[Bot] Update Inference Providers documentation

[HuggingFaceInfra]

This PR automatically upgrades the @huggingface/tasks and @huggingface/inference packages and regenerates the Inference Providers documentation by running:

cd scripts/inference-providers
pnpm update @huggingface/tasks@latest @huggingface/inference@latest
pnpm run generate

This PR was automatically created by the Update Inference Providers Documentation workflow.

Please review the changes before merging.

---

Note

Low Risk
Low risk documentation-only update; main risk is inaccurate example modelId/providerModelId mappings if the regenerated data is wrong.

Overview
Updates generated Inference Providers docs to reflect new default example models and provider mappings across several tasks.

Notable doc changes: switches hf-inference question-answering examples to deepset/roberta-base-squad2; updates Novita LLM/VLM example models; adds SambaNova chat completion (LLM/VLM) snippets and includes SambaNova in the chat-completion/image-text-to-text task provider mappings.

Removes Together’s text-to-image example from provider docs and from the text-to-image task mapping, and fixes a formatting issue in text-classification recommended models list.

^{Reviewed by Cursor Bugbot for commit 171be0e. Bugbot is set up for automated code reviews on this repo. Configure here.}

[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.