Speech to text decoding of text

What is the decoding method used in OpenNMT-py to decode the target text?
I learned about

  1. Language model
    2)Acoustic model
    3)Lexicon model

How to implement the lexicon model in OpenNMT-py?

I reckon you’d better have a look at Kaldi if you’re interested in such models.

This is written in C++.
Any other which is written in python?