When prompted, enter your token to log in:

from huggingface_hub import notebook_login
notebook_login()

Load the dataset
VoxPopuli is a large-scale multilingual speech corpus consisting of 
data sourced from 2009-2020 European Parliament event recordings.