IRC logs of #shogun for Friday, 2014-06-20

--- Log opened Fri Jun 20 00:00:40 2014
-!- khalednasr [~k.nasr92@41.69.187.45] has joined #shogun00:16
-!- PirosB3 [~pirosb3@host76-224-dynamic.52-79-r.retail.telecomitalia.it] has joined #shogun00:21
shogun-buildbotbuild #135 of osx2 - python is complete: Failure [failed python modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx2%20-%20python/builds/135  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>00:49
-!- khalednasr [~k.nasr92@41.69.187.45] has quit [Ping timeout: 244 seconds]00:58
-!- PirosB3 [~pirosb3@host76-224-dynamic.52-79-r.retail.telecomitalia.it] has quit [Quit: PirosB3]02:29
-!- kislay [~Abhijeet@27.116.52.184] has joined #shogun03:25
-!- zxtx_ [~zv@bcdc654f.skybroadband.com] has joined #shogun04:26
-!- soumyaC [uid15286@gateway/web/irccloud.com/x-yznspsolqjmzcomh] has quit [Quit: Connection closed for inactivity]04:26
shogun-buildbotbuild #831 of nightly_default is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/831  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>04:33
shogun-buildbotbuild #137 of osx2 - python is complete: Failure [failed examples and unit tests]  Build details are at http://buildbot.shogun-toolbox.org/builders/osx2%20-%20python/builds/137  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>05:39
-!- pickle27 [~pickle27@192-0-136-118.cpe.teksavvy.com] has joined #shogun05:40
kislayhey pickle2706:19
-!- pickle27 [~pickle27@192-0-136-118.cpe.teksavvy.com] has quit [Remote host closed the connection]06:28
-!- zxtx_ [~zv@bcdc654f.skybroadband.com] has quit [Ping timeout: 255 seconds]06:54
@besser82wiking, see: http://www.cmake.org/cmake/help/v3.0/prop_tgt/MACOSX_RPATH.html07:41
@besser82wiking, what's the beef with it?07:41
-!- thor|alf [~thoralf@91-65-141-60-dynip.superkabel.de] has joined #shogun08:23
thor|alfHappy Friday everybody.08:24
-!- thor|alf [~thoralf@91-65-141-60-dynip.superkabel.de] has quit [Quit: Konversation terminated!]08:29
-!- HeikoS [~heiko@90.201.233.229] has joined #shogun08:48
-!- mode/#shogun [+o HeikoS] by ChanServ08:48
-!- abinash_panda [~abinash@117.200.236.101] has joined #shogun09:05
-!- abinash_panda [~abinash@117.200.236.101] has quit [Quit: leaving]09:19
-!- abinash_panda [~abinash@117.200.236.101] has joined #shogun09:20
thoralfHey abinash_panda09:42
-!- zxtx_ [~zv@194.69.98.191] has joined #shogun10:17
-!- zxtx_ [~zv@194.69.98.191] has quit [Ping timeout: 255 seconds]10:22
-!- HeikoS [~heiko@90.201.233.229] has quit [Quit: Leaving.]10:30
-!- abinash_panda [~abinash@117.200.236.101] has quit [Ping timeout: 244 seconds]10:32
-!- zxtx_ [~zv@userq187.uk.uudial.com] has joined #shogun10:36
-!- PirosB3 [~pirosb3@host76-224-dynamic.52-79-r.retail.telecomitalia.it] has joined #shogun10:59
-!- abinash_panda [~abinash@117.200.237.89] has joined #shogun11:03
@wikingbesser82: 3.0 started to scream about it11:26
@besser82MACOSX_RPATH11:27
@besser82Whether to use rpaths on Mac OS X.11:27
@besser82When this property is set to true, the directory portion of the "install_name" field of shared libraries will be @rpath unless overridden by INSTALL_NAME_DIR. Runtime paths will also be embedded in binaries using this target and can be controlled by the INSTALL_RPATH target property. This property is initialized by the value of the variable CMAKE_MACOSX_RPATH if it is set when a target is created.11:27
@besser82Policy CMP0042 was introduced to change the default value of MACOSX_RPATH to ON. This is because use of @rpath is a more flexible and powerful alternative to @executable_path and @loader_path.11:27
@besser82wiking, ^^11:27
@wikingbesser82: oh yeah one more thing shogun-ca.cloud.fedoraproject.org11:27
@wikingbesser82: how can i enable new ports?11:27
@besser82wiking, open them in firewalld  ;)11:27
@wikingbesser82: i've tried to update/etc/sysconfig/iptables11:27
@besser82wiking, open them in firewalld  ;)11:28
@wikingps aux|grep firewall11:28
@wikingroot     24751  0.0  0.0 112624   936 pts/0    S+   09:28   0:00 grep --color=auto firewall11:28
@wikingbesser82: i've added the rule by hand to iptables11:28
@wikingno luck :(11:28
@besser82wiking, no firewalld runing?11:29
@wiking-A INPUT -m conntrack --ctstate NEW -m tcp -p tcp --dport 2375 -j ACCEPT11:29
@besser82mhhh...11:29
@wikingbut that didn't help at all11:29
@wikingi have a feeling that it's filtered out somewhere else11:29
@wiking2014/06/20 09:29:27 Listening for HTTP on tcp (0.0.0.0:2375)11:29
@wikingbut if you try to telnet shogun-ca.cloud.fedoraproject.org 237511:30
@besser82wiking, let me discuss this on #fedora-admin...11:30
@wikingTrying 209.132.184.157...11:30
@wikingand that's all11:30
@wikingso imo it's really filtered somewhere else11:30
@wikingwould be great to have any port enabled other than 22 ;P11:30
@wikingi'm good with 80, 443 or basically any port11:30
@besser82you want to have httpd on that?11:31
@wikingbesser82: no i want dockerd on it11:31
@besser82or which kind of service?11:31
@wikingdocker11:31
@besser82allrighty...11:31
@wikingdocker's rest api11:31
@wikingany tcp port is good for me11:31
@wikingother than 2211:31
@besser82lemme check with the guys in #fedora-admin11:31
@wikingbesser82: kk thnx heaps11:31
@besser82wiking, np  ;)11:33
-!- zxtx_ [~zv@userq187.uk.uudial.com] has quit [Ping timeout: 264 seconds]12:01
-!- sejDino [~dino@gatsby.ucl.ac.uk] has left #shogun []12:32
-!- khalednasr [~k.nasr92@41.69.219.168] has joined #shogun13:02
-!- kislay [~Abhijeet@27.116.52.184] has quit [Ping timeout: 240 seconds]13:16
-!- abinash_panda [~abinash@117.200.237.89] has quit [Ping timeout: 272 seconds]13:40
-!- asdf1111 [~user@81.180.113.221] has joined #shogun14:21
-!- asdf1111 [~user@81.180.113.221] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)]14:26
-!- PirosB3 [~pirosb3@host76-224-dynamic.52-79-r.retail.telecomitalia.it] has quit [Quit: PirosB3]14:26
-!- Saurabh7 [~Saurabh7@117.200.188.221] has joined #shogun14:29
-!- pickle27 [~pickle27@192-0-136-118.cpe.teksavvy.com] has joined #shogun14:50
-!- HeikoS [~heiko@90.201.233.229] has joined #shogun14:59
-!- mode/#shogun [+o HeikoS] by ChanServ14:59
-!- pickle27 [~pickle27@192-0-136-118.cpe.teksavvy.com] has quit [Remote host closed the connection]15:04
-!- zxtx_ [~zv@userq187.uk.uudial.com] has joined #shogun15:10
-!- zxtx_ [~zv@userq187.uk.uudial.com] has quit [Ping timeout: 240 seconds]15:15
-!- zxtx_ [~zv@userq135.uk.uudial.com] has joined #shogun15:15
-!- PirosB3 [~pirosb3@host76-224-dynamic.52-79-r.retail.telecomitalia.it] has joined #shogun15:16
@HeikoSthoralf: around?15:19
-!- kislay [~Abhijeet@103.240.206.171] has joined #shogun15:21
-!- PirosB3 [~pirosb3@host76-224-dynamic.52-79-r.retail.telecomitalia.it] has quit [Quit: PirosB3]15:21
thoralfHeikoS: Yes.15:22
@HeikoSthoralf: hi!15:22
@HeikoScurrently checking the next steps for workshop15:22
@HeikoSwiking: around?15:22
@HeikoSthoralf: next steps are: registration and preliminary schedule15:23
-!- lisitsyn [~qdrgsm@80.252.20.67] has quit [Ping timeout: 264 seconds]15:23
@HeikoSthoralf: Let's face it, the two of us will have to design everything15:23
@HeikoSthoralf: so we should decide on the number of talk slots and make people vote again, as last year15:24
@HeikoSbut less talks this time15:24
@HeikoSthoralf: do you have a minute now to discuss this?15:24
thoralfHeikoS: There is no such thing as a good timing for this. ;)15:25
thoralfSo let's do it.15:25
@HeikoSthoralf: ok :)15:25
@HeikoSso here is the preliminary program:15:26
@HeikoSthree parts:15:26
@HeikoS1.) into 2.)science 3.) cummunity15:26
@HeikoS1.) a) speed dating. b)welcome, introduction, overview, c) shogun overview15:27
@HeikoS2) a) ML intro, b) external talk c) ml1 d) ml215:27
@HeikoS3) a) Shogun's last year b) GSoC & cummunity c) hacking Shogun15:27
@HeikoSdiscussions between the three blocks15:27
@HeikoSthoralf: what do you think?15:27
@HeikoSThen we would have to do the welcome, overview, Shogun's last year, hacking GSoC talks15:28
@HeikoS(maybe fernando can also help=)15:28
-!- lisitsyn [~qdrgsm@80.252.20.67] has joined #shogun15:28
@HeikoSand there are 2 science talks (what to do?) and maybe one external one15:28
@HeikoSah sonney might also be there15:28
thoralfHeikoS: 2) is a big vague - any ideas about this already?15:28
@HeikoSthoralf: last year we had people voting on topics15:29
@HeikoSthoralf: basically, every developer added an item to a list that people could vote on15:29
@HeikoSthoralf: for external talk, we could call for contributions (also for science talks)15:29
thoralfHeikoS: every developer [that plans to attend the workshop]?15:29
@HeikoSthoralf: well yes15:29
@HeikoSyou, me, maybe fernando, mayber sonney15:30
thoralfHeikoS: Hmm.  Okay.15:30
thoralfI can talk about something, but only overview, since my current work doesn't involve shogun.15:31
@HeikoSthoralf: maybe some things on research you have done before?15:31
@HeikoSthoralf: we actually dont need people to vote this time as we are quite constrained on what to talk about15:31
thoralfHeikoS: Of course... as we're currently moving MultiLabel stuff into shogun.15:31
@HeikoSthoralf: so each of us can just choose what to present15:31
@HeikoSthoralf: yeah, so a 40 min talk about this stuff would be cool, ill put that in15:31
@HeikoSthoralf: can you put in a descriptive title into the googledoc or tell me now?15:32
@HeikoSthoralf: we just need rough things for now to send out a program draft to the people to register15:33
thoralfHeikoS: Intro to MultiLabel learning with shogun? ;)15:33
@HeikoSthoralf: nice!15:34
@HeikoSDo you want to talk about other things?15:34
@HeikoSLike GSoC?15:34
@HeikoSOr Shogun's last year?15:34
@HeikoSor both?15:34
@HeikoSsince you did a lot over that yera15:34
@HeikoSyear15:34
thoralfWell, I was fighting with different limitations of SHOGUN, like memory-violations, (old) make builds, solving my problems with shogun without hitting memory limits. ;)15:36
thoralfDoing hashing.15:36
thoralfNothing cutting-edge.15:36
thoralfJust solving "some" "problems"...15:36
thoralfOh, well.  Fighting with sparse interfaces.15:37
thoralfBrr.15:37
thoralfStill big construction site.15:37
@HeikoSthoralf:  I know, what about some more high level things?15:38
@HeikoSlike what kind of GSoC projects we had15:38
@HeikoSwhy15:38
@HeikoSdevelopers who joined15:39
@HeikoSthings being implemented15:39
@HeikoSlike the notebooks, the cloud server, the demos, etc15:39
thoralfWell, I didn't have the time to follow the list or gsoc discussions.  So my view is really limited.15:39
@HeikoSthoralf: the point is, if we want to run the workshop, we should provide som ekind of high-level overview (obviously need to prepare this) to have some story in case none of the others can do this. If we don't have time/mindset to do this, it might be quite difficult ot run the workshop15:42
@HeikoSthoralf: I cannot lead this alone, so would depend on your help15:43
thoralfHeikoS: I can only offer lower levels.  Setting up shogun; how to create own project/classes outside shogun (most of the people are hacking their stuff in examples, but I don't like this approach).15:46
@HeikoSthoralf: what about talking a bit about GSoC from your side? Introduce the program, your project, how things work, present the results? And maybe some of the other studen'ts projects (like one or two more)15:47
thoralfSetting up IDE with shogun is a bit bumpy and finally didn't work out for me (doesn't support my workflows).15:47
@HeikoSthoralf: but the above is already great15:47
@HeikoSthoralf: basic intro to hacking  Shognu would already be amazing15:47
@HeikoSthoralf: like walk through a few examples, explain interfaces, blablabla15:47
thoralfTools for debugging shogun is also on my expertise list ;)15:47
@HeikoSthoralf: what was the time for hands-on again?15:47
thoralf10-1815:47
thoralfMonday the 27th iirc15:48
thoralfSunday.15:48
thoralfYes.  Sunday the 27th iirc15:48
@HeikoSok15:49
@HeikoSthoralf: just writing the call for participation15:49
@HeikoSwiking: could you set up a registration page as last year?15:50
thoralfSo I can do the hacking part.  Setting up a project, developing a leaking toy project, debugging it, maybe also telling something about refcounting.  Things like this.15:50
@HeikoSthoralf: great!15:50
@HeikoSthoralf: ok we dont need to fix details now, but first have a rough idea of the program in order  to send out the announcement15:50
@HeikoSthoralf: then we can figure details15:50
@HeikoSbut for now just need something to attract people15:50
@HeikoSI goota go to a talk now, but back later to discss. Editing the document now, please also do some updates on it on whatever you feel is good15:51
thoralfI'm a bit limited today.. but I'll try to do something tonight.15:52
thoralfWill you be here later?  When?15:52
@HeikoSthoralf: maybe later today, if not Sunday.15:53
@HeikoSthoralf: lets aim to send out the mail this weekend15:53
@HeikoSthoralf: I have some more time on Sunday15:54
@HeikoSand later today15:54
@HeikoSin 2 hrs15:54
@HeikoSgotta go now, see you15:54
thoralfOkay, Sunday is fine as well.15:54
thoralfOr just use mail.15:54
-!- HeikoS [~heiko@90.201.233.229] has quit [Ping timeout: 264 seconds]16:00
-!- pickle27 [~pickle27@216.191.231.218] has joined #shogun16:06
-!- soumyaC [uid15286@gateway/web/irccloud.com/x-gqgxttctpstzgzti] has joined #shogun16:07
-!- Saurabh7 [~Saurabh7@117.200.188.221] has quit [Ping timeout: 264 seconds]16:10
-!- Saurabh7 [~Saurabh7@117.222.12.36] has joined #shogun16:21
@wikingbesser82: news on the fw?16:26
-!- Saurabh7 [~Saurabh7@117.222.12.36] has quit [Ping timeout: 264 seconds]16:39
kislay what am I doing wrong here http://pastebin.com/xD8YCHR2 ? can anyone point me to it.16:52
-!- PirosB3 [~pirosb3@host76-224-dynamic.52-79-r.retail.telecomitalia.it] has joined #shogun16:52
kislayhey lisitsyn1 , there?16:54
lisitsyn1kislay: yeah16:55
kislaylisitsyn, if possible..could you see whats wrong in here http://pastebin.com/xD8YCHR216:56
kislaylisitsyn1,                                                                     ^^^^^^^^^^^^^^^^^^^^16:56
lisitsyn1kislay: the matrix is not reversible as its rows are linearly dependent16:57
kislaylisitsyn1, But why does matlab gives me a nice output?16:57
lisitsyn1what kind of output?16:58
kislayit is giving me pinv() output16:58
kislaywhich is definites16:58
lisitsyn1http://www.wolframalpha.com/input/?i=inverse+%5B%5B1%2C2%2C3%5D%2C%5B4%2C5%2C6%5D%2C%5B7%2C8%2C9%5D%5D16:58
kislay0.0051    0.0102    0.0153 ...16:58
lisitsyn1ah so you mean our pinv doesn't work16:59
lisitsyn1well it could be :)16:59
kislayyeah.16:59
lisitsyn1I guess it would be better if it had det that is small but not zero16:59
kislayum, How to calculate inverse then? , any idea.17:00
lisitsyn1kislay: why do you need an inverse of a singular matrix? :)17:01
kislaylisitsyn1, actually.. I am eigenizing the LDA class.17:02
lisitsyn1kislay: hehe well you've got to regularize it or so17:02
kislaylisitsyn1, and there, I am not able to calculate the inverse of the scatter matrix.17:03
lisitsyn1kislay: well it won't work for these features then :)17:03
kislaylisitsyn1,  Even EIgen doesn't have a direct api for calculating pseduinv.17:04
lisitsyn1kislay: I wouldn't bother with that17:04
lisitsyn1kislay: actually they have moore-penrose thing in their FAQ17:04
kislaylisitsyn1, can it be used here!17:05
kislaylisitsyn1, okay. I saw that before. Let me try that then. thanks.17:06
kislaylisitsyn, you think I put this pinv thing as a issue?17:06
lisitsyn1kislay: well not really an issue but you don't really need pseudo inverse here I think17:08
lisitsyn1I mean that's ok your covariance is singular17:08
lisitsyn1you just reduce these features then ;)17:08
kislayokay.17:08
-!- lisitsyn1 [~lisitsin@mxs.kg.ru] has quit [Quit: Leaving.]17:14
pickle27kislay, I remember the days of fighting eigen vs matlab17:19
kislaypickle27, what is that? :)17:19
pickle27matlab does some stuff sometimes and its not super clear what its doing behind the scenes17:19
pickle27to get the same result in eigen you need to understand the math more - there is a lot of ways to do eigen values and inv in eigen17:20
pickle27like all those crazy classes whose names I don't remember17:20
kislaypickle27, more maths!17:20
pickle27yeah lol17:20
pickle27unfortunately Im not that math savvy anymore17:20
pickle27hopefully what lisitsyn mentioned is helpful to you right now17:21
kislaypickle27, I am hoping!17:21
pickle27otherwise I'd try one of the other eigen classes and see if it works for you17:21
kislaypickle27, that would be great.17:21
-!- HeikoS [~heiko@nat-160-69.internal.eduroam.ucl.ac.uk] has joined #shogun17:26
-!- mode/#shogun [+o HeikoS] by ChanServ17:26
@HeikoSthoralf: back17:38
-!- pickle27 [~pickle27@216.191.231.218] has quit [Quit: This computer has gone to sleep]17:59
-!- khalednasr [~k.nasr92@41.69.219.168] has quit [Quit: Leaving]18:35
-!- pickle27 [~pickle27@216.191.231.218] has joined #shogun18:42
@wikingHeikoS: http://www.cs.cmu.edu/~sahong/docs/jegonzal_sahong_final.pdf18:43
-!- PirosB3 [~pirosb3@host76-224-dynamic.52-79-r.retail.telecomitalia.it] has quit [Quit: PirosB3]18:44
-!- PirosB3 [~pirosb3@host76-224-dynamic.52-79-r.retail.telecomitalia.it] has joined #shogun18:53
-!- PirosB3 [~pirosb3@host76-224-dynamic.52-79-r.retail.telecomitalia.it] has quit [Ping timeout: 240 seconds]18:57
@HeikoSwiking: checking19:00
thoralfHeikoS: http://help.eventbrite.de/customer/de/portal/articles/42860419:00
@HeikoSwiking: best thing is to do variational inference19:00
@HeikoSwiking: fastest, also SGD possible19:00
@HeikoSthoralf: nice!19:00
@HeikoSso its free19:00
@HeikoSthoralf: gotta run off now, will work on document later today, and then also be available tomorrow19:00
@HeikoSthanks for the chat, was nice talking after a year ;)19:01
@HeikoShave a nice evening19:01
-!- HeikoS [~heiko@nat-160-69.internal.eduroam.ucl.ac.uk] has left #shogun []19:01
-!- zxtx_ [~zv@userq135.uk.uudial.com] has quit [Ping timeout: 264 seconds]19:08
-!- iglesiasg [524ae0a7@gateway/web/freenode/ip.82.74.224.167] has joined #shogun20:19
-!- iglesiasg [524ae0a7@gateway/web/freenode/ip.82.74.224.167] has quit [Quit: Page closed]21:13
-!- kislay [~Abhijeet@103.240.206.171] has quit [Ping timeout: 264 seconds]22:02
-!- pickle27 [~pickle27@216.191.231.218] has quit [Quit: This computer has gone to sleep]22:14
-!- thor|alf [~thoralf@91-65-141-60-dynip.superkabel.de] has joined #shogun22:48
thor|alfHeyhey.22:50
-!- khalednasr [~k.nasr92@41.69.219.168] has joined #shogun23:07
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun23:21
shogun-notifier-shogun: jiaolong :develop * c189eaa / / (7 files): https://github.com/shogun-toolbox/shogun/commit/c189eaa3feaecfb84a790b3ee8aa9cce08fffeb923:21
shogun-notifier-shogun: added graph cut and unit-test23:21
shogun-notifier-shogun: tklein23 :develop * ea676cb / / (7 files): https://github.com/shogun-toolbox/shogun/commit/ea676cb949e427bf3085190c6751d45228b3118723:21
shogun-notifier-shogun: Merge pull request #2319 from Jiaolong/graph_cut23:21
shogun-notifier-shogun:23:21
shogun-notifier-shogun: added graph cut and unit-test23:21
shogun-buildbotbuild #2386 of bsd1 - libshogun is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/2386  blamelist: jiaolong <jiaolongxu@gmail.com>, tklein23 <tklein23@users.noreply.github.com>23:29
-!- travis-ci [~travis-ci@ec2-23-22-116-186.compute-1.amazonaws.com] has joined #shogun23:37
travis-ci[travis-ci] it's tklein23'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/2807794723:37
-!- travis-ci [~travis-ci@ec2-23-22-116-186.compute-1.amazonaws.com] has left #shogun []23:37
shogun-buildbotbuild #359 of deb4 - python3 is complete: Failure [failed test python modular]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb4%20-%20python3/builds/359  blamelist: jiaolong <jiaolongxu@gmail.com>, tklein23 <tklein23@users.noreply.github.com>23:46
shogun-buildbotbuild #2376 of deb3 - modular_interfaces is complete: Failure [failed configure]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/2376  blamelist: jiaolong <jiaolongxu@gmail.com>, tklein23 <tklein23@users.noreply.github.com>23:47
--- Log closed Sat Jun 21 00:00:41 2014

Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!