[Auto] Make processor subclasses overridable on load time

[vasqu]

As per title, discovered in #39585 (went through a few versions until we found this :p)

[github-actions]

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

run-slow: auto

[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.

[zucchini-nlp]

Nice! Let's make sure that the tests are passing, check a few important models with slow tests. Shouldn't be a problem, just for safety

Feel free to merge when everything passes :)

[zucchini-nlp]

meant to ✅

[vasqu]

run-slow: vit,clip,detr,table_transformer,got_ocr2,whisper,wav2vec2,qwen2_audio,speech_t5,csm,gemma3,qwen2_5_vl,llava,smolvlm,gemma3n,qwen2_5_omni

[github-actions]

This comment contains run-slow, running the specified jobs:

models: ["models/clip", "models/csm", "models/detr", "models/gemma3", "models/gemma3n", "models/got_ocr2", "models/llava", "models/qwen2_5_omni", "models/qwen2_5_vl", "models/qwen2_audio", "models/smolvlm", "models/table_transformer", "models/vit", "models/wav2vec2", "models/whisper"]
quantizations: []

[github-actions]

CI Results

Workflow Run ⚙️

✅ No failing test specific to this PR 🎉 !