--- Log opened Tue Nov 18 00:00:13 2014 | ||
-!- HeikoS [~heiko@0545399b.skybroadband.com] has quit [Quit: Leaving.] | 00:07 | |
-!- HeikoS [~heiko@0545399b.skybroadband.com] has joined #shogun | 00:10 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 00:10 | |
-!- shaochuan [~shaochuan@c-50-184-81-180.hsd1.ca.comcast.net] has joined #shogun | 00:18 | |
-!- shaochuan [~shaochuan@c-50-184-81-180.hsd1.ca.comcast.net] has quit [Ping timeout: 265 seconds] | 00:22 | |
-!- HeikoS [~heiko@0545399b.skybroadband.com] has quit [Quit: Leaving.] | 00:24 | |
-!- iglesiasg [~iglesias@524B8E0B.cm-4-4c.dynamic.ziggo.nl] has quit [Quit: Lost terminal] | 00:24 | |
-!- pickle27 [~pickle27@192-0-136-118.cpe.teksavvy.com] has joined #shogun | 01:26 | |
-!- shaochuan [~shaochuan@c-50-184-81-180.hsd1.ca.comcast.net] has joined #shogun | 04:21 | |
-!- shaochuan [~shaochuan@c-50-184-81-180.hsd1.ca.comcast.net] has quit [Ping timeout: 240 seconds] | 04:25 | |
-!- Phawx [~Phawx@cpe-68-203-5-71.austin.res.rr.com] has joined #shogun | 04:27 | |
-!- Phawx [~Phawx@cpe-68-203-5-71.austin.res.rr.com] has quit [Ping timeout: 244 seconds] | 05:16 | |
-!- pickle27 [~pickle27@192-0-136-118.cpe.teksavvy.com] has quit [Remote host closed the connection] | 07:01 | |
-!- glongo [sid17594@gateway/web/irccloud.com/x-ovbkwllxfqtfzfwc] has quit [Ping timeout: 244 seconds] | 09:27 | |
-!- glongo [sid17594@gateway/web/irccloud.com/x-ndkyoaysdhnzatpj] has joined #shogun | 09:28 | |
-!- HeikoS [~heiko@0545399b.skybroadband.com] has joined #shogun | 09:58 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 09:58 | |
-!- PirosB3 [~pirosb3@195.101.111.125] has joined #shogun | 10:51 | |
-!- HeikoS [~heiko@0545399b.skybroadband.com] has quit [Quit: Leaving.] | 11:06 | |
-!- HeikoS [~heiko@0545399b.skybroadband.com] has joined #shogun | 12:54 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 12:54 | |
@lisitsyn | HeikoS: https://docs.google.com/forms/d/1kckcq_uv8dk9-W5rIdtqRwCHN4Uh209ELPUjTEZJDxc/viewform | 14:44 |
---|---|---|
@lisitsyn | wiking: ^ | 14:44 |
@wiking | ahahah rtofl | 14:47 |
@lisitsyn | wiking: I desperately fail on that | 14:50 |
@HeikoS | hahaha same here, great | 14:50 |
@lisitsyn | HeikoS: hey what about your priors? | 14:50 |
@lisitsyn | wiking: did you know HeikoS priors are too wide? | 14:51 |
@HeikoS | lisitsyn: your mothers prior beliefs are so wide she has to use a noninformative one | 14:52 |
@lisitsyn | hahah | 14:52 |
@lisitsyn | HeikoS: I just seen this link in twitter | 14:52 |
@lisitsyn | you -> dino -> and some guy I follow -> me again | 14:53 |
@HeikoS | jaja | 14:54 |
@HeikoS | nice | 14:54 |
@HeikoS | dino posted it | 14:54 |
@lisitsyn | HeikoS: there should be some index like | 14:55 |
@lisitsyn | how much twitter users did it pass through to get back to you | 14:55 |
@lisitsyn | err :D | 14:56 |
@lisitsyn | how many I mean :D | 14:56 |
@HeikoS | hehe | 14:56 |
@HeikoS | i did not even use twietter | 14:56 |
@HeikoS | 14:56 | |
@HeikoS | but email :D | 14:56 |
@lisitsyn | yeah true | 14:56 |
@lisitsyn | then any kind of communication works | 14:56 |
-!- HeikoS [~heiko@0545399b.skybroadband.com] has quit [Quit: Leaving.] | 15:12 | |
-!- pickle27 [~pickle27@192-0-136-118.cpe.teksavvy.com] has joined #shogun | 15:29 | |
-!- HeikoS [~heiko@pat-191-250.internal.eduroam.ucl.ac.uk] has joined #shogun | 15:48 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 15:48 | |
-!- pickle27 [~pickle27@192-0-136-118.cpe.teksavvy.com] has quit [Remote host closed the connection] | 15:54 | |
@HeikoS | wiking: around? | 16:00 |
@HeikoS | lisitsyn: around? | 16:00 |
@lisitsyn | HeikoS: ja | 16:00 |
@HeikoS | lisitsyn: how to disable libcurl in shogun compilation? | 16:00 |
@HeikoS | whats the switch? | 16:01 |
@HeikoS | FIND_PACKAGE(CURL) | 16:01 |
@HeikoS | IF (CURL_FOUND) | 16:01 |
@HeikoS | SET(HAVE_CURL 1) | 16:01 |
@HeikoS | LIST(APPEND DEFINES HAVE_CURL) | 16:01 |
@HeikoS | LIST(APPEND INCLUDES ${CURL_INCLUDE_DIRS}) | 16:01 |
@HeikoS | SET(POSTLINKFLAGS ${POSTLINKFLAGS} ${CURL_LIBRARIES}) | 16:01 |
@HeikoS | ENDIF(CURL_FOUND) | 16:01 |
@HeikoS | thats in CMakeLists.txt | 16:01 |
@HeikoS | google doens thelp me | 16:01 |
@lisitsyn | HeikoS: hmm | 16:01 |
@lisitsyn | well you can comment it out :D | 16:02 |
@lisitsyn | the thing is that once you're into that IF | 16:02 |
@lisitsyn | you have includes etc | 16:02 |
@HeikoS | yeah I know thats why I wanna disable | 16:02 |
@lisitsyn | HeikoS: let me check curl finder cmake script | 16:02 |
@HeikoS | ah got it | 16:03 |
@lisitsyn | HeikoS: what? | 16:04 |
@lisitsyn | HeikoS: I see no option to fail FIND_PACKAGE if it founds one | 16:04 |
@HeikoS | cmake -DEnableCURL=Off .. | 16:05 |
@lisitsyn | hmm really? | 16:05 |
@lisitsyn | HeikoS: I wonder where this option is handled :D | 16:06 |
@HeikoS | find .. -type f -name '*.txt' | xargs grep 'CURL' | 16:06 |
@HeikoS | #../build/CMakeCache.txt:EnableLibCURL:UNINITIALIZED=Off | 16:07 |
@lisitsyn | hmm cool to know | 16:07 |
@HeikoS | lisitsyn: btw see esbens email | 16:08 |
@HeikoS | wiking: ^ | 16:08 |
@HeikoS | about meta language project | 16:08 |
@lisitsyn | HeikoS: oh sh I forgot to write a mail about examples of dsl :D | 16:08 |
@lisitsyn | sorry | 16:08 |
@HeikoS | lisitsyn: thats optional anyways | 16:08 |
@HeikoS | lisitsyn: we are now starting to prototype | 16:08 |
@HeikoS | so see email :) | 16:08 |
@lisitsyn | heh | 16:09 |
@lisitsyn | okay well I don't mind actually | 16:09 |
@lisitsyn | I mean writing parsers is like | 16:09 |
@lisitsyn | useless in general but useful for him | 16:09 |
@lisitsyn | HeikoS: I am now reading bishop | 16:11 |
@lisitsyn | breaks my mind actually | 16:11 |
@HeikoS | lisitsyn: I think this will be super useful for us | 16:12 |
@lisitsyn | HeikoS: yeah absolutely | 16:12 |
@HeikoS | lisitsyn: let me know if you have questions, but its a very good book, almost the best on the topic | 16:12 |
@lisitsyn | my point was just like | 16:12 |
@lisitsyn | you can just skip the parsing thing | 16:12 |
@lisitsyn | and use flexible syntax of some language | 16:12 |
@lisitsyn | HeikoS: e.g. in scala you don't even write ( ) for functions | 16:12 |
@lisitsyn | legal to write "do something" | 16:13 |
@lisitsyn | HeikoS: like you describe code and description is transformed into examples :D | 16:14 |
@HeikoS | lisitsyn: but how do you want to do this then? | 16:14 |
@HeikoS | so you mean skip inventing own language? | 16:14 |
@HeikoS | and use say scala | 16:14 |
@lisitsyn | yeah that was my idea | 16:15 |
@HeikoS | I see | 16:15 |
@HeikoS | mmmh | 16:15 |
@lisitsyn | I mean all these languages are pretty flexible | 16:15 |
@HeikoS | not sure | 16:15 |
@HeikoS | how to represent the example? | 16:15 |
@HeikoS | how to translate? | 16:15 |
@HeikoS | then we need to write converter for every target | 16:15 |
@lisitsyn | it is easier to do inside language | 16:15 |
@lisitsyn | I think | 16:15 |
@lisitsyn | HeikoS: I am now thinking like | 16:15 |
@lisitsyn | did you consider really textual description of example? | 16:16 |
@lisitsyn | I mean | 16:16 |
@lisitsyn | initialize LibLinear as svm with C=1.0 ----> svm = LibLinear(1.0) | 16:16 |
@HeikoS | no | 16:17 |
@HeikoS | we will have | 16:17 |
@HeikoS | LibLinear ll = LibLinear(1.0= | 16:18 |
@HeikoS | will be typed | 16:18 |
@HeikoS | since also translating into c++ | 16:18 |
@HeikoS | and java etc | 16:18 |
@HeikoS | see my gist | 16:18 |
@HeikoS | http://www.shogun-toolbox.org/page/Events/gsoc2014_ideas#meta_language | 16:18 |
@HeikoS | https://gist.github.com/karlnapf/9079385 | 16:19 |
@lisitsyn | HeikoS: oh that's just C++ :) | 16:19 |
@HeikoS | yeah basically yes | 16:19 |
@HeikoS | thing is | 16:19 |
@HeikoS | if we do this in c++ | 16:19 |
@HeikoS | how to transform? | 16:19 |
@HeikoS | we need to parse | 16:19 |
@HeikoS | I mean we can use c++ parser maybe | 16:19 |
@lisitsyn | I don't really get it | 16:20 |
@HeikoS | but how to extract the things we want? | 16:20 |
@HeikoS | I dont see what you mean, maybe describe an example how to do things | 16:20 |
@lisitsyn | HeikoS: extract things from? | 16:20 |
@HeikoS | or better prototype | 16:20 |
@lisitsyn | hmm yeah ok easy | 16:20 |
@lisitsyn | HeikoS: just some kind of 10 minutes let me set up some regexps :) | 16:23 |
@HeikoS | ok! | 16:23 |
@HeikoS | this might be useful for esben, | 16:23 |
@HeikoS | (should have come earlier, now we already got a concept) | 16:23 |
@HeikoS | but happy to change | 16:23 |
@lisitsyn | HeikoS: I am sorry you know I am like failing everywhere :) | 16:24 |
@HeikoS | haha | 16:24 |
@lisitsyn | HeikoS: ok still here? | 16:58 |
@lisitsyn | https://gist.github.com/lisitsyn/38662774c3f594cc6580 | 16:58 |
@lisitsyn | HeikoS: put output there as comment as well | 17:00 |
@HeikoS | checking | 17:15 |
@HeikoS | lisitsyn: thats fine, but how to do do this for java? | 17:15 |
@lisitsyn | HeikoS: can't see the problem | 17:16 |
@HeikoS | ok? | 17:16 |
@lisitsyn | HeikoS: well I mean | 17:16 |
@HeikoS | let me look at this a bit more | 17:16 |
@lisitsyn | just wrap it into some class | 17:16 |
@lisitsyn | public static main | 17:16 |
@lisitsyn | blabla | 17:16 |
@HeikoS | how to do UNREF at the end of scope? | 17:17 |
@HeikoS | how to do the type inference? | 17:17 |
@HeikoS | declarations? | 17:17 |
@HeikoS | can you do a c++ example? | 17:17 |
@HeikoS | I mean this is great | 17:17 |
@HeikoS | just want to be sure this works fine | 17:17 |
@lisitsyn | HeikoS: what type inference? | 17:18 |
@lisitsyn | HeikoS: it is still typed actually | 17:18 |
@lisitsyn | so I see no type inference involved | 17:19 |
@HeikoS | lisitsyn: mmmh | 17:19 |
@HeikoS | what I mean is | 17:19 |
@HeikoS | say | 17:19 |
@lisitsyn | HeikoS: as for unref and declarations - some matchers here are to initialize variable so it is possible to track whether it is first declaration or second (assignment) | 17:19 |
@lisitsyn | same for unref | 17:19 |
@lisitsyn | ha btw | 17:20 |
@lisitsyn | we can also draw graphs | 17:20 |
@lisitsyn | :D | 17:20 |
@lisitsyn | what is used by what | 17:20 |
@HeikoS | thats cool | 17:20 |
@lisitsyn | so | 17:20 |
@lisitsyn | what did you mean? | 17:20 |
@HeikoS | lisitsyn: so you suggest to go this way: a python script that does things, and one just changes the dictionaries? | 17:20 |
@HeikoS | yes I see | 17:20 |
@lisitsyn | HeikoS: oh no actually no | 17:20 |
@lisitsyn | it is just how I'd prototype that :) | 17:21 |
@lisitsyn | it is still better to do with parsing library | 17:21 |
@HeikoS | lisitsyn: what would be your suggestion for esbens project? | 17:21 |
@HeikoS | lisitsyn: I see | 17:21 |
@HeikoS | good idea | 17:21 |
@lisitsyn | just write some kind of allowed sentences | 17:21 |
@HeikoS | kk | 17:21 |
@HeikoS | can you send this to esben and explain your reasoning? | 17:21 |
@lisitsyn | initialize [class] from [parameters] as [name] | 17:21 |
@lisitsyn | yes | 17:21 |
@HeikoS | maybe we want to change the syntax of the meta language then to your idea | 17:22 |
@HeikoS | as its way cleaner/easier to write, less cluttered | 17:22 |
@lisitsyn | HeikoS: yeah may be it is worth it | 17:22 |
@HeikoS | totally | 17:22 |
@HeikoS | make sure you explain it properly that the poor guy does not get confused ;) | 17:22 |
@lisitsyn | HeikoS: and put VW on top of it to predict programs | 17:22 |
@HeikoS | haha | 17:22 |
@lisitsyn | or RNN | 17:23 |
@lisitsyn | have you seen they are doing pretty well generating text | 17:23 |
@lisitsyn | with RNNs | 17:23 |
@HeikoS | lisitsyn: btw esben is a quite smart guy, we should drag him in while he is doing his project ;) | 17:23 |
@lisitsyn | HeikoS: no doubt :) | 17:23 |
@HeikoS | yeah we had a talk on that here recently by a guy from M$ | 17:23 |
@HeikoS | quite cool | 17:23 |
@HeikoS | for integrated IDEs | 17:23 |
@HeikoS | they learn from code what to suggest to complete | 17:23 |
@HeikoS | autocomplete | 17:23 |
@HeikoS | like whole program snippets | 17:24 |
@lisitsyn | HeikoS: well if you coded in intellij idea :D | 17:24 |
@lisitsyn | it wouldn't be of real surprise | 17:24 |
@HeikoS | no for java I think | 17:24 |
@HeikoS | or c# | 17:24 |
@lisitsyn | for java it is really easy | 17:24 |
@lisitsyn | the language itself is kind of designed for monkeys | 17:24 |
@HeikoS | lisitsyn: can you send the stuff today? esben already started a few days ago | 17:24 |
@HeikoS | its good to get input as early as possible | 17:24 |
@lisitsyn | I am just writing the mail | 17:24 |
@HeikoS | awesome thanks a lot! | 17:25 |
@lisitsyn | HeikoS: ooh it seems I confuse his first/last name | 17:33 |
@lisitsyn | it is written like Soerig, Esben | 17:33 |
@lisitsyn | :D | 17:33 |
@lisitsyn | it is like confusing when you don't really understand first/last name | 17:36 |
@lisitsyn | what is last and what is first i mean | 17:36 |
@HeikoS | haha | 18:15 |
@HeikoS | esben is first, lisitsyn | 18:15 |
@lisitsyn | HeikoS: yeah I got it | 18:15 |
@lisitsyn | HeikoS: I sent him a letter explaining that :D | 18:15 |
@lisitsyn | HeikoS: can you please update me on what are relations of shogun and stan? | 18:16 |
@HeikoS | lisitsyn: shogun and stan, why? | 18:17 |
@lisitsyn | HeikoS: I am pretty sure I heard sth about stan from you | 18:17 |
@lisitsyn | but I can't really remember what | 18:17 |
@HeikoS | lisitsyn: yes, I want to have shogun have bindings to stan | 18:17 |
@HeikoS | for two things: | 18:17 |
@lisitsyn | HeikoS: how? | 18:17 |
@lisitsyn | ehmm | 18:17 |
@HeikoS | autodiff, and mcmc | 18:17 |
@lisitsyn | what are bindings? | 18:18 |
@HeikoS | lisitsyn: like use stan from within shogun | 18:18 |
@lisitsyn | aha I see | 18:18 |
@HeikoS | say implement shoguns kernel functions in stan, then we get the gradients for free | 18:18 |
@HeikoS | and probabilistic models in shogun can also be learned using stan's great hmc sampler | 18:18 |
@lisitsyn | HeikoS: so we just need a way to IO data? | 18:19 |
@HeikoS | lisitsyn: no its more complicated | 18:19 |
@lisitsyn | is it like you just execute some stan program with your data and get it in shogun format as well? | 18:19 |
@HeikoS | stan has its own language | 18:19 |
@HeikoS | that is translated into c++ code | 18:19 |
@HeikoS | take say a GP | 18:20 |
@HeikoS | we have lots of different GP models in Shogun | 18:20 |
@HeikoS | we can write those models in stan, in the stan language | 18:20 |
@lisitsyn | yeah | 18:21 |
@HeikoS | then we can offer inference for those models under the Shogun interface | 18:21 |
@HeikoS | but internally, stan is used | 18:21 |
@HeikoS | but thats hidden from user | 18:21 |
@HeikoS | thats one use case | 18:21 |
@lisitsyn | I see | 18:21 |
@HeikoS | kind of cool to be able to switch between approximate inference and MCMC under the same interface | 18:21 |
@HeikoS | the other thing is | 18:21 |
@HeikoS | autodiff | 18:21 |
@HeikoS | we have lots of code that just implements gradient evaluataions in Shogun | 18:22 |
@lisitsyn | how does it work? | 18:22 |
@HeikoS | if we would write the functions we want the gradients of (say derivative of Gaussian kernel wrt its length-scale sigma), in stan, then we would get this for free | 18:22 |
@HeikoS | it takes the stan language snippet, transforms it into some kind of symbolic representation | 18:22 |
@lisitsyn | I am now browsing stan's code | 18:23 |
@HeikoS | and then generates c++ code for both evaluation and gradient | 18:23 |
@HeikoS | check wikipedia of automatic differentiation | 18:23 |
@lisitsyn | yeah I know about auto diff | 18:23 |
@HeikoS | so the cool thing is | 18:23 |
@HeikoS | you get c++ code that evaluates the gradient | 18:23 |
@lisitsyn | I just meant how it works in stan | 18:23 |
@lisitsyn | yeah I see | 18:23 |
@HeikoS | so its as fast as evaluation the function itself | 18:23 |
@lisitsyn | I am like confused | 18:23 |
@lisitsyn | :D | 18:24 |
@lisitsyn | its pretty good engineered | 18:24 |
@HeikoS | so if we put this in our build process, i.e. generating the c++ code from stan code, and the compiling c++ code and expose it via shogun | 18:24 |
@HeikoS | thats cool | 18:24 |
@HeikoS | yes | 18:24 |
@HeikoS | they have a full time guy | 18:24 |
@HeikoS | for a couple of years now | 18:24 |
@HeikoS | its a great great project | 18:24 |
@lisitsyn | I mean it is really like good C++ code :D | 18:24 |
@HeikoS | yes | 18:24 |
@HeikoS | I know | 18:24 |
@lisitsyn | would not be possible for science people to write such code | 18:24 |
@HeikoS | they use a lot of template programming | 18:24 |
@HeikoS | no the have a software engineering guy (that also knows about mcmc) | 18:25 |
@HeikoS | lisitsyn: actually, most of stan is the autodiff | 18:25 |
@HeikoS | the mcmc is just a very small part | 18:25 |
@lisitsyn | I see | 18:25 |
@HeikoS | lisitsyn: they generalised eigen3 for the autodiff | 18:26 |
@HeikoS | thats why it is so cool | 18:26 |
@HeikoS | all the autodiff happens via template programming | 18:26 |
@HeikoS | I asked them a few times to release the autodiff seperately | 18:26 |
@HeikoS | they always told me to officially request this on the mailing list, and that its quite easy | 18:26 |
@lisitsyn | HeikoS: so you need gcc to do that? | 18:27 |
@lisitsyn | or clang whatever | 18:27 |
@HeikoS | I think they use boost something | 18:27 |
@HeikoS | not sure about internals | 18:27 |
@lisitsyn | HeikoS: I mean is it really compiling C++ code on the run? | 18:27 |
@HeikoS | lisitsyn: no | 18:28 |
@HeikoS | the stanc program takes a stan language listing and produces standard c++ code | 18:28 |
@HeikoS | this can then be compiled as normal | 18:28 |
@lisitsyn | ah I see | 18:28 |
@HeikoS | so two steps to generate a stan program | 18:28 |
@lisitsyn | got it | 18:28 |
@HeikoS | 1. use stanc | 18:28 |
@HeikoS | 2. use gcc | 18:28 |
@lisitsyn | but you still need compiler around | 18:29 |
@HeikoS | yeah sure | 18:29 |
@HeikoS | stanc generate c++ code | 18:29 |
@lisitsyn | that's pretty cool | 18:29 |
@HeikoS | indeed | 18:29 |
@HeikoS | its great to have this layer, as allows for autodiff | 18:29 |
@HeikoS | would be good to do this in shogun too | 18:29 |
@lisitsyn | shogunc? | 18:29 |
@lisitsyn | :D | 18:29 |
@HeikoS | also you get backend independence on the fly | 18:29 |
@HeikoS | yeah some kind of | 18:29 |
@lisitsyn | well | 18:29 |
@HeikoS | like say you have CKernel | 18:30 |
@lisitsyn | HeikoS: we are too general still | 18:30 |
@HeikoS | CGaussianKernel's compute method can be generated | 18:30 |
@lisitsyn | no clear vision what we focus on | 18:30 |
@HeikoS | no i dont agree | 18:30 |
@HeikoS | but certainly not this idea | 18:30 |
@HeikoS | this is just a cool idea | 18:30 |
@HeikoS | but focus of other things is pretty clear | 18:30 |
@HeikoS | see all the latest discussions :) | 18:30 |
@lisitsyn | yeah we can add some template stuff here | 18:30 |
@lisitsyn | but actually | 18:30 |
@lisitsyn | it is not worth it :D | 18:30 |
@lisitsyn | I mean people use clusters and stuff | 18:31 |
@HeikoS | ? | 18:31 |
@lisitsyn | HeikoS: I mean templates are good to get crazy performance | 18:31 |
@HeikoS | ah i see | 18:32 |
@HeikoS | yeah sure sure | 18:32 |
@lisitsyn | but people just throw more power usually | 18:32 |
@HeikoS | this is all not for this or next year | 18:32 |
@HeikoS | we got other problems | 18:32 |
@HeikoS | lambdays ideas for cleaning up classes are good | 18:32 |
@lisitsyn | HeikoS: yeah sure | 18:32 |
@HeikoS | but we need to release first | 18:32 |
@HeikoS | and then next gsoc also starts soon | 18:32 |
@HeikoS | so need ideas, mentory, etc | 18:33 |
@HeikoS | mentors | 18:33 |
@HeikoS | and documentation | 18:33 |
@HeikoS | binareis | 18:33 |
@HeikoS | all that | 18:33 |
@HeikoS | hackathon | 18:33 |
@HeikoS | funding | 18:33 |
@HeikoS | argh | 18:33 |
@HeikoS | too much | 18:33 |
@HeikoS | foundation | 18:33 |
@lisitsyn | :D | 18:33 |
@lisitsyn | crazzzy | 18:33 |
@HeikoS | its all feasible | 18:34 |
@HeikoS | just some work | 18:34 |
@HeikoS | get rid of static interfaces | 18:34 |
@HeikoS | make swig matlab work first | 18:34 |
@HeikoS | clena up testing framework | 18:34 |
@HeikoS | clean up base classes | 18:34 |
@HeikoS | drop serialisation | 18:34 |
@HeikoS | drop migration | 18:34 |
@HeikoS | drop legacy code that is not tested | 18:34 |
@HeikoS | make binaries work smoothly | 18:35 |
@HeikoS | mac build | 18:35 |
@HeikoS | if all that was in | 18:35 |
@HeikoS | would be super awesome | 18:35 |
@HeikoS | as we got enough ML algos ;) | 18:35 |
@HeikoS | and then next step: speed | 18:35 |
@HeikoS | swig build split | 18:35 |
@HeikoS | (actually that one also soon as it blows up already) | 18:35 |
-!- shaochuan [~shaochuan@c-50-184-81-180.hsd1.ca.comcast.net] has joined #shogun | 18:57 | |
-!- PirosB3 [~pirosb3@195.101.111.125] has quit [Quit: PirosB3] | 19:00 | |
-!- shaochuan [~shaochuan@c-50-184-81-180.hsd1.ca.comcast.net] has quit [Ping timeout: 240 seconds] | 19:03 | |
-!- HeikoS [~heiko@pat-191-250.internal.eduroam.ucl.ac.uk] has quit [Ping timeout: 256 seconds] | 19:43 | |
-!- shaochuan [~shaochuan@c-50-184-81-180.hsd1.ca.comcast.net] has joined #shogun | 22:27 | |
-!- shaochuan [~shaochuan@c-50-184-81-180.hsd1.ca.comcast.net] has quit [Ping timeout: 255 seconds] | 22:32 | |
-!- pickle27 [~pickle27@192-0-136-118.cpe.teksavvy.com] has joined #shogun | 23:32 | |
--- Log closed Wed Nov 19 00:00:15 2014 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!