TokenizerTrainer#

class montreal_forced_aligner.tokenization.trainer.TokenizerTrainer(oov_count_threshold=5, **kwargs)[source]#

Bases: PyniniTrainerMixin, TokenizerMixin

property data_source_identifier#

Corpus name

export_model(output_model_path)[source]#

Export tokenizer model to specified path

Parameters:

output_model_path (Path) – Path to export model

finalize_training()[source]#

Finalize training

initialize_training()[source]#

Initialize training tokenizer model

property meta#

Metadata for exported tokenizer model

property sym_path#

Internal temporary symbol file

train()[source]#

Train a tokenizer model