运行infer_diarization.py报错，pyannote.audio需要下载什么版本的

[Cyl147369991]

Traceback (most recent call last):
File "e:\publicproject\3D-Speaker\speakerlab\bin\infer_diarization.py", line 40, in
from pyannote.audio import Inference, Model
File "E:\publicproject\3D-Speaker\venv\lib\site-packages\pyannote\audio_init_.py", line 29, in
from .core.inference import Inference
File "E:\publicproject\3D-Speaker\venv\lib\site-packages\pyannote\audio\core\inference.py", line 35, in
from pyannote.audio.core.io import AudioFile
File "E:\publicproject\3D-Speaker\venv\lib\site-packages\pyannote\audio\core\io.py", line 60, in
) -> torchaudio.AudioMetaData:
AttributeError: module 'torchaudio' has no attribute 'AudioMetaData'

[JiJiJiang]

可以试试3.2.0~3.3.1, 亲测3.3.1可以 python3.10