Function _encode_reference_audio_codes

clis/moss_ttsd_app.py:357–368

(
    processor,
    clone_wavs: list[torch.Tensor],
    cloned_speakers: list[int],
    speaker_count: int,
    sample_rate: int,
)

def _encode_reference_audio_codes(
    processor,
    clone_wavs: list[torch.Tensor],
    cloned_speakers: list[int],
    speaker_count: int,
    sample_rate: int,
) -> list[Optional[torch.Tensor]]:
    encoded_list = processor.encode_audios_from_wav(clone_wavs, sampling_rate=sample_rate)
    reference_audio_codes: list[Optional[torch.Tensor]] = [None for _ in range(speaker_count)]
    for speaker_id, audio_codes in zip(cloned_speakers, encoded_list):
        reference_audio_codes[speaker_id - 1] = audio_codes
    return reference_audio_codes


def build_conversation(