| --- Log opened Mon Aug 15 00:00:08 2011 | ||
| --- Day changed Mon Aug 15 2011 | ||
| @sonney2k | ahh and enough coverage of examples for new languages like csharp, lua, java, ruby, ... | 00:00 |
|---|---|---|
| @sonney2k | if we don't manage all $LANG we will include the lang nevertheless but not mention it in release notes :D | 00:00 |
| @sonney2k | uh oh last release was in december last year | 00:02 |
| @sonney2k | it is really about time to have a new release | 00:02 |
| @sonney2k | one should actually release every 1-2 months | 00:03 |
| @sonney2k | with not too big changes of course | 00:03 |
| alesis-novik | I'm making a few changes, I'll commit them a bit later | 00:07 |
| blackburn | whoaa | 00:07 |
| CIA-87 | shogun: Sergey Lisitsyn master * rcfe4efb / (13 files in 2 dirs): United Landmark/Classic MDS and Isomap stuff - http://bit.ly/qHb4FC | 00:07 |
| CIA-87 | shogun: Sergey Lisitsyn master * r5e6a160 / (22 files in 5 dirs): Fixed Classic/Landmark modular examples - http://bit.ly/o53qvh | 00:07 |
| blackburn | shogun-buildbot: go! :D | 00:08 |
| blackburn | sonney2k: didn't catch the commit? | 00:09 |
| @sonney2k | blackburn, I think it is still busy | 00:10 |
| blackburn | got | 00:10 |
| blackburn | hey almost all tests are OK | 00:11 |
| blackburn | hmm | 00:12 |
| blackburn | serialization of svmlight | 00:12 |
| blackburn | and heiko's stuff | 00:12 |
| blackburn | everything else is ok | 00:12 |
| @sonney2k | my poor machine rebooted again | 00:22 |
| -!- in3xes [~in3xes@180.149.49.227] has quit [Remote host closed the connection] | 00:46 | |
| @sonney2k | f-x, why don't use use get_vector and return float32_t too? | 00:48 |
| @sonney2k | for the online case | 00:48 |
| f-x | sonney2k: for which problem? the get_w() thing? | 00:49 |
| CIA-87 | shogun: Sergey Lisitsyn master * r953fa3a / (2 files): Got graphical dimreduct examples to work again - http://bit.ly/n8J3Kp | 00:57 |
| blackburn | one more parallel algo | 00:57 |
| CIA-87 | shogun: Sergey Lisitsyn master * r81cc6a2 / (2 files): Made LLE parallel - http://bit.ly/pamH0N | 00:57 |
| blackburn | sonney2k: can't get any speed burst :( | 01:00 |
| blackburn | 4s for LLE on 1000 examples with k=200 and 3s for 2-threaded with same params | 01:03 |
| blackburn | makes no sense :( | 01:03 |
| blackburn | 32s vs 24s for k=500 | 01:05 |
| blackburn | okay probably it can impact on ##@%@$% threaded system | 01:06 |
| CIA-87 | shogun: Sergey Lisitsyn master * rc02be2d / (2 files): Added pthread_exit for Isomap and MDS - http://bit.ly/n3Mkil | 01:14 |
| CIA-87 | shogun: Sergey Lisitsyn master * r4c5331c / (2 files): Removed unnecessary pthread_exit - http://bit.ly/nXtVJq | 01:56 |
| CIA-87 | shogun: Sergey Lisitsyn master * rb219442 / (2 files): Made LTSA parallel - http://bit.ly/q65kYb | 01:56 |
| CIA-87 | shogun: Sergey Lisitsyn master * r9f86e64 / (8 files): Added parameters addition to dimensionality reduction algos - http://bit.ly/pav0MP | 02:20 |
| -!- blackburn [~blackburn@188.122.238.50] has quit [Quit: Leaving.] | 02:29 | |
| -!- alesis-novik [~alesis@cnat038.wlan.net.ed.ac.uk] has quit [Ping timeout: 276 seconds] | 03:29 | |
| -!- alesis-novik [~alesis@cnat038.wlan.net.ed.ac.uk] has joined #shogun | 03:33 | |
| -!- in3xes [~in3xes@180.149.49.227] has joined #shogun | 04:21 | |
| -!- f-x [~user@117.192.192.146] has quit [Ping timeout: 260 seconds] | 04:24 | |
| -!- in3xes_ [~in3xes@180.149.49.227] has joined #shogun | 04:56 | |
| -!- in3xes [~in3xes@180.149.49.227] has quit [Ping timeout: 240 seconds] | 04:59 | |
| -!- in3xes_ is now known as in3xes | 05:00 | |
| -!- alesis-novik [~alesis@cnat038.wlan.net.ed.ac.uk] has quit [Quit: Leaving] | 05:25 | |
| -!- alesis-novik [~alesis@188.74.87.206] has joined #shogun | 06:13 | |
| -!- in3xes_ [~in3xes@180.149.49.227] has joined #shogun | 07:38 | |
| -!- in3xes [~in3xes@180.149.49.227] has quit [Ping timeout: 258 seconds] | 07:41 | |
| -!- in3xes__ [~in3xes@180.149.49.227] has joined #shogun | 07:49 | |
| -!- in3xes_ [~in3xes@180.149.49.227] has quit [Ping timeout: 240 seconds] | 07:52 | |
| -!- in3xes__ [~in3xes@180.149.49.227] has quit [Remote host closed the connection] | 07:56 | |
| -!- in3xes__ [~in3xes@180.149.49.227] has joined #shogun | 08:01 | |
| -!- in3xes [~in3xes@27.107.155.20] has joined #shogun | 08:04 | |
| -!- in3xes [~in3xes@27.107.155.20] has quit [Remote host closed the connection] | 08:05 | |
| -!- in3xes__ [~in3xes@180.149.49.227] has quit [Ping timeout: 240 seconds] | 08:06 | |
| -!- in3xes [~in3xes@27.107.155.20] has joined #shogun | 08:08 | |
| -!- in3xes [~in3xes@27.107.155.20] has quit [Ping timeout: 258 seconds] | 09:32 | |
| -!- in3xes [~in3xes@180.149.49.230] has joined #shogun | 09:53 | |
| -!- in3xes [~in3xes@180.149.49.230] has quit [Ping timeout: 258 seconds] | 10:04 | |
| CIA-87 | shogun: Alesis Novik master * r4560951 / (10 files in 4 dirs): Multiple fixes (SG*, log-likelihood instead of likelihood), improved documentation - http://bit.ly/ojI74s | 10:05 |
| CIA-87 | shogun: Soeren Sonnenburg master * r0a08046 / (10 files in 4 dirs): | 10:05 |
| CIA-87 | shogun: Merge pull request #296 from alesis/gmm | 10:05 |
| CIA-87 | shogun: Multiple fixes (SG*, log-likelihood instead of likelihood), improved docu - http://bit.ly/p9XMcf | 10:05 |
| -!- blackburn [~blackburn@188.122.238.50] has joined #shogun | 11:02 | |
| -!- in3xes [~in3xes@49.201.38.6] has joined #shogun | 11:12 | |
| -!- VojtechFranc [~quassel@2001:718:2:1634:221:6aff:fe58:151a] has joined #shogun | 11:21 | |
| VojtechFranc | Hi alesis-novik | 11:21 |
| blackburn | VojtechFranc: have a minute for lame questions about kPCA? | 11:22 |
| VojtechFranc | of course, if I can answer it | 11:23 |
| blackburn | in your stprtoolbox I have found kpca.m for hmm fitting | 11:24 |
| blackburn | and what to use for transforming data? | 11:24 |
| VojtechFranc | for hmm ?? | 11:24 |
| blackburn | hehe sorry | 11:25 |
| VojtechFranc | there is a function kernelproj(X, model) where model is obtained from kpca | 11:25 |
| blackburn | hmm is hmm not markov model | 11:25 |
| blackburn | kpcarec? | 11:25 |
| VojtechFranc | haha | 11:25 |
| blackburn | what is the purpose of kpcarec? | 11:25 |
| VojtechFranc | kpcarec it does preimage computation | 11:25 |
| VojtechFranc | i.e. given a point in the kernel space it finds the closes preimage in the input space | 11:26 |
| blackburn | ah | 11:26 |
| VojtechFranc | it is useful e.g. in image denoising | 11:26 |
| blackburn | I see | 11:26 |
| blackburn | so for transforming data | 11:26 |
| blackburn | it is sufficient to fit kPCA | 11:26 |
| blackburn | and then do kernelproj on its model? | 11:27 |
| VojtechFranc | exactly, there is also example | 11:27 |
| blackburn | I'm asking because I have to implement it in shogun | 11:27 |
| blackburn | soeren did init thing already | 11:27 |
| blackburn | but not transform | 11:27 |
| blackburn | hmm kernelproj is even .c | 11:28 |
| VojtechFranc | it is easy | 11:28 |
| VojtechFranc | yes | 11:28 |
| blackburn | oh, getting grasp of it | 11:28 |
| blackburn | many thanks | 11:29 |
| VojtechFranc | computing the preimage is however a bit more difficult | 11:29 |
| blackburn | I don't really need it now | 11:29 |
| VojtechFranc | then ok, but it is useful to generate nice figures ... :) | 11:29 |
| VojtechFranc | just a joke | 11:30 |
| blackburn | well of course but I simply can't put it into shogun preprocessor | 11:30 |
| blackburn | no inverse transform here :) | 11:30 |
| VojtechFranc | hmm, apparantly there is a large room for improvements in shogun :) | 11:31 |
| blackburn | yeah it was yesterday's topic here | 11:32 |
| alesis-novik | Hello VojtechFranc. I can't really chat right now, I was wondering if you would be available later today? I free up at 13:30 UTC, could we chat at say 14:00 UTC? | 11:41 |
| VojtechFranc | yes, no problem. 14:00 UTC is OK | 11:42 |
| -!- VojtechFranc [~quassel@2001:718:2:1634:221:6aff:fe58:151a] has quit [Remote host closed the connection] | 11:42 | |
| -!- alesis-novik [~alesis@188.74.87.206] has quit [Quit: Leaving] | 11:42 | |
| -!- VojtechFranc [~quassel@2001:718:2:1634:221:6aff:fe58:151a] has joined #shogun | 11:42 | |
| -!- in3xes_ [~in3xes@182.156.163.13] has joined #shogun | 11:48 | |
| -!- in3xes [~in3xes@49.201.38.6] has quit [Read error: Connection reset by peer] | 11:50 | |
| -!- in3xes_ [~in3xes@182.156.163.13] has quit [Read error: Connection reset by peer] | 12:39 | |
| @sonney2k | blackburn, sure preimage would fit - it would be just a function preimage(features) | 12:45 |
| @sonney2k | but VojtechFranc doesn't preimage only work for real valued data? | 12:45 |
| blackburn | huh last algo to parallelize | 12:46 |
| @sonney2k | blackburn, btw you need huge data set to see any gains | 12:50 |
| @sonney2k | it doesn't matter if sth trains just a few seconds | 12:50 |
| @sonney2k | if you have sth that takes an hour then it is useful to parallelize | 12:51 |
| blackburn | yeah but I like to feel it is scalable (at least a little) | 12:51 |
| -!- in3xes_ [~in3xes@180.149.49.230] has joined #shogun | 12:57 | |
| CIA-87 | shogun: Sergey Lisitsyn master * rcb38bbf / src/shogun/preprocessor/LocalTangentSpaceAlignment.cpp : Some clean for LTSA parallelization - http://bit.ly/pRyGR9 | 12:57 |
| CIA-87 | shogun: Sergey Lisitsyn master * r62e6aee / (2 files): Made HLLE parallel - http://bit.ly/okRUGq | 12:57 |
| blackburn | sonney2k: all parallel! | 12:58 |
| blackburn | except Laplacian Eigenmaps and MultidimensionalScaling - makes no sense | 13:01 |
| -!- in3xes__ [~in3xes@180.149.49.230] has joined #shogun | 13:14 | |
| -!- in3xes_ [~in3xes@180.149.49.230] has quit [Ping timeout: 276 seconds] | 13:18 | |
| VojtechFranc | sonney2k, it is even worse. you need to solve the pre-image problem for each kernel separately. there are feasible solutions for Gaussian kernel and polynomial 2 kernel (also implemented in my toolbox) | 13:20 |
| -!- in3xes__ is now known as in3xes | 13:25 | |
| -!- heiko [~heiko@134.91.55.189] has joined #shogun | 13:58 | |
| -!- heiko [~heiko@134.91.55.189] has quit [Ping timeout: 258 seconds] | 14:04 | |
| -!- heiko [~heiko@main.uni-duisburg.de] has joined #shogun | 14:19 | |
| -!- f-x [~user@117.192.192.146] has joined #shogun | 14:23 | |
| blackburn | HLLE for parkinsons dataset: http://dl.dropbox.com/u/10139213/shogun/parkinsons.pdf | 14:25 |
| -!- in3xes_ [~in3xes@180.149.49.230] has joined #shogun | 14:31 | |
| -!- in3xes [~in3xes@180.149.49.230] has quit [Ping timeout: 250 seconds] | 14:34 | |
| serialhex | shogun-buildbot status ruby_modular | 14:39 |
| serialhex | hmm.... it is not my slave :( | 14:39 |
| serialhex | shogun-buildbot status all | 14:40 |
| serialhex | nope, it dosnt listen to me... oh well :( | 14:41 |
| -!- blackburn [~blackburn@188.122.238.50] has quit [Quit: Leaving.] | 15:05 | |
| -!- in3xes__ [~in3xes@180.149.49.230] has joined #shogun | 15:11 | |
| -!- in3xes_ [~in3xes@180.149.49.230] has quit [Ping timeout: 258 seconds] | 15:14 | |
| -!- in3xes__ is now known as in3xes | 15:41 | |
| -!- alesis-novik [~alesis@cpat001.wlan.net.ed.ac.uk] has joined #shogun | 16:04 | |
| alesis-novik | Hello VojtechFranc | 16:04 |
| VojtechFranc | hi alesis, give me 10 minutes please | 16:06 |
| -!- in3xes [~in3xes@180.149.49.230] has quit [Ping timeout: 250 seconds] | 16:07 | |
| VojtechFranc | alesis-novik, ok I'm ready | 16:15 |
| VojtechFranc | what is the situation with the experiments and documentation? | 16:15 |
| alesis-novik | Well, I documented the code, not going into maths though | 16:15 |
| alesis-novik | I had some problems with the image benchmarks | 16:16 |
| alesis-novik | I think it's fixed now, but I need to rerun the tiger benchmark | 16:16 |
| alesis-novik | I also changed examples and clustering a bit | 16:16 |
| alesis-novik | so that it would return log-likelihood instead of likelihood | 16:16 |
| VojtechFranc | so now have completed the USPS and synthetic data benchmark? | 16:17 |
| alesis-novik | No, I have completed USPS and 2 of the images | 16:17 |
| VojtechFranc | what about the synthetic? | 16:17 |
| alesis-novik | I haven't run them yet. Like I said, it takes a lot of time, which I can't always dedicate on my machine for benchmarking | 16:18 |
| VojtechFranc | ok, have you included the benchmarking scripts to shogun? | 16:19 |
| alesis-novik | no, I'm not sure where I should put them either | 16:19 |
| -!- in3xes [~in3xes@182.156.163.54] has joined #shogun | 16:20 | |
| VojtechFranc | me neither, we can ask sonney2k. | 16:20 |
| VojtechFranc | sonney2k, alesis wrote some python scripts to benchmark the EM and SMEM algorithm. What is the proper place in shogun for these codes? | 16:21 |
| VojtechFranc | alesis-novik, I'd like to pass through all your codes onec again. Could you pls send me list of all files you modified/wrote? | 16:22 |
| alesis-novik | VojtechFranc, I'm also going to comment the examples and add them to examples/documented | 16:22 |
| VojtechFranc | it would be nice | 16:22 |
| alesis-novik | VojtechFranc, the ones to do with the project are distributions/Gaussian.*, clustering/GMM.* and a change in features/DotFeatures.* | 16:23 |
| VojtechFranc | and also the examples I guess | 16:24 |
| alesis-novik | currently undocumented/python_modular/clustering_gmm_modular.py | 16:25 |
| alesis-novik | undocumented/python_modular/graphical/em_1d_gmm.py | 16:26 |
| alesis-novik | undocumented/python_modular/graphical/em_2d_gmm.py | 16:26 |
| alesis-novik | undocumented/python_modular/graphical/smem_1d_gmm.py | 16:26 |
| alesis-novik | undocumented/python_modular/graphical/smem_2d_gmm.py | 16:26 |
| VojtechFranc | ok | 16:26 |
| VojtechFranc | I'm just looking at the benchmark results you sent me: benchmark_usps_em/smem.txt | 16:27 |
| VojtechFranc | could you pls give me an exmplanation of the values | 16:27 |
| VojtechFranc | ? | 16:27 |
| VojtechFranc | it seems tha for each run there are 3x the log-likelihoods and one value that might be std? | 16:28 |
| alesis-novik | I did in the letter, its average, std, max and min I believe | 16:28 |
| VojtechFranc | I see | 16:29 |
| VojtechFranc | I guess the log-likelihood values are not normalized to the number of examples? | 16:32 |
| alesis-novik | no, I didn't | 16:33 |
| VojtechFranc | it is interesting that both EM and SMEM achived the same maximal log-likelihood. this means that the SMEM is not very useful in this case... | 16:36 |
| alesis-novik | VojtechFranc, I believe it's because the initialization conditions are very good (k-means and 1-nn), so EM doesn't converge to local optima | 16:36 |
| VojtechFranc | probably | 16:39 |
| VojtechFranc | alesis-novik, what remains to be done? to finish documenting the examples and the benchmarks? | 16:40 |
| alesis-novik | VojtechFranc, I guess. I think any further work should be left for post-release/GSoC | 16:40 |
| alesis-novik | VojtechFranc, do you think I should add mathematical detail to documentation. So far I've only added the PDF formula for Gaussian | 16:41 |
| VojtechFranc | you mean the Oxygen documentation? | 16:42 |
| alesis-novik | yes | 16:42 |
| VojtechFranc | I don't think it is necessary. There might be a link to relevant papers. | 16:42 |
| VojtechFranc | I did git pull from your repository and try to compile. I got AL_KERNELCACHE -DUSE_HMMPARALLELSTRUCTURES -DUSE_HMMPARALLEL -DUSE_BIGSTATES -DUSE_HMMCACHE -DUSE_REFERENCE_COUNTING -DUSE_SVMLIGHT -DSWIG_TYPE_TABLE=shogun -DSHOGUN -DLINUX -DHAVE_POWL -DHAVE_SQRTL -DHAVE_LOG2 -DHAVE_XML -DHAVE_ATLAS -DHAVE_LAPACK -DUSE_GZIP -DHAVE_LARGEFILE -DUSE_SHORTREAL_KERNELCACHE -DUSE_HMMPARALLELSTRUCTURES -DUSE_HMMPARALLEL -DUSE_BIGSTATES - | 16:43 |
| VojtechFranc | DUSE_HMMCACHE -DUSE_REFERENCE_COUNTING -DUSE_SVMLIGHT -c -I. -I.. -I/usr/include/libxml2 -I../libshogun -I.. -o base/Version.cpp.o base/Version.cpp | 16:43 |
| VojtechFranc | base/Version.cpp:24: error: ‘VERSION_PARAMETER’ was not declared in this scope | 16:43 |
| VojtechFranc | make[1]: *** [base/Version.cpp.o] Error 1 | 16:43 |
| VojtechFranc | make[1]: Leaving directory `/home/vojta/shogun/alesis_git_branch/shogun/src/shogun' | 16:43 |
| VojtechFranc | alesi-novik: don't you have an idea what might be wrong? | 16:44 |
| alesis-novik | which branch did you use? | 16:44 |
| alesis-novik | The updated code is in the gmm branch, though you might as well use the shogun main branch, since I've pushed everything | 16:45 |
| VojtechFranc | I guess I'm using git://github.com/alesis/shogun | 16:47 |
| alesis-novik | yeah, it's a very old version there | 16:48 |
| VojtechFranc | maybe I'm in a wrong dir I have several branches downloded... | 16:49 |
| VojtechFranc | then I'll try the main shogun branch | 16:50 |
| VojtechFranc | ok, I figured out that I was using the gmm branch at the end | 16:54 |
| alesis-novik | Is it working now? | 17:03 |
| VojtechFranc | yes it compiles. I tried the main shogun branch. | 17:09 |
| VojtechFranc | however, it seems you didn't push the example scripts. did you? | 17:10 |
| alesis-novik | the examples should be there | 17:14 |
| VojtechFranc | ok, I need to go. I'll try it tomorrow ... cheers | 17:16 |
| alesis-novik | Ok, tell me if anything goes wrong | 17:17 |
| VojtechFranc | ok, cheers | 17:19 |
| -!- VojtechFranc [~quassel@2001:718:2:1634:221:6aff:fe58:151a] has quit [Remote host closed the connection] | 17:19 | |
| -!- in3xes [~in3xes@182.156.163.54] has quit [Remote host closed the connection] | 17:22 | |
| -!- f-x` [~user@117.192.213.170] has joined #shogun | 17:34 | |
| -!- f-x [~user@117.192.192.146] has quit [Ping timeout: 260 seconds] | 17:36 | |
| -!- heiko [~heiko@main.uni-duisburg.de] has quit [Ping timeout: 258 seconds] | 18:03 | |
| -!- in3xes [~in3xes@180.149.49.227] has joined #shogun | 18:32 | |
| -!- f-x`` [~user@117.192.213.170] has joined #shogun | 19:05 | |
| -!- f-x`` [~user@117.192.213.170] has left #shogun [] | 19:05 | |
| -!- f-x [~user@117.192.213.170] has joined #shogun | 19:06 | |
| -!- f-x` [~user@117.192.213.170] has quit [Ping timeout: 260 seconds] | 19:06 | |
| -!- in3xes_ [~in3xes@180.149.49.227] has joined #shogun | 19:24 | |
| -!- in3xes [~in3xes@180.149.49.227] has quit [Ping timeout: 258 seconds] | 19:27 | |
| -!- blackburn [~blackburn@188.122.238.50] has joined #shogun | 19:32 | |
| -!- f-x [~user@117.192.213.170] has quit [Ping timeout: 260 seconds] | 19:38 | |
| -!- in3xes_ is now known as in3xes | 19:38 | |
| -!- in3xes_ [~in3xes@180.149.49.227] has joined #shogun | 19:48 | |
| -!- in3xes [~in3xes@180.149.49.227] has quit [Ping timeout: 240 seconds] | 19:51 | |
| -!- in3xes_ is now known as in3xes | 19:51 | |
| -!- alesis-novik [~alesis@cpat001.wlan.net.ed.ac.uk] has quit [Quit: Leaving] | 19:57 | |
| -!- in3xes_ [~in3xes@180.149.49.227] has joined #shogun | 20:26 | |
| blackburn | today is a soft pencils down? | 20:27 |
| -!- in3xes [~in3xes@180.149.49.227] has quit [Ping timeout: 240 seconds] | 20:29 | |
| -!- in3xes_ is now known as in3xes | 20:29 | |
| blackburn | serialhex: around? | 20:39 |
| CIA-87 | shogun: Shashwat Lal Das master * r692f79f / (24 files in 5 dirs): Partial transition to float32_t for online framework. - http://bit.ly/r1T3ut | 20:44 |
| CIA-87 | shogun: Shashwat Lal Das master * r2976834 / (25 files in 6 dirs): Working float64_t to float32_t transition for online framework. - http://bit.ly/pw2iHp | 20:44 |
| CIA-87 | shogun: Shashwat Lal Das master * re034620 / examples/undocumented/libshogun/streaming_from_simple.cpp : Remove a doc typo. - http://bit.ly/qGVHyz | 20:44 |
| CIA-87 | shogun: Soeren Sonnenburg master * r2f343fb / (13 files in 7 dirs): | 20:44 |
| CIA-87 | shogun: Merge pull request #295 from frx/conv_to_32 | 20:44 |
| CIA-87 | shogun: Conversion of weight vector and related parameters to float in Online Framework - http://bit.ly/nAp8wr | 20:44 |
| shogun-buildbot | build #28 of libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org:5347/builders/libshogun/builds/28 blamelist: fraggerx@gmail.com | 20:46 |
| @sonney2k | grmpf | 20:48 |
| shogun-buildbot | build #29 of libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org:5347/builders/libshogun/builds/29 | 20:48 |
| blackburn | grmpf is something bad | 20:48 |
| blackburn | :D | 20:48 |
| @sonney2k | why does the buildbot sometimes checkout stone age old versions and then fail? | 20:49 |
| @sonney2k | and then directly afterwards use the correct version? | 20:49 |
| blackburn | no idea | 20:49 |
| blackburn | sonney2k: what do you think, should I add reconstruction error option for MDS/LLE/..? | 20:51 |
| @sonney2k | blackburn, this is nice to have but nothing essential - so yes but no high prio and of course up to you :) | 20:52 |
| blackburn | I think about integrating heiko's thing and my algos.. | 20:52 |
| @sonney2k | blackburn, ok then please don't do that yet | 20:53 |
| @sonney2k | this is post 1.0 stuff | 20:53 |
| blackburn | it is very inconvenient to get nice k value with only tries | 20:53 |
| blackburn | but no automagic stuff for that | 20:53 |
| -!- in3xes [~in3xes@180.149.49.227] has quit [Ping timeout: 240 seconds] | 20:54 | |
| @sonney2k | blackburn, I understand that but recall we have soft pencils down today | 20:54 |
| @sonney2k | and 1 more week of gsoc | 20:54 |
| @sonney2k | prio should be fixing bugs / documentation | 20:54 |
| blackburn | I'll finish doc stuff for my algos today | 20:54 |
| @sonney2k | blackburn, that other stuff is something for fall 2011 :) | 20:55 |
| blackburn | yeah sure | 20:56 |
| @sonney2k | blackburn, but it is certainly possible to connect CMethod CPreprocessors | 20:56 |
| @sonney2k | and as soon as you do this you will have 'heiko' support :D | 20:57 |
| blackburn | CMethod? | 20:57 |
| @sonney2k | CMachine | 20:57 |
| blackburn | I should think about it | 20:58 |
| -!- in3xes [~in3xes@182.156.163.54] has joined #shogun | 20:58 | |
| @sonney2k | blackburn, btw the buildbot was right | 20:58 |
| blackburn | about? | 20:58 |
| @sonney2k | about the failure | 20:58 |
| @sonney2k | build failure | 20:58 |
| @sonney2k | it was trying to compile an older version from f-x | 20:58 |
| blackburn | nice | 20:59 |
| @sonney2k | and that one was really broken | 20:59 |
| @sonney2k | now we only need stable build-slaves | 20:59 |
| * blackburn have already spent 5 minutes to come up with sentence for doc | 20:59 | |
| blackburn | it could be harder than I thought haha | 21:00 |
| @sonney2k | I think I will utilize the google money to get a more powerful server/hosting for shogun-toolbox.org | 21:00 |
| blackburn | so will you get it? | 21:00 |
| @sonney2k | blackburn, just write - it doesn't have to be perfect in the first iteration. recall that serialhex has your back | 21:00 |
| @sonney2k | blackburn, I don't know | 21:00 |
| blackburn | it is not fair TU get it | 21:00 |
| @sonney2k | in the worst case I pay these 120 EUR / year myself | 21:01 |
| blackburn | I could donate a part | 21:01 |
| blackburn | not fair again | 21:02 |
| blackburn | sonney2k: I've read that instead of "class CDistance;" stuff here we could write "using class CDistance;" | 21:05 |
| @sonney2k | blackburn, I hope the project gets the money at some stage. I mean if we can afford a more powerful server to do all trac/git/ etc hosting | 21:05 |
| @sonney2k | that would be perfectly fine | 21:06 |
| blackburn | oh no | 21:06 |
| @sonney2k | I hope we can convince some people to be build slaves | 21:06 |
| blackburn | we can't | 21:06 |
| blackburn | :D | 21:06 |
| blackburn | I can but I have notebook - will not really help | 21:06 |
| blackburn | sonney2k: so what's up with our projects? | 21:09 |
| @sonney2k | blackburn, which porjects? | 21:09 |
| blackburn | I finished | 21:09 |
| blackburn | alesis? | 21:09 |
| @sonney2k | vojtech seemed happy | 21:09 |
| blackburn | f-x? | 21:09 |
| @sonney2k | sploving did all the $lang | 21:09 |
| @sonney2k | so also good | 21:09 |
| blackburn | so it is all ok? | 21:10 |
| @sonney2k | f-x did much more than I thought already before midterm | 21:10 |
| @sonney2k | the framework was a lot of work | 21:10 |
| @sonney2k | so nice that he even started to port over vw | 21:10 |
| @sonney2k | and converted SGD etc | 21:10 |
| @sonney2k | heiko also did a great job | 21:11 |
| @sonney2k | to conclude - 100% success. however we still need some polishing to do a release | 21:11 |
| blackburn | yeah, f-x and heiko did a lot. I don't really think I did as much as they:) | 21:12 |
| blackburn | sonney2k: description is given in / description is given at? | 21:14 |
| @sonney2k | blackburn, ? | 21:15 |
| blackburn | need english grammar advice hah | 21:15 |
| @sonney2k | for examples? | 21:15 |
| @sonney2k | argh | 21:21 |
| @sonney2k | nightly_all still doesn't compile | 21:21 |
| @sonney2k | http://www.shogun-toolbox.org:5347/builders/nightly_all/builds/3/steps/compile/logs/stdio | 21:21 |
| @sonney2k | but that very fast :D | 21:22 |
| blackburn | ?? | 21:22 |
| blackburn | why? | 21:22 |
| @sonney2k | blackburn, compiling was fast because it failed - I was just joking :) | 21:30 |
| blackburn | why it fails? | 21:30 |
| @sonney2k | CIA-87, tell us please | 21:31 |
| @sonney2k | hmmhh why is CIA-87 quiet? | 21:31 |
| @sonney2k | 9ca73170fb0f87974a3fcd3342136130d6b16544 fix build failure in ui trying to access protected member variable | 21:31 |
| @sonney2k | (just pushed) | 21:32 |
| CIA-87 | shogun: Soeren Sonnenburg master * r9ca7317 / (src/shogun/mathematics/Math.h src/shogun/ui/SGInterface.cpp): fix build failure in ui trying to access protected member variable - http://bit.ly/pUtVdo | 21:33 |
| @sonney2k | blackburn, ^ there he is | 21:37 |
| blackburn | I see | 21:37 |
| -!- sonney2k is now known as sonney2k|bughunt | 21:39 | |
| blackburn | haha | 21:39 |
| -!- sonney2k|bughunt is now known as sonney2k | 21:39 | |
| blackburn | wtf was it | 21:39 |
| @sonney2k | I am fixing one bug per day for this nightly_all | 21:39 |
| @sonney2k | lets see if it will compile at some point or not | 21:40 |
| @sonney2k | blackburn, look we only have 853 warnings now ;D | 21:40 |
| blackburn | yeah I will spend some time today to fix some of them | 21:41 |
| shogun-buildbot | build #3 of nightly_none is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org:5347/builders/nightly_none/builds/3 | 21:42 |
| -!- f-x [~user@117.192.199.77] has joined #shogun | 21:48 | |
| -!- in3xes [~in3xes@182.156.163.54] has quit [Ping timeout: 260 seconds] | 22:03 | |
| shogun-buildbot | build #30 of libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org:5347/builders/libshogun/builds/30 blamelist: sonne@debian.org | 22:03 |
| blackburn | sonney2k: github down? | 22:04 |
| blackburn | I can't pull from it | 22:04 |
| blackburn | hm all ok | 22:06 |
| blackburn | false alarm | 22:06 |
| -!- in3xes [~in3xes@49.201.194.3] has joined #shogun | 22:16 | |
| blackburn | ehh | 22:18 |
| blackburn | CIA-87: what's up with you? | 22:18 |
| shogun-buildbot | build #32 of libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org:5347/builders/libshogun/builds/32 | 22:22 |
| -!- in3xes [~in3xes@49.201.194.3] has quit [Quit: Leaving] | 22:28 | |
| CIA-87 | shogun: Sergey Lisitsyn master * r9d2dd54 / (8 files): Updated doc for dimensionality reduction algorithms - http://bit.ly/pNjvMO | 22:28 |
| blackburn | sonney2k: I like things a little more today :D | 22:33 |
| shogun-buildbot | build #26 of python_modular is complete: Failure [failed git] Build details are at http://www.shogun-toolbox.org:5347/builders/python_modular/builds/26 blamelist: blackburn91@gmail.com | 22:39 |
| CIA-87 | shogun: Sergey Lisitsyn master * r4826d03 / (9 files in 5 dirs): Moar doc - http://bit.ly/n4ZfJn | 22:41 |
| @sonney2k | blackburn, what is wrong with JL's covertree? | 23:38 |
| blackburn | sonney2k: nothing but later | 23:39 |
| blackburn | I will restore it when we have shogun 1.* milestone | 23:39 |
| @sonney2k | blackburn, yeah but then just create a new milestone and assign it there | 23:39 |
| @sonney2k | or mark it unassigned | 23:39 |
| blackburn | okay | 23:40 |
| blackburn | sonney2k: I sent an email to heiko about reported bug | 23:40 |
| @sonney2k | blackburn, do you think it would be worth to work on a pthread based parallelization framework? | 23:40 |
| @sonney2k | thanks | 23:40 |
| @sonney2k | I mean for shogun | 23:41 |
| blackburn | can you mind any framework for it? | 23:41 |
| @sonney2k | blackburn, yeah it is probably not so easy | 23:41 |
| shogun-buildbot | build #27 of python_modular is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org:5347/builders/python_modular/builds/27 | 23:41 |
| @sonney2k | there usually is very custom 'user' data | 23:41 |
| blackburn | whoaa! | 23:42 |
| * sonney2k checks the warnings count | 23:42 | |
| blackburn | 819 | 23:42 |
| @sonney2k | yeah | 23:42 |
| @sonney2k | going down baby | 23:42 |
| blackburn | okay I'll make it <800 now for feeling more happy | 23:43 |
| CIA-87 | shogun: Shashwat Lal Das master * rf2cce5f / (28 files in 7 dirs): Fixed most online framework warnings. - http://bit.ly/p4A7p6 | 23:57 |
| CIA-87 | shogun: Shashwat Lal Das master * r693acf9 / doc/Doxyfile : Reverted Doxyfile. - http://bit.ly/qCJkoR | 23:57 |
| CIA-87 | shogun: Soeren Sonnenburg master * rd05ad85 / (27 files in 6 dirs): | 23:57 |
| CIA-87 | shogun: Merge pull request #297 from frx/streaming_vw | 23:57 |
| CIA-87 | shogun: Missing/incorrect doc additions and corrections, plus load_regressor() for VW - http://bit.ly/oHbTn7 | 23:57 |
| @sonney2k | blackburn, maybe f-x just did | 23:57 |
| @sonney2k | we will see :) | 23:58 |
| blackburn | I'm about to remove some more | 23:58 |
| --- Log closed Tue Aug 16 00:00:38 2011 | ||
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!