BatchFeature: Convert List[np.ndarray] to np.ndarray before converting to pytorch tensors

[eladsegal]

What does this PR do?

Fixes #14307

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you read the contributor guideline,
  Pull Request section?
• Was this discussed/approved via a Github issue or the forum? Please add a link
  to it if that's the case.
• Did you make sure to update the documentation with your changes? Here are the
  documentation guidelines, and
  here are tips on formatting docstrings.
• Did you write any new necessary tests?

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.

[sgugger]

Thanks for your PR!
I think the check needs to be a bit more thorough: we need to also check all the elements of the list are np.ndarray (or at least the first one), according to the warning at least.

[eladsegal]

Thanks @sgugger,
I added a check for the first element, but here are my thoughts about it:
Iterating through the full list doesn't seem helpful - what would you do if half of the list is List and the other half is np.array? It would still be more efficient to call np.array in such a case.
It's the same thing for checking just the first element, but at least the check doesn't have the cost of iterating through the full list.
Therefore, I think that it's best to always pass the list to np.array without checking it.

[sgugger]

Failure is unrelated and already fixed on master.
Thanks again for your work on this PR!