Fix Wav2Vec2Config.vocab_size type to allow None

[jiqing-feng]

Description

Some wav2vec2 models (e.g. audio classification variants) have vocab_size: null in their config.json. The current type annotation vocab_size: int = 32 causes huggingface_hub's strict dataclass validation to reject None, making these models fail to load.

The fix changes the type annotation to int | None.

Reproduction

from transformers import pipeline

pipe = pipeline("audio-classification", "audeering/wav2vec2-large-robust-12-ft-emotion-msp-dim")

huggingface_hub.errors.StrictDataclassFieldValidationError: Validation error for field 'vocab_size':
    TypeError: Field 'vocab_size' expected int, got NoneType (value: None)

Fix

- vocab_size: int = 32
+ vocab_size: int | None = 32

[Rocketknight1]

Seems safe so I'm happy to accept!

[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.

[jiqing-feng]

Hi @ydshieh . Would you please review this PR? Thanks!

[jiqing-feng]

Hi @Rocketknight1 . Would you please review this PR? The auto-merge is failed because of the CI which is not related to my changes.

[Rocketknight1]

@jiqing-feng yes, I'll try to merge now.

[github-actions]

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

run-slow: wav2vec2