convert List into np.array and put it into Dataset in python

Question

I have 3 list like this

path = ['path/to/file-1', 'path/to/file-2', 'path/to/file-3', ...]
sentence = ['some text-1', 'some text-2', 'some text-3', ...]
audio = [
    {
        'path': 'path/to/file-1'
        'array': array([-0.00036621, ...,  0.00015259])
        'sampling_rate': 16000
    }, ...
]

I want to create datasets.arrow_dataset.Dataset of these, I use this code

import datasets

audio_arrays = []
for audio_file in tqdm(audio_paths[0:10]):
    data, sample_rate = librosa.load(audio_file, sr=None)
    audio_arrays.append(np.array(data, dtype=np.float32))

audio_dict = []
for i, k in tqdm(enumerate(audio_arrays)):
    audio_dict.append({
    'path': audio_paths[i],
    'array': np.array(k),
    'sampling_rate': 16000
})

data = {"path": ok_path, 'audio': audio_dict, "sentence": ok_sentence}
dataset = Dataset.from_dict(data)

but when I try to get the type of dataset[0]['audio']['array'] I got class 'list'

print(dataset[0]['audio']['array']) #output is <class 'list'>