| FROM ghcr.io/ggerganov/llama.cpp:server | |
| ARG MODEL_URL=https://huggingface.co/EleutherAI/llemma_7b-GGUF/resolve/main/llemma_7b-Q4_K_M.gguf | |
| RUN mkdir -p /models && wget -q -O /models/llemma.gguf "$MODEL_URL" | |
| EXPOSE 8080 | |
| ENTRYPOINT ["/llama-server", "-m", "/models/llemma.gguf", "-c", "4096", "--host", "0.0.0.0", "--port", "8080"] | |