The non utf-8 characters are random as far as I can tell. Seems like corrupted memory, most likely an issue with the bindings. I cannot share the entire code base but could share small snippets. You'd better localize the problem on the server with a simple reproducible example.
Sorry, your questions are too basic, you can answer them yourself.
Cross compiling to arm : error after doing make
Already asked at https://github.com/cmusphinx/sphinxbase/issues/49
However, I could only display the Langugae HMM if it is statically produced, correct? There are several linguists, FlatLinguist is static, others are dynamic and create states on the fly. Usually you can not display a HMM because it is a very large structure, you can do it only for very simple cases. Or, finally, am I all wrong and Java clears up the entire HMM after decoding occurs? Core linguist structure stay alive, temporary objects are released.
It is better to write words into the grammar like "one", "thirty" and wrote the code to convert the text into actual number. This is what rasa is doing for example.
Please suggest me know that after applying the adaptaion I am getting accuracy for the words I was applying adaptaion for but it is side affecting other sentences. It could happen. Speech recognition algorithms are statistical and changes in parameters could cause side effects. Usually the more adaptation data you use, the more accurate system is. Please let me know where I am making a mistake. I do not see any mistake here.
I'm sorry, your understanding is very rough, lattices are not created from HMM, they are constructed during search process in SearchManager. You can read it all in the code of course. If you want an overview, start with a whitepaper: https://www.cs.cmu.edu/~rsingh/homepage/papers/sphinx4_eurospeech2003.pdf and, if you need more details, with a textbook: https://www.amazon.com/Spoken-Language-Processing-Algorithm-Development/dp/0130226165