Change nn.dropout to layer.Dropout in TFBart

[astariul]

What does this PR do?

This PR changes all the tf.nn.dropout calls in modeling_tf_bart.py and use tf.keras.layers.Dropout instead.
More consistent with modeling_tf_roberta.py

Fixes #9045

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?

@patrickvonplaten

[patrickvonplaten]

Hey @astariul-colanim

Thanks for the fix! Looks good to me if it solves the error on TPU. Also cc @jplu

[jplu]

Thanks!!

LGTM!

[patrickvonplaten]

Awesome, I checked and TFBart is the only model that made use of tf.nn.dropout so after this PR it'll be more consistent again. Thanks a lot!

Merging to avoid too many conflicts with TFBart refactor: #9029

@sgugger @LysandreJik -> we should make sure to always use tf.keras.layers.Dropout then and not tf.nn.dropout