError: cannot import name Isomap"; i got the recent version from git, compiled "python_modular" | 20:42 |
prabhakarr | This is from the example code given: "../examples/documented/python_modular/converter_isomap_modular.py" | 20:44 |
CIA-64 | shogun: Soeren Sonnenburg master * r25f5321 / (5 files): | 20:46 |
CIA-64 | shogun: Merge pull request #403 from pluskid/gsoc-pullrequest-1 | 20:46 |
CIA-64 | shogun: Make the testsuite data generation script run again. - http://git.io/KKePKg | 20:46 |
blackburn | prabhakarr: did you install lapack? | 20:46 |
-!- flxb [~cronor@g225030062.adsl.alicedsl.de] has joined #shogun | 20:47 | |
prabhakarr | blackburn: thanks for the pointers: this is my first time in any IRC. lapack is installed | 20:49 |
blackburn | prabhakarr: was it detected by shogun configure? | 20:49 |
prabhakarr | blackburn: thats a good catch: ./configure gives: "Checking for Vanilla lapack and blas support ... no (missing some features: LDA, liblinear, ...)" ; | 20:55 |
blackburn | prabhakarr: what is the os you use? | 20:55 |
prabhakarr | blackburn: ubuntu 11.04 x86_64. installed lapack. now ./configure recognises and gives "yes". now retrying the install | 20:56 |
blackburn | prabhakarr: try to install liblapack-dev | 20:57 |
prabhakarr | blackburn: thats exactly what I did. [realised as soon as you pointed out the lapack configure "no", that i was missing the dev package]. install in progress, fingers crossed :) | 20:58 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 21:14 | |
-!- romi_ [~mizobe@187.74.1.223] has joined #shogun | 21:18 | |
prabhakarr | blackburn: with your pointer for lapack, it seems to be working. Thanks a ton. | 21:24 |
blackburn | no problem | 21:24 |
-!- prabhakarr [0e8b5206@gateway/web/freenode/ip.14.139.82.6] has left #shogun [] | 21:25 | |
blackburn | sonney2k: hey | 21:30 |
@sonney2k | blackburn, hey | 21:30 |
blackburn | sonney2k: what is # of proposals? ;) | 21:31 |
@sonney2k | I am swamped in stuff | 21:31 |
@sonney2k | let me check | 21:31 |
blackburn | and how are you? | 21:31 |
blackburn | oh that's bad | 21:31 |
@sonney2k | 11 | 21:31 |
blackburn | sonney2k: will be 12 in a min | 21:31 |
@sonney2k | this extrapolates to 30 or so :) | 21:32 |
@sonney2k | I wouldn't mind to see a lot less this year | 21:32 |
blackburn | sonney2k: 12 | 21:33 |
@sonney2k | indeed | 21:34 |
@sonney2k | blackburn, what do you think of http://help.github.com/line-endings/ ? | 21:34 |
blackburn | sonney2k: I have probably no opinion on thta | 21:35 |
-!- PhilTillet_ [~Philippe@tillet-p42154.maisel.int-evry.fr] has quit [Remote host closed the connection] | 21:35 | |
@sonney2k | we should be on LF only anyways | 21:35 |
blackburn | are we on CRLF right now? | 21:35 |
@sonney2k | or do we have any developer doing any work on win crap? | 21:35 |
@sonney2k | linux is just LF | 21:35 |
@sonney2k | osx too | 21:35 |
blackburn | I know no one | 21:35 |
@sonney2k | so there shouldn't be any danger | 21:36 |
blackburn | aha | 21:36 |
-!- PhilTillet [~Philippe@tillet-p42154.maisel.int-evry.fr] has joined #shogun | 21:37 | |
@sonney2k | blackburn, reviewing https://github.com/shogun-toolbox/shogun/pull/402/files is a pain | 21:37 |
@sonney2k | many changes and a lot of them just whitspace | 21:37 |
blackburn | sonney2k: whoaa | 21:37 |
blackburn | sonney2k: does evan want to participate in gsoc? | 21:37 |
@sonney2k | but the whitespace fixes seem rightful | 21:38 |
@sonney2k | fixes lots of trailing spaces | 21:38 |
blackburn | yeah trailing spaces is shit | 21:38 |
blackburn | sonney2k: any option for vim for that? | 21:38 |
PhilTillet | sonney2k, can the source file i sent to you be considered as the "small patch"? :D | 21:39 |
-!- gsomix [~gsomix@188.168.128.120] has quit [Ping timeout: 276 seconds] | 21:39 | |
@sonney2k | blackburn, btw https://github.com/shogun-toolbox/shogun/pull/397/files also seems nice work | 21:40 |
@sonney2k | and a neat example fetching data from uci :) | 21:40 |
blackburn | yeah I thought I suggested it | 21:41 |
@sonney2k | at least we never had clustering metrics and I have never heard of munkres before but like the idea of using it to resolve label ambiguity | 21:41 |
blackburn | sonney2k: what is idea he applying for? | 21:42 |
n4nd0 | Munkres ... I used that for my sokoban solver, nice algorithm | 21:42 |
n4nd0 | blackburn: he is applying for SO :S | 21:42 |
blackburn | damn | 21:43 |
blackburn | I feel pain right in my ass thinking about it | 21:44 |
@sonney2k | PhilTillet, you need to get into shogun too - the isolated code is very drafty (which is ok at this stage) but it would be a big big plus to have the kernel machine opencl parallelization within shogun | 21:48 |
@sonney2k | PhilTillet, ahh btw the code still does not do the 1000 support vectors / 100000 test examples case | 21:49 |
@sonney2k | which is the more common case | 21:49 |
@sonney2k | PhilTillet, I know you are very busy with conferences and stuff but there is still 1 week... | 21:50 |
PhilTillet | sonney2k, I am done with the conference actually, back in France :) What do you mean with "does not do" ? infinite runtime? | 21:51 |
@sonney2k | PhilTillet, well you currently use a case where you have 1000 SVs and 1000 test examples | 21:52 |
PhilTillet | yep | 21:52 |
PhilTillet | and tested it with 10000test example too | 21:52 |
PhilTillet | but the execution time was very long :p | 21:52 |
PhilTillet | i mean i can test it with 100 000 but then I have to take a coffee break lol | 21:52 |
PhilTillet | and I don't like coffee, so let's say a water break | 21:53 |
-!- gsomix [~gsomix@188.168.4.150] has joined #shogun | 21:54 | |
PhilTillet | ok the test case for 100000 test example is launched | 21:54 |
@sonney2k | PhilTillet, well 100 times longer | 21:55 |
PhilTillet | I'll run it on that Benchmarking machine, should be faster :p | 21:55 |
@sonney2k | yeah just some 30 seconds or so | 21:56 |
@sonney2k | PhilTillet, btw I am not happy with the numerical precision though | 21:56 |
@sonney2k | I don't see any reason why results should deviate by more than 1e-6 | 21:56 |
@sonney2k | 1e-3 is a lot IMO | 21:57 |
PhilTillet | well, it's 1e-3 over 1000example | 21:57 |
PhilTillet | so 1e-6 per example | 21:57 |
PhilTillet | the norm2 of the difference is still zero from the tests I did | 21:58 |
@sonney2k | norm2 is unfair I really mean absdiff | 22:03 |
PhilTillet | And okay i'll try to integrate the OpenCL work into CKernelMachine | 22:03 |
PhilTillet | yes | 22:03 |
@sonney2k | PhilTillet, shouldn't your benchmark machine be able to do double btw? | 22:03 |
PhilTillet | it can | 22:04 |
PhilTillet | but not on my mobile GPU | 22:04 |
PhilTillet | :) | 22:04 |
@sonney2k | then you could directly compare for the 1000x1000 problem | 22:04 |
PhilTillet | the problem is that the GPU has to be double compatible | 22:04 |
PhilTillet | which is not always the case :p | 22:04 |
blackburn | anyone wanting to port some matlab code? | 22:05 |
@sonney2k | yeah yeah but it is important to check if there is a bug (which is what I suspsect) | 22:05 |
@sonney2k | blackburn, ??? | 22:05 |
blackburn | sonney2k: what is ??? for ? ;) | 22:05 |
@sonney2k | matlab code? what are you doing? | 22:06 |
blackburn | sonney2k: easy task - port some dim reduction method | 22:06 |
blackburn | anyway, n4nd0 will care about stochastic proximity embedding next days | 22:07 |
n4nd0 | blackburn: let's see how it goes with it ;) | 22:08 |
PhilTillet | sonney2k, the SVM is still training with 100 000 examples :p | 22:13 |
PhilTillet | the training seems quite computationally expensive :p | 22:47 |
blackburn | PhilTillet: what is d and N? | 22:47 |
PhilTillet | N = 100 000, d=1000 | 22:48 |
blackburn | oh | 22:48 |
blackburn | and kernel? | 22:48 |
PhilTillet | Gaussian | 22:48 |
blackburn | oh that can take a while :) | 22:48 |
@sonney2k | PhilTillet, training? why training? | 22:48 |
@sonney2k | or do you mean apply? | 22:48 |
PhilTillet | well the SVM is trained before apply | 22:48 |
PhilTillet | does not need to be? | 22:48 |
@sonney2k | why that? | 22:48 |
blackburn | PhilTillet: you probably can set alphas manually.. | 22:49 |
@sonney2k | just set alphas to 1 | 22:49 |
PhilTillet | oh. | 22:49 |
@sonney2k | (or random numbers again) | 22:49 |
PhilTillet | :D | 22:49 |
@sonney2k | but yes training is very expensive... | 22:50 |
PhilTillet | i'm sure training can also be done on GPU :P | 22:50 |
PhilTillet | but not in 1 week :D | 22:50 |
blackburn | that's something complex | 22:50 |
blackburn | take a look on libsvm source | 22:50 |
PhilTillet | So I have to set the ids of the support vectors manually too? | 22:53 |
blackburn | PhilTillet: yeah just create KernelMachine using custom alphas and indices | 22:53 |
PhilTillet | okay XD | 22:54 |
shogun-buildbot | build #405 of python_modular is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/python_modular/builds/405 blamelist: pluskid@gmail.com | 22:55 |
blackburn | ehm?! | 22:55 |
n4nd0 | I got a graphical example for multiclass kernel machines | 23:06 |
n4nd0 | http://dl.dropbox.com/u/11020840/multiclass_svm.png | 23:06 |
n4nd0 | what do you think guys? | 23:06 |
shogun-buildbot | build #401 of ruby_modular is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/ruby_modular/builds/401 blamelist: pluskid@gmail.com | 23:06 |
blackburn | WOW | 23:07 |
n4nd0 | it is a least curious :P | 23:07 |
blackburn | impressive | 23:07 |
blackburn | n4nd0: what is the svm? | 23:07 |
n4nd0 | my gf asks why I am doing flowers | 23:07 |
blackburn | tell her because of gay polygon | 23:08 |
n4nd0 | LibSVM with GaussianKernel | 23:08 |
blackburn | hmm | 23:08 |
n4nd0 | one vs rest | 23:08 |
n4nd0 | I am going to try with o.vs.o | 23:08 |
n4nd0 | but I think it will be the same | 23:08 |
blackburn | interesting | 23:08 |
blackburn | yeah a little difference | 23:09 |
n4nd0 | oh shit, seg fault with one vs. one, got to debug | 23:10 |
PhilTillet | sonney2k, so I should add a clean apply_opencl function to CKernelMachine ? | 23:10 |
PhilTillet | and if possible make it templated so that both apply_opencl<float> and apply_opencl<double> work? | 23:11 |
CIA-64 | shogun: Evan Shelhamer master * rb763e52 / (337 files in 23 dirs): Trailing whitespace cleanup - http://git.io/sidz_g | 23:18 |
CIA-64 | shogun: Soeren Sonnenburg master * r17ea108 / (337 files in 23 dirs): | 23:18 |
CIA-64 | shogun: Merge pull request #404 from shelhamer/whitespace-cleanup | 23:18 |
CIA-64 | shogun: Trailing whitespace cleanup - http://git.io/j1EoZQ | 23:18 |
@sonney2k | PhilTillet, well first I would be very interested in seeing the results (much lower error) and comparison to 1000 SVs - 100k test | 23:26 |
@sonney2k | then a apply_gpu makes sense where the old apply code is put into apply_cpu | 23:28 |
-!- Vuvu [~Vivan_Ric@115.248.130.148] has joined #shogun | 23:28 | |
@sonney2k | and then one can set a flag such that apply() uses the cpu or gpu function | 23:29 |
-!- flxb [~cronor@g225030062.adsl.alicedsl.de] has quit [Read error: Connection reset by peer] | 23:29 | |
-!- flxb [~cronor@g225030062.adsl.alicedsl.de] has joined #shogun | 23:30 | |
shogun-buildbot | build #410 of java_modular is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/java_modular/builds/410 | 23:36 |
shogun-buildbot | build #406 of python_modular is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/python_modular/builds/406 | 23:43 |
shogun-buildbot | build #402 of ruby_modular is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/ruby_modular/builds/402 | 23:56 |
-!- harshit_ [~harshit@182.68.160.94] has quit [Quit: Leaving] | 23:57 | |
-!- jckrz [~jacek@89-69-164-5.dynamic.chello.pl] has quit [Quit: Ex-Chat] | 23:58 | |
--- Log closed Sat Mar 31 00:00:19 2012 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!