batiai/Qwen3-Embedding-4B-GGUF

Swap default to Q6_K — drift measured indistinguishable, saves disk on Mac

Remove wrongly-prefixed filename

Remove wrongly-prefixed filename

Rename Q6_K to match card convention

Rename Q8_0 to match card convention

Add tier comparison across 0.6B/4B/8B

Add README with quality benchmark + RAG integration guide

Add Q6_K quantization

Add Q8_0 quantization from Qwen BF16

initial commit