Building Jarvis as an App for Local Languages

The jarvis app that is currently under development is in English. I want to configure it to use the local language. How to develop such an application for local languages? what programming languages โ€‹โ€‹should i know to move on to development? I tested the english version of jarvis, it works well for me. How to link C # to HTK for development purposes?

+3


source to share


2 answers


How to develop such an application for local languages?

You don't need to develop from scratch, use existing software and build it. For example, you can consider https://github.com/jasperproject/jasper-client , it is being developed quite actively.

what programming languages โ€‹โ€‹should i know to start developing?

Most of the NLP libraries are in Python or Java. You also need a shell script (awk / perl) because models are often built using Linux tools.

For speech recognition the easiest is to use CMUSphinx, a tutorial on adding your language to CMUSphinx is at http://cmusphinx.sourceforge.net/wiki/tutorialam .

I have tested the english version of jarvis, it works well for me. How to link C # to HTK for development purposes?



There are many ways to interact:

1) C # can call HTK tools as binaries via Process.Start http://msdn.microsoft.com/en-us/library/system.diagnostics.process.start(v=vs.110).aspx

2) You can create a library from HTK and call it with PInvoke via interop framework

3) You can create a TCP or HTTP server with HTK tools and connect to that server from a C # application to get speech recognition results.

All in all, you could probably use the existing solutions as mentioned above, they have all the tricky stuff, you only need to set up your local language.

+2


source


I would advise you to go for HTK or if you have a lot of training data then go to kaldi for one of the best speech recognition tools for a local language that uses deep learning.



0


source







All Articles