learn_to_pronounce.fst.FSTEvaluator

class learn_to_pronounce.fst.FSTEvaluator(fst_path: str)[source]

Evaluates FST given lexicon. Prints WER and PER, computed from comparing ground truth pronunciations and generated one

__init__(fst_path: str)[source]

constructor of fst evaluator

Parameters

fst_path: str

path to FST model to evaluate

evaluate(lexicon: PronunciationDictionary)[source]

Runs evaluation

Parameters

lexicon: PronunciationDictionary

words and ground truth pronunciations to evaluate on