[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 (and merged) by the Update Inference Providers Documentation workflow.

---

Note

Low Risk
Generated documentation and dependency pins in the inference-providers script only; no application runtime or API behavior changes in this repo.

Overview
Bumps @huggingface/inference to ^4.13.20 and @huggingface/tasks to ^0.21.17 under scripts/inference-providers, then regenerates Inference Providers markdown.

Provider / task doc updates reflect newer package metadata: SambaNova gains a Chat Completion (LLM) snippet (openai/gpt-oss-120b) and is added to the shared chat-completion LLM providersMapping. Z.ai VLM examples move from GLM-4.5V to GLM-4.6V-Flash in provider and task pages.

HF Inference example models change for image segmentation (fashn-ai/fashn-human-parser instead of mattmdjaga/segformer_b2_clothes) and zero-shot classification (MoritzLaurer/mDeBERTa-v3-base-xnli-multilingual-nli-2mil7 instead of facebook/bart-large-mnli), with matching updates on task pages.

^{Reviewed by Cursor Bugbot for commit ca89650. 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.