--- Log opened Fri May 16 00:00:41 2014 | ||
--- Day changed Fri May 16 2014 | ||
-!- pickle27f [~pickle27@216.191.231.218] has quit [Quit: This computer has gone to sleep] | 00:00 | |
-!- juliohm [~julio@unaffiliated/juliohm] has quit [Quit: Leaving] | 02:18 | |
shogun-buildbot | build #5 of nightly_fedora is complete: Failure [failed git] Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_fedora/builds/5 blamelist: lambday <heavensdevil6909@gmail.com>, Soumyajit De <heavensdevil6909@gmail.com> | 03:00 |
---|---|---|
-!- pickle27 [~pickle27@192-0-136-118.cpe.teksavvy.com] has joined #shogun | 03:23 | |
-!- kislay [~Abhijeet@223.186.134.119] has quit [Read error: Connection reset by peer] | 03:30 | |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 03:44 | |
shogun-notifier- | shogun-web: Kevin Hughes :master * b7e6feb / static/css/styles.css: https://github.com/shogun-toolbox/shogun-web/commit/b7e6feb100a1c8385da2653a7536d39913128d66 | 03:44 |
shogun-notifier- | shogun-web: comment out gsoc image in css while the box is about europython | 03:44 |
-!- PirosB3 [~pirosb3@cpc4-newt31-2-0-cust631.newt.cable.virginm.net] has quit [Quit: PirosB3] | 04:01 | |
-!- soumyaC [uid15286@gateway/web/irccloud.com/x-pvvvkzahjfllvupe] has quit [Quit: Connection closed for inactivity] | 04:16 | |
shogun-buildbot | build #803 of nightly_default is complete: Failure [failed notebooks] Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/803 blamelist: lambday <heavensdevil6909@gmail.com>, Soumyajit De <heavensdevil6909@gmail.com> | 04:34 |
-!- pickle27 [~pickle27@192-0-136-118.cpe.teksavvy.com] has quit [Remote host closed the connection] | 06:30 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 06:44 | |
-!- kislay [~Abhijeet@106.221.159.105] has joined #shogun | 07:19 | |
-!- kislay [~Abhijeet@106.221.159.105] has quit [Read error: Connection reset by peer] | 08:00 | |
-!- jiaolong [9e6d09e4@gateway/web/freenode/ip.158.109.9.228] has joined #shogun | 09:21 | |
-!- soumyaC [uid15286@gateway/web/irccloud.com/x-enlorqhudqemkzis] has joined #shogun | 10:29 | |
-!- zxtx [~zxtx@sunnythinking.org] has quit [Ping timeout: 258 seconds] | 11:25 | |
-!- zxtx [~zxtx@sunnythinking.org] has joined #shogun | 11:43 | |
-!- PirosB3 [~pirosb3@cpc4-newt31-2-0-cust631.newt.cable.virginm.net] has joined #shogun | 11:59 | |
-!- zxtx [~zxtx@sunnythinking.org] has quit [Ping timeout: 258 seconds] | 12:01 | |
-!- kislay [~Abhijeet@110.226.208.72] has joined #shogun | 12:15 | |
-!- zxtx [~zxtx@sunnythinking.org] has joined #shogun | 12:17 | |
-!- PirosB3 [~pirosb3@cpc4-newt31-2-0-cust631.newt.cable.virginm.net] has quit [Quit: PirosB3] | 12:18 | |
-!- PirosB3 [~pirosb3@cpc4-newt31-2-0-cust631.newt.cable.virginm.net] has joined #shogun | 12:48 | |
-!- kislay [~Abhijeet@110.226.208.72] has quit [Quit: Leaving] | 13:32 | |
-!- pickle27 [~pickle27@192-0-136-118.cpe.teksavvy.com] has joined #shogun | 14:44 | |
-!- pickle27 [~pickle27@192-0-136-118.cpe.teksavvy.com] has quit [Remote host closed the connection] | 14:49 | |
-!- khalednasr [~k.nasr92@41.69.208.89] has joined #shogun | 15:14 | |
@besser82 | thoralf, hey! | 15:42 |
thoralf | Hey! | 15:43 |
-!- khalednasr [~k.nasr92@41.69.208.89] has quit [Quit: Leaving] | 15:43 | |
--- Log closed Fri May 16 15:46:00 2014 | ||
--- Log opened Fri May 16 15:46:09 2014 | ||
-!- shogun-toolbox [~shogun@7nn.de] has joined #shogun | 15:46 | |
-!- Irssi: #shogun: Total of 14 nicks [3 ops, 0 halfops, 0 voices, 11 normal] | 15:46 | |
-!- Irssi: Join to #shogun was synced in 6 secs | 15:46 | |
-!- naywhayare [~ryan@spoon.lugatgt.org] has joined #shogun | 15:46 | |
-!- pickle27f [~pickle27@216.191.231.218] has joined #shogun | 15:53 | |
-!- kislay [~Abhijeet@106.221.144.144] has joined #shogun | 16:28 | |
sonne|work | pickle27f: can you upload a fixed website? | 16:33 |
pickle27f | I pushed the fix for the css | 16:34 |
pickle27f | I don't know what text is supposed to be there | 16:34 |
pickle27f | besser82, you around? | 16:36 |
pickle27f | kislay, has a PR that adds some stuff to cmake | 16:36 |
@besser82 | pickle27f, yo! | 16:37 |
@besser82 | pickle27f, which one? | 16:37 |
pickle27f | besser82, https://github.com/shogun-toolbox/shogun/pull/2221#issuecomment-43144697 | 16:37 |
@besser82 | pickle27f, k, thx! ^^ | 16:37 |
pickle27f | np thanks for taking a look! | 16:37 |
@besser82 | pickle27f, np =) | 16:37 |
pickle27f | sonne|work, what should the text be? | 16:40 |
sonne|work | pickle27f: there is some text in the .html | 16:40 |
sonne|work | just flickering and going away | 16:40 |
sonne|work | I guess that but fernando would know | 16:40 |
pickle27f | its loading fine for me | 16:41 |
pickle27f | sonne|work, http://monosnap.com/image/k8SgLMgnkTNBYjzUVf0s6ykCl4uX9b here is what I see | 16:43 |
@besser82 | pickle27f, just commented on the PR | 16:44 |
pickle27f | awesome thanks! | 16:46 |
pickle27f | I'll have kislay address your comments and then anyone can merge it | 16:46 |
@besser82 | pickle27f, allrighty! thx! Lemme know, when it's ready. I'll have another look then && merge | 16:47 |
pickle27f | kk! | 16:47 |
@besser82 | pickle27f, btw. you did do anything related to NLP? | 16:47 |
pickle27f | can't say I have | 16:48 |
@besser82 | pickle27f, kk | 16:48 |
@besser82 | pickle27f, I'm just looking for someone, who can show me how to conv LDA-model into lLDA / sLDA... | 16:48 |
@besser82 | and clarify the most signficant differences... | 16:49 |
@besser82 | read lots of papers but I'm not smarter than before :( | 16:49 |
@besser82 | just eye-sore :( | 16:49 |
sonne|work | pickle27f: c'mon look at the source code | 16:49 |
@besser82 | sonne|work, hey! | 16:50 |
@besser82 | sonne|work, can you tell me the significant difference between lLDA and sLDA? | 16:50 |
pickle27f | sonne|work, pooh how did I mess that before ... | 16:50 |
@besser82 | and possibly MedLDA? | 16:50 |
pickle27f | I'll have to fix that when I get home later slash maybe tomorrow but yeah I'll push something up | 16:51 |
kislay | pickle27f, see if it's okay. I have removed the system include for opencv from CMakeLists.txt | 17:22 |
pickle27f | kislay, looks good and it still works right? - did you try a clean build? | 17:23 |
kislay | pickle27f, yeah! np with that! | 17:24 |
pickle27f | great! | 17:25 |
pickle27f | besser82, I think we are good now! | 17:25 |
pickle27f | just waiting on travis | 17:25 |
@besser82 | pickle27f, lgtm ;) will merge after travis run is clean | 17:34 |
@besser82 | kislay, could you remove the mentioned cluttering whitespace, too? | 17:36 |
@besser82 | kislay, plz... | 17:36 |
kislay | besser82, sure sure. | 17:36 |
@besser82 | kislay, fine =) let's w8 on travis then | 17:36 |
-!- thoralf [~thoralf@91-66-177-150-dynip.superkabel.de] has quit [Quit: Konversation terminated!] | 17:38 | |
-!- pickle27f [~pickle27@216.191.231.218] has quit [Quit: This computer has gone to sleep] | 18:07 | |
-!- pickle27f [~pickle27@216.191.231.218] has joined #shogun | 18:23 | |
-!- pickle27f [~pickle27@216.191.231.218] has quit [Quit: This computer has gone to sleep] | 18:51 | |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 18:58 | |
shogun-notifier- | shogun: Abhijeet :develop * ece3607 / CMakeLists.txt,tests/integration/opencv/opencv_test.cpp: https://github.com/shogun-toolbox/shogun/commit/ece3607e5eaa4b13511cd72342e7ad5ce5fa8ff8 | 18:58 |
shogun-notifier- | shogun: added opencv as the optional dependency and a basic test in tests/integration/opencv. | 18:58 |
shogun-notifier- | shogun: Abhijeet :develop * 7968225 / CMakeLists.txt: https://github.com/shogun-toolbox/shogun/commit/79682251489732a0f9e0c6f8fc2fb91e68b98731 | 18:58 |
shogun-notifier- | shogun: added flags for opencv integration in CMakeLists.txt. | 18:58 |
shogun-notifier- | shogun: Abhijeet :develop * 2dc8e4d / CMakeLists.txt: https://github.com/shogun-toolbox/shogun/commit/2dc8e4db5bf191b20ae3b7935ee129d0b2ed5a32 | 18:58 |
shogun-notifier- | shogun: shortened the flag name for OpenCV integration | 18:58 |
shogun-notifier- | shogun: Abhijeet :develop * 1210365 / tests/integration/opencv/opencv_test.cpp: https://github.com/shogun-toolbox/shogun/commit/121036515a5224644a90b5cda88563da3d424006 | 18:58 |
shogun-notifier- | shogun: opencv integration test cleaned_. | 18:58 |
shogun-notifier- | shogun: Abhijeet :develop * 177c519 / CMakeLists.txt: https://github.com/shogun-toolbox/shogun/commit/177c5199184c97b34c4f8227fee9e85acbf4b606 | 18:58 |
shogun-notifier- | shogun: removed system include for OpenCV from CMakeLists.txt | 18:58 |
shogun-notifier- | shogun: Abhijeet :develop * 5aa2b89 / CMakeLists.txt: https://github.com/shogun-toolbox/shogun/commit/5aa2b89e7cd5c80e0ecadf6ae01e46d92b0e41bb | 18:58 |
shogun-notifier- | shogun: cleaned whitespaces. | 18:58 |
shogun-notifier- | shogun: Bj?rn Esser :develop * 96b815f / CMakeLists.txt,tests/integration/opencv/opencv_test.cpp: https://github.com/shogun-toolbox/shogun/commit/96b815fd1fa9769a24122f9016ff5a685a8a6944 | 18:58 |
shogun-notifier- | shogun: Merge pull request #2221 from kislayabhi/develop | 18:58 |
@besser82 | kislay, merged ^^ | 18:59 |
kislay | besser82, yeah! :) | 18:59 |
shogun-buildbot | build #2946 of deb1 - libshogun is complete: Success [build successful] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/2946 | 19:04 |
-!- pickle27f [~pickle27@216.191.231.218] has joined #shogun | 19:34 | |
pickle27f | thanks for the merge! | 19:34 |
pickle27f | who do kislay and I talk to about getting an OpenCV build bot? slash can we just add to Travis? | 19:35 |
kislay | yeah. wo do we talk :) | 19:35 |
pickle27f | I think we can add to Travis our selves if you want to look into that - you'll need to edit the travis.yml file in the repo | 19:40 |
pickle27f | I don't know about the other bots | 19:41 |
pickle27f | slash I don't even know how or who configures those | 19:41 |
-!- thoralf [~thoralf@46.115.128.90] has joined #shogun | 20:19 | |
thoralf | Hello! | 20:19 |
-!- PirosB3 [~pirosb3@cpc4-newt31-2-0-cust631.newt.cable.virginm.net] has quit [Quit: PirosB3] | 20:45 | |
-!- thoralf [~thoralf@46.115.128.90] has quit [Read error: Connection reset by peer] | 20:45 | |
-!- besser82 [quassel@fedora/besser82] has left #shogun ["Verlassend"] | 20:45 | |
-!- besser82 [quassel@fedora/besser82] has joined #shogun | 20:46 | |
-!- mode/#shogun [+o besser82] by ChanServ | 20:46 | |
-!- thoralf [~thoralf@46.115.128.90] has joined #shogun | 20:46 | |
-!- kislay [~Abhijeet@106.221.144.144] has quit [Ping timeout: 265 seconds] | 20:52 | |
-!- thoralf [~thoralf@46.115.128.90] has quit [Quit: Konversation terminated!] | 20:54 | |
shogun-buildbot | build #301 of debian wheezy - memcheck is complete: Failure [failed memory check] Build details are at http://buildbot.shogun-toolbox.org/builders/debian%20wheezy%20-%20memcheck/builds/301 blamelist: Abhijeet <abhijeetkislay@gmail.com>, Bj?rn Esser <bjoern.esser@gmail.com> | 21:35 |
-!- PirosB3 [~pirosb3@cpc4-newt31-2-0-cust631.newt.cable.virginm.net] has joined #shogun | 21:57 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 21:58 | |
-!- pickle27f [~pickle27@216.191.231.218] has quit [Quit: This computer has gone to sleep] | 22:04 | |
-!- pickle27f [~pickle27@216.191.231.218] has joined #shogun | 22:05 | |
-!- PirosB3 [~pirosb3@cpc4-newt31-2-0-cust631.newt.cable.virginm.net] has quit [Quit: PirosB3] | 22:32 | |
-!- abinash_panda [0e8b2996@gateway/web/freenode/ip.14.139.41.150] has joined #shogun | 23:05 | |
-!- pickle27f [~pickle27@216.191.231.218] has quit [Quit: This computer has gone to sleep] | 23:07 | |
-!- pickle27f [~pickle27@216.191.231.218] has joined #shogun | 23:10 | |
-!- thoralf [~thoralf@p4FECB862.dip0.t-ipconnect.de] has joined #shogun | 23:11 | |
abinash_panda | thoralf: Hi ! | 23:11 |
thoralf | Yo. | 23:11 |
thoralf | Hey :) | 23:11 |
abinash_panda | thoralf: So, I was going through the implementation of CMulticlassModel for getting to know more about the implementation of get_joint_feature_vector method. | 23:14 |
abinash_panda | thoralf: Do the implementation of get_joint_feature_vector would be somewhat like this is CMultilabelModel ? | 23:16 |
thoralf | abinash_panda: Of course - because get_joint_feature_vector() is very central in defining a structured model. | 23:17 |
thoralf | We're interested in very efficient implementation, so suggestions are welcome. | 23:18 |
abinash_panda | thoralf: Ok. | 23:18 |
thoralf | abinash_panda: Any ideas? | 23:19 |
abinash_panda | thoralf: I was going through the paper http://machinelearning.wustl.edu/mlpapers/paper_files/TsochantaridisJHA05.pdf for getting to know more about joint feature map. | 23:20 |
abinash_panda | thoralf: It has some special cases of structured output learning as examples. | 23:21 |
thoralf | Yeah, this paper is great. | 23:21 |
abinash_panda | thoralf: So, I what I think that the implementation of get_joint_feature_vector() would be the tensor product of input and output labels. Right ? | 23:22 |
thoralf | tensor product of label-vector and feature-vector, but yes. | 23:23 |
abinash_panda | thoralf: Yeah. I meant the same. | 23:23 |
thoralf | Okay :) | 23:24 |
abinash_panda | thoralf: Now, regarding the argmax function. | 23:24 |
abinash_panda | thoralf: I am still confused. Could you please help me out | 23:24 |
thoralf | Shoot. :) | 23:25 |
abinash_panda | thoralf: You have mentioned in the mail, for the initial implementation we can treat the labels independently. | 23:28 |
thoralf | Yes. | 23:28 |
abinash_panda | thoralf: I am still unable to understand its implementation. | 23:30 |
thoralf | Hmm. Okay. | 23:34 |
thoralf | Imagine you have one input vector and want to predict a binary label vector. | 23:34 |
abinash_panda | thoralf: Ok | 23:34 |
thoralf | Let's say 3 labels, like (0,0,1) | 23:34 |
abinash_panda | Ok | 23:36 |
thoralf | For one input x you have 8 possible outputs (000), (001), (010), ..., (111). You have a scoring function f(x,y) which tells you a score for (input, possible outputs)... and you want to find the "best" output labeling. | 23:36 |
thoralf | best means, highest score. | 23:36 |
thoralf | But you don't want to "brute force" them. | 23:36 |
thoralf | You want some "clever" decoding. | 23:37 |
thoralf | Something that is cheaper than trying all 2^n combinations. | 23:37 |
thoralf | And that has to be implemented in argmax. | 23:37 |
thoralf | "Find the best labeling" | 23:37 |
abinash_panda | Ok. | 23:38 |
thoralf | Do you have an idea how to implement this multilabel stuff with binary classifiers? | 23:38 |
abinash_panda | We can use binary relevance method, training one binary classifier for each label. Right? | 23:40 |
thoralf | Yes. | 23:42 |
thoralf | And then we ask the i-th classifier for the i-th label. | 23:42 |
thoralf | The cost is linear in the number of labels, which is totally fine. :) | 23:43 |
abinash_panda | Ok. So what I think that we can implement the same in the argmax function. | 23:43 |
thoralf | Exact. | 23:45 |
thoralf | But instead of training n classifiers, we train one classifier over the joint-feature-space. | 23:46 |
thoralf | You know what I mean? | 23:46 |
-!- pickle27f [~pickle27@216.191.231.218] has quit [Quit: This computer has gone to sleep] | 23:46 | |
abinash_panda | Yeah. We would be training only one classifier. | 23:47 |
thoralf | Then you probably know how to transform label-vector and feature-vector to form the joint-feature-vector? | 23:48 |
thoralf | Sorry for asking dumb questions. I'm trying cover all possible problems. ;) | 23:49 |
abinash_panda | In the argmax method, for N labels we can go for computing the argmax N number of times and would return only those labels for whose the output is 1. | 23:51 |
abinash_panda | Am I able to make myself clear | 23:52 |
thoralf | Yeah, I think so. | 23:53 |
abinash_panda | Ok. :) | 23:53 |
thoralf | It's even easier: Output only labels which decision value is positive. | 23:53 |
abinash_panda | Yeah, thats better :) | 23:53 |
thoralf | ">= 0" for true, "< 0" for false. | 23:53 |
abinash_panda | Ok. | 23:54 |
thoralf | Btw., if you like to, you can also implement a multilabel class which uses N independent binary SVMs. This sounds strange, but you can learn them all in parallel. :) | 23:55 |
abinash_panda | Yeah, sure. I would try to implement it. | 23:55 |
thoralf | You get parallel training on many cores for free. | 23:56 |
thoralf | Who many cores does your biggest computer have? ;) | 23:56 |
* thoralf was able to train multilabel with 1000 classes on a weekend. | 23:57 | |
abinash_panda | Currently, my computer have 4 cores ;) | 23:57 |
abinash_panda | thoralf: On how many cores you have implemented this multilabel classification? | 23:58 |
thoralf | Well, the computer had 8 cores (with hyperthreading it was 16 threads) | 23:59 |
thoralf | So scaling by factor 16 for free. | 23:59 |
thoralf | No worrying about how to parallelize SVM solvers. | 23:59 |
--- Log closed Sat May 17 00:00:02 2014 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!