ChanServ changed the topic of #mlpack to: "mlpack: a fast, flexible machine learning library :: We don't always respond instantly, but we will respond; please be patient :: Logs at http://www.mlpack.org/irc/
yslking has joined #mlpack
yslking has quit [Remote host closed the connection]
ImQ009 has joined #mlpack
< NippunSharmaGitt> Hey, I have a doubt. [here](https://github.com/mlpack/mlpack/blob/28d03e5dd73245e29212caae31a75b9874661eb0/CMakeLists.txt#L12) we have a defined an option to build matlab bindings ? Does mlpack has bindings in matlab ?
< shrit[m]> Not yet
< NippunSharmaGitt> okay, thanks
william has joined #mlpack
william is now known as Guest38022
Guest38022 has quit [Quit: Connection closed]
< anjishnu[m]> Is the improvisation and implementation of Ann modules project there for this year's gsoc as well?
< ABHINAVANAND[m]> Yes.
< anjishnu[m]> Okay. Who is mentoring this year?
< ABHINAVANAND[m]> The potential mentor given on the page is Sreenik Seal.
< anjishnu[m]> Hmmm I just checked.
< anjishnu[m]> I think the project hasn't been updated from last year
< ABHINAVANAND[m]> Does mlpack have archives of previous years proposals? It can be used as reference while writing a proposal.
< anjishnu[m]> I don't think you can find complete proposals archived, but you can probably find brief proposals from the mailing archives
< zoq> ABHINAVANAND[m]: The improvisation and implementation of ann modules, is open-ended so if you have something you like to work on under the project feel free to propose something :)
< ABHINAVANAND[m]> Yes I have made a first draft of the proposal.<https://docs.google.com/document/d/11nGFwhl658RUwX9S8QQvrvDU6g-Yh-dEeEEMmA8TsUM/edit?usp=sharing|https://docs.google.com/document/d/11nGFwhl658RUwX9S8QQvrvDU6g-Yh-dEeEEMmA8TsUM/edit?usp=sharing>
< ABHINAVANAND[m]> I wanted to know if the description is sufficient or should I add more details of how will I implement the function.
< zoq> NippunSharmaGitt: If you are okay, with using an older version of mlpack you can find MATLAB bindings.
< zoq> ABHINAVANAND[m]: THis year GSoC is going to be shorter as previous years, looking at the draft, I would recommend to focus on two or three layers.
< zoq> ABHINAVANAND[m]: Keep in mind testing and documentation takes time.
< zoq> ABHINAVANAND[m]: In case you haven't seen it already - https://github.com/mlpack/mlpack/wiki/Google-Summer-of-Code-Application-Guide provides some pointers as well.
< zoq[m]> Hello https://www.mlpack.org/gsoc.html mainly the first thing todo is to get familiar with codebase.
lazaros65 has joined #mlpack
lazaros65 has quit [Ping timeout: 240 seconds]
< rcurtin[m]> Nippun Sharma (Gitter): nice catch, want to open a PR to remove that line? we haven't had any matlab bindings in quite a long time
< NippunSharmaGitt> Sure, I will open the PR.
< rcurtin[m]> 👍️ thanks!
< rcurtin[m]> there might be a reference to that option in `doc/guide/build.hpp` too, not sure
< NippunSharmaGitt> Okay, I will check that 👍
< jonpsy[m]> hi zoq I think there's a potential mess-up
< jonpsy[m]> in our ensmallen, problems module. The ```Shuffle()``` module shuffles the indices of the population
< jonpsy[m]> it expands to ```arma::shuffle(arma::linspace<arma::Col<size_t> >(0, populationSize -1))```
< jonpsy[m]> thing is, ```arma::linspace``` generates 100 sample by default. So, say if you have popsize of 150, you're missing 50 of your population
< jonpsy[m]> * in our ensmallen, problems module. The `Shuffle()` method shuffles the indices of the population
< rcurtin[m]> jonpsy: interesting, nice catch---if you make the change and add `populationSize` as the last argument, do all the tests still pass?
< jonpsy[m]> Lemme check and get back to you :)
< jonpsy[m]> I'll probably create an issue in ensmallen. That'd be better, no?
< rcurtin[m]> sure, that works, you can also just create a PR... I'm pretty sure adding `populationSize` would be a fix, we just need to make sure the tests don't strangely fail or something first :)
< rcurtin[m]> (there could be some reason someone didn't put the `populationSize` there, not sure... if the tests fail, maybe there was a reason, if the tests pass, then maybe the `populationSize` was just forgotten :))
< jonpsy[m]> Yeah, that's why I didn't create an issue an wanted to consult first xD
< rcurtin[m]> 👍️
< jonpsy[m]> > it expands to `arma::shuffle(arma::linspace<arma::Col<size_t> >(0, populationSize -1))`
< jonpsy[m]> Ah, I think the user who's PR I'm continuing did this mistake. This code isn't there in master, I should've known better.
gitter_rituraj-c has joined #mlpack
< gitter_rituraj-c> Hello everyone, myself Rituraj Dutta , from Gauhati University of Science and Technology pursuing my Btech in Information Technology with minors in Artificial Intelligence ,and I'm interested in the project mentioned in the organization's idea list which is "Algorithm Optimization". I have a good grasp over most of the Machine Learning Algorithmic Techniques and their underlying working criteria. I have a good
< gitter_rituraj-c> knowledge behind their intuition as I have read some articles written by famous people such as Tom Mitchell and many others. I also have a decent skill in problem solving and a good control over C/C++. I am looking forward to contribute to this library and expand my knowledge further upon Machine Learning. It would be really helpful for me from the mentors and community side to guide and support me.
< rcurtin[m]> Hi gitter_rituraj-commits , welcome to the community! A good starting point might be to look at http://www.mlpack.org/community.html and http://www.mlpack.org/gsoc.html
< rcurtin[m]> hmm, I guess that needs to be updated to point out that we were in GSoC 2020 too 😃
< RishabhGargGitte> Hey @rcurtin @zoq , can you please take a look at #2746 . I think it's complete now. All the suggested changes have been applied. Thanks a lot :)
< RishabhGargGitte> Hey all, can someone please explain me what do we have to write in tests for bindings? I am working on adding binding for a new feature. Now I have to write tests for it and I am not sure what do we have to test for in the tests for bindings ?
< rcurtin[m]> Basically in the tests, we just want to ensure that all the options work as desired, and that reasonable errors are thrown when invalid options are passed. There's no need to test the actual algorithms themselves, as that should be done in the test for the algorithm
< rcurtin[m]> so, e.g., `mlpack_knn`'s options are tested in `src/mlpack/tests/main_tests/knn_test.cpp` but the actual KNN algorithm is tested in `src/mlpack/tests/knn_test.cpp`
< kaushal07[m]> can anyone help??
< jonpsy[m]> kaushal07: i suppose you get this error during runtime?
< kaushal07[m]> yes
< jonpsy[m]> undefined reference means it couldn't find the definition
< jonpsy[m]> my naive guess would be you haven't installed boost
< jonpsy[m]> have you?
< kaushal07[m]> yes it is installed
< kaushal07[m]> I should make the build file in the examples folder
< jonpsy[m]> whats happening is your linker is unable to find the right definition for ```boost::.....::is_equal```
< jonpsy[m]> this could be due to, boost version mismatch, among other things
< jonpsy[m]> maybe you could try creating an issue with all the specifications? :)
< kaushal07[m]> <jonpsy[m] "this could be due to, boost vers"> Even I think so
< kaushal07[m]> <jonpsy[m] "maybe you could try creating an "> ok will see
< kaushal07[m]> maybe I should try to use linker command to link with the boost
< jonpsy[m]> https://stackoverflow.com/a/31920258 kaushal07 try this
< kaushal07[m]> kaushal_choudhary@DESKTOP-GSRINP3:~/examples/lstm_stock_prediction$ make
< kaushal07[m]> make: Nothing to be done for 'default'.
< kaushal07[m]> > <@kaushal07:matrix.org> kaushal_choudhary@DESKTOP-GSRINP3:~/examples/lstm_stock_prediction$ make
< kaushal07[m]> > make: Nothing to be done for 'default'.
< kaushal07[m]> now its showing like this
< jonpsy[m]> Make is super messy :)
< jonpsy[m]> You most probably are missing a tab space
< jonpsy[m]> I'd suggest use vim for makefiles
< kaushal07[m]> <jonpsy[m] "You most probably are missing a "> where??
< jonpsy[m]> A common mistake which happens is
< jonpsy[m]> Space to tab. Vim needs tab only in commands, you can't use space.
< kaushal07[m]> ok
< jonpsy[m]> * Space to tab. Make needs tab only in commands, you can't use space.
< jonpsy[m]> Try [this](https://tuxthink.blogspot.com/2013/04/make-nothing-to-be-done-for-default.html?m=1#:~:text=make%3A%20Nothing%20to%20be%20done%20for%20%60default'.,-While%20writing%20make&text=The%20common%20cause%20for%20the,the%20target%20from%20the%20prerequisite)
< kaushal07[m]> still not working
< jonpsy[m]> Bummer.., I suggest creating an issue. Also, post all your specs (including boost ver.) And the commands you used. Good luck!
< kaushal07[m]> okay will try something else a little bit
< kaushal07[m]> and then open an issue
kaushal07[m] is now known as kaushal[m]
ImQ009 has quit [Quit: Leaving]
< zoq> kaushal[m]: What version of mlpack are you using to build the example? From the message I can see you are not linking against boost serialization, which is not necessary if you build like the latest master branch.