nemo_curator.stages.audio

View as Markdown

Audio curation stages for NeMo Curator.

This module provides stages for processing and curating audio data, including ASR inference, quality assessment, ALM data preparation, audio preprocessing (mono conversion, segment concatenation, timestamp mapping), audio quality filtering (SIGMOS, UTMOS, bandwidth classification filtering), VAD segmentation, speaker diarization/separation, and advanced audio processing pipelines.

Subpackages

Submodules

Package Contents

Data

__all__

API

nemo_curator.stages.audio.__all__ = ['ALMDataBuilderStage', 'ALMDataOverlapStage', 'AudioDataFilterStage', 'BandFilt...