fix(models): Resolve regressions in Wav2Vec2PhonemeCTCTokenizer (wav2vec2-lv-60-espeak-cv-ft)

[harshaljanjani]

What does this PR do?

The following Wav2Vec2PhonemeCTC use cases were identified and fixed in this PR:

→ 05c0e1d ("rm slow tokenizers") added self.backend = kwargs.pop("backend", None). Wav2Vec2PhonemeCTCTokenizer already used self.backend for its phonemizer EspeakBackend object set in init_backend. Regardless of call order, one clobbers the other; either the base class overwrites the phonemizer object with None (breaking phonemize()), or the phonemizer object overwrites the base class's serializable value (breaking save_pretrained with EspeakBackend is not JSON serializable). Renamed to self._phonemizer_backend so both attributes coexist. Followed the same naming convention used for _word_delimiter_token and _phone_delimiter_token in the same file.
→ Same commit consolidated tokenization_utils.py into tokenization_python.py. In the old code, _encode_plus had return_offsets_mapping as a named param and raised NotImplementedError before reaching tokenize(). After the refactor, return_offsets_mapping is no longer a named param in _encode_plus, so it flows through **kwargs → tokenize() → prepare_for_tokenization(), which had a fixed signature. Added **kwargs to match the base class contract at tokenization_python.py#L836-L838. No other models are affected; Wav2Vec2PhonemeCTCTokenizer is the only override of prepare_for_tokenization that was missing **kwargs :)
→ For more details on reproducing the bug and the output screenshots, please visit the linked issue!

Fixes #45198

cc: @Rocketknight1 @itazap

CI coverage fixed by this PR (as suggested for inclusion in the PR):

CI run test coverage of this behavior:

→ models/wav2vec2_phoneme/test_tokenization_wav2vec2_phoneme.py::Wav2Vec2PhonemeCTCTokenizerTest:

test_batch_encode_dynamic_overflowing, test_batch_encode_plus_batch_sequence_length, test_batch_encode_plus_padding, test_call, test_case_insensitive, test_change_phonemizer_lang, test_chat_template, test_chat_template_batched, test_decode_with_del, test_empty_input_string, test_encode, test_encode_basic_padding, test_encode_decode, test_encode_decode_with_del, test_encode_decode_with_del_filter, test_encode_plus_with_padding_0, test_encode_plus_with_padding_1, test_encode_with_del, test_mask_output, test_maximum_encoding_length_pair_input, test_maximum_encoding_length_single_input, test_number_of_added_tokens, test_offsets, test_offsets_batch, test_padding_to_multiple_of, test_phonemize, test_phonemize_with_word_del, test_prepare_seq2seq_batch, test_pretokenized_inputs, test_right_and_left_truncation, test_save_and_load_tokenizer, test_special_tokens_mask, test_special_tokens_mask_input_pairs, test_token_type_ids, test_tokenizer_add_new_tokens

Repro output after the fixes (feel free to cross-check):

Code Agent Policy

• 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.
• 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 fix any necessary existing tests?

[itazap]

Thank you! good catch, apologies for the overlap in naming

[itazap]

let's just add a small test so that we catch this next time 😉

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[harshaljanjani]

@itazap Added, please check and let me know if it's alright :)

[harshaljanjani]

Removed as per the review.

[harshaljanjani]

Good day @itazap; just checking in to see if there have been any updates :)

[itazap]

hey sorry, good news is actually the existing tests like test_phenomize, test_encode, etc. already call _phenomize so we don't need a new test 😅 the bad news these tests are failing with and without this change

[harshaljanjani]

@itazap Makes sense, removed :)

[itazap]

I'd prefer to wait to merge a fix that allows these tests to pass! Looking into it as well

[harshaljanjani]

Ah, sorry I missed what you truly meant in the prev. comment. I've completed the investigation and expanded the PR coverage. Dropping an explanation of why this is a bit more involved now, could you double check if this fixes all the tests and if the fixes make sense?

→ tokenization_wav2vec2_phoneme.py; dropped word_delimiter_token / phone_delimiter_token from super().init() (they're delimiters not vocab tokens), so V5 auto-promotion no longer adds " " into the vocab. Round-trip is still preserved by reinjecting them into init_kwargs and re-reading from model_specific_special_tokens on load.
→ Found a get_tokenizer() precedence bug: kwargs.update(cls.special_tokens_map) was clobbering user kwargs (e.g. pad_token="[PAD]" → <pad>). Copied the correct wav2vec2 pattern verbatim :))

[github-actions]

[For maintainers] Suggested jobs to run (before merge)

run-slow: wav2vec2_phoneme

[itazap]

perfect, thanks a lot! 🤗