Is this datasets audio classification notebook bugged?

In this tutorial (Audio classification) I’ve read about a pytorch audio classifiaction notebook -

“For a more in-depth example of how to finetune a model for audio classification, take a look at the corresponding PyTorch notebook.”

but when I tried to run it in google colab something got real messy somehow and a lot of errors occured … maybe it is a library conflict.

The first two cells run normally, but the third cell can’t be compiled when using datasets 1.14, because the notebook login wants username + password but results in an URL error after logging in somehow …

404 Client Error: Not Found for url: hf/api/login
Error Cannot POST /api/login

I think this is caused by notebook_login() in datsets 1.14. If I change datasets 1.14 to just

!pip install datasets

Then the notebook login leads to a login with just the huggingface token and after logging in no error occurs … I proceeded with the following cells but then when I try to run the sixth cell with

from datasets import load_dataset, load_metric

I get an error that states

AttributeError: readonly attribute

So I guess this error is caused by using the wrong datasets version … I can’t do the notebook login when using datasets 1.14 but it seems that at the same time I get an error when using a datasets version that is newer than 1.14 when trying to import the functions that are needed for running the notebook … can anoyone help me fix this issue?