fix: use converted token_ids for alignment for sensevoice model with timest...

Han Zhang

2025-03-18 3c349ac0531b07239f37b81254f8568ab80e3f6a

 fun_text_processing/inverse_text_normalization/de/verbalizers/verbalize.py

@@ -1,18 +1,21 @@

from fun_text_processing.inverse_text_normalization.de.verbalizers.cardinal import CardinalFst
from fun_text_processing.inverse_text_normalization.de.verbalizers.decimal import DecimalFst
from fun_text_processing.inverse_text_normalization.de.verbalizers.measure import MeasureFst
from fun_text_processing.inverse_text_normalization.de.verbalizers.money import MoneyFst
from fun_text_processing.inverse_text_normalization.de.verbalizers.time import TimeFst
from fun_text_processing.text_normalization.de.verbalizers.cardinal import CardinalFst as TNCardinalVerbalizer
from fun_text_processing.text_normalization.de.verbalizers.decimal import DecimalFst as TNDecimalVerbalizer
from fun_text_processing.text_normalization.de.verbalizers.cardinal import (
    CardinalFst as TNCardinalVerbalizer,
)
from fun_text_processing.text_normalization.de.verbalizers.decimal import (
    DecimalFst as TNDecimalVerbalizer,
)
from fun_text_processing.text_normalization.en.graph_utils import GraphFst


class VerbalizeFst(GraphFst):
    """
    Composes other verbalizer grammars.
    For deployment, this grammar will be compiled and exported to OpenFst Finate State Archiv (FAR) File. 
    For deployment, this grammar will be compiled and exported to OpenFst Finate State Archiv (FAR) File.
    More details to deployment at NeMo/tools/text_processing_deployment.
    """