fix: Display embedding models in settings with visual distinction

[bittoby]

Describe Changes

• Show embedding models in Settings → Model Providers → Llama.cpp (previously hidden)
• Add distinctive amber "Embedding" badge to identify embedding models
• Filter embedding models from chat model dropdown (they can't be used for chat)
• Add tooltip: "Embedding Model (for RAG/vectors, not chat)"

Problem

Embedding models (e.g., all-MiniLM-L6-v2, bge-small-en) were imported successfully but not visible in the UI. When users tried to import again, they got "model already exists" error, causing confusion.

Fixes Issues

• Closes bug: Can not show loaded embedding model #7357

Self Checklist

• Added relevant comments, esp in complex areas
• Updated docs (for bug fixes / features)
• Created issues for follow-up changes or refactoring needed

[bittoby]

@louis-jan @urmauur Could you please review this PR?