--- Log opened Wed Apr 24 00:00:28 2013 | ||
@iglesiasg | wiking: around? | 00:13 |
---|---|---|
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 00:25 | |
-!- van511 [~van51@athedsl-319014.home.otenet.gr] has joined #shogun | 00:50 | |
-!- van51 [~van51@athedsl-407661.home.otenet.gr] has quit [Ping timeout: 276 seconds] | 00:51 | |
-!- van511 [~van51@athedsl-319014.home.otenet.gr] has left #shogun [] | 00:59 | |
-!- van511 [~van51@athedsl-319014.home.otenet.gr] has joined #shogun | 01:00 | |
-!- kevin [820f1060@gateway/web/freenode/ip.130.15.16.96] has joined #shogun | 01:23 | |
-!- kevin is now known as Guest87738 | 01:24 | |
-!- Guest87738 is now known as pickle27 | 01:24 | |
-!- van511 [~van51@athedsl-319014.home.otenet.gr] has left #shogun [] | 01:25 | |
-!- vikram360 [~vikram360@117.221.124.81] has quit [Ping timeout: 240 seconds] | 02:53 | |
-!- iglesiasg [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 03:06 | |
-!- sids_aquarius [~sidi_@li400-124.members.linode.com] has quit [Remote host closed the connection] | 03:21 | |
-!- sids_aquarius [~sidi_@li400-124.members.linode.com] has joined #shogun | 03:22 | |
sijin | @lisitsyn I see your submit, thx. | 03:55 |
shogun-buildbot | build #371 of nightly_default is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/371 | 04:26 |
-!- pickle27 [820f1060@gateway/web/freenode/ip.130.15.16.96] has quit [Quit: Page closed] | 04:36 | |
-!- votjakovr [~chatzilla@host-46-241-3-209.bbcustomer.zsttk.net] has joined #shogun | 05:13 | |
-!- rgirish28 [~rgirish28@61.17.226.66] has quit [Ping timeout: 260 seconds] | 05:48 | |
-!- saket [0e8b5206@gateway/web/freenode/ip.14.139.82.6] has joined #shogun | 06:14 | |
-!- rgirish28 [~rgirish28@61.17.21.199] has joined #shogun | 06:17 | |
-!- rgirish28 [~rgirish28@61.17.21.199] has quit [Ping timeout: 255 seconds] | 06:31 | |
-!- rgirish28 [~rgirish28@111.92.71.249] has joined #shogun | 06:36 | |
-!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has joined #shogun | 06:44 | |
-!- votjakovr [~chatzilla@host-46-241-3-209.bbcustomer.zsttk.net] has quit [Quit: ChatZilla 0.9.90 [Firefox 20.0.1/20130409194949]] | 07:26 | |
-!- rgirish28 [~rgirish28@111.92.71.249] has quit [Read error: Connection reset by peer] | 08:11 | |
-!- rgirish28_ [~rgirish28@111.92.71.249] has joined #shogun | 08:11 | |
-!- saket [0e8b5206@gateway/web/freenode/ip.14.139.82.6] has quit [Quit: Page closed] | 08:16 | |
-!- jiuding [dbe09f12@gateway/web/freenode/ip.219.224.159.18] has joined #shogun | 08:36 | |
-!- ishan [~ishan@14.139.122.114] has joined #shogun | 08:58 | |
-!- ishan [~ishan@14.139.122.114] has quit [Ping timeout: 268 seconds] | 09:09 | |
-!- vgorbati [4d2f8429@gateway/web/freenode/ip.77.47.132.41] has joined #shogun | 09:19 | |
-!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has quit [Ping timeout: 245 seconds] | 09:33 | |
-!- van51 [~van51@athedsl-319014.home.otenet.gr] has joined #shogun | 09:36 | |
-!- van51 [~van51@athedsl-319014.home.otenet.gr] has quit [Client Quit] | 09:37 | |
-!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has joined #shogun | 09:38 | |
-!- van51 [~van51@athedsl-319014.home.otenet.gr] has joined #shogun | 09:38 | |
vgorbati | lisitsyn: hello, are you here? | 09:48 |
@lisitsyn | vgorbati: yes but will leave in a minute | 09:48 |
vgorbati | lisitsyn: when do you return, if you do? | 09:49 |
@lisitsyn | vgorbati: ~20 minutes or 4 bus stops | 09:49 |
@lisitsyn | :D | 09:49 |
vgorbati | lisitsyn: I wlll catch you then, if you don't mind | 09:50 |
@lisitsyn | yeah sure | 09:50 |
-!- rgirish28_ [~rgirish28@111.92.71.249] has quit [Ping timeout: 255 seconds] | 09:51 | |
-!- lisitsyn [~blackburn@46.20.65.94] has quit [Quit: Leaving.] | 09:54 | |
-!- vgorbati [4d2f8429@gateway/web/freenode/ip.77.47.132.41] has quit [Ping timeout: 245 seconds] | 09:58 | |
-!- vgorbati [2535754b@gateway/web/freenode/ip.37.53.117.75] has joined #shogun | 10:14 | |
-!- lisitsyn [~lisitsin@mxs.kg.ru] has joined #shogun | 10:27 | |
-!- mode/#shogun [+o lisitsyn] by ChanServ | 10:27 | |
@lisitsyn | vgorbati: re | 10:27 |
vgorbati | lisitsyn: well, I tried to think of good data to test LLE, and one of tries was a straight 3D line. The results were kind a weird - http://www.youtube.com/watch?v=ZAMqMLsTVBY (as usual, initial - left, embedded - right). | 10:29 |
vgorbati | not that link | 10:29 |
vgorbati | wait | 10:29 |
vgorbati | lisitsyn: http://www.youtube.com/watch?v=KES-ZlvSFAw | 10:29 |
@lisitsyn | haha the first link is kinda strange! :D | 10:30 |
vgorbati | lisitsyn: yeah, I just randomly clicked at first suggestion when was looking my vid) | 10:31 |
vgorbati | lisitsyn: what do you think about the second link?) | 10:31 |
@lisitsyn | vgorbati: well it doesn't look wrong for me | 10:32 |
vgorbati | lisitsyn: could you please explain? | 10:32 |
@lisitsyn | and actually well with k=2 distances are preserved | 10:32 |
@lisitsyn | aren't they? | 10:32 |
@lisitsyn | err | 10:32 |
@lisitsyn | not distances | 10:32 |
@lisitsyn | neighborhoods | 10:32 |
vgorbati | lisitsyn: not totally | 10:33 |
vgorbati | lisitsyn: minimal similarity is 16/20 | 10:33 |
vgorbati | lisitsyn: I though dimensionality reduction should capture some dependency in data, and it does not look like the dependency was captured in that case | 10:34 |
@lisitsyn | vgorbati: PCA would put your line to one axis | 10:35 |
@lisitsyn | vgorbati: well LLE captures it | 10:35 |
@lisitsyn | I wouldn't say it lost the dependency | 10:35 |
-!- HeikoS [~heiko@nat-180-153.internal.eduroam.ucl.ac.uk] has joined #shogun | 10:36 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 10:36 | |
@HeikoS | lisitsyn: MultidimensionaScalingTest.distance_preserving fails | 10:36 |
@lisitsyn | HeikoS: yeah I have seen | 10:36 |
@lisitsyn | sorry forgot to check that | 10:36 |
@lisitsyn | will fix tonight | 10:36 |
vgorbati | lisitsyn: about 'with k=2 neighbors are preserved' - but the k was set to 20.. | 10:36 |
-!- iglesiasg [~test@n153-p56.kthopen.kth.se] has joined #shogun | 10:36 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 10:36 | |
@lisitsyn | vgorbati: yes I mean it doesn't break neighborhoods that much | 10:37 |
-!- foulwall [~foulwall@li379-21.members.linode.com] has joined #shogun | 10:38 | |
vgorbati | lisitsyn: but what then should I test?:) I can try to test that when k==20 that at least 2-3 neighbors are preserved) | 10:38 |
-!- votjakovr [~chatzilla@host-46-241-3-209.bbcustomer.zsttk.net] has joined #shogun | 10:42 | |
vgorbati | lisitsyn: ok, I just tested for k==5 on the line with small noise added, it preserves pretty much all the neighbors (sometimes 1 misses) | 10:43 |
-!- jiuding [dbe09f12@gateway/web/freenode/ip.219.224.159.18] has quit [Ping timeout: 245 seconds] | 10:44 | |
vgorbati | lisitsyn: is this ok for the fix? | 10:44 |
deerishi | hi everyone.i want to discuss about the gsoc proposal.i want to implement speech recognition algorithims in the shogun toolbox | 10:46 |
@iglesiasg | deerishi: aham you were thinking of it as a GSoC project | 10:47 |
@HeikoS | deerishi: hi, is this on our ideas page? | 10:47 |
@iglesiasg | deerishi: I see it difficult to find a mentor for that at this moment -- too late already I think | 10:47 |
@lisitsyn | vgorbati: yes | 10:48 |
deerishi | @HeikoS: this is not on the ideas page. | 10:48 |
@HeikoS | deerishi: I agree with iglesiasg, you need a mentor for this | 10:48 |
sijin | does anybody know why CDenseFeatures->get_num_features is not const method? | 10:49 |
@HeikoS | sijin: this probably has no reason :) | 10:49 |
@iglesiasg | hehe indeed | 10:49 |
@HeikoS | I tried to change this once and gives many many follow up compile errors :D | 10:49 |
@HeikoS | but feel free to change :) | 10:49 |
deerishi | does shogun has any speech recognotion projects implemented? | 10:50 |
@lisitsyn | HeikoS: if it was not virtual it would make some great sense | 10:50 |
sijin | But when i pass CFeatures* with const, I can not use this method | 10:50 |
@lisitsyn | sijin: why to pass pointer with const? | 10:50 |
@lisitsyn | CFeatures* const? | 10:50 |
@lisitsyn | or const CFeatures*? | 10:50 |
sijin | yes I pass it as Const CFeatures* | 10:51 |
@lisitsyn | what is the advantage? | 10:51 |
deerishi | @iglesiasg : can speech recognition be a gsoc proposal?i mean is it feasible? | 10:52 |
@iglesiasg | sijin: I thought you meant CDenseFeatures::get_num_features() const | 10:52 |
@iglesiasg | deerishi: I think it is late to find a mentor for that | 10:52 |
vgorbati | lisitsyn: btw, I just tried to change to Isomap, and it preserves all neighbors - i even tried manipulating n_neighbors, but still - all neighbors are preserved) Maybe let's test Isomap for now? | 10:52 |
@lisitsyn | vgorbati: hah! yes isomap should preserve it totally | 10:52 |
@lisitsyn | feel free to transform it actually | 10:53 |
@lisitsyn | I don't mind | 10:53 |
sijin | @lisitsyn I am writing function to wrap Projectingfunction | 10:53 |
vgorbati | lisitsyn: I mean, there's no test for Isomap right now, isn't it? | 10:53 |
@lisitsyn | vgorbati: yes | 10:53 |
@lisitsyn | as you can see | 10:53 |
@lisitsyn | sijin: what should that function do? | 10:53 |
sijin | And the projection operator receive const Densevector | 10:53 |
@HeikoS | deerishi: its all about the mentor, and we currently dont have one. We have HMM implemented but not specially for speec recognition. If you don't know a mentor this project is unfeasible, yes | 10:54 |
sijin | @lisitsyn Since I want to define a shogun::projectionfunction, so that I can use tapkee with tapkee_shogun.hpp included | 10:54 |
sijin | or I will need to include all the definition in tapkee | 10:55 |
deerishi | is it that late!!!! @HeikoS ,@iglesiasg: basically i am interested in ML,NLP and signal processing.i want to give a proposal that somehow combines all these elements | 10:55 |
vgorbati | lisitsyn: ok, then I will transform:) And I have another question - I was looking at tapkee's embedtDistributedStochasticNeighborEmbedding method, and there's a check for perplexity: 0<=perplexity<=(n_vectors-1)/3.0 + 1e-6. I can't understand the reasons for such ranges | 10:56 |
@lisitsyn | vgorbati: 1e-6 or N-1/3? | 10:56 |
vgorbati | N-1/3 | 10:56 |
@iglesiasg | deerishi: the deadline for student proposals is next week | 10:56 |
@iglesiasg | you would need | 10:56 |
@lisitsyn | vgorbati: it comes from the original implementation of t-sne | 10:56 |
@lisitsyn | which I integrated to my lib | 10:57 |
@iglesiasg | deerishi: 1) Find someone who can mentor this project and want to be gsoc mentor for shogun. | 10:57 |
@iglesiasg | deerishi: 2) describe the project similar to the others that appear in the list | 10:57 |
@iglesiasg | deerishi: 3) prepare your application | 10:57 |
@iglesiasg | deerishi: that's roughly all the process. Providing that, yeah I think it is a bit too late :) | 10:58 |
@iglesiasg | you may still try if you wish of course, for my part there's no inconvenient | 10:58 |
sijin | @lisitsyn in my current implementation, shogun::ProjectingFunction has only one member, which is a pointer of tapkee::ProjectingFunciton. | 10:59 |
-!- foulwall [~foulwall@li379-21.members.linode.com] has quit [Remote host closed the connection] | 11:01 | |
deerishi | @iglesiag: is Andreas Ziehe a mentor this time? | 11:03 |
deerishi | he worked on ICA last time | 11:03 |
-!- gsomix [~Miranda@r206-10.smr.ru] has joined #shogun | 11:03 | |
gsomix | hello | 11:03 |
deerishi | i wanna go along the lines of ICA(since it is used in signal processing) | 11:04 |
@iglesiasg | deerishi: then it seems reasonable to apply for the ICA project, don't you think so? | 11:05 |
deerishi | yes : ICA is very useful in signal processing,like the cocktail party problem. | 11:06 |
sijin | @lisitsyn that function will operate on (CFeatures*) and output the projected (CFeatures*) | 11:08 |
-!- Yanglittle [deb20af4@gateway/web/freenode/ip.222.178.10.244] has joined #shogun | 11:15 | |
@iglesiasg | wiking: around? | 11:17 |
Yanglittle | hey, i run the classifier_multiclaslogisticregression from python_modular, it says "can't import name multiclasslogisticregression" at line 5. | 11:29 |
@lisitsyn | Yanglittle: do you have eigen3 installed? | 11:30 |
Yanglittle | oh, not yet. | 11:30 |
@lisitsyn | will work once you install and recompile | 11:31 |
Yanglittle | ok. | 11:31 |
-!- iglesiasg [~test@n153-p56.kthopen.kth.se] has quit [Quit: leaving] | 11:32 | |
-!- nando [~nando@n162-p214.kthopen.kth.se] has joined #shogun | 11:35 | |
deerishi | @HeikoS : i want to give my gsoc proposal on ICA.is Andreas Ziehe still a mentor for this year's gsoc? | 11:38 |
@lisitsyn | deerishi: please check the ideas list | 11:39 |
@HeikoS | deerishi: ^ | 11:41 |
deerishi | @lisitsyn: i checked it.i meant is he on irc.i had sent him a mail a while ago,he did not reply | 11:41 |
@lisitsyn | that happens, he will answer | 11:42 |
@HeikoS | deerishi: he should be still in, might be busy, maybe try again. | 11:42 |
@HeikoS | deershi, make sure not to just ask him to tell you what to do, but rather try to propose something | 11:44 |
deerishi | @HeikoS : k.thanks. | 11:50 |
deerishi | will start working on it right away | 11:50 |
deerishi | :) | 11:50 |
@HeikoS | some initial code specifically for your project will be very useful | 11:50 |
deerishi | could you elaborate on that.i mean i have done some coding work on ICA before. | 11:51 |
@HeikoS | deerishi: a simple thing that relates to the project would already be enough | 11:54 |
@HeikoS | or a preliminary plan on how you want to implement it | 11:54 |
@HeikoS | best to discuss with the mentor | 11:55 |
-!- erlenda [~erlenda@fw-oslo.intelcom.no] has joined #shogun | 11:58 | |
-!- saket [0e8b5206@gateway/web/freenode/ip.14.139.82.6] has joined #shogun | 11:59 | |
-!- vgorbati [2535754b@gateway/web/freenode/ip.37.53.117.75] has quit [Ping timeout: 245 seconds] | 11:59 | |
sijin | @lisitsyn the Projectingfunction.implentation is still 0 | 12:21 |
sijin | ProjectingFunction.implementation | 12:23 |
-!- abinash [75ef5e6e@gateway/web/freenode/ip.117.239.94.110] has joined #shogun | 12:33 | |
-!- makokal [~bokal@212.201.44.246] has joined #shogun | 12:37 | |
@lisitsyn | sijin: still? what did you expect? | 12:41 |
-!- makokal [~bokal@212.201.44.246] has quit [Quit: makokal] | 12:46 | |
@lisitsyn | I didn't add a projection function for kpca yet ;) | 12:47 |
saket | lisitsyn: there? | 12:55 |
@lisitsyn | saket: yes | 13:01 |
saket | Hi in the KNN code, if you don't pass testfeatures while apply_multiclass ...it is using the original train data on;ly! | 13:02 |
saket | instead of distance->rhs | 13:02 |
saket | I tested it by printing SG_PRINT(num_lab) and also results->get_num_labels() | 13:04 |
saket | where | 13:05 |
saket | CMulticlassLabels* result = knn->apply_multiclass(); | 13:05 |
saket | printf("testing size : %d\n", result->get_num_labels()) ; | 13:05 |
saket | lisitsyn: can you please check? | 13:07 |
nando | saket: what do you mean? | 13:07 |
nando | if you don't pass test data, it only uses the train data yeah | 13:07 |
-!- makokal [~bokal@212.201.44.246] has joined #shogun | 13:08 | |
nando | upps wrong nick | 13:08 |
-!- nando is now known as iglesiasg | 13:08 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 13:08 | |
saket | iglesiasg: it should use the distance->rhs as per the code | 13:09 |
@iglesiasg | saket: mmm I don't understand sorry | 13:09 |
saket | /* create distance */ | 13:11 |
saket | CEuclideanDistance* distance=new CEuclideanDistance(features, testfeatures); | 13:12 |
saket | // create KNN CKNN* knn = new CKNN(3, distance, labels ) ; | 13:12 |
saket | CKNN* knn = new CKNN(3, distance, labels ) ; | 13:12 |
saket | knn->train(); | 13:12 |
saket | CMulticlassLabels* result = knn->apply_multiclass(); | 13:12 |
saket | printf("testing size : %d\n", result->get_num_labels()) ; | 13:12 |
saket | I am expecting testing size = |testfeatures| | 13:13 |
saket | But I get testing size = |features| | 13:13 |
@iglesiasg | ok, I understand what you mean | 13:15 |
@iglesiasg | let me check | 13:15 |
saket | iglesiasg: :) | 13:15 |
saket | iglesiasg: do you want me to send this testing code ? | 13:16 |
@iglesiasg | saket: the result is right | 13:16 |
@iglesiasg | you should pass the test features to KNN if you want to apply them to the test features | 13:16 |
@iglesiasg | reason why? | 13:17 |
saket | iglesiasg: reason why? I didn't get you | 13:17 |
@iglesiasg | a classifier has *one* set of features I believe | 13:17 |
@iglesiasg | independently that it is associated or not with distance or kernels | 13:18 |
@iglesiasg | which do have two set of features | 13:18 |
@iglesiasg | therefore you have to either set the features you want to test before calling apply | 13:18 |
@iglesiasg | or give them to apply | 13:19 |
saket | oh | 13:19 |
saket | I wil do so | 13:20 |
saket | *will | 13:20 |
saket | iglesiasg: also you had asked me to send a PR to develop in https://github.com/shogun-toolbox/shogun/pull/994 | 13:21 |
saket | iglesiasg: I had sent one to develop a few days ago (https://github.com/shogun-toolbox/shogun/pull/995 ) ...please take a look :) | 13:22 |
-!- iglesiasg [~nando@n162-p214.kthopen.kth.se] has quit [Read error: Operation timed out] | 13:23 | |
-!- van51 [~van51@athedsl-319014.home.otenet.gr] has left #shogun ["QUIT :Leaving."] | 13:32 | |
-!- ishan [~ishan@14.139.122.114] has joined #shogun | 13:36 | |
-!- ishan_ [~ishan@14.139.122.114] has joined #shogun | 13:38 | |
-!- ishan [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 13:38 | |
sijin | @lisitsyn sorry, i went out for a while. i am back | 13:40 |
sijin | @lisitsyn I intend to warp Kernel PCA using the projecting function | 13:41 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Remote host closed the connection] | 13:41 | |
-!- ishan__ [~ishan@14.139.122.114] has joined #shogun | 13:42 | |
-!- ishan_ [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 13:42 | |
-!- iglesiasg [~nando@n162-p214.kthopen.kth.se] has joined #shogun | 13:43 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 13:43 | |
@iglesiasg | saket: about the PR | 13:44 |
saket | iglesiasg: yes | 13:44 |
@iglesiasg | saket: as I told you in the other PR | 13:45 |
@iglesiasg | I cannot merge PRs so another person will take a look :) | 13:45 |
@iglesiasg | saket: I remember though that you had some problems with multiple definitions of v_array | 13:45 |
@iglesiasg | did you solve that? | 13:45 |
saket | Kinda yes, | 13:46 |
saket | I switched to vector from STL | 13:46 |
@iglesiasg | hehe ok | 13:46 |
saket | instead of v_array from /lib/v_array.h | 13:46 |
@iglesiasg | so JLCoverTree is using now stl::vector? | 13:46 |
saket | no | 13:46 |
saket | Mine (BallTree) is | 13:47 |
@iglesiasg | ah ok, got it | 13:47 |
@iglesiasg | saket: how is the ball tree working? | 13:47 |
@iglesiasg | is it already finished? | 13:48 |
saket | Its working good | 13:48 |
saket | for small dimensions | 13:48 |
sijin | @lisitsyn The current Kernel PCA calculate the transformation matrix using "computer_eigenvector", and preserve the transformation_matrix. | 13:48 |
@iglesiasg | saket: how fast is it? | 13:48 |
saket | it has very high relative speedup | 13:48 |
saket | but as dimensions increase | 13:48 |
@iglesiasg | saket: compare to cover tree? | 13:48 |
saket | it decreases to 1.2 near 40 | 13:48 |
saket | I will compare to cover tree .. | 13:48 |
saket | currently comparing with naive knn | 13:48 |
@iglesiasg | aham ok | 13:49 |
@iglesiasg | our naive knn is pretty slow IIRC | 13:49 |
-!- ishan [~ishan@14.139.122.114] has joined #shogun | 13:49 | |
-!- ishan__ [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 13:49 | |
saket | shall i use sort from algorithms STL to check then ? | 13:50 |
saket | *algorithm STL | 13:50 |
@iglesiasg | it is not a matter of the quicksort | 13:50 |
@iglesiasg | sonney2k even claims his is faster than STL's one :) | 13:50 |
@iglesiasg | it is a matter of KNN | 13:51 |
@iglesiasg | let me take a look to remember what was it | 13:51 |
sijin | @lisitsyn I just felt that so little work is done with tapkee if tapkee lib is only used for calculationg the transformation(embeded data). | 13:51 |
@iglesiasg | saket: https://github.com/shogun-toolbox/shogun/blob/master/src/shogun/multiclass/KNN.cpp | 13:53 |
@iglesiasg | lines 155 to 185 | 13:53 |
saket | iglesiasg: looking .. | 13:53 |
-!- abinash [75ef5e6e@gateway/web/freenode/ip.117.239.94.110] has quit [Ping timeout: 245 seconds] | 13:55 | |
saket | iglesiasg: its not *that* slow either... perhaps the only speed up i can think here is somehow making labels and distances as pair in line 174 | 13:56 |
@iglesiasg | saket: why would that be faster? | 13:57 |
saket | iglesiasg: saving cpu operations | 13:57 |
saket | faster in cpu utilisation by saving the loop not order (big-O ) wise | 13:58 |
@iglesiasg | saket: mmm I don't see why it would save cpu operations | 13:59 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 14:00 | |
saket | iglesiasg: because you won't need the executions in lines 176 && 177 | 14:01 |
@sonney2k | iglesiasg, than qsort in libc - I don't know about std::sort | 14:01 |
saket | like | 14:01 |
saket | say you have an array of nodes { float64_t dist, int32_t label } | 14:01 |
saket | and the distance class could directly fill this array | 14:01 |
saket | then lines 176 && 177 would not be needed | 14:02 |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 14:03 | |
shogun-notifier- | shogun: Soeren Sonnenburg :develop * aa9f749 / src/NEWS: https://github.com/shogun-toolbox/shogun/commit/aa9f749d6bbbc39e4473f6e4ce45ec9098f7fb99 | 14:03 |
shogun-notifier- | shogun: update news | 14:03 |
@iglesiasg | saket: mmm | 14:04 |
@wiking | sonney2k: have u managed to check buildbot? | 14:04 |
@iglesiasg | saket: I think you need the train_lab array for every example | 14:04 |
@iglesiasg | saket: so my guess is that you actually need to do that | 14:04 |
@iglesiasg | saket: maybe you can take lines 176 and 177 out of the loop | 14:04 |
@iglesiasg | saket: but you would need to create n arrays like that in any case | 14:05 |
@iglesiasg | I may be wrong though :) | 14:05 |
@iglesiasg | saket: anyway, I think I remembered now why I was saying our KNN is slow | 14:05 |
@sonney2k | wiking, just trying it now | 14:06 |
@iglesiasg | saket: so what is the complexity of our naive KNN? O() of what? | 14:06 |
saket | iglesiasg: Yes it won't work for multiple train examples.. | 14:06 |
saket | O(nlogn) | 14:06 |
@iglesiasg | saket: for each tet example isn't it? | 14:07 |
saket | yes | 14:07 |
@iglesiasg | test* | 14:07 |
@iglesiasg | ok | 14:07 |
@iglesiasg | and we are doing *K*NN not *N*NN :) | 14:07 |
@iglesiasg | what I mean is K << N normally | 14:07 |
@iglesiasg | saket: agree? | 14:08 |
saket | generally it is | 14:08 |
@iglesiasg | saket: indeed | 14:08 |
@iglesiasg | saket: now this is the part I like | 14:08 |
@iglesiasg | saket: do you need O(N log(N)) to get the first K elements of an array of N elements? | 14:08 |
-!- ishan [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 14:08 | |
saket | no | 14:09 |
@iglesiasg | iglesiasg: with the first I mean that you need to order them according to something | 14:09 |
@iglesiasg | saket: indeed, how much do you need? | 14:09 |
-!- ishan [~ishan@14.139.122.114] has joined #shogun | 14:09 | |
saket | you can get them in a method to that of quick sort in O(n) | 14:09 |
saket | *method similar | 14:09 |
@iglesiasg | saket: in O(N)?? how | 14:09 |
saket | using Order statistics | 14:09 |
saket | it is similar to the way in which we get median in O(n) | 14:10 |
@iglesiasg | mmm not sure what that is | 14:10 |
@iglesiasg | but my claim is that you can't do that faster than O(N log(K)) | 14:10 |
@iglesiasg | O(N) is too optimistic hehe | 14:10 |
saket | I think the algo is more commonly known as "median of medians" | 14:10 |
saket | so basically using "median of median" technique one can get the k largest elements in an array in O(n) | 14:11 |
saket | and then O(klogk) to sort them ...but k << N so it is O(n) only | 14:11 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Remote host closed the connection] | 14:11 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 14:12 | |
@lisitsyn | http://www.cplusplus.com/reference/algorithm/nth_element/ | 14:12 |
@lisitsyn | I guess that? | 14:12 |
saket | iglesiasg: the name of algorithm is "Selection algorithm" | 14:12 |
saket | http://en.wikipedia.org/wiki/Selection_algorithm | 14:12 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Remote host closed the connection] | 14:13 | |
saket | lisitsyn: yes exactly that one | 14:13 |
@iglesiasg | saket: yeah, I am checking | 14:13 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 14:13 | |
@iglesiasg | saket: I am not sure but I think they are just forgetting about the log(K) part | 14:13 |
@iglesiasg | well forgetting | 14:14 |
@iglesiasg | one may say that N log(K) \approx N for N >> K | 14:14 |
saket | no..there is no log(k) | 14:14 |
saket | if you want the k largest/smallest elements sorted then its O(N + klogk) | 14:14 |
sijin | the algorithm is just O(log(n)) + O(n) | 14:15 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Remote host closed the connection] | 14:15 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 14:15 | |
sijin | You use log(n) to find the k-th largest ,after that , you only need to select the numbers that is smaller than the k-th largest | 14:15 |
saket | iglesiasg: http://en.wikipedia.org/wiki/Partial_sorting | 14:16 |
@iglesiasg | it looks promising | 14:17 |
saket | I can do this as an extension to the project LMNN :-) | 14:18 |
@sonney2k | saket, idea was that we want kNN results for all possible k=1...K | 14:18 |
@iglesiasg | saket: the approach I knew is the one in data structure-based solutions in the last link | 14:19 |
-!- ishan_ [~ishan@14.139.122.114] has joined #shogun | 14:19 | |
-!- ishan [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 14:19 | |
saket | sonney2k: once we have the top K {distance, label} pairs then the code in current KNN is good enough for knn results for all possible 1...k | 14:20 |
saket | imo | 14:20 |
sijin | @lisitsyn do you have time now? | 14:23 |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 14:24 | |
-!- travis-ci [~travis-ci@ec2-50-19-58-94.compute-1.amazonaws.com] has joined #shogun | 14:24 | |
travis-ci | [travis-ci] it's Soeren Sonnenburg's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/6596846 | 14:24 |
-!- travis-ci [~travis-ci@ec2-50-19-58-94.compute-1.amazonaws.com] has left #shogun [] | 14:24 | |
@wiking | doh | 14:24 |
@wiking | travis needs fix :( | 14:24 |
@iglesiasg | wiking: yes, I was looking around for you :) | 14:25 |
@wiking | iglesiasg: here | 14:25 |
@sonney2k | wiking, I am totally clueless why it doesn't work | 14:25 |
@iglesiasg | wiking: so it seems that what is failing are the tests | 14:26 |
@wiking | iglesiasg: which, now new ones? | 14:26 |
@iglesiasg | no idea, let me see | 14:26 |
@iglesiasg | some in converter IIRC | 14:26 |
@iglesiasg | and yesterday something in serialization | 14:27 |
@iglesiasg | wiking: anyway, can we set a different exit status if tests fail than if the build fails? | 14:28 |
@wiking | iglesiasg: i'll rebuild now shogun locally | 14:28 |
@wiking | iglesiasg: and tell u what's happening here with me | 14:29 |
@wiking | since in osx a lot of integration tests failed | 14:29 |
@iglesiasg | wiking: the problem that I see is that now all the pull requests to develop appear as travis build failed | 14:29 |
@wiking | iglesiasg: yes | 14:29 |
@wiking | i know | 14:29 |
@iglesiasg | wiking: even though sometimes it is "only" because I test failed | 14:29 |
@wiking | iglesiasg: not only | 14:29 |
@wiking | it should fail if it's really something bad | 14:29 |
@wiking | the thing is that there's this serialization error | 14:29 |
@wiking | ok i see there's now even more | 14:30 |
@iglesiasg | wiking: mmm I have checked a bit some traces and didn't find anything but tests broken | 14:30 |
@iglesiasg | maybe I missed sth | 14:30 |
@wiking | [ FAILED ] MultidimensionaScalingTest.distance_preserving | 14:30 |
@wiking | who broke this? | 14:30 |
@iglesiasg | yes | 14:30 |
@wiking | or rather, which commit broke this? | 14:30 |
@iglesiasg | no idea | 14:30 |
@wiking | ok well it's time to find out | 14:30 |
@iglesiasg | wiking: indeed | 14:31 |
@wiking | well it's easy | 14:31 |
@wiking | https://travis-ci.org/shogun-toolbox/shogun/builds | 14:31 |
@wiking | and find a build where only python_modular fails | 14:31 |
@wiking | https://travis-ci.org/shogun-toolbox/shogun/builds/6573069 | 14:31 |
@wiking | there | 14:31 |
@wiking | here it was still ok | 14:31 |
@wiking | here it broke | 14:31 |
@wiking | https://travis-ci.org/shogun-toolbox/shogun/builds/6578855 | 14:31 |
@wiking | so lisitsyn ;) | 14:31 |
@wiking | lisitsyn: your tapkee update brok MultidimensionaScalingTest.distance_preserving unit test | 14:32 |
@iglesiasg | wiking: python_modular is broken in both there, isn't it? | 14:32 |
@wiking | iglesiasg: ignore python_modular error | 14:32 |
@wiking | now just let's fix first MultidimensionaScalingTest.distance_preserving unitttest | 14:32 |
@iglesiasg | wiking: all right, why? | 14:32 |
@wiking | iglesiasg: i have no idea | 14:33 |
@wiking | do a git bisect | 14:33 |
@wiking | and you'll know why | 14:33 |
@wiking | i mean for MultidimensionaScalingTest.distance_preserving failure | 14:33 |
@iglesiasg | wiking: I mean why to ignore python_modular | 14:33 |
@wiking | for python_modular i dont know why that exact serialization test fails | 14:33 |
@iglesiasg | ok | 14:33 |
@wiking | iglesiasg: because that's broken ever since i've enabled python_modular on travis | 14:33 |
@iglesiasg | got it | 14:33 |
@wiking | none of us knows why that doesn't work | 14:33 |
@wiking | but it just doesn't | 14:33 |
@wiking | but everything else was fine | 14:34 |
@wiking | i.e. libshogun build both with gcc and clang compiled well and all the unit tests ran as well w/o errors | 14:34 |
@iglesiasg | yes | 14:34 |
@wiking | and it got broken with lisitsyn's tapkee upgrade | 14:35 |
@wiking | so it's somewhere there... i mean the error | 14:35 |
@iglesiasg | yeah | 14:35 |
@wiking | so please try to fix that | 14:35 |
@iglesiasg | I understand it now :) | 14:35 |
@wiking | i'll try to debug now that integration test problem | 14:35 |
@iglesiasg | wiking: it is sort a fucking huge commit :) | 14:35 |
@wiking | as i had much more troubles with them on osx | 14:35 |
@iglesiasg | let's better wait until lisitsyn can take a look | 14:35 |
@wiking | iglesiasg: ask lisitsyn then wtf has happened :) | 14:35 |
@iglesiasg | wiking: nah no worries | 14:36 |
@iglesiasg | it will be fixed soon probably | 14:36 |
@wiking | yeah right but until then we don't know anything for sure about the PRs | 14:37 |
@wiking | so it's rather crucial that code that is commited into the repo | 14:37 |
@wiking | is actually running w/o errors | 14:37 |
@iglesiasg | wiking: yes, I totally agree | 14:37 |
@iglesiasg | let's not merge stuff until that is fixed | 14:37 |
@wiking | that's a lot of help when finally GSoC starts | 14:37 |
@iglesiasg | wiking: mmmm one thing | 14:38 |
@wiking | and then PRs starts to land | 14:38 |
@iglesiasg | wiking: introducing the fix in the develop | 14:38 |
@iglesiasg | wiking: wont't change anything in the current state of the PRs, will it? | 14:38 |
@wiking | shoot | 14:38 |
@wiking | iglesiasg: unfortunately no | 14:38 |
@wiking | it won't try to rebuild it :( | 14:38 |
-!- ishan_ [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 14:38 | |
* iglesiasg wonders if rebasing and pushing might be a solution | 14:39 | |
-!- ishan_ [~ishan@14.139.122.114] has joined #shogun | 14:39 | |
@lisitsyn | wiking: iglesiasg: re | 14:41 |
@lisitsyn | wiking: I know it fails will fix today | 14:41 |
@iglesiasg | lisitsyn: :) | 14:42 |
@lisitsyn | I totally forgot to run tests yesterday | 14:43 |
@iglesiasg | wiking: I think a possible solution for the PRs after the fix is in develop is | 14:43 |
@iglesiasg | rebase upsatram develop | 14:43 |
@iglesiasg | and push to origin develop/feature | 14:44 |
@iglesiasg | although I think the second step would require forcing the push or deleting the branch in the repo first | 14:44 |
sijin | lisitsyn: would you talk about your expectation of kpca warper? | 14:44 |
@lisitsyn | sijin: yes, what's up? | 14:47 |
sijin | I would like to hear your expecation of this warper so that I can develop the warper which can agree with your interface in tapkee | 14:49 |
-!- ishan_ [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 14:49 | |
@iglesiasg | sonney2k: how is it going in Paris BTW? nice talks? | 14:49 |
@lisitsyn | nice | 14:49 |
-!- ishan_ [~ishan@14.139.122.114] has joined #shogun | 14:49 | |
@lisitsyn | sijin: alright - as I see that | 14:49 |
@lisitsyn | we can just store the embedding matrix and later use it in apply | 14:50 |
@lisitsyn | IIRC kpca is just K_{train,test} * embedding_{train} | 14:50 |
sijin | yeah, i know that. in current implementation, nxn matrix is stored, which is of course wasting. | 14:51 |
@lisitsyn | sijin: https://github.com/shogun-toolbox/shogun/blob/master/src/shogun/preprocessor/KernelPCA.cpp what tapkee does is lines 72-96 | 14:55 |
sijin | Yes. I have read this code | 14:55 |
vgorbati | lisitsyn: 'will fix today' - I guess it would be better if I send my PR's after your fix? | 14:55 |
@lisitsyn | vgorbati: your pr with isomap test? | 14:56 |
vgorbati | lisitsyn: both: with isomap test and with t-SNE wrapper | 14:57 |
@lisitsyn | feel free to send anytime - should not break my fix | 14:57 |
sijin | The reason why I start from Projecting function is that I found many code are doing the same thing. I would like to have a simple function to do that | 14:57 |
@lisitsyn | sijin: just implement it as kernel pca member function | 14:58 |
sijin | Fine. I will do that. | 14:58 |
-!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has quit [Ping timeout: 245 seconds] | 14:59 | |
vgorbati | lisitsyn: I meant that if I wait for your fix, then send a PR and then a travis build fails it would be clear that's because of my code. Otherwise it wont be that clear) | 15:00 |
@lisitsyn | ah | 15:01 |
sijin | lisitsyn: btw, i have implemented the interface to use ProjectingFunciton. If any other embedding method need that I can commit that file also | 15:01 |
@lisitsyn | alright | 15:01 |
@lisitsyn | sijin: I don't know - lets see | 15:01 |
-!- vikram360 [~vikram360@117.216.147.147] has joined #shogun | 15:01 | |
@lisitsyn | vgorbati: I'd say ETA 20-21 kiev time | 15:02 |
vgorbati | lisitsyn: :D | 15:02 |
sijin | I mean I have written a interface in "tapkee_shogun.cpp", It can return the projecting function as output | 15:02 |
vgorbati | lisitsyn: thanks for localization) | 15:03 |
@lisitsyn | sijin: alright that may appear useful | 15:03 |
sijin | Fine. I will commit them together with the warper. Or if you want to see now, I can commit that file right now, since the warper will not use Projecting funciton. | 15:06 |
@lisitsyn | sijin: just do a PR once you are ready | 15:07 |
sijin | lisitsyn: :), ok | 15:08 |
-!- ishan_ [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 15:09 | |
-!- ishan_ [~ishan@14.139.122.114] has joined #shogun | 15:09 | |
-!- abinash [75ef5e6e@gateway/web/freenode/ip.117.239.94.110] has joined #shogun | 15:16 | |
-!- ishan__ [~ishan@14.139.122.114] has joined #shogun | 15:20 | |
-!- ishan_ [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 15:20 | |
-!- abinash [75ef5e6e@gateway/web/freenode/ip.117.239.94.110] has quit [Ping timeout: 245 seconds] | 15:27 | |
-!- ishan [~ishan@14.139.122.114] has joined #shogun | 15:39 | |
-!- ishan__ [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 15:39 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Remote host closed the connection] | 15:41 | |
@wiking | iglesiasg: ok | 15:45 |
@wiking | iglesiasg: i've recompiled a lot of things | 15:46 |
@wiking | iglesiasg: but still | 15:46 |
@wiking | iglesiasg: the same things fail as yesterday | 15:47 |
@wiking | lisitsyn: yo | 15:47 |
@iglesiasg | wiking: the tapkee one? | 15:47 |
@wiking | iglesiasg: no, i mean integration test of python_modular | 15:48 |
@iglesiasg | ok | 15:48 |
-!- ishan [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 15:49 | |
-!- ishan [~ishan@14.139.122.114] has joined #shogun | 15:50 | |
@lisitsyn | wiking: yes | 15:50 |
@lisitsyn | yo! | 15:50 |
@lisitsyn | :D | 15:50 |
@lisitsyn | what? | 15:50 |
@wiking | lisitsyn: your tapkee update broke a unitttest | 15:51 |
@lisitsyn | wiking: bahh! really? I said I'll fix it at least two times :D | 15:51 |
@wiking | fix it! | 15:54 |
@wiking | lisitsyn: [ RUN ] MultidimensionaScalingTest.distance_preserving | 15:55 |
@wiking | converter/MultidimensionalScaling_unittest.cc:45: Failure | 15:55 |
@wiking | The difference between euclidean_distance_matrix(i,j) and euclidean_distance_for_embedding_matrix(i,j) is 275.89303860676392, which exceeds 1e-9, where | 15:55 |
@wiking | euclidean_distance_matrix(i,j) evaluates to 4.5450052246886221, | 15:55 |
@wiking | euclidean_distance_for_embedding_matrix(i,j) evaluates to 280.43804383145255, and | 15:55 |
@wiking | 1e-9 evaluates to 1.0000000000000001e-09. | 15:55 |
@wiking | [ FAILED ] MultidimensionaScalingTest.distance_preserving (6 ms) | 15:55 |
@lisitsyn | wiking: come on I said I'll fix it tonight | 15:56 |
@wiking | lisitsyn: :) | 15:56 |
@lisitsyn | wiking: I do not know the reason it fails but should be some easy fix | 15:58 |
@wiking | can we disable integration test? | 16:00 |
@wiking | at least with python | 16:00 |
@lisitsyn | wiking: all integration tests? | 16:00 |
@lisitsyn | or what? | 16:00 |
@wiking | because it fails big time on my OS | 16:00 |
@wiking | lisitsyn: at least python_modular | 16:00 |
@lisitsyn | which test? | 16:00 |
@lisitsyn | disable where? | 16:01 |
@wiking | lisitsyn: these: http://pastebin.com/sAbq05kN | 16:01 |
@lisitsyn | wiking: this should be called reverse engineering - when you engineer a solution that always fails | 16:02 |
@wiking | lisitsyn: so obviously with integration test for python something is very weird | 16:02 |
@wiking | it works for some default setups | 16:02 |
@lisitsyn | wiking: kinda strange it always fails | 16:03 |
@wiking | but for example on osx it's just fails big time | 16:03 |
@wiking | and travis build machine create some weird shit as well | 16:03 |
@wiking | and now because of faulty integration tests we are not having successful builds on travis | 16:03 |
@wiking | lisitsyn: but of course it might be that the seeded random for osx is tooootaly different than the one on linux | 16:04 |
@wiking | as obviously the random generator in osx is different from linux | 16:04 |
@wiking | and this way | 16:04 |
@wiking | of course we'll have some troubles... | 16:04 |
@wiking | relying on seed+random | 16:05 |
@wiking | hmmm i guess the intergration test would fail as well on freebsd | 16:06 |
@lisitsyn | wiking: but we don't really on that that much? | 16:07 |
@lisitsyn | rely* | 16:07 |
@lisitsyn | :D | 16:07 |
@lisitsyn | not really | 16:07 |
@wiking | lisitsyn: apart from those tests that are failing on osx :P | 16:07 |
@lisitsyn | wiking: have you checked or? | 16:08 |
@wiking | lisitsyn: just checking whether it's consistent on freebsd ;P | 16:08 |
@wiking | i would guess so | 16:08 |
@wiking | as the random machine on osx and freebsd are the same | 16:08 |
@wiking | but different from linux | 16:08 |
-!- ishan_ [~ishan@14.139.122.114] has joined #shogun | 16:10 | |
-!- ishan [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 16:10 | |
@wiking | lisitsyn: and indeed the one failing on travis as well based on a seeded random input | 16:10 |
@lisitsyn | wiking: argh - any idea how to resolve that? | 16:10 |
-!- saket [0e8b5206@gateway/web/freenode/ip.14.139.82.6] has quit [Ping timeout: 245 seconds] | 16:11 | |
@wiking | well ... dont rely on seeded random input? :) | 16:13 |
@lisitsyn | wiking: sounds like 'write high-quality code' | 16:13 |
@wiking | lisitsyn: indeed :D | 16:14 |
@lisitsyn | wiking: you've been to industry for years haven't you? | 16:14 |
@wiking | lisitsyn: yes :D | 16:14 |
@wiking | i wanted to say 'unfortunately yes' :D | 16:14 |
@lisitsyn | wiking: we don't differ that much | 16:15 |
@wiking | so | 16:15 |
@lisitsyn | so we need real solution :) | 16:15 |
@wiking | i'd like to propose that we remove integration testing for the time being when calling make tests | 16:15 |
@lisitsyn | converting all examples would take some time | 16:15 |
@wiking | yeah i know | 16:15 |
@wiking | but until then | 16:15 |
@wiking | let's not try to run them | 16:15 |
@lisitsyn | I don't mind, but ask sonney2k | 16:16 |
@wiking | sonney2k: ^ | 16:16 |
@wiking | but let me double-check my assumption about random here | 16:16 |
@wiking | just trying to do integration test on fbsd | 16:16 |
@wiking | let's hope the same tests fails | 16:17 |
@wiking | lisitsyn: i'm sure we can even break integration tests within linux as well | 16:17 |
@wiking | it's just a matter of tweaking the random machine | 16:17 |
@wiking | hahaha | 16:19 |
@wiking | should i add 'rewrite integration tests' as an entrance bug in github? :))) | 16:20 |
@wiking | if somebody does that, imo he passed :P | 16:20 |
-!- ishan__ [~ishan@14.139.122.114] has joined #shogun | 16:20 | |
-!- ishan_ [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 16:20 | |
@lisitsyn | wiking: heh well feel free | 16:20 |
@lisitsyn | but not like rewrite | 16:21 |
-!- iglesiasg [~nando@n162-p214.kthopen.kth.se] has quit [Read error: Operation timed out] | 16:21 | |
@lisitsyn | get rid of random tests I think | 16:21 |
@wiking | anybody seen this baby before: http://pastebin.com/W0xKUCMZ | 16:22 |
-!- gsomix [~Miranda@r206-10.smr.ru] has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org] | 16:26 | |
@wiking | lol man somebody still uses miranda IM | 16:26 |
@wiking | amaaaazing gsomix! :) | 16:26 |
@lisitsyn | wiking: old school :D | 16:27 |
@lisitsyn | wiking: I used it in 2006 I think | 16:28 |
@lisitsyn | and 2007 probably | 16:28 |
@wiking | uyeah | 16:28 |
@wiking | somewhere there | 16:28 |
-!- erlenda [~erlenda@fw-oslo.intelcom.no] has quit [Quit: Leaving] | 16:37 | |
-!- ishan__ [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 16:39 | |
-!- ishan__ [~ishan@14.139.122.114] has joined #shogun | 16:40 | |
-!- ishan [~ishan@14.139.122.114] has joined #shogun | 16:50 | |
-!- ishan__ [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 16:50 | |
-!- Kirsh [~kirshanth@175.157.212.180] has joined #shogun | 16:57 | |
-!- gsomix [~Miranda@37.61.181.167] has joined #shogun | 17:01 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 17:03 | |
-!- lisitsyn [~lisitsin@mxs.kg.ru] has quit [Quit: Leaving.] | 17:09 | |
-!- ishan [~ishan@14.139.122.114] has quit [Ping timeout: 252 seconds] | 17:11 | |
-!- ishan [~ishan@14.139.122.114] has joined #shogun | 17:19 | |
-!- Yanglittle [deb20af4@gateway/web/freenode/ip.222.178.10.244] has quit [Quit: Page closed] | 17:36 | |
* wiking hates recompiling atlas... | 17:43 | |
-!- gsomix [~Miranda@37.61.181.167] has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org] | 17:45 | |
-!- ishan_ [~ishan@14.139.122.114] has joined #shogun | 17:50 | |
-!- ishan [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 17:50 | |
-!- ishan__ [~ishan@14.139.122.114] has joined #shogun | 18:20 | |
-!- ishan_ [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 18:20 | |
-!- lisitsyn [~blackburn@46.20.65.68] has joined #shogun | 18:35 | |
-!- mode/#shogun [+o lisitsyn] by ChanServ | 18:35 | |
-!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has joined #shogun | 18:37 | |
-!- ishan__ [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 18:50 | |
-!- ishan__ [~ishan@14.139.122.114] has joined #shogun | 18:50 | |
-!- gsomix [~Miranda@46.20.65.68] has joined #shogun | 18:59 | |
gsomix | good evening | 18:59 |
-!- sids_aquarius [~sidi_@li400-124.members.linode.com] has quit [Ping timeout: 245 seconds] | 19:04 | |
-!- votjakovr [~chatzilla@host-46-241-3-209.bbcustomer.zsttk.net] has quit [Quit: ChatZilla 0.9.90 [Firefox 20.0.1/20130409194949]] | 19:11 | |
-!- ishan [~ishan@14.139.122.114] has joined #shogun | 19:20 | |
-!- ishan__ [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 19:20 | |
-!- sids_aquarius [~sidi_@14.139.122.114] has joined #shogun | 19:37 | |
sids_aquarius | join #tux4kids | 19:37 |
-!- saket [0e8b5206@gateway/web/freenode/ip.14.139.82.6] has joined #shogun | 19:39 | |
deerishi | @HeikoS: hi | 19:49 |
@HeikoS | deerishi: hi | 19:49 |
deerishi | just sent my pull request | 19:49 |
@HeikoS | ah nice! | 19:49 |
-!- ishan [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 19:49 | |
-!- ishan [~ishan@14.139.122.114] has joined #shogun | 19:50 | |
deerishi | hope this one has everything you asked for . :) | 19:50 |
deerishi | i reduced the number of loops. | 19:50 |
deerishi | also now | 19:50 |
deerishi | the data from tools | 19:50 |
deerishi | is being imported directed | 19:51 |
deerishi | *directly | 19:51 |
@HeikoS | deerishi: ok nice | 19:51 |
@HeikoS | still the while loops | 19:51 |
deerishi | kindly give me your feedback . :) | 19:51 |
@HeikoS | I dont like while loops in python | 19:51 |
@HeikoS | you can easily replace them by vector operations | 19:51 |
deerishi | just the two because , sqrt function takes in only 1 element at a time | 19:51 |
deerishi | rest all are the vector operations | 19:52 |
@HeikoS | [sqrt(x) for x in bla] | 19:52 |
@HeikoS | no need for loops | 19:52 |
@HeikoS | also those guys here: | 19:53 |
@HeikoS | for i in range(lx0):+ modified_d1[i][:]=X2[i][0:number_of_points_for_circle1] | 19:53 |
@HeikoS | if you use the variable you are iterating over only for indexing, you dont need a loop | 19:53 |
@HeikoS | esp no while loop | 19:54 |
@HeikoS | slow - danger of non-termination etc | 19:54 |
-!- pararth [0e8b614d@gateway/web/freenode/ip.14.139.97.77] has joined #shogun | 19:57 | |
@HeikoS | deerishi: no big deal to remove them, everything else is fine. Once this is done we can renew the integration test :) | 20:04 |
deerishi | just give me a couple of minutes, i will remove them | 20:05 |
@HeikoS | deerishi: no worries | 20:05 |
deerishi | so its good enough to merge ? | 20:07 |
@HeikoS | deerishi: once the loops are hone, we can do the integration test which is a minor thing, and then merge | 20:08 |
deerishi | k | 20:08 |
@lisitsyn | booya | 20:11 |
-!- ishan_ [~ishan@14.139.122.114] has joined #shogun | 20:20 | |
-!- ishan [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 20:20 | |
@lisitsyn | wiking: bug in custody | 20:26 |
@lisitsyn | :D | 20:26 |
-!- vikram360 [~vikram360@117.216.147.147] has quit [Ping timeout: 252 seconds] | 20:26 | |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 20:29 | |
-!- saket [0e8b5206@gateway/web/freenode/ip.14.139.82.6] has quit [Quit: Page closed] | 20:35 | |
-!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has quit [Ping timeout: 245 seconds] | 20:36 | |
-!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has joined #shogun | 20:40 | |
deerishi | @HeikoS: just sent another PR | 20:40 |
@HeikoS | deerishi: saw it! | 20:41 |
@HeikoS | now this looks nicer! :) | 20:41 |
deerishi | can you check it? | 20:41 |
deerishi | finally.... :) | 20:41 |
@HeikoS | deerishi: do the results still make sense? | 20:41 |
@HeikoS | if you run it? | 20:41 |
deerishi | results are still the same | 20:41 |
@HeikoS | okay | 20:41 |
@HeikoS | so then lets make the integration test work | 20:41 |
@HeikoS | what you need to do is to go to tests/integration/python_modular | 20:42 |
@HeikoS | then run python generator.py preprocessor_kpca.py | 20:42 |
@HeikoS | this will generate two files | 20:42 |
@HeikoS | which contain the things that your example returns, one file for each eleent of the parmeter list | 20:42 |
@HeikoS | they are in tests/data | 20:42 |
@HeikoS | you have to commit those | 20:43 |
@HeikoS | once this is done, you have to commit the data version in the development repository | 20:43 |
@HeikoS | then push everything | 20:43 |
@HeikoS | to update PR | 20:43 |
@HeikoS | then the travis build will work | 20:43 |
@HeikoS | and one more small thing, have a look in examples/documented/ there are descriptions for the examples, please add one for yours, just a few lines that exaplain what is being done in the example | 20:44 |
-!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has quit [Ping timeout: 245 seconds] | 20:45 | |
-!- ishan__ [~ishan@14.139.122.114] has joined #shogun | 20:50 | |
-!- ishan_ [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 20:50 | |
-!- iglesiasg [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 20:54 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 20:54 | |
-!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has joined #shogun | 20:59 | |
-!- pararth [0e8b614d@gateway/web/freenode/ip.14.139.97.77] has quit [Ping timeout: 245 seconds] | 21:02 | |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 21:03 | |
shogun-notifier- | shogun: Sergey Lisitsyn :develop * 2fcf2ce / src/shogun/lib/tapkee/tapkee_methods.hpp: https://github.com/shogun-toolbox/shogun/commit/2fcf2ce03098cf6efd22a02fedf817566754f1f9 | 21:03 |
shogun-notifier- | shogun: Fixed bug in MDS | 21:03 |
@iglesiasg | lisitsyn: yaaay bug fixed :) | 21:04 |
@iglesiasg | lisitsyn: do you think we should do sth about the pull requests with travis? | 21:09 |
@iglesiasg | https://travis-ci.org/shogun-toolbox/shogun/jobs/6609074 LLE neighbors preserving | 21:11 |
-!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has quit [Ping timeout: 245 seconds] | 21:12 | |
-!- iglesiasg [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 21:15 | |
-!- iglesiasg [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 21:18 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 21:18 | |
-!- iglesiasg [~nando@s83-179-44-135.cust.tele2.se] has quit [Client Quit] | 21:19 | |
-!- ishan [~ishan@14.139.122.114] has joined #shogun | 21:21 | |
-!- ishan__ [~ishan@14.139.122.114] has quit [Read error: Connection reset by peer] | 21:21 | |
-!- travis-ci [~travis-ci@ec2-174-129-147-163.compute-1.amazonaws.com] has joined #shogun | 21:23 | |
travis-ci | [travis-ci] it's Sergey Lisitsyn's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/6609072 | 21:23 |
-!- travis-ci [~travis-ci@ec2-174-129-147-163.compute-1.amazonaws.com] has left #shogun [] | 21:23 | |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 21:24 | |
-!- ishan [~ishan@14.139.122.114] has quit [Remote host closed the connection] | 21:49 | |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 264 seconds] | 22:02 | |
-!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has joined #shogun | 22:04 | |
deerishi | @HeikoS: i ran the integration test but it gave the following error 'module' object has no attribute 'preprocessor_kpca' | 22:05 |
deerishi | ? | 22:05 |
deerishi | ERROR generating '../../../tests/integration/tests/preprocessor_kpca0.txt' using 'preprocessor_kpca.py' | 22:07 |
deerishi | @HeikoS : what does it mean? | 22:07 |
@HeikoS | pls show me the code again | 22:07 |
-!- gsomix [~Miranda@46.20.65.68] has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org] | 22:09 | |
deerishi | https://github.com/deerishi/shogun/blob/shogun3kpca/examples/undocumented/python_modular/tools/generate_circle_data.py | 22:10 |
deerishi | https://github.com/deerishi/shogun/blob/shogun3kpca/examples/undocumented/python_modular/graphical/preprocessor_kpca_graphical.py | 22:10 |
@HeikoS | deerishi: and the call you did? | 22:11 |
deerishi | https://github.com/deerishi/shogun/blob/shogun3kpca/examples/undocumented/python_modular/preprocessor_kpca.py | 22:11 |
@HeikoS | deerishi: please pastebin the command and the output | 22:11 |
deerishi | python generator.py preprocessor_kpca.py | 22:12 |
@HeikoS | output? | 22:13 |
@HeikoS | deerishi: sorry, but you have to ask somebody else on this, I will go home now, | 22:14 |
deerishi | http://pastebin.com/fJ7BCAtw | 22:14 |
deerishi | can we discuss this tomorrow? | 22:14 |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun | 22:14 | |
@HeikoS | deerishi: busy tomorrow, but we can try | 22:14 |
@HeikoS | you can also ask lisitsyn on this | 22:15 |
deerishi | k | 22:15 |
@HeikoS | but please create a full pastebin on the error | 22:15 |
@HeikoS | have a good day! :) bye | 22:15 |
deerishi | bye. :) | 22:15 |
deerishi | @lisitsyn : der? | 22:16 |
-!- HeikoS [~heiko@nat-180-153.internal.eduroam.ucl.ac.uk] has left #shogun [] | 22:18 | |
deerishi | @lisitsyn : could you help me in the integration test | 22:20 |
deerishi | i am getting an error ERROR generating '../../../tests/integration/tests/preprocessor_kpca0.txt' using 'preprocessor_kpca.py' | 22:20 |
deerishi | what does it mean ? and how can i solv eit | 22:20 |
deerishi | *solve it ? | 22:21 |
deerishi | the code is working fine | 22:22 |
shogun-notifier- | shogun: vladislav.horbatiuk@gmail.com :develop * 43dea2e / tests/unit/converter/ (2 files): https://github.com/shogun-toolbox/shogun/commit/43dea2e27009320993af8531ba646142d7da3ca8 | 22:26 |
shogun-notifier- | shogun: Switched the unit test for any neighborhood preserving DR method from testing Locally Linear Embedding method to testing Isomap (since LLE seems to be not very robust to noise). | 22:26 |
shogun-notifier- | shogun: vladislav.horbatiuk@gmail.com :develop * ffc40a3 / tests/unit/converter/Isomap_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/ffc40a3ff9f985f475900c5b0fe31e874efe6d9c | 22:26 |
shogun-notifier- | shogun: Fixed missing #include's in test Isomap_unittest.cc. | 22:26 |
shogun-notifier- | shogun: Sergey Lisitsyn :develop * c9153ac / tests/unit/converter/ (2 files): https://github.com/shogun-toolbox/shogun/commit/c9153ac8d9517db6cd3292810b5f9482236b20ff | 22:26 |
shogun-notifier- | shogun: Merge pull request #1016 from vladislav-horbatiuk/lleTest-fix | 22:26 |
shogun-notifier- | shogun: | 22:26 |
shogun-notifier- | shogun: Switched the unit test for any neighborhood preserving DR method from sting Locally Linear Embedding method to testing Isomap. | 22:26 |
-!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has quit [Quit: Page closed] | 22:30 | |
-!- sids_aquarius [~sidi_@14.139.122.114] has quit [Read error: Connection reset by peer] | 22:36 | |
-!- sids_aquarius [~sidi_@14.139.122.114] has joined #shogun | 22:39 | |
-!- sids_aquarius [~sidi_@14.139.122.114] has left #shogun [] | 22:42 | |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 22:42 | |
-!- sids_aquarius_ [~sidi_@14.139.122.114] has joined #shogun | 22:47 | |
-!- travis-ci [~travis-ci@ec2-174-129-147-163.compute-1.amazonaws.com] has joined #shogun | 23:04 | |
travis-ci | [travis-ci] it's Sergey Lisitsyn's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/6611377 | 23:04 |
-!- travis-ci [~travis-ci@ec2-174-129-147-163.compute-1.amazonaws.com] has left #shogun [] | 23:04 | |
-!- makokal [~bokal@212.201.44.246] has quit [Quit: makokal] | 23:06 | |
-!- sids_aquarius_ [~sidi_@14.139.122.114] has quit [Read error: Connection reset by peer] | 23:13 | |
-!- sids_aquarius_ [~sidi_@li400-124.members.linode.com] has joined #shogun | 23:13 | |
-!- iglesiasg [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 23:17 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 23:17 | |
-!- Kirsh [~kirshanth@175.157.212.180] has quit [Ping timeout: 256 seconds] | 23:19 | |
-!- makokal [~bokal@pD9EF09DC.dip0.t-ipconnect.de] has joined #shogun | 23:34 | |
--- Log closed Thu Apr 25 00:00:29 2013 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!