[DOCS] Adds DeBERTA v2 to the tokenizers list in API docs

[szabosteve]

Overview

This PR expands the Tokenization properties section on the PUT trained models API doc page and the Infer trained model API doc page with the DeBERTa v2 tokenizer reference docs. The updates also effect the Inference processor reference docs.

Preview

PUT trained models - Tokenization properties
Infer trained model
Inference processor

[github-actions]

Documentation preview:

• ✨ Changed pages

[elasticsearchmachine]

Pinging @elastic/ml-core (Team:ML)

[elasticsearchmachine]

Pinging @elastic/es-docs (Team:Docs)

[maxhniebergall]

Looks good, thanks Istvan!

overall, I think that the deberta_v2 tokenizer is usable for any ML task, but it doesn't appear in all of the tasks in https://elasticsearch_bk_112752.docs-preview.app.elstc.co/guide/en/elasticsearch/reference/master/inference-processor.html#inference-processor-fill-mask-opt

I also noticed that page doesn't have a secontion for text_similarity, which is the main thing we will be using deberta for. I don't think that needs to be a part of this PR, but we should consider adding that.

[szabosteve]

Hey @maxhniebergall,
I've added DeBERTa to all the ML tasks on the inference pipeline reference doc page.
I also opened an issue to document text_similarity and add it to the page: https://github.com/elastic/search-docs-team/issues/188

[maxhniebergall]

LGTM

Just wondering why some of the new sections in docs/reference/ingest/processors/inference.asciidoc have span and some dont?

edit: I see that zero-shot classification doesn't support span https://github.com/elastic/elasticsearch/blob/main/x-pack/plugin/core/src/main/java/org/elasticsearch/xpack/core/ml/inference/trainedmodel/ZeroShotClassificationConfig.java#L145

[elasticsearchmachine]

💚 Backport successful

Status	Branch	Result
✅	8.x