[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
Docs-only, script-generated changes to example snippets and recommended-model lists; no application or API runtime code.

Overview
Automated refresh of Inference Providers docs after bumping @huggingface/tasks from 0.21.32 to 0.21.33 (and lockfile sync); @huggingface/inference stays at ^4.13.24.

Regenerated InferenceSnippet provider/model mappings shift example models: Fireworks AI chat LLM uses deepseek-ai/DeepSeek-V4-Flash-0731 instead of zai-org/GLM-5.2; PublicAI LLM uses speakleash/Bielik-11B-v3.0-Instruct instead of Swiss Apertus 8B; Z.ai VLM example uses zai-org/GLM-4.6V instead of GLM-4.6V-Flash; HF Inference image-segmentation example uses mattmdjaga/segformer_b2_clothes instead of the NVIDIA SegFormer ADE model. Matching updates appear on shared task pages (chat-completion, image-text-to-text, image-segmentation).

Recommended models lists on chat-completion and text-generation no longer include zai-org/GLM-4.5 (sourced from tasks package / API).

^{Reviewed by Cursor Bugbot for commit 05ab11e. 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.