--- Log opened Mon Dec 19 00:00:19 2011 | ||
15SAAI18M | shogun: Soeren Sonnenburg master * rf4dcb8a / .gitignore : ignore Converter.i - http://git.io/6H18Tw | 00:23 |
---|---|---|
15SAAI18M | shogun: Soeren Sonnenburg master * r7644ddc / (src/shogun/io/File.cpp src/shogun/io/File.h): add function to read whole file into buffer - http://git.io/lbW7vg | 00:24 |
blackburn | sonney2k: what did you implement? | 00:33 |
blackburn | a lot of changes, can't catch idea | 00:33 |
@sonney2k | serialization support for swig -builtin | 00:33 |
blackburn | python only? | 00:34 |
@sonney2k | yes | 00:35 |
blackburn | sonney2k: I see. why did you focused on it? | 00:35 |
@sonney2k | I don't understand | 00:35 |
@sonney2k | we need serialization support | 00:36 |
blackburn | wasn't it working before? | 00:36 |
@sonney2k | so I needed to fix this for the new -builtin | 00:36 |
@sonney2k | no | 00:36 |
blackburn | no = wasn't? | 00:36 |
@sonney2k | we had it in the old code w/o -builtin | 00:36 |
@sonney2k | but that was python | 00:36 |
blackburn | what is the difference? | 00:36 |
@sonney2k | converting to -builtin made it no longer work | 00:36 |
@sonney2k | so I fixed that | 00:36 |
blackburn | oh okay | 00:36 |
@sonney2k | still no buffer protocol support though... too tough... | 00:37 |
@sonney2k | have to leave the train now | 00:37 |
@sonney2k | cu | 00:37 |
blackburn | train? now? | 00:38 |
blackburn | :D | 00:38 |
@sonney2k | Re | 01:22 |
blackburn | sonney2k: strange time to be in train :) | 01:23 |
@sonney2k | just came back from granada... | 01:23 |
blackburn | hmm thought you came on friday | 01:24 |
blackburn | sonney2k: can GMNP SVM be considered as state of art still? | 01:25 |
@sonney2k | in which respect? | 01:25 |
@sonney2k | it is just a multiclass svm... | 01:26 |
blackburn | I understand, but I wonder if there is any better formulation (in any means) | 01:26 |
@sonney2k | one-vs-rest :) | 01:27 |
blackburn | sonney2k: one-vs-rest is worse on my dataset in means of accuracy | 01:27 |
blackburn | and performance is awful | 01:27 |
blackburn | how can you treat it as better? | 01:27 |
@sonney2k | usually a lot faster and same perf | 01:28 |
blackburn | sonney2k: hmm 903 svms are not so fast :) | 01:28 |
@sonney2k | YMMV | 01:29 |
blackburn | sure | 01:29 |
blackburn | I haz no opulence :( | 01:31 |
blackburn | :D | 01:31 |
blackburn | sonney2k: have you any experience of random forests? | 01:32 |
blackburn | I would implement it for shogun some day I guess | 01:32 |
blackburn | btw we already haZZ tasks for at least 3 gsoc students | 01:33 |
@sonney2k | no never did that | 01:36 |
@sonney2k | I have to sleep now... | 01:36 |
@sonney2k | but yes - lots of nice things :) | 01:37 |
@sonney2k | btw you should not call your toolbox largescale | 01:37 |
@sonney2k | people associate with largescale petabytes of data | 01:37 |
blackburn | sonney2k: middlescale? | 01:38 |
blackburn | :D | 01:38 |
@sonney2k | so maybe just Dimension Reduction Toolbox and mention that it is fast | 01:38 |
blackburn | sonney2k: at least our thing doesn't crash on 40000d datasets | 01:38 |
blackburn | hmm okay | 01:38 |
blackburn | sonney2k: how do you think can it be called just | 01:38 |
@sonney2k | but it will crash on 400000 examples already | 01:39 |
blackburn | ah nevermind | 01:39 |
blackburn | sonney2k: it is not possible to embed 400000 vectors theoretically :) | 01:39 |
@sonney2k | it is :) | 01:40 |
@sonney2k | anyway bed time! | 01:40 |
@sonney2k | cu | 01:40 |
blackburn | sonney2k: see you | 01:40 |
blackburn | btw shogun failed to compile on my machine | 01:40 |
blackburn | but I need to sleep as well | 01:40 |
blackburn | /home/blackburn/shogun/shogun/src/shogun/regression/svr/LibSVR.cpp:109: undefined reference to `shogun::CKernelMachine::create_new_model(int)' | 01:40 |
blackburn | /home/blackburn/shogun/shogun/src/shogun/regression/svr/LibSVR.cpp:109: undefined reference to `shogun::CKernelMachine::create_new_model(int)' | 01:40 |
blackburn | /home/blackburn/shogun/shogun/src/shogun/regression/svr/LibSVR.cpp:109: undefined reference to `shogun::CKernelMachine::create_new_model(int)' | 01:40 |
blackburn | I said it is not possible cause it is not possible to solve 400000x400000 eigenproblem :) | 01:43 |
blackburn | especially dense | 01:43 |
blackburn | on typical machine | 01:44 |
-!- blackburn [~blackburn@31.28.51.215] has quit [Quit: Leaving.] | 02:02 | |
-!- ishaanmlhtr [~ishaan@115.240.122.177] has joined #shogun | 05:50 | |
-!- ishaanmlhtr [~ishaan@115.240.122.177] has quit [Quit: Leaving] | 06:24 | |
-!- Ram108 [~amma@14.99.110.249] has joined #shogun | 07:17 | |
-!- axitkhurana [3ba16b46@gateway/web/freenode/ip.59.161.107.70] has joined #shogun | 09:27 | |
sonne|work | ahh great you have that problem now too | 09:35 |
sonne|work | please resolve it | 09:35 |
sonne|work | I cannot trigger that... | 09:35 |
-!- sonne|work [~sonnenbu@194.78.35.195] has quit [Quit: Leaving.] | 10:16 | |
-!- sonne|work [~sonnenbu@194.78.35.195] has joined #shogun | 10:21 | |
15SAAI18M | shogun: Soeren Sonnenburg master * re787f2d / src/shogun/machine/KernelMachine.cpp : remove 'inline' declaration from .cpp file - http://git.io/FiUeww | 10:56 |
15SAAI18M | shogun: Soeren Sonnenburg master * rdaf4856 / (2 files in 2 dirs): fix warnings about nk not being used - http://git.io/M5z9nQ | 10:56 |
shogun-buildbot | build #444 of libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/libshogun/builds/444 | 10:59 |
shogun-buildbot | build #388 of r_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/r_static/builds/388 | 11:04 |
shogun-buildbot | build #392 of octave_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/octave_static/builds/392 | 11:08 |
sonne|work | blackburn, btw what was the issue with java and memory leaks? | 11:11 |
shogun-buildbot | build #380 of python_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/python_static/builds/380 | 11:12 |
-!- ishaanmlhtr [~chatzilla@115.241.182.169] has joined #shogun | 11:17 | |
-!- ishaanmlhtr [~chatzilla@115.241.182.169] has quit [Client Quit] | 11:19 | |
-!- axitkhurana [3ba16b46@gateway/web/freenode/ip.59.161.107.70] has left #shogun [] | 11:47 | |
-!- axitkhurana [~akshit@59.161.107.70] has joined #shogun | 11:47 | |
-!- Ram108 [~amma@14.99.110.249] has quit [Ping timeout: 252 seconds] | 12:22 | |
-!- axitkhurana [~akshit@59.161.107.70] has left #shogun [] | 13:46 | |
-!- axitkhurana [~akshit@59.161.107.70] has joined #shogun | 13:46 | |
-!- sonne|work1 [~sonnenbu@194.78.35.195] has joined #shogun | 13:46 | |
-!- sonne|work [~sonnenbu@194.78.35.195] has quit [Ping timeout: 244 seconds] | 13:47 | |
-!- 15SAAI18M [~CIA@cia.atheme.org] has quit [Ping timeout: 244 seconds] | 13:47 | |
-!- sonne|work1 is now known as sonne-work | 13:49 | |
-!- CIA-1 [~CIA@cia.atheme.org] has joined #shogun | 13:52 | |
sonne-work | welcome back CIA-1 | 13:56 |
-!- Netsplit *.net <-> *.split quits: CIA-1 | 13:57 | |
-!- Netsplit over, joins: CIA-1 | 14:00 | |
-!- axitkhurana1 [~akshit@14.98.200.133] has joined #shogun | 14:09 | |
-!- axitkhurana [~akshit@59.161.107.70] has quit [Ping timeout: 255 seconds] | 14:11 | |
-!- ishaanmlhtr [~chatzilla@115.241.216.146] has joined #shogun | 14:15 | |
-!- axitkhurana1 [~akshit@14.98.200.133] has quit [Ping timeout: 252 seconds] | 15:21 | |
-!- ishaanmlhtr [~chatzilla@115.241.216.146] has quit [Quit: ChatZilla 0.9.87 [Firefox 3.6.24/20111103063747]] | 16:21 | |
-!- Ram108 [~amma@14.96.173.148] has joined #shogun | 16:41 | |
-!- puneetgoyal [~puneetgoy@117.197.167.250] has joined #shogun | 18:13 | |
-!- Ram108 [~amma@14.96.173.148] has quit [Ping timeout: 252 seconds] | 18:47 | |
-!- ishaanmlhtr [~chatzilla@115.241.157.83] has joined #shogun | 18:56 | |
-!- blackburn [~blackburn@188.168.4.230] has joined #shogun | 19:15 | |
-!- ishaanmlhtr [~chatzilla@115.241.157.83] has quit [Quit: ChatZilla 0.9.87 [Firefox 3.6.24/20111103063747]] | 19:22 | |
-!- blackburn1 [~blackburn@188.168.4.189] has joined #shogun | 19:23 | |
-!- blackburn [~blackburn@188.168.4.230] has quit [Ping timeout: 252 seconds] | 19:23 | |
-!- ishaanmlhtr [~ishaan@115.241.201.123] has joined #shogun | 19:27 | |
ishaanmlhtr | blackburn1:2055376946 ,226418,1634299437 | 19:28 |
blackburn1 | ishaanmlhtr: strange numbers. what is it? | 19:28 |
ishaanmlhtr | number , rows ,cols | 19:29 |
blackburn1 | ehm | 19:29 |
blackburn1 | check if we have used same files | 19:29 |
ishaanmlhtr | ya,checked..train-images.idx3-ubyte - this is what i use.. | 19:29 |
ishaanmlhtr | and i think even you used the same | 19:30 |
blackburn1 | strange.. | 19:30 |
ishaanmlhtr | why did you use '>L' in your unpack format? | 19:36 |
blackburn1 | ishaanmlhtr: it is a format of that file, I guess it is 64/32 issue | 21:10 |
-!- ishaanmlhtr [~ishaan@115.241.201.123] has quit [Read error: Operation timed out] | 21:45 | |
-!- blackburn1 [~blackburn@188.168.4.189] has quit [Ping timeout: 252 seconds] | 21:46 | |
-!- blackburn [~blackburn@188.168.4.250] has joined #shogun | 21:47 | |
-!- puneetgoyal [~puneetgoy@117.197.167.250] has quit [Ping timeout: 240 seconds] | 23:13 | |
-!- blackburn [~blackburn@188.168.4.250] has quit [Quit: Leaving.] | 23:28 | |
--- Log closed Tue Dec 20 00:00:19 2011 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!