Dockerfile

FROM python:3.10-slim

# Install necessary packages
RUN apt-get update && apt-get install -y \
    build-essential \
    curl \
    && rm -rf /var/lib/apt/lists/*

# Install llama-cpp-python server and huggingface-cli
RUN pip install llama-cpp-python[server] huggingface_hub

# Download Llama-3.2-3B-Instruct Q4_K_M model from bartowski
RUN huggingface-cli download bartowski/Llama-3.2-3B-Instruct-GGUF Llama-3.2-3B-Instruct-Q4_K_M.gguf --local-dir /app --local-dir-use-symlinks False

# Expose the API port
EXPOSE 7860

# Start the API server
CMD ["python3", "-m", "llama_cpp.server", "--model", "/app/Llama-3.2-3B-Instruct-Q4_K_M.gguf", "--host", "0.0.0.0", "--port", "7860"]