hotchpotch/bekko-embedding-v1-a8m

Document BF16 and Q8 GGUF usage

Document llama.cpp, Ollama, and GGUF usage

docs: discourage ONNX Runtime for native CPU

docs: clarify recommended inference backends

docs: add Bekko blog and paper links

docs: link Bekko Embedding paper and add citation

Keep GPU attention backend optional in quickstart

Clarify GPU attention setup in quickstart

Document OpenVINO dependencies and Node.js CPU usage

Describe bekko as ultra-compact

Use canonical benchmark chart URL

Update model card lineage and license

Simplify quickstart model loading

Add Raspberry Pi 5 row to selection table

Drop width attributes from images

Translate Japanese snippets and simplify prose punctuation

Add search example, model selection guide, FAQ, and browser demo link

Polish tables and inline code in model card

Rewrite model card prose in a natural tone

Refine model card inference guidance

Update release model card comparisons

Update ONNX and OpenVINO exports for prefix-none a8m

Update ONNX and OpenVINO exports for prefix-none a8m

Upload prefix-none a8m model artifacts for 20260707

initial commit