paddlespeech.vector.utils.vector_utils module

paddlespeech.vector.utils.vector_utils.Q_from_tokens(token_num)

Get prior model, data from uniform, would support others(guassian) in future

paddlespeech.vector.utils.vector_utils.get_chunks(seg_dur, audio_id, audio_duration)

Get all chunk segments from a utterance

Args:

seg_dur (float): segment chunk duration, seconds audio_id (str): utterance name, audio_duration (float): utterance duration, seconds

Returns:

List: all the chunk segments