--- Log opened Tue Mar 25 00:00:53 2014 | ||
-!- jiaolong_ [553a5634@gateway/web/freenode/ip.85.58.86.52] has quit [Quit: Page closed] | 00:33 | |
-!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has quit [Quit: Leaving] | 00:40 | |
nikolis | Hey guys I have puted the export PYTHONPATH and export LD_LIBRARY_PATH in .bashrc and everything runs smoothly using python from terminal but I am not able to use modshogun when using an IDE for python anybody got any idea how to make it work ? | 00:43 |
---|---|---|
shogun-buildbot | Hey! build deb3 - modular_interfaces #2252 is complete: Success [build successful] | 00:45 |
shogun-buildbot | Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/2252 | 00:45 |
thoralf | nikolis: You're not starting your python ide from the command line, right? | 00:53 |
nikolis | no cause it is eclipse | 01:15 |
nikolis | but I guess that would be the answer to my question | 01:15 |
nikolis | thoralf: everything runs smooth now thanks a lot for you assistance good night | 01:21 |
thoralf | Starting eclipse from the command line worked? | 01:21 |
nikolis | yeap | 01:22 |
nikolis | just like that :) | 01:22 |
-!- nikolis [~nikolis@213.16.163.237.dsl.dyn.forthnet.gr] has quit [Quit: Leaving] | 01:24 | |
-!- thoralf [~thoralf@91-66-33-156-dynip.superkabel.de] has quit [Quit: Konversation terminated!] | 01:24 | |
-!- kislay [~kislay@117.239.94.112] has quit [Quit: Leaving] | 02:30 | |
-!- PirosB3 [~pirosb3@cpc4-newt31-2-0-cust631.newt.cable.virginm.net] has quit [Quit: PirosB3] | 02:35 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 02:43 | |
-!- andrew___ [~andrew@c-50-184-110-141.hsd1.ca.comcast.net] has quit [Ping timeout: 246 seconds] | 03:24 | |
-!- andrew__ [~andrew@c-50-184-110-141.hsd1.ca.comcast.net] has joined #shogun | 03:25 | |
-!- andrew__ [~andrew@c-50-184-110-141.hsd1.ca.comcast.net] has quit [Ping timeout: 264 seconds] | 03:31 | |
-!- srichakradhar [srichakrad@gateway/shell/bnc4free/x-zgfvhrcxanycceqf] has joined #shogun | 03:39 | |
-!- zxtx [~zv@23-114-213-191.lightspeed.irvnca.sbcglobal.net] has quit [Ping timeout: 268 seconds] | 03:40 | |
-!- chakri [srichakrad@2001:41d0:2:b81b::cafe:0] has left #shogun ["Leaving"] | 03:40 | |
shogun-buildbot | build #757 of nightly_default is complete: Success [build successful] Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/757 | 03:44 |
-!- andrew__ [~andrew@c-50-184-110-141.hsd1.ca.comcast.net] has joined #shogun | 03:51 | |
-!- soumyaC [uid15286@gateway/web/irccloud.com/x-txzwqkotkaqemdvr] has quit [Quit: Connection closed for inactivity] | 05:15 | |
-!- besser82 [quassel@fedora/besser82] has quit [Remote host closed the connection] | 05:17 | |
-!- srichakradhar [srichakrad@gateway/shell/bnc4free/x-zgfvhrcxanycceqf] has quit [Ping timeout: 265 seconds] | 05:19 | |
-!- dhruv13j [dhruv13j@gateway/shell/bnc4free/x-dojotckqqmfmtwvw] has quit [Ping timeout: 265 seconds] | 05:19 | |
-!- besser82 [quassel@fedora/besser82] has joined #shogun | 05:26 | |
-!- mode/#shogun [+o besser82] by ChanServ | 05:26 | |
-!- dhruv13j [dhruv13j@gateway/shell/bnc4free/x-rhdsnafgovxwiqrk] has joined #shogun | 05:27 | |
-!- chakri [srichakrad@gateway/shell/bnc4free/x-ickxjwezybzlwrby] has joined #shogun | 05:35 | |
dhruv13j | wiking: hey! if you are free, I need some assistance with cmake stuff | 06:07 |
-!- abinash_panda [01261312@gateway/web/freenode/ip.1.38.19.18] has joined #shogun | 09:35 | |
-!- lisitsyn1 [~lisitsin@mxs.kg.ru] has joined #shogun | 09:37 | |
-!- travis-ci [~travis-ci@ec2-54-221-152-237.compute-1.amazonaws.com] has joined #shogun | 09:51 | |
travis-ci | [travis-ci] it's Fernando Iglesias'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/21466585 | 09:51 |
-!- travis-ci [~travis-ci@ec2-54-221-152-237.compute-1.amazonaws.com] has left #shogun [] | 09:51 | |
-!- nikolis [~nikolis@213.16.163.237.dsl.dyn.forthnet.gr] has joined #shogun | 09:59 | |
-!- soumyaC [uid15286@gateway/web/irccloud.com/x-srorwsudpqcgzjld] has joined #shogun | 10:32 | |
-!- kislay [~kislay@210.212.61.185] has joined #shogun | 10:34 | |
-!- PirosB3 [~pirosb3@cpc4-newt31-2-0-cust631.newt.cable.virginm.net] has joined #shogun | 10:56 | |
-!- PirosB3 [~pirosb3@cpc4-newt31-2-0-cust631.newt.cable.virginm.net] has quit [Client Quit] | 10:56 | |
nikolis | if anybody can give me a hint on how to use this mycsvObject.get_matrix from python modular or direct me to a documentation page that would be great | 11:22 |
-!- abinash_panda [01261312@gateway/web/freenode/ip.1.38.19.18] has quit [Ping timeout: 245 seconds] | 11:28 | |
-!- DarkRoom [~darkroom@14.139.82.6] has quit [Ping timeout: 264 seconds] | 11:49 | |
-!- DarkRoom [~darkroom@14.139.82.6] has joined #shogun | 11:51 | |
-!- HeikoS1 [~heiko@pat-191-250.internal.eduroam.ucl.ac.uk] has joined #shogun | 12:23 | |
-!- nikolis [~nikolis@213.16.163.237.dsl.dyn.forthnet.gr] has quit [Quit: Leaving] | 12:57 | |
-!- sonney2k [~shogun@94.23.237.10] has quit [Read error: Operation timed out] | 13:34 | |
-!- sonney2k [~shogun@7nn.de] has joined #shogun | 13:34 | |
-!- lambday [67157d54@gateway/web/freenode/ip.103.21.125.84] has joined #shogun | 14:43 | |
-!- lisitsyn [~lisitsyn@80.252.20.67] has quit [Ping timeout: 268 seconds] | 14:59 | |
-!- sunil1337 [0e8bd057@gateway/web/freenode/ip.14.139.208.87] has joined #shogun | 15:42 | |
sunil1337 | lambday: hi! where do we start on this https://github.com/shogun-toolbox/shogun/issues/1973 | 15:43 |
-!- lambday_ [67157d54@gateway/web/freenode/ip.103.21.125.84] has joined #shogun | 15:45 | |
lambday_ | sunil1337: hi | 15:45 |
sunil1337 | lambday_: hi! we were discussing https://github.com/shogun-toolbox/shogun/issues/1973 | 15:46 |
-!- samal [~gosurf@24.213.108.63] has quit [Read error: Connection reset by peer] | 15:46 | |
lambday_ | sunil1337: yeah.. sorry been caught up into some stuffs in between.. | 15:47 |
lambday_ | sunil1337: umm.. I think according to our discussion... we need to change LinearOperator a bit | 15:47 |
sunil1337 | lambday_: :) me too | 15:47 |
-!- lambday [67157d54@gateway/web/freenode/ip.103.21.125.84] has quit [Ping timeout: 245 seconds] | 15:48 | |
lambday_ | sunil1337: so, to put VectorDotOperator under LinearOperator interface, we need to keep LinearOperator's operand and return type templates | 15:48 |
lambday_ | sunil1337: please note that there were a few mistakes in the gist that I pasted | 15:48 |
lambday_ | sunil1337: anyway, so it would be like | 15:49 |
lambday_ | template <class Operand, class RetType> LinearOperator.... | 15:49 |
lambday_ | with RetType apply(Operand operand) | 15:49 |
sunil1337 | lambday_: okay! we were also discussing about introducing a new VectorOperator instead.. as linear was kind of odd for dot product | 15:49 |
lambday_ | sunil1337: VectorDotOperator can be put under LinearOperator as well.. so for vector dot operator, return type would be a scalar T, and operand should be another vector | 15:51 |
lambday_ | and for matrix linear operators, it return type is vector and operand type is also vector | 15:51 |
sunil1337 | lambday_: okay :) | 15:52 |
lambday_ | sunil1337: what do you think about the set_backend() thing? HeikoS1 suggested that we should have different set of backends for dot and other solvers | 15:52 |
lambday_ | the idea is to have a LinearAlgebra class which upon calling set_backend or sort, initializes functors with proper objects... and the linear algebra interface uses those inside | 15:53 |
lambday_ | but I haven't though much about how to take care of the different types.. | 15:54 |
sunil1337 | lambday_: I was thinking of first implementing this in dot product and later extend this to other solvers. | 15:55 |
lambday_ | sunil1337: yeah... but the interface should be ready for all I think.. | 15:55 |
lambday_ | I mean, having the option to set backends and all | 15:55 |
lambday_ | sunil1337: may be wiking has some comments regarding thsi | 15:56 |
sunil1337 | lambday_: yes! we could have a default for each functor like you did in the first gist for eigen3 and when set_backend is called it modifies the functors. | 15:57 |
lambday_ | sunil1337: yeah... please note that this should initialize things globally | 16:00 |
lambday_ | as in, a global instance of LinearAlgebra, say, sg_linalg | 16:01 |
-!- nikolis [~nikolis@213.16.163.237.dsl.dyn.forthnet.gr] has joined #shogun | 16:01 | |
lambday_ | and apply() methods use those to compute things | 16:02 |
lambday_ | sunil1337: I will try to come up with another gist for proof of concept for this.. | 16:02 |
lambday_ | sunil1337: and you're most welcome to give your inputs | 16:03 |
lambday_ | sunil1337: oh I almost forgot, the constructor that you added on LogDetEstimator, it breaks in absense of lapack... basically the lanczos eigensolver requires both eigen3 and lapack.. so I think the whole constructor should be put under both checks | 16:04 |
sunil1337 | lambday_: so when sg_linalg->set_backend is called it calls similar set_backend method in all of these classes(dot product) setting a enum which then uses switch case for calling the appropriate functor keeping the interface simple | 16:04 |
sunil1337 | lambday_: i had added eigen3 check.. will add the lapack check as well:) | 16:04 |
lambday_ | sunil1337: I mean in the class itself :) | 16:05 |
lambday_ | sunil1337: nono... ummm | 16:05 |
lambday_ | sunil1337: I am not entirely sure but I was thinking that may be this LinearAlgebra class has base class pointers for those functors (let's not make those functors but plain classes with just a compute method)... | 16:06 |
lambday_ | sunil1337: and LinearAlgebra::set_backend(ELinAlgBackendType backend_type) initializes these things with appropriate derived instances | 16:07 |
lambday_ | and then inside the apply() or compute() methods in our linear algebra routines, they rely on those pointers instead | 16:08 |
sunil1337 | lambday_: okay! yeah that would avoid all the switch case checking :) | 16:09 |
lambday_ | sunil1337: yeah you just do it once when setting the backend | 16:09 |
lambday_ | later on everything else follows based on what you initialized | 16:09 |
-!- PirosB3 [~pirosb3@cpc4-newt31-2-0-cust631.newt.cable.virginm.net] has joined #shogun | 16:10 | |
lambday_ | sunil1337: the only problem I see in this is how to make this generic...other than making LinearAlgebra generic | 16:10 |
lambday_ | sunil1337: cause we'll be needing these functors for float64, int32, complex128... etc | 16:11 |
sunil1337 | lambday_: hmm... | 16:11 |
lambday_ | sunil1337: and we also may want to differentiate between dot backends and other (eigensolver or so) backends.. | 16:13 |
lambday_ | ah this is getting messy | 16:13 |
sunil1337 | lambday_: yeah!.... but what if we used the switch method instead? that way we could choose what to initialize | 16:13 |
lambday_ | sunil1337: switch method as in? for different types? | 16:14 |
sunil1337 | lambday_: have a enum for different types of backends and use a switch case for calling different backends from VectorDot(v1, v2) | 16:15 |
sunil1337 | lambday_: this way we could make this generic :) | 16:16 |
lambday_ | sunil1337: yeah that's alright but how do you plan to make the functors available for different datatypes that actual apply() or compute() methods use? I mean having different pointers for different datatypes is just horrible :( | 16:18 |
lambday_ | and if you make LinearAlgebra a template, then setting backend will work only for one datatype at a time.. | 16:19 |
sunil1337 | lambday_: i am not sure i get you...i meant LinearAlgebra::set_backend would call the corresponding set_backend of vectordot which will set the enum variable then the apply() would use switch to choose a method for computing the dot product which would be a template | 16:22 |
lambday_ | sunil1337: what I was thinking was to avoid switches inside apply().. | 16:24 |
lambday_ | sunil1337: I mean, the enum thing doesn't have to be stored anyway... | 16:25 |
lambday_ | sunil1337: its just for initializing the base ptr with proper instance | 16:25 |
lambday_ | those baes ptrs are data members of the LinearAlgebra class.. | 16:26 |
lambday_ | base* | 16:26 |
lambday_ | sunil1337: let me think of something and explain it properly on the thread... | 16:28 |
sunil1337 | lambday_: okay! :) will be back after dinner! | 16:28 |
lambday_ | sunil1337: okie | 16:29 |
-!- sunil1337 [0e8bd057@gateway/web/freenode/ip.14.139.208.87] has left #shogun [] | 16:29 | |
-!- kislay [~kislay@210.212.61.185] has quit [Read error: Connection reset by peer] | 16:56 | |
-!- kislay [~kislay@220.227.97.99] has joined #shogun | 16:56 | |
-!- thoralf [~thoralf@91-66-33-156-dynip.superkabel.de] has joined #shogun | 17:13 | |
thoralf | Hey! | 17:13 |
thoralf | jiaolong: Ping? | 17:14 |
jiaolong | thoralf: Hi | 17:15 |
jiaolong | thoralf: Have you seen my email? | 17:15 |
thoralf | jiaolong: Yes, that's why I'm here. | 17:15 |
thoralf | jiaolong: On which system are you developing? | 17:16 |
jiaolong | thoralf: Ubuntu 12.04 | 17:16 |
thoralf | Okay, it's not the newest, but should work find. | 17:16 |
thoralf | Did you try to uninstall the json-dev packages? | 17:17 |
jiaolong | thoralf: No | 17:17 |
thoralf | As we have a suspect now, we can try to isolate the problem. | 17:17 |
jiaolong | thoralf: OK, I can try it | 17:17 |
jiaolong | thoralf: Yes, it seems JASON is the reason | 17:18 |
thoralf | Oh, then remove jason, sorry. ;) | 17:18 |
jiaolong | thoralf: The knn module has been fixed | 17:18 |
jiaolong | thoralf: OK, let me try | 17:18 |
-!- PirosB3 [~pirosb3@cpc4-newt31-2-0-cust631.newt.cable.virginm.net] has quit [Read error: Connection reset by peer] | 17:38 | |
-!- PirosB3 [~pirosb3@cpc4-newt31-2-0-cust631.newt.cable.virginm.net] has joined #shogun | 17:38 | |
thoralf | wiking, besser82: Seems like our current cmake setup doesn't list "json" within the summary of config variables. | 17:41 |
-!- lambday_ [67157d54@gateway/web/freenode/ip.103.21.125.84] has quit [Ping timeout: 245 seconds] | 17:42 | |
thoralf | wiking, besser82: I guess this happens because we use "pkg_search_module(JSON libjson>=0.11 json>=0.11 json-c>=0.11)" instead of FIND_PACKAGE(JSON). | 17:42 |
@besser82 | thoralf, there is no FindJSON.cmake skript ;) | 17:43 |
thoralf | besser82: How hard is it to write one? | 17:43 |
@besser82 | thoralf, more or less... | 17:43 |
jiaolong | thoralf: I removed libjson0 from synaptic and the openjdk is also removed | 17:44 |
thoralf | It's only minor issue, because it's only affecting debug output... so we might just ignore it. | 17:44 |
@besser82 | thoralf, but I can try to make the pkg_search_module more verbose with some switch ;) | 17:44 |
jiaolong | thoralf: I have to install openjdk again otherwise cmake reports error Could NOT find Java | 17:45 |
thoralf | besser82: Seems to be the case that every package that's located using FIND_PACKAGE gets mentioned in the "configuration variables section" | 17:45 |
thoralf | jiaolong: What's the problem with java? | 17:45 |
@besser82 | thoralf, so I'll try to add those pkg_conf using packages to that output, too. ;) | 17:46 |
jiaolong | thoralf: while removing libjson0 I found jdk was also removed | 17:46 |
thoralf | besser82: Cool! :) | 17:46 |
@besser82 | thoralf, np :D Will do me best tomorrow in da train to leipzig ;) | 17:46 |
-!- sunil1337 [0e8bd057@gateway/web/freenode/ip.14.139.208.87] has joined #shogun | 17:46 | |
thoralf | besser82: If it's not working out, then just skip it... but in any case, thank you for checking. :) | 17:47 |
@besser82 | thoralf, allrighty! | 17:47 |
thoralf | jiaolong: Java is not what we're fighting with, so you could easily skip this for now... but it doesn't hurt to enable it. | 17:48 |
jiaolong | thoralf: OK | 17:49 |
thoralf | besser82: Did you check why travis keeps running the tests even if the build failed? | 17:50 |
thoralf | besser82: We talked about this a week ago. | 17:50 |
-!- sunil1337 [0e8bd057@gateway/web/freenode/ip.14.139.208.87] has quit [Ping timeout: 245 seconds] | 17:51 | |
jiaolong | thoralf: Hi, I finished the testing and got the same error: ImportError: cannot import name SerializableJsonFile | 18:03 |
thoralf | jiaolong: Okay, let us check some parameters. | 18:04 |
thoralf | jiaolong: Did you remove the build directory before starting the new build? | 18:04 |
jiaolong | thoralf: Yes | 18:04 |
thoralf | jiaolong: did you "make install" before running the failing test? | 18:05 |
jiaolong | thoralf: sudo make install | 18:05 |
thoralf | jiaolong: You're installing it in a system location? You know that you also may install it in your user home? | 18:06 |
thoralf | jiaolong: Did you remove the directory where shogun has been installed to before building? Maybe some old files are interfering with your new installation? | 18:07 |
jiaolong | thoralf: I used your scripts and I still need root permission | 18:07 |
jiaolong | thoralf: No, I didn't uninstall shogun | 18:08 |
thoralf | jiaolong: What command was failing without root? | 18:08 |
thoralf | Probably only a delete-command, right? | 18:08 |
jiaolong | thoralf: But I think if the problem is from libjson, remove it may still cause the testing error | 18:08 |
jiaolong | thoralf: delete need root and executing your scripts also need when it runs to install | 18:09 |
thoralf | jiaolong: For sure not. I'm using it without root. ;) | 18:09 |
thoralf | -DCMAKE_INSTALL_PREFIX="$BUILDDIR/install" <-- This installs the shogun libraries into a user directory. ;) | 18:10 |
jiaolong | thoralf: yes, I also thought no need to sudo | 18:10 |
jiaolong | thoralf: but in fact I need | 18:11 |
thoralf | https://gist.github.com/tklein23/9730737 <-- Check the newest script there. I added one line a few days ago (look for GoogleMock). | 18:11 |
thoralf | That's the line that may complain about missing permissions... but it's only because all files in GoogleMock are created read-only. | 18:12 |
@besser82 | thoralf, not yet, unfortunately; but I still have that on my do's | 18:12 |
jiaolong | thoralf: OK, thanks | 18:12 |
thoralf | jiaolong: But still, the the error happens with and without JSON, we're either missing something or JSON is not the problem. | 18:13 |
thoralf | s/the the/if the/ | 18:13 |
jiaolong | thoralf: If this is the only error in my local, it won't happen in Travis, right? | 18:18 |
thoralf | jiaolong: Yes. | 18:18 |
jiaolong | thoralf: For the multilabel libsvmfile, it is the only error now | 18:19 |
jiaolong | thoralf: I am running the new script again, let's see if it asks me for sudo ;) | 18:20 |
jiaolong | thoralf: I will just push to my origin repo before solving these testing errors in my local | 18:22 |
-!- khalednasr [~k.nasr92@41.69.141.226] has joined #shogun | 18:22 | |
jiaolong | thoralf: I have changed int16_t into index_t in your MultilabelLabel class, do you agree? | 18:23 |
thoralf | jiaolong: Oh, wait. I did this as well in a PR I didn't merge. | 18:24 |
thoralf | jiaolong: If I merge this, you may get a conflict | 18:24 |
jiaolong | thoralf: no problem, I can fetch your version | 18:24 |
jiaolong | thoralf: OK, it is better not to touch your MultilabelLabel class? | 18:25 |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 18:25 | |
shogun-notifier- | shogun: Thoralf Klein :develop * ec90ad9 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/ec90ad920f5a6fdc0edd08db6a0dbec34e06dcce | 18:25 |
shogun-notifier- | shogun: Changed int16_t to int32_t in MultilabelLabels. | 18:25 |
shogun-notifier- | shogun: tklein23 :develop * b493737 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/b493737d0436ee590d558cd25369bc517c58a4ef | 18:25 |
shogun-notifier- | shogun: Merge pull request #2077 from tklein23/change_multilabel_labeltype | 18:25 |
shogun-notifier- | shogun: | 18:25 |
shogun-notifier- | shogun: Changed int16_t to int32_t in MultilabelLabels. | 18:25 |
thoralf | jiaolong: Of course you can touch it. We somethimes may interfer, but that's okay. ;) | 18:25 |
jiaolong | thoralf: OK, I am still doing this task, so you can do changes as much as you can | 18:26 |
-!- nikolis [~nikolis@213.16.163.237.dsl.dyn.forthnet.gr] has quit [Quit: Leaving] | 18:26 | |
jiaolong | thoralf: no sudo this time, :) | 18:27 |
thoralf | jiaolong: :D | 18:27 |
thoralf | jiaolong: Shall we continue debugging your problem? | 18:28 |
jiaolong | thoralf: if I do git ci --amend, I cannot push it to my origin, do you have suggestions about this? I want to keep just a single commit in my last PR | 18:28 |
thoralf | jiaolong: git push --force origin your_branch ;) | 18:29 |
thoralf | This might confuse people who cloned your local repository, but that's all. | 18:29 |
jiaolong | thoralf: I finished the testing again, the problems are still there | 18:29 |
thoralf | Sometime travis is confused by this. | 18:29 |
thoralf | But git push --force is okay as soon as it only affects your repository. | 18:30 |
jiaolong | thoralf: I did that last time and I guess that's the reason I got Travis error, so now, I really hesitate to do that | 18:30 |
jiaolong | thoralf: Yes, but I need to push force evey time afterwards | 18:31 |
thoralf | But something else: You can start collecting commits, and if your done, you do "git rebase -i HEAD~10", squash them to one commit, and then do a "git push --force". | 18:31 |
thoralf | HEAD~10 is the number of commits you want to squash. | 18:31 |
jiaolong | thoralf: Yes, OK | 18:31 |
thoralf | jiaolong: Yes, you have to do --force afterwards, because the branches are not compatible any more (different hashes on client/server side and no fast-forward). | 18:32 |
jiaolong | thoralf:OK, but it is allowed to add multiple commits in a single PR in shogun? I know it is not nice. | 18:33 |
jiaolong | thoralf: I can try to add as few as possible | 18:34 |
thoralf | jiaolong: I'm doing it some times, mostly to isolate things. For example, if you're refactoring and fixing indentation in one PR, then it would be hard to see what changed. In this case I first commit the reindented version, then do my refactoring. | 18:35 |
jiaolong | thoralf: OK, perfect | 18:35 |
jiaolong | thoralf: Since my current error has nothing to do with my changed code, I would like to focus on the entrance task | 18:36 |
jiaolong | thoralf: I would like to ignore the Json issue for a while | 18:36 |
thoralf | jiaolong: Find something in between. Often only one commit is not well-suited for what you did, but having big number of commits is bad as well. Nothing religious. | 18:37 |
thoralf | jiaolong: Is the json thing only affecting the R interface? | 18:37 |
jiaolong | thoralf: OK, perfec, I really hate push force | 18:37 |
jiaolong | thoralf: For R, I have no idea, I don't think their related | 18:38 |
jiaolong | thoralf: It seems most R module test are failed | 18:38 |
thoralf | jiaolong: Do the errors disappear when you disable R? | 18:38 |
jiaolong | thoralf: I disabled R in the last testing | 18:39 |
jiaolong | thoralf: Json errors are still there | 18:39 |
jiaolong | thoralf: Should I use SGVector<float64_t>* as the multilabel type in libsvmfile? If you confirm, I can do it for the next commit | 18:42 |
thoralf | That's fine. | 18:42 |
jiaolong | thoralf: OK, then for the IO I only need to discuss with wiking about the file-driven issue | 18:43 |
-!- sunil1337 [0e8bd057@gateway/web/freenode/ip.14.139.208.87] has joined #shogun | 18:43 | |
jiaolong | thoralf: I hope the LibSVMFile for multilabel problem can be done as soon as possible | 18:44 |
thoralf | jiaolong: What I suggest. Having two methods. One for reading single-labels, which will fail if somebody tries to read a multilabel file. The other one is capable of reading multilabels, but also accepts single-labels, because they're just a special case. | 18:46 |
-!- khalednasr [~k.nasr92@41.69.141.226] has quit [Remote host closed the connection] | 18:46 | |
thoralf | jiaolong: Otherwise we need to have something to "detect" the content of a label file using a header or something like that. | 18:46 |
thoralf | jiaolong: Does this collide with what you discussed with wiking so far? | 18:47 |
jiaolong | thoralf: Yes, what you are suggesting is actually what I have done | 18:47 |
jiaolong | thoralf: I am not sure, but maybe he has more considerations | 18:47 |
jiaolong | thoralf: one thing, I need to regularize that for single-label, -1, 0, +1, ... is not allowed | 18:49 |
jiaolong | thoralf: but it was allowed in the previous unit test | 18:50 |
thoralf | jiaolong: Why is -1, +1, 0 not allowed? | 18:50 |
jiaolong | thoralf: I convert -1 into 0, if -1 exists | 18:50 |
thoralf | 0 is for "has no label", +1 for "is in class", -1 for "is not in class" | 18:50 |
thoralf | Oh, no magic conversions. | 18:51 |
thoralf | Please. | 18:51 |
jiaolong | thoralf: then I need to reconsider the reader | 18:51 |
thoralf | Then we will never be able to read parially labeled files. | 18:51 |
thoralf | jiaolong: As I said: The semantic for multi-labeled and single-labeled is different. | 18:52 |
jiaolong | thoralf: OK, it is fine since I will remove SGSparseVector | 18:52 |
-!- sunil1337 [0e8bd057@gateway/web/freenode/ip.14.139.208.87] has quit [Ping timeout: 245 seconds] | 18:53 | |
jiaolong | thoralf: OK, I simply use convert the label into sparse index previously | 18:53 |
thoralf | jiaolong: Sorry, I don't understand what's the plan... not changing semantics of single-label files, but what happens to multi-label files? | 18:55 |
jiaolong | thoralf: Thanks for your discussion, I think it is very important before I make the next commit | 18:55 |
jiaolong | thoralf: I will change the code soon | 18:55 |
thoralf | reading "1,2,3 1:1 2:1 7:1" ends how? | 18:55 |
jiaolong | thoralf: Yes | 18:55 |
thoralf | or better: reading "0,2,5 1:1 2:1 7:1" ends how? | 18:55 |
jiaolong | thoralf: That was the easieast way to implement but now I realized the problem | 18:56 |
thoralf | It will return a SGVector<> =[0, 2, 5] and a sparse vector with features, right? | 18:56 |
jiaolong | thoralf: 0, 2, 5 -> 1:1 2:1 7:1 | 18:57 |
thoralf | Okay. | 18:57 |
thoralf | Both reading modes should not change what they read. So NEVER implicitly replace something if a magic condition is fulfilled. | 18:57 |
thoralf | No magic. ;) | 18:57 |
thoralf | But, I still see potential for reusing the multi-label reader from within the single-label reader. | 18:58 |
jiaolong | thoralf: Sure, sure :) | 18:58 |
thoralf | a multilabel with only one value out of {-1, 0, +2} can easily converted to a single label and you're done. | 18:59 |
thoralf | Just throwing an error if number of labels != 1. | 18:59 |
jiaolong | thoralf: well, but what if this is a multclass problem? I means three classes -1, 0, +2 | 19:00 |
jiaolong | thoralf: It is not common but it should be allowed, right? | 19:01 |
thoralf | jiaolong: Let me describe it differently: "label1,label2,label3 features..." is multi-label with 3 labels. | 19:01 |
thoralf | "label1 features" is multi-label with 1 label | 19:01 |
thoralf | And single-label as well. | 19:01 |
thoralf | Since we're readling floats, it's also possible to have real-value labels (for specifying regression problems instead of classification) | 19:02 |
thoralf | So everything is possible | 19:03 |
thoralf | Just read a sequence of numbers. | 19:03 |
jiaolong | thoralf: OK, yes | 19:03 |
jiaolong | thoralf: OK, I see. it is time to remove the SGSpaseVector<float64_t>* | 19:04 |
thoralf | But as multi-labels reads a sequence of numbers, then we could make the single-label case a sequence of numbers with sequence length one, so we can reuse the multi-class code. | 19:04 |
thoralf | Yes, please remove the SparseVector. :) | 19:04 |
thoralf | Or replace it by SGVector<>, however. | 19:05 |
jiaolong | thoralf: OK, perfect | 19:05 |
jiaolong | thoralf: OK, thank you so much! Now I think I can start the work. | 19:07 |
thoralf | jiaolong: Good luck. And sorry for the confusion. | 19:10 |
thoralf | The details haven't been mature enough... | 19:11 |
jiaolong | thoralf: Now it is very clear. I should consult you in the early. Thank you very much! | 19:12 |
thoralf | No problem. Good luck! | 19:12 |
jiaolong | thoralf: Cheers! See you! | 19:13 |
-!- zxtx [~zv@129-79-241-148.dhcp-bl.indiana.edu] has joined #shogun | 19:16 | |
-!- zxtx [~zv@129-79-241-148.dhcp-bl.indiana.edu] has quit [Ping timeout: 246 seconds] | 19:21 | |
-!- jiaolong [9e6d09e4@gateway/web/freenode/ip.158.109.9.228] has quit [] | 19:22 | |
-!- lisitsyn [~lisitsyn@80.252.20.67] has joined #shogun | 19:23 | |
thoralf | Hey Sergey | 19:26 |
-!- zxtx [~zv@129-79-241-164.dhcp-bl.indiana.edu] has joined #shogun | 19:37 | |
-!- nikolis [~nikolis@213.16.163.237.dsl.dyn.forthnet.gr] has joined #shogun | 19:37 | |
-!- andrew__ [~andrew@c-50-184-110-141.hsd1.ca.comcast.net] has quit [Ping timeout: 265 seconds] | 19:38 | |
-!- newlido [~quassel@41.218.177.77] has joined #shogun | 19:43 | |
-!- andrew__ [~andrew@50-0-142-100.dsl.dynamic.sonic.net] has joined #shogun | 19:45 | |
-!- travis-ci [~travis-ci@ec2-54-204-147-63.compute-1.amazonaws.com] has joined #shogun | 19:49 | |
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/21526227 | 19:49 |
-!- travis-ci [~travis-ci@ec2-54-204-147-63.compute-1.amazonaws.com] has left #shogun [] | 19:49 | |
dhruv13j | wiking: hey! you there? | 19:51 |
dhruv13j | wiking: I need some help with cmake stuff... | 19:51 |
dhruv13j | thoralf: hey! can you help? | 19:52 |
thoralf | dhruv13j: Depends. | 19:53 |
thoralf | Try it. | 19:53 |
dhruv13j | thoralf: I am trying to bundle nanoflann | 19:53 |
dhruv13j | thoralf: I got it to detect nanoflann when it's installed... | 19:55 |
andrew__ | In my install directory, lib/pythonX.Y/dist-packages/ is empty. I'm pretty sure I ran cmake with -DPythonModular=ON and also ran sudo make install. Ideas? | 19:56 |
dhruv13j | thoralf: but I'm missing something, and I can't figure out what | 19:56 |
dhruv13j | thoralf: I'll pastebin my additions... | 19:56 |
thoralf | dhruv13j: I will not check/debug your solution. If you have a particular question, then I might help. | 19:57 |
dhruv13j | thoralf: okay... | 19:58 |
thoralf | dhruv13j: Really - I'm not into this task, so if I debug your solution, I'd be able to solve it as well... ;) | 19:58 |
dhruv13j | thoralf: okay :-) I will try and debug some more, when I'm really really stuck, I'll come back... | 19:59 |
thoralf | dhruv13j: Maybe besser82 can help you. He's more familiar with cmake. | 19:59 |
thoralf | andrew__: you did a "make all", "make install" and still the files are missing? | 20:00 |
dhruv13j | thoralf: I'll ask him if i can't do it myself... I'm probably missing something crucial.... | 20:00 |
andrew__ | i'm not sure i ran make all, so I'll try that | 20:01 |
thoralf | andrew__: I'll quickly build here with python enabled and check where to expect the files. | 20:01 |
dhruv13j | andrew__: are you on linux? | 20:06 |
andrew__ | yeah | 20:14 |
andrew__ | okay i just rebuilt everything and still nothing | 20:15 |
dhruv13j | andrew__: hey! if you have not set -DCMAKE_INSTALL_PREFIX, then you should expect it in /usr/local/lib/pythonX.Y/dist-packages/shogun | 20:19 |
dhruv13j | andrew__: is that where you are looking? | 20:19 |
andrew__ | yeah /dist-packages/ is empty | 20:22 |
andrew__ | okay might as well start poking around the src while i'm getting everything running. Where am I going to find SGVector? | 20:24 |
andrew__ | /src/shogun/mathematics ? | 20:25 |
thoralf | andrew__: shogun/lib | 20:26 |
thoralf | andrew__: Try "grep -ri src/shogun SGVector" | 20:26 |
thoralf | or find src/hosgun -inaname "*SGVector*" | 20:26 |
-!- HeikoS1 [~heiko@pat-191-250.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.] | 20:27 | |
dhruv13j | andrew__: is pythonX.Y your default python? if not, check the cmake readme for how to build it for the desired pythonX.Y | 20:27 |
andrew__ | druhv13j: great suggestion. In fact my default python is canopy 2.7.3 | 20:29 |
dhruv13j | andrew__: and are you trying to build for 3.3? | 20:31 |
dhruv13j | andrew__: if so, you need to specify some stuff for cmake | 20:32 |
-!- lambday [67157f54@gateway/web/freenode/ip.103.21.127.84] has joined #shogun | 20:32 | |
andrew__ | I don't really have a preference. 2.7 is fine. I'm thinking my current default is a funky python distribution though | 20:34 |
-!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has joined #shogun | 20:34 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 20:34 | |
@iglesiasg | Hi thoralf! | 20:36 |
thoralf | Hey iglesiasg | 20:36 |
@iglesiasg | thoralf, do you think it would be possible to run this check you have been using to detect wrong includes after we merge this PR? github.com/shogun-toolbox/shogun/pull/2063 | 20:36 |
thoralf | iglesiasg: Of course. It's fully scripted. | 20:37 |
thoralf | iglesiasg: I'm planning to integrate it into the buildbot. ;) | 20:37 |
@iglesiasg | thoralf, great. For the moment you just have it locally, or is it somewhere in the repo? | 20:38 |
thoralf | iglesiasg: Only locally, since the complicated part is getting the right clang compiled with the right plugins. | 20:38 |
shogun-notifier- | shogun: khalednasr :develop * 77716f2 / / (12 files): https://github.com/shogun-toolbox/shogun/commit/77716f257d09c1de32be601faf7de2ff1bc8dca8 | 20:39 |
shogun-notifier- | shogun: basic neural networks | 20:39 |
shogun-notifier- | shogun: khalednasr :develop * 804c399 / src/shogun/neuralnets/NeuralLayer.cpp,src/shogun/neuralnets/NeuralNetwork.cpp: https://github.com/shogun-toolbox/shogun/commit/804c3991df3bd8267c24f1e78e952b7cc1aaa196 | 20:39 |
shogun-notifier- | shogun: small bugfix | 20:39 |
shogun-notifier- | shogun: khalednasr :develop * d825264 / src/shogun/neuralnets/NeuralLayer.cpp: https://github.com/shogun-toolbox/shogun/commit/d82526468028191065922b024874cb91fc4874a1 | 20:39 |
shogun-notifier- | shogun: small bugfix | 20:39 |
shogun-notifier- | shogun: khalednasr :develop * f590f9a / / (8 files): https://github.com/shogun-toolbox/shogun/commit/f590f9ad66d6e485f47de8340c2fd526b0445d6d | 20:39 |
shogun-notifier- | shogun: fixed indents | 20:39 |
shogun-notifier- | shogun: khalednasr :develop * 74ef966 / src/shogun/neuralnets/ (4 files): https://github.com/shogun-toolbox/shogun/commit/74ef966b3c51c6e7b331b4f8a11f419d7ab21ca0 | 20:39 |
shogun-notifier- | shogun: fixed indents | 20:39 |
shogun-notifier- | shogun: khalednasr :develop * a4d0fba / src/shogun/neuralnets/ (6 files): https://github.com/shogun-toolbox/shogun/commit/a4d0fbacf6eff9c51c9abfe7ef3e4c5a1ebb3b3a | 20:39 |
shogun-notifier- | shogun: removed unnecessary includes and typedefs | 20:39 |
shogun-notifier- | shogun: khalednasr :develop * 9f959b3 / src/shogun/neuralnets/ (4 files): https://github.com/shogun-toolbox/shogun/commit/9f959b3e566136a68e0d999ed1f6f40e51538d81 | 20:39 |
shogun-notifier- | shogun: removed some unnecessary stuff | 20:39 |
shogun-notifier- | shogun: khalednasr :develop * c58e077 / src/shogun/neuralnets/ (4 files): https://github.com/shogun-toolbox/shogun/commit/c58e0775f984fc40ef525fe7a0a64c79ddfc4c8a | 20:39 |
shogun-notifier- | shogun: removed some more unnecessary stuff | 20:39 |
shogun-notifier- | shogun: khalednasr :develop * 80b9ce5 / / (11 files): https://github.com/shogun-toolbox/shogun/commit/80b9ce55a906d94574ef6057e8d2abd8fe341c9d | 20:39 |
shogun-notifier- | shogun: made CNeuralNetwork inherit from CMachine | 20:39 |
@iglesiasg | thoralf, ok. I am going to merge right away. Is it too much to ask you to run it afterwards? | 20:39 |
lisitsyn | sorry guys :D | 20:39 |
lisitsyn | spammed it a bit | 20:39 |
@iglesiasg | thoralf, I want to be sure that this including config.h doesn't do anything | 20:39 |
@iglesiasg | lisitsyn, hahaha :P | 20:39 |
thoralf | iglesiasg: Well, you only included config.h, so it didn't hurt. | 20:39 |
thoralf | iglesiasg: And btw., I added guards to config.h anyway. | 20:40 |
@iglesiasg | lisitsyn, you were waiting for a conversation to start to merge, right? :P | 20:40 |
@iglesiasg | lisitsyn, just kidding :) | 20:40 |
lisitsyn | sure thing | 20:40 |
lisitsyn | :D | 20:40 |
@iglesiasg | thoralf, ok. It sounds like 0% danger then | 20:40 |
thoralf | iglesiasg: Aaaand: config.h is implicitly included with common.h, so it's available everywhere anyway. | 20:40 |
@iglesiasg | thoralf, mmm wouldn't that mean that there's no need to include in the places where common.h is included then? | 20:41 |
thoralf | iglesiasg: Sounds like, eh? ;) | 20:42 |
thoralf | Let me check. | 20:42 |
thoralf | Yes, common.h includes config.h. | 20:42 |
-!- andrew__1 [~andrew@50-0-142-100.dsl.dynamic.sonic.net] has joined #shogun | 20:43 | |
@iglesiasg | thoralf, I think we still need this patch though, as not all the headers include common.h | 20:44 |
-!- andrew__ [~andrew@50-0-142-100.dsl.dynamic.sonic.net] has quit [Ping timeout: 246 seconds] | 20:45 | |
shogun-notifier- | shogun: frank0523 :develop * 615a1ea / src/interfaces/ (3 files): https://github.com/shogun-toolbox/shogun/commit/615a1ea33858edd6dde7830f49c4d1cf3513fd7e | 20:46 |
shogun-notifier- | shogun: add config.h in all .h files in interfaces | 20:46 |
shogun-notifier- | shogun: Yuyu Zhang :develop * ebea39a / src/shogun/base/ (9 files): https://github.com/shogun-toolbox/shogun/commit/ebea39aa405221ad90fde1a410828cd566ae8e7b | 20:46 |
shogun-notifier- | shogun: add config.h in all .h files in shogun/base | 20:46 |
shogun-notifier- | shogun: Yuyu Zhang :develop * 6e11d94 / src/shogun/clustering/ (4 files): https://github.com/shogun-toolbox/shogun/commit/6e11d943590e8d8a34b7f90adeeeb4d154deb22a | 20:46 |
shogun-notifier- | shogun: add config.h in all .h files in shogun/clustering | 20:46 |
shogun-notifier- | shogun: Yuyu Zhang :develop * 398d7b4 / src/shogun/converter/ (3 files): https://github.com/shogun-toolbox/shogun/commit/398d7b4fe65566e31993732450afed30050a1896 | 20:46 |
shogun-notifier- | shogun: add config.h in all .h files in shogun/converter | 20:46 |
shogun-notifier- | shogun: Yuyu Zhang :develop * 6ff3446 / src/shogun/distance/ (26 files): https://github.com/shogun-toolbox/shogun/commit/6ff344690f8179b7ea6adb58de65081227b5438c | 20:46 |
shogun-notifier- | shogun: add config.h in all .h files in shogun/distance | 20:46 |
shogun-notifier- | shogun: Yuyu Zhang :develop * fb1933b / src/shogun/distributions/ (7 files): https://github.com/shogun-toolbox/shogun/commit/fb1933b914e2b085d6df76cfefb81bb4d6caf919 | 20:46 |
shogun-notifier- | shogun: add config.h in all .h files in shogun/distributions | 20:46 |
shogun-notifier- | shogun: Yuyu Zhang :develop * 51e5087 / src/shogun/ensemble/ (4 files): https://github.com/shogun-toolbox/shogun/commit/51e50879ddd799054f4eb7bbc6ba5b3fb4d02b57 | 20:46 |
shogun-notifier- | shogun: add config.h in all .h files in shogun/ensemble | 20:46 |
shogun-notifier- | shogun: Yuyu Zhang :develop * 67dd251 / src/shogun/evaluation/ (30 files): https://github.com/shogun-toolbox/shogun/commit/67dd2516810a297fe1d642ce63d445f90f5a941a | 20:46 |
shogun-notifier- | shogun: add config.h in all .h files in shogun/evaluation | 20:46 |
shogun-notifier- | shogun: Yuyu Zhang :develop * 204b191 / src/shogun/features/ (41 files): https://github.com/shogun-toolbox/shogun/commit/204b1910fc09cbd1bb2077b102cca5feb91741d8 | 20:46 |
shogun-notifier- | shogun: add config.h in all .h files in shogun/features | 20:46 |
shogun-notifier- | shogun: Yuyu Zhang :develop * 824bbc8 / src/shogun/preprocessor/ (16 files): https://github.com/shogun-toolbox/shogun/commit/824bbc896cd538a60e87ca49bb959829792da9dc | 20:46 |
shogun-notifier- | shogun: add config.h in all .h files in shogun/preprocessor | 20:46 |
shogun-notifier- | shogun: Yuyu Zhang :develop * 1158210 / src/shogun/regression/ (3 files): https://github.com/shogun-toolbox/shogun/commit/1158210dbc72b1302905c6abd97a86220b70d610 | 20:46 |
shogun-notifier- | shogun: add config.h in all .h files in shogun/regression | 20:46 |
shogun-notifier- | shogun: Yuyu Zhang :develop * d22e148 / src/shogun/statistics/ (15 files): https://github.com/shogun-toolbox/shogun/commit/d22e14852f1512091244bfb3344ec97a00523a0a | 20:46 |
shogun-notifier- | shogun: add config.h in all .h files in shogun/statistics | 20:46 |
shogun-notifier- | shogun: Yuyu Zhang :develop * d47ed75 / src/shogun/structure/ (30 files): https://github.com/shogun-toolbox/shogun/commit/d47ed75e4de6ad27ee6d24cf3d187b32a7391782 | 20:46 |
shogun-notifier- | shogun: add config.h in all .h files in shogun/structure | 20:46 |
shogun-notifier- | shogun: Yuyu Zhang :develop * 277e3ef / src/shogun/transfer/ (16 files): https://github.com/shogun-toolbox/shogun/commit/277e3ef47c8601b1ff2564902ffca42338450193 | 20:46 |
shogun-notifier- | shogun: add config.h in all .h files in shogun/transfer | 20:46 |
shogun-notifier- | shogun: Fernando Iglesias :develop * ac88728 / src/ (498 files): https://github.com/shogun-toolbox/shogun/commit/ac887280e054a15c0e4662a432d7f193abc28592 | 20:46 |
shogun-notifier- | shogun: Merge pull request #2063 from frank0523/develop | 20:46 |
shogun-notifier- | shogun: | 20:46 |
shogun-notifier- | shogun: Continue adding include config.h | 20:46 |
shogun-buildbot | build #2811 of deb1 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/2811 blamelist: khalednasr <k.nasr92@gmail.com> | 20:48 |
lisitsyn | damn! | 20:49 |
shogun-notifier- | shogun: Abhijeet :develop * b29c8cb / doc/ipython-notebooks/gaussian_process/gaussian_processes.ipynb: https://github.com/shogun-toolbox/shogun/commit/b29c8cbb0bab9dc89c6f5436450d2630a883049d | 20:50 |
shogun-buildbot | build #2812 of deb1 - libshogun is complete: Success [build successful] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/2812 | 20:50 |
shogun-notifier- | shogun: wiki links added to GP notebook's abstract | 20:50 |
shogun-notifier- | shogun: Abhijeet :develop * 388541d / doc/ipython-notebooks/ (2 files): https://github.com/shogun-toolbox/shogun/commit/388541dfccd5d6b449729c1933a226bf7b852ffd | 20:50 |
shogun-notifier- | shogun: add wiki links to pca_notebook's abstract | 20:50 |
shogun-notifier- | shogun: Abhijeet :develop * 91e4c07 / doc/ipython-notebooks/clustering/KMeans.ipynb: https://github.com/shogun-toolbox/shogun/commit/91e4c07f711b9df9dfc30d6a8da576134d5ddba6 | 20:50 |
shogun-notifier- | shogun: abstract revised for KMeans.ipynb | 20:50 |
shogun-notifier- | shogun: Abhijeet :develop * e770dc5 / doc/ipython-notebooks/classification/SupportVectorMachines.ipynb: https://github.com/shogun-toolbox/shogun/commit/e770dc56e4f99bd3b4d6e0f1f28f87c10813749a | 20:50 |
shogun-notifier- | shogun: abstract revised for SupportVectorMachines.ipynb | 20:50 |
shogun-notifier- | shogun: Abhijeet :develop * ac75ebf / doc/ipython-notebooks/classification/HashedDocDotFeatures.ipynb: https://github.com/shogun-toolbox/shogun/commit/ac75ebf21966f5e7675b0a20df13c6c01fd8ea2a | 20:50 |
shogun-notifier- | shogun: abstract revised for HashedDocDotFeatures.ipynb | 20:50 |
shogun-notifier- | shogun: Abhijeet :develop * fa41eaf / doc/ipython-notebooks/converter/Tapkee.ipynb: https://github.com/shogun-toolbox/shogun/commit/fa41eaf475b1d65fdb9eaba5e3db75d1ad0afbd0 | 20:50 |
shogun-notifier- | shogun: links added to Tapkee.ipynb | 20:50 |
shogun-notifier- | shogun: Abhijeet :develop * 9aabe96 / doc/ipython-notebooks/evaluation/xval_modelselection.ipynb: https://github.com/shogun-toolbox/shogun/commit/9aabe96fd5cc2cf2db586720924966ff865a75c9 | 20:50 |
shogun-notifier- | shogun: abstract revised for xval_modelselection.ipynb | 20:50 |
shogun-notifier- | shogun: Abhijeet :develop * dd73421 / doc/ipython-notebooks/ica/bss_image.ipynb: https://github.com/shogun-toolbox/shogun/commit/dd73421ad6ab6330a015228cec621cca2a3c147a | 20:50 |
shogun-notifier- | shogun: abstract revised for bss_image.ipynb | 20:50 |
shogun-notifier- | shogun: Abhijeet :develop * 0942bb0 / doc/ipython-notebooks/ica/bss_audio.ipynb: https://github.com/shogun-toolbox/shogun/commit/0942bb035f4bbf10ae8b774d7512483ce3feff51 | 20:50 |
shogun-notifier- | shogun: abstract revised for bss_audio.ipynb | 20:50 |
shogun-notifier- | shogun: Abhijeet :develop * 6354925 / doc/ipython-notebooks/ (2 files): https://github.com/shogun-toolbox/shogun/commit/63549258653a6da4cf94dec027985e079295f28f | 20:50 |
shogun-notifier- | shogun: add wiki links to pca_notebook's abstract | 20:50 |
shogun-notifier- | shogun: Abhijeet :develop * f46c81d / doc/ipython-notebooks/clustering/KMeans.ipynb: https://github.com/shogun-toolbox/shogun/commit/f46c81d7287ff847c2305a5a2940b3e5083e53bb | 20:50 |
shogun-notifier- | shogun: abstract revised for KMeans.ipynb | 20:50 |
shogun-notifier- | shogun: Abhijeet :develop * c8d1a3c / doc/ipython-notebooks/classification/SupportVectorMachines.ipynb: https://github.com/shogun-toolbox/shogun/commit/c8d1a3c64c235aca8d2e7a1a4a37c658d820cd24 | 20:50 |
shogun-notifier- | shogun: abstract revised for SupportVectorMachines.ipynb | 20:50 |
shogun-notifier- | shogun: Abhijeet :develop * b05f046 / doc/ipython-notebooks/classification/HashedDocDotFeatures.ipynb: https://github.com/shogun-toolbox/shogun/commit/b05f046c4e16b8bd3b1b0fd58fd2db67b501ff8e | 20:50 |
shogun-notifier- | shogun: abstract revised for HashedDocDotFeatures.ipynb | 20:50 |
shogun-notifier- | shogun: Abhijeet :develop * b38936b / doc/ipython-notebooks/converter/Tapkee.ipynb: https://github.com/shogun-toolbox/shogun/commit/b38936b0e5df9381ee00e5448799e383afa63dcc | 20:50 |
shogun-notifier- | shogun: links added to Tapkee.ipynb | 20:50 |
shogun-notifier- | shogun: Abhijeet :develop * e875ca9 / doc/ipython-notebooks/evaluation/xval_modelselection.ipynb: https://github.com/shogun-toolbox/shogun/commit/e875ca9669e97513cdea02debbad51fcf9924d40 | 20:50 |
shogun-notifier- | shogun: abstract revised for xval_modelselection.ipynb | 20:50 |
shogun-notifier- | shogun: Abhijeet :develop * 34b09c9 / doc/ipython-notebooks/ica/bss_image.ipynb: https://github.com/shogun-toolbox/shogun/commit/34b09c9c75e2d86108634a68ea4f8b7c87b2da69 | 20:50 |
shogun-notifier- | shogun: abstract revised for bss_image.ipynb | 20:50 |
shogun-notifier- | shogun: Abhijeet :develop * e125f1c / doc/ipython-notebooks/ica/bss_audio.ipynb: https://github.com/shogun-toolbox/shogun/commit/e125f1c366c908f42a4970958c2d179d76f9c961 | 20:50 |
shogun-notifier- | shogun: abstract revised for bss_audio.ipynb | 20:50 |
shogun-notifier- | shogun: Abhijeet :develop * 18e608f / doc/ipython-notebooks/ica/ecg_sep.ipynb: https://github.com/shogun-toolbox/shogun/commit/18e608ffc79c14cbfbe4fd929cc59ddc64ecd225 | 20:50 |
shogun-notifier- | shogun: abstract revised for ecg_sep.ipynb | 20:50 |
-!- andrew__1 [~andrew@50-0-142-100.dsl.dynamic.sonic.net] has quit [Ping timeout: 264 seconds] | 20:52 | |
shogun-notifier- | shogun: Sergey Lisitsyn :develop * 558f824 / tests/unit/base/clone_unittest.cc.jinja2: https://github.com/shogun-toolbox/shogun/commit/558f824443ef73ef1af38d86e285ee3f56da5546 | 20:56 |
shogun-notifier- | shogun: Ignore failing neural clone tests | 20:56 |
-!- sonne|osx [~sonne@e178249094.adsl.alicedsl.de] has joined #shogun | 20:58 | |
-!- andrew__ [~andrew@50-0-142-100.dsl.dynamic.sonic.net] has joined #shogun | 21:00 | |
-!- khalednasr [~k.nasr92@41.69.141.226] has joined #shogun | 21:00 | |
lisitsyn | khalednasr: hey | 21:01 |
khalednasr | lisitsyn, hey | 21:01 |
lisitsyn | khalednasr: I merged your pr | 21:02 |
lisitsyn | though it has some test failure | 21:02 |
lisitsyn | but now I realize it could be just first commits of your pr | 21:02 |
lisitsyn | and it will change with newer ones :D | 21:02 |
lisitsyn | khalednasr: is that true? | 21:02 |
khalednasr | lisitsyn, I had that error with older commits | 21:02 |
lisitsyn | argh | 21:03 |
lisitsyn | okay my bad | 21:03 |
shogun-notifier- | shogun: Sergey Lisitsyn :develop * a5685e5 / tests/unit/base/clone_unittest.cc.jinja2: https://github.com/shogun-toolbox/shogun/commit/a5685e53091b59d49e4d384906c5c9fba417acdc | 21:03 |
shogun-notifier- | shogun: Revert "Ignore failing neural clone tests" | 21:03 |
shogun-notifier- | shogun: | 21:03 |
shogun-notifier- | shogun: This reverts commit 558f824443ef73ef1af38d86e285ee3f56da5546. | 21:03 |
-!- lambday [67157f54@gateway/web/freenode/ip.103.21.127.84] has quit [Ping timeout: 245 seconds] | 21:04 | |
-!- khalednasr [~k.nasr92@41.69.141.226] has quit [Quit: Leaving] | 21:16 | |
-!- HeikoS [~heiko@0546d087.skybroadband.com] has joined #shogun | 21:20 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 21:20 | |
lisitsyn | wiking: I am freaked out by this last mail in the mentors list | 21:41 |
thoralf | lisitsyn: Did I miss something? | 21:56 |
shogun-notifier- | shogun-data: pl8787 :master * 9e523df / toy/ml-100k/ (23 files): https://github.com/shogun-toolbox/shogun-data/commit/9e523df22a5506095acda656889015153c095985 | 22:04 |
shogun-notifier- | shogun-data: Add Movielens data to data toy. | 22:04 |
shogun-notifier- | shogun-data: pl8787 :master * 9a5a985 / toy/ml-100k/allbut.pl,toy/ml-100k/mku.sh: https://github.com/shogun-toolbox/shogun-data/commit/9a5a985d201da36090c4087f1555700a7ea86110 | 22:04 |
shogun-notifier- | shogun-data: Remove the two scripts. | 22:04 |
shogun-notifier- | shogun-data: Heiko Strathmann :master * c8732bc / toy/ml-100k/ (21 files): https://github.com/shogun-toolbox/shogun-data/commit/c8732bcf586029b7483bbf2125f6f7387c2ecaa3 | 22:04 |
shogun-notifier- | shogun-data: Merge pull request #41 from pl8787/master | 22:04 |
shogun-notifier- | shogun-data: | 22:04 |
shogun-notifier- | shogun-data: Add Movielens data to data toy. Shogun issue #2068 | 22:04 |
-!- khalednasr [~k.nasr92@41.69.141.226] has joined #shogun | 22:05 | |
andrew__ | I could use some guidance for #1889. Is it as simple as modifying lib/SGVector::equals to call CMath::fequals with some small accuracy | 22:06 |
kislay | HeikoS, Hi! | 22:08 |
@HeikoS | kislay: hi! | 22:08 |
@HeikoS | andrew__: yes, and this should be optional | 22:09 |
andrew__ | what should be optional? | 22:10 |
kislay | HeikoS, Is there any referencex/tutos you may wanna share regarding GP notebook. I went through the notebook already but just wanna delve deep :) | 22:10 |
andrew__ | is there a style guide to follow? | 22:10 |
@HeikoS | andrew__: the accuracy | 22:11 |
@HeikoS | kislay: you are looking for literature on GPs? | 22:11 |
@HeikoS | kislay: from what perspective? theory or implementation? | 22:11 |
-!- khalednasr [~k.nasr92@41.69.141.226] has quit [Quit: Leaving] | 22:11 | |
@HeikoS | andrew__: its just changing the existing method, add an accuracy parameter and give it a default value (of zero) | 22:12 |
kislay | HeikoS, err..You remember what you told me yesterday? | 22:12 |
shogun-notifier- | shogun: Sunil Mahendrakar :develop * 7bff52b / / (3 files): https://github.com/shogun-toolbox/shogun/commit/7bff52bbc5cb94c1c946846c662b1171dc757a2f | 22:12 |
shogun-notifier- | shogun: fixed compile error when lapack not installed for logdet constructor and its unit test | 22:12 |
shogun-notifier- | shogun: Heiko Strathmann :develop * 5b7df93 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/5b7df93672c93caa2153535f4fb0c2512fc60f86 | 22:12 |
shogun-notifier- | shogun: Merge pull request #2083 from sunil1337/lapack_fix | 22:12 |
shogun-notifier- | shogun: | 22:12 |
shogun-notifier- | shogun: fixed compile error when lapack not installed for logdet constructor and its unit test | 22:13 |
@HeikoS | kislay: haha...no :D | 22:13 |
kislay | HeikoS, that I can work on extending a button on the GP demo!!! | 22:13 |
@HeikoS | kislay: ah yeah :) | 22:13 |
kislay | HeikoS, | 22:13 |
@HeikoS | kislay: I remember | 22:13 |
@HeikoS | so that is already implemented | 22:14 |
kislay | HeikoS, :D | 22:14 |
@HeikoS | and also being done in the notebook | 22:14 |
lisitsyn | thoralf: are you on mentor list? ;) | 22:14 |
lisitsyn | of gsoc | 22:14 |
@HeikoS | so you just need to extend the web-demo | 22:14 |
kislay | HeikoS, you quoted 24 hours back "extend a web demo with model-selection for GP's. | 22:14 |
kislay | add a button ("learn parameters with ML2")" | 22:14 |
@HeikoS | kislay: exactly | 22:14 |
@HeikoS | kislay: that is being done in the GP notebook | 22:15 |
@HeikoS | it learns kernel and likelihood parmaeters | 22:15 |
@HeikoS | using ML2 | 22:15 |
andrew__ | Is there a way for me to run tests locally on my new changes? | 22:15 |
kislay | HeikoS, yes it is. I read it. But it always feels good to get 100% confidence on what I am implementing. My imagination flares up :D | 22:16 |
@HeikoS | kislay: so you want to read on background? | 22:16 |
kislay | HeikoS, yeah. | 22:16 |
@HeikoS | kislay: ok, try chapter 5 of this book http://www.gaussianprocess.org/gpml/ | 22:16 |
@HeikoS | thats the best book on the subject | 22:17 |
kislay | HeikoS, That's what I wanted. :) | 22:18 |
@HeikoS | kislay: let me know if you have any questions | 22:18 |
@HeikoS | kislay: but the book is pretty self-exüplaining | 22:18 |
@HeikoS | also the chatper on regression and classification might be useful if you are interested in those things | 22:18 |
lisitsyn | I felt asleep on page 5 | 22:18 |
kislay | lisitsyn, you tried that book?? | 22:19 |
lisitsyn | yeah sure | 22:19 |
kislay | kislay, oh!! | 22:19 |
lisitsyn | gmpl is famous thing | 22:19 |
-!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has quit [Ping timeout: 246 seconds] | 22:19 | |
-!- travis-ci [~travis-ci@ec2-75-101-210-159.compute-1.amazonaws.com] has joined #shogun | 22:20 | |
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/21536228 | 22:20 |
-!- travis-ci [~travis-ci@ec2-75-101-210-159.compute-1.amazonaws.com] has left #shogun [] | 22:20 | |
kislay | HeikoS, can you add this thing in a issue. That way I can make you remember what I am working on with few lines :D | 22:20 |
@HeikoS | kislay: sure! | 22:20 |
kislay | lisitsyn, so you think there are better options right? | 22:21 |
lisitsyn | I didn't say this :) | 22:21 |
lisitsyn | though gpml is full of mathy math :D | 22:21 |
lisitsyn | that's what my joke was about | 22:21 |
kislay | lisitsyn, hmm..... you are jeopardizing my situation :D | 22:23 |
-!- zxtx [~zv@129-79-241-164.dhcp-bl.indiana.edu] has quit [Ping timeout: 268 seconds] | 22:28 | |
-!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has joined #shogun | 22:32 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 22:32 | |
@HeikoS | lisitsyn haha ;) | 22:35 |
andrew__ | C++ question: I'm changing line 397 of SGVector to call fequals of CMath | 22:35 |
andrew__ | What's the proper way to do this? CMath.fequals? CMath::fequals? I would appreciate a link to C++ for java programmers if such a thing exists | 22:37 |
andrew__ | mostly I'm confused how to insure CMath.fequals is in the local scope | 22:37 |
@HeikoS | andrew__: it is CMath:: but please ask google for such questions | 22:40 |
-!- PirosB3 [~pirosb3@cpc4-newt31-2-0-cust631.newt.cable.virginm.net] has quit [Quit: PirosB3] | 22:40 | |
-!- andrew__ [~andrew@50-0-142-100.dsl.dynamic.sonic.net] has quit [Ping timeout: 252 seconds] | 22:54 | |
-!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has quit [Quit: Leaving] | 23:07 | |
thoralf | lisitsyn: Hmm, seems that I'm not on the mentor list. | 23:13 |
thoralf | lisitsyn: Never received a mail from it. | 23:13 |
-!- sonne|osx [~sonne@e178249094.adsl.alicedsl.de] has quit [Quit: sonne|osx] | 23:16 | |
-!- travis-ci [~travis-ci@ec2-54-221-152-237.compute-1.amazonaws.com] has joined #shogun | 23:29 | |
travis-ci | [travis-ci] it's Fernando Iglesias'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/21536780 | 23:29 |
-!- travis-ci [~travis-ci@ec2-54-221-152-237.compute-1.amazonaws.com] has left #shogun [] | 23:29 | |
-!- andrew__ [~andrew@c-50-184-110-141.hsd1.ca.comcast.net] has joined #shogun | 23:44 | |
-!- khalednasr [~k.nasr92@41.69.141.226] has joined #shogun | 23:45 | |
--- Log closed Wed Mar 26 00:00:54 2014 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!