TrainSpeakerLmArguments#
- class montreal_forced_aligner.language_modeling.multiprocessing.TrainSpeakerLmArguments(job_name, db_string, log_path, model_path, order, method, target_num_ngrams, hclg_options)[source]#
Bases:
MfaArgumentsArguments for
TrainSpeakerLmFunction- Parameters:
job_name (int) – Integer ID of the job
db_string (str) – String for database connections
log_path (str) – Path to save logging information during the run
model_directory (str) – Path to model directory
word_symbols_paths (dict[int, str]) – Per dictionary words symbol table paths
speaker_mapping (dict[int, str]) – Mapping of dictionaries to speakers
speaker_paths (dict[int, str]) – Per speaker output LM paths
oov_word (str) – OOV word
order (int) – Ngram order of the language models
method (str) – Ngram smoothing method
target_num_ngrams (int) – Target number of ngrams