[mlpack] Fwd: GSOC Dataset and Experimentation tools

Ryan Curtin ryan at ratml.org
Fri Mar 18 01:00:41 EDT 2016


On Thu, Mar 17, 2016 at 08:15:35PM +0530, Chinmaya Bhat wrote:
> Hi,
> 
> I am a 2nd year undergraduate student in Computer Science from BMS
> Institute of Technology, Bangalore, India and am currently doing a course
> on Machine Learning on Coursera. I can code well in C and C++.
> 
> I had subscribed to the mailing lists and irc channels (nickname on irc :
> csbqwerty or Chinmaya) but I could not do much in these 2 weeks due to my
> exams.
> Now that they are complete, I want to contribute to mlpack.
> 
> 
> I am very interested in sending my proposals on 2 topics :
> 
>    1. Dataset and experimentation tools
>    2. More diverse build slaves (OSX an Windows)
> 
> I loved the idea for dataset and experimentation tools as I had some
> difficulty while using mlpack and using it and I think it would be highly
> used and helpful to many users.

Hi Chinmaya,

>    - How do I make the interface in the command line program? Can you
>    direct me to something similar?

Take a look at one of the other programs mlpack implements, and you can
see the PARAM_STRING() and PARAM_DOUBLE() and other macros and also the
PROGRAM_INFO() macro that are used:

https://github.com/mlpack/mlpack/blob/master/src/mlpack/methods/hmm/hmm_train_main.cpp

Then there is the CLI class which is used for accessing parameters:

http://mlpack.org/docs/mlpack-git/doxygen.php?doc=classmlpack_1_1CLI.html

Hopefully that clarifies things a little bit.  Let me know if it
doesn't.

>    - Also, is it too late for me to apply as I really wanted to do a lot
>    more and I feel that I dont have the time ?

The application deadline is March 25, so you still have a week to
prepare an application.  If you have more time now that your exams are
complete (hopefully they went well!), it's definitely possible to do
enough research and reading and prepare a good application.

Thanks,

Ryan

-- 
Ryan Curtin    | "Think!  What to do!"
ryan at ratml.org |   - The Master



More information about the mlpack mailing list