set correct model input names for gptsw3tokenizer

[DarioSucic]

What does this PR do?

Makes it so the tokenizer doesn't output token_type_ids, as these break generation.

Seems like a harmless change, but I'm not sure what these are used for so let me know if this is the wrong approach!

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.

@ArthurZucker @amyeroberts @ekgren @Apsod

[HuggingFaceDocBuilderDev]

The documentation is not available anymore as the PR was closed or merged.

[Apsod]

LGTM
We're not using token type ids during training, so there is no reason for the tokenizer to output them, and doing so just leads to unintended behaviour.

[ArthurZucker]

Fine with me too if this is a fix, otherwise might affect other users who might be using them. Not being used in training would tell me that it's a fix! Thanks both for opening the issue and reviewing

[amyeroberts]

Thanks for fixing!