--- Log opened Sun Apr 03 00:00:36 2011 | ||
@knrrrd | Nite guys | 01:23 |
---|---|---|
@knrrrd | Time to sleep, bye | 01:24 |
-!- blackburn [~qdrgsm@62.106.119.77] has quit [Quit: Leaving.] | 02:21 | |
-!- serialhex [~quassel@99-101-149-136.lightspeed.wepbfl.sbcglobal.net] has quit [Remote host closed the connection] | 02:28 | |
-!- dvevre [b49531e3@gateway/web/freenode/ip.180.149.49.227] has quit [Ping timeout: 252 seconds] | 02:45 | |
-!- serialhex [~quassel@99-101-149-136.lightspeed.wepbfl.sbcglobal.net] has joined #shogun | 04:41 | |
-!- siddharth_k [~siddharth@117.211.88.150] has joined #shogun | 06:31 | |
-!- aifargonos [~aifargono@46.18.27.35] has joined #shogun | 08:48 | |
-!- akhil_ [75d35896@gateway/web/freenode/ip.117.211.88.150] has joined #shogun | 09:04 | |
-!- siddharth_k [~siddharth@117.211.88.150] has quit [Quit: Leaving] | 09:13 | |
-!- akhil_ [75d35896@gateway/web/freenode/ip.117.211.88.150] has quit [Quit: Page closed] | 09:22 | |
-!- akhil_ [75d35896@gateway/web/freenode/ip.117.211.88.150] has joined #shogun | 09:23 | |
-!- siddharth [~siddharth@117.211.88.150] has joined #shogun | 09:36 | |
-!- Ryaether [~Ryaether@50-80-170-245.client.mchsi.com] has quit [Ping timeout: 276 seconds] | 09:41 | |
* sonney2k opens his eyes | 09:59 | |
-!- akhil_ [75d35896@gateway/web/freenode/ip.117.211.88.150] has quit [Ping timeout: 252 seconds] | 10:34 | |
@knrrrd | ;) | 10:56 |
@bettyboo | hihi | 10:56 |
@sonney2k | bettyboo, are you always smiling? | 10:58 |
@bettyboo | sonney2k: but you are interested in shogun? | 10:58 |
@sonney2k | bettyboo, yes of course - now I understand the reason why you are smiling. | 10:59 |
@bettyboo | sonney2k: input set of strings, output n-dimensional representation | 10:59 |
@sonney2k | bettyboo, You are talkng about kPCA or KFA right? | 10:59 |
@bettyboo | sonney2k: shogun: Implement snappy compression support (http://code.google.com/p/snappy/). | 10:59 |
Tanmoy | @sonney2k some of the preprocessing methods for example norm won't it be better to use GSL's code | 11:24 |
Tanmoy | especially BLAS | 11:24 |
-!- blackburn [~qdrgsm@62.106.119.77] has joined #shogun | 11:43 | |
-!- akhil_ [75d35896@gateway/web/freenode/ip.117.211.88.150] has joined #shogun | 11:55 | |
-!- Ziyuan [~Ziyuan@116.22.36.234] has joined #shogun | 12:00 | |
-!- mank [75d35895@gateway/web/freenode/ip.117.211.88.149] has joined #shogun | 12:24 | |
@sonney2k | Tanmoy, I am not sure I understand: Did you mean we should use blas whenever possible? | 12:30 |
@sonney2k | because - that is what we are trying | 12:35 |
@sonney2k | (though the preproc code is probably among the oldest parts of shogun and not yet utilizing all of that) | 12:36 |
Tanmoy | @sonney2k yes thts wat i mean | 12:42 |
Tanmoy | for example why would this be required http://www.shogun-toolbox.org/doc/classshogun_1_1CNormOne.html | 12:43 |
Tanmoy | http://www.gnu.org/softhttp://www.gnu.org/software/gsl/manual/html_node/Level-1-GSL-BLAS-Interface.html | 12:43 |
Tanmoy | when already implemented as a norm in BLAS | 12:43 |
Tanmoy | @sonney2k however for the missing algorithms part i think the desgn issues needs to be a bit better sorted out | 12:44 |
@sonney2k | Tanmoy, well one is a preprocessor that you can attach to features and one is just the method to compute the norm of a vector | 12:44 |
Tanmoy | but i belive the work is to find norm right ? | 12:45 |
@sonney2k | no, it is that when you have a set of vectors to ensure that they all have norm one. | 12:46 |
@sonney2k | so divide each vector by its norm | 12:46 |
@sonney2k | it is actually using blas already... | 12:47 |
Tanmoy | yup | 12:47 |
Tanmoy | the link i sent was for norm | 12:47 |
Tanmoy | maybe i shld have said divide by norm :) | 12:48 |
@bettyboo | rahaha | 12:48 |
CIA-30 | shogun: Soeren Sonnenburg master * rb0614d4 / .gitignore : improve gitignore for modular interfaces - http://bit.ly/eaES8Z | 12:49 |
CIA-30 | shogun: Soeren Sonnenburg master * r085319f / src/Makefile.template : Add dummy make install targets for java, lua and ruby - http://bit.ly/gprCcw | 12:49 |
CIA-30 | shogun: Soeren Sonnenburg master * r56b366c / .gitignore : improve gitignore to exclude doxy files etc - http://bit.ly/eAnNhO | 12:49 |
@knrrrd | i think i need to fix something here | 12:49 |
-!- bettyboo [~bettyboo@bane.ml.tu-berlin.de] has quit [Remote host closed the connection] | 12:50 | |
-!- siddharth [~siddharth@117.211.88.150] has quit [Ping timeout: 250 seconds] | 12:54 | |
-!- bettyboo [~bettyboo@bane.ml.tu-berlin.de] has joined #shogun | 12:54 | |
-!- mode/#shogun [+o bettyboo] by ChanServ | 12:54 | |
mank | @sonney2k, Hi. I had sent you an email regarding GSoC Project ( i/o in .hdf5 format ) | 13:00 |
@sonney2k | mank, I recall answering :) | 13:00 |
mank | yeah, I wanted to discuss more about it. | 13:01 |
@knrrrd | woah, truecrypt really improved in terms of GUI and usability | 13:01 |
@knrrrd | generating a encrypted harddisk is really easy now | 13:01 |
@knrrrd | and its cross-platform (in contrast to linux and osx specific drivers) | 13:02 |
mank | @sonney2k, Is it a good time to discuss ? | 13:08 |
@sonney2k | mank, if it does not take too long - ask! | 13:09 |
mank | Ok, couple of quick questions. | 13:09 |
mank | 1. Are we going to use HDF5 Api's for i/o ? | 13:10 |
blackburn | sonney2k: I have implemented a fast nearest neighbor classification (in case of k=1) in CKNN class. what should I do after cleaning and commenting to get it commited? | 13:10 |
mank | 2. where should i start from ? | 13:11 |
-!- siddharth__ [~siddharth@117.211.88.150] has joined #shogun | 13:11 | |
@sonney2k | mank, hdf5 api's are already used for i/o look at CHDF5File and the CSerializableHDF5File classes | 13:12 |
@sonney2k | mank, look at the CFile API | 13:13 |
@sonney2k | blackburn, submit a patch to the mailing list (attach what git format-patch produces ) | 13:14 |
blackburn | sonney2k: okay, thank you | 13:14 |
mank | @sonney2k: Thank you. I am really interested in working on this. I will have a look at this and maybe discuss more after going through these resources. | 13:16 |
@sonney2k | mank, it would really make sense to do the octave / and older matlab binary formats too... for that look also at the source of octave (it's linked to at the gsoc page) | 13:17 |
mank | ok. I will have a look at that too. | 13:17 |
mank | @sonney2k, Thanks for your time! | 13:18 |
-!- siddharth__ [~siddharth@117.211.88.150] has quit [Read error: Connection reset by peer] | 13:41 | |
-!- akhil_ [75d35896@gateway/web/freenode/ip.117.211.88.150] has quit [Quit: Page closed] | 14:03 | |
-!- siddharth__ [~siddharth@117.211.88.150] has joined #shogun | 14:14 | |
-!- siddharth__ [~siddharth@117.211.88.150] has quit [Ping timeout: 250 seconds] | 14:20 | |
* knrrrd work on the balcony. cleaning chairs and table. | 14:20 | |
-!- siddharth__ [~siddharth@117.211.88.150] has joined #shogun | 14:35 | |
-!- mank [75d35895@gateway/web/freenode/ip.117.211.88.149] has quit [Quit: Page closed] | 14:46 | |
-!- akhil_ [75d35896@gateway/web/freenode/ip.117.211.88.150] has joined #shogun | 15:30 | |
-!- akhil_ [75d35896@gateway/web/freenode/ip.117.211.88.150] has quit [Quit: Page closed] | 16:00 | |
siddharth__ | sonney2k, hi | 16:16 |
@bettyboo | siddharth__, hey | 16:16 |
siddharth__ | hey | 16:16 |
@bettyboo | ho | 16:16 |
siddharth__ | Dimensionality reduction methods like LLE,ISOMAP not implemented in Shogun right? | 16:17 |
blackburn | I will send a proposal about it | 16:17 |
siddharth__ | you working on this? | 16:17 |
blackburn | not working, want to | 16:18 |
siddharth__ | so can two people do and divide the work? | 16:18 |
blackburn | I don't know, even don't know any chances to get into this program :) | 16:19 |
@bettyboo | funny ;> | 16:19 |
siddharth__ | :P | 16:19 |
@bettyboo | lol ;D | 16:19 |
blackburn | btw, there is a kernel pca not implemented | 16:20 |
siddharth__ | actually there are many Dimensionality Reduction methods yet to be implemented | 16:20 |
blackburn | or implemented.. I forgot | 16:20 |
blackburn | I'll try to send a proposal about LLE, SNE, ISOMAP and MDS | 16:20 |
siddharth__ | well lets see...I am also interested in this | 16:22 |
siddharth__ | what about boosting algorithms? | 16:22 |
blackburn | what is the question? :) | 16:23 |
@bettyboo | :> | 16:23 |
siddharth__ | I mean they are also not implemented right? | 16:23 |
blackburn | I don't pretty sure, but there is an classifier/boosting | 16:24 |
siddharth__ | yeah that i went through | 16:25 |
siddharth__ | most of them implemented | 16:25 |
blackburn | I thought you already chose idea to work with | 16:27 |
siddharth__ | yeah I am interested in implementing primal SVM solver...though if this finishes quickly,I want to do work on Dim Red | 16:29 |
blackburn | ah, I see | 16:30 |
@knrrrd | there is also the kfa task. | 16:30 |
@knrrrd | if you are interested. | 16:30 |
siddharth__ | ok I will look into that | 16:31 |
@knrrrd | lle, sne and isomap would also be interesting, as they all require a knn backend | 16:32 |
-!- serialhex is now known as gsoc_serialhex | 16:32 | |
@knrrrd | where mds can be build ontop of pca easily | 16:32 |
blackburn | isomap is very similar to mds too | 16:33 |
@knrrrd | blackburn: yes? how that? | 16:34 |
@knrrrd | ah. yes, isomap employs mds is the last step | 16:35 |
@knrrrd | i recall | 16:35 |
blackburn | knrrrd: it is like mds, but with graph distance | 16:35 |
blackburn | yeap | 16:35 |
-!- serialhex [~quassel@99-101-149-136.lightspeed.wepbfl.sbcglobal.net] has joined #shogun | 16:37 | |
-!- gsoc_serialhex [~quassel@99-101-149-136.lightspeed.wepbfl.sbcglobal.net] has left #shogun ["http://quassel-irc.org - Chat comfortably. Anywhere."] | 16:38 | |
-!- siddharth__ [~siddharth@117.211.88.150] has quit [Remote host closed the connection] | 17:21 | |
-!- siddharth__ [~siddharth@117.211.88.150] has joined #shogun | 17:30 | |
@knrrrd | Hmmm | 18:12 |
blackburn | that hmmm reminded me a http://www.zazzle.com/hmm_tshirt-235879303522160289 | 18:16 |
blackburn | :) | 18:16 |
@bettyboo | ;D blackburn | 18:16 |
-!- iLucian [~tc@145.116.231.246] has joined #shogun | 18:50 | |
-!- pece [~pczerkas@sub213-112.elpos.net] has joined #shogun | 20:23 | |
pece | hello, does LaRank implementation used in shogun support online training? | 20:25 |
-!- hiteshk [~hitesh_na@117.211.88.150] has joined #shogun | 20:52 | |
@sonney2k | blackburn, that hmmh is pretty cool :D | 20:58 |
@sonney2k | pece, well we don't have online features | 20:58 |
@sonney2k | so no not really | 20:58 |
@sonney2k | but it could if we had them... | 20:58 |
hiteshk | hi | 22:05 |
@bettyboo | ho | 22:05 |
hiteshk | sonney2k , thr | 22:05 |
-!- hiteshk [~hitesh_na@117.211.88.150] has left #shogun [] | 22:53 | |
-!- pece [~pczerkas@sub213-112.elpos.net] has quit [] | 23:05 | |
-!- iLucian [~tc@145.116.231.246] has quit [Quit: Leaving] | 23:17 | |
-!- nidhelplis [~nickname@46.115.19.28] has joined #shogun | 23:27 | |
nidhelplis | about running toolbox on xp | 23:29 |
-!- nidxelplis [2e73131c@gateway/web/freenode/ip.46.115.19.28] has joined #shogun | 23:30 | |
nidxelplis | aint one hir? | 23:31 |
-!- nidxelplis [2e73131c@gateway/web/freenode/ip.46.115.19.28] has quit [Quit: Page closed] | 23:36 | |
-!- nidhelplis [~nickname@46.115.19.28] has quit [] | 23:39 | |
--- Log closed Mon Apr 04 00:00:36 2011 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!