How to create your own language model for pocketsphinx?

Question

linux

Good afternoon!
There was a need to recognize "stop words" in wav files.
To do this, I decided to use the cmusphinx/pocketsphinx library.
I set everything up and put the Russian model off. site "zero_ru_cont_8k_v3.tar.gz".
And everything seems to be OK, but you need to recognize words that are not included in this model (for example, the name of the company).
I understand what is needed:
1) Add a word to the pronunciation dictionary (.dic) - done!
2) Adding a word to the language model (.lm) is a problem... Tell me
how to generate .lm for your words (in Russian)?
The .lm strings look something like this:
-5.7145 peacemaker -0.2282
-6.4940 peacemaker -0.2713
-5.8390 peacemaker -0.2622
-6.4512 peacekeepers -0.2025 Many thanks
in advance!