Dockerfile

FROM ghcr.io/ggml-org/llama.cpp:server@sha256:a9b230ae078a6dd947ede5825f227af9fc5fc9e126c4717bebd4762c02601285

RUN useradd -m -u 1000 user
USER user

ENV HOME=/home/user
ENV PATH=/home/user/.local/bin:$PATH

WORKDIR /app

ADD --chown=user https://huggingface.co/Felladrin/LittleLamb-293M-RP/resolve/main/LittleLamb-293M-RP.Q8_0.gguf /app/model.gguf

EXPOSE 7860

CMD ["--port", "7860", "--no-mmap", "--flash-attn", "on", "--fit", "on", "--ctx-size", "40960", "--cache-ram", " 768", "--temperature", "0.6", "--top-p", "0.9", "--min-p", "0.005", "--top-k", "30", "--repeat-penalty", "1.1", "--repeat-last-n", "128", "--backend-sampling", "--webui-mcp-proxy", "-m", "/app/model.gguf", "--chat-template-kwargs", "{ \"enable_thinking\": false }", "--no-mmproj"]