Fix type of n_windows in cross_validation

[Riklia]

As described in the documentation and in the implementation, when test_size is defined, we set n_windows to None. However, the type of n_windows is specified as int. This causes a warning on the caller side.

[CLAassistant]

All committers have signed the CLA.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[marcopeix]

Hi @Riklia, I want to make a release of neuralforecast soon and would like to include your PR. Can you solve the conflict please? Thanks!

[Riklia]

Hi @marcopeix! I rebased, and it should be fine now.

[marcopeix]

Hello @Riklia , can you address these comments please? Also, make sure to install the pre-commit hooks with:

pre-commit install
pre-commit run --files neuralforecast/*

Linting errors should be picked up before you make a PR.

Thanks!

[Riklia]

Thanks for the detailed instructions! Addressed all the comments.