Speech.
load_language_model
Load the N-gram language model.
Specifies the absolute server-side path of the acoustic model file.