将声音文件导入为 sf 但在 jupyter notebook sf 上给出 NameError

Question

I installed Soundfile by pip install soundfile. Here is my code:

import soundfile as sf
def speech_file_to_array_fn(batch):
    speech_array, sampling_rate = sf.read(batch["file"])
    batch["speech"] = speech_array
    batch["sampling_rate"] = sampling_rate
    batch["target_text"] = batch["text"]
    return batch

timit = timit.map(speech_file_to_array_fn, remove_columns=timit.column_names["train"], num_proc=4)

but it gives an NameError. NameError: name 'sf' is not defined

I could not see the problem, I work on jupyter notebook.

Answer 1

Downgrading to version 0.10.2 solved to issue for me (pip install SoundFile==0.10.2).