I also had this idea and gave it a thought before 😀. Theoretically, mlpack libraries can be seamlessly used in android studio using android ndk.
Guys at tensorflow have achieved it (i guess the whole library is not fully working but classification works​. It can act as a reference point). I've also read that Snapdragon is working on optimizing tensorflow in their upcoming processor(Snapdragon 835?) which mlpack might be able benefit from in the future! And Marcus, I've skimmed through the projects ideas, "Essential Deep Learning modules" caught my attention mostly. Reinforced learning also seemed interesting(I am more interested in implementation of algorithms) but this Android idea is great! Thanks! On Mar 16, 2017 12:57 AM, "Ryan Curtin" <[email protected]> wrote: On Wed, Mar 15, 2017 at 05:02:24PM +0100, Marcus Edel wrote: > Hello Shoaib, > > welcome and thanks for getting in touch! > > > A little overview of my background, I am a CS major student in my 3rd year. I > > have also taken AI's course (which was majority focused on machine learning, and > > involved implementing algorithms). I've been also exploring online courses > > (coursera and udacity). Apart from this, I have been closely dealing and > > observing open source community at xda-developers(Link to profile). > > Maybe it's time to bring mlpack on android :) Also, I heard really good things > about the coursera and udacity courses. I think a very long time ago James Cline had mlpack programs running on his phone. I know that mlpack compiles fine on ARM (I have an ARM Chromebook I sometimes work from) so it should definitely still be possible. I guess, it could be an exciting application to train a neural network on a desktop, serialize it, and then load it on a phone to do local classification directly on the phone. -- Ryan Curtin | "This is how Number One works!" [email protected] | - Number One
_______________________________________________ mlpack mailing list [email protected] http://knife.lugatgt.org/cgi-bin/mailman/listinfo/mlpack
