Fix local_files_only tokenizer fallback when tokenizer files are missing (Issue 45538)

[Brianzhengca]

What does this PR do?

Fixes #45538

Root Cause

from_pretrained(..., local_files_only=True) let missing tokenizer files resolve to None and still proceeded into _from_pretrained(...). That allowed a stub tokenizer to initialize instead of raising, which led to the bogus huge model_max_length.

Describe the Fix

The fix adds a fail-closed check in PreTrainedTokenizerBase.from_pretrained.

After resolving files, it now checks whether all real tokenizer assets for that class resolved to None.
If they did, it raises the existing Can't load tokenizer... OSError instead of continuing into
_from_pretrained(...) and constructing a stub tokenizer

Local Tests

• Check with BertTokenizer and CLIPTokenizer using a nonexistent model id and
  local_files_only=True.
  Before the fix, they loaded unexpectedly with the huge fallback max length.
  After the fix, they raised the expected OSError.

Code Agent Policy

The Transformers repo is currently being overwhelmed by a large number of PRs and issue comments written by
code agents. We are currently bottlenecked by our ability to review and respond to them. As a result,
we ask that new users do not submit pure code agent PRs at this time.
You may use code agents in drafting or to help you diagnose issues. We'd also ask autonomous "OpenClaw"-like agents
not to open any PRs or issues for the moment.

PRs that appear to be fully agent-written will probably be closed without review, and we may block users who do this
repeatedly or maliciously.

This is a rapidly-evolving situation that's causing significant shockwaves in the open-source community. As a result,
this policy is likely to be updated regularly in the near future. For more information, please read CONTRIBUTING.md.

• I confirm that this is not a pure code agent PR.

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. (CLIPTokenizer uses 10**30 as model_max_length #45538)
• 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 @Cyrilvallez

[github-actions]

View the CircleCI Test Summary for this PR:

https://huggingface.co/spaces/transformers-community/circle-ci-viz?pr=45541&sha=249d2e