Replace Gemma embedding model with BAAI/bge-small-en-v1.5 and cache at build time

Fix Gemma model repository name to google/embeddinggemma-300m

Swap Qwen3-Embedding-0.6B for distilled C10X/Qwen3-Embedding-TurboX.v2

Configure for 100% self-contained CPU deployment

local: phase 5 arena and text highlighter