Dockerfile

FROM python:3.10-slim

# Install system dependencies required to build llama-cpp
RUN apt-get update && apt-get install -y \
    build-essential \
    cmake \
    & rm -rf /var/lib/apt/lists/*

WORKDIR /app

# Set environment variables for llama-cpp optimization (CPU)
ENV LLAMA_ARG_MMAP=1

COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

COPY . .

# Expose port 7860 as required by Hugging Face Spaces
EXPOSE 7860

CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]