Do I need to create my own Language Model if I generate a ASR model?
Created by: PHDZheng
I used a data with 8000 Mandarin characters to train a the DeepSpeech model. I wonder if I need to generate an own Language Model based on the data, or I can still use the baidu released Language Model?
I also want to know what is the relationship between the 'vocab.txt' of the DeepSpeech model and the Language Model?