Dockerfile

FROM python:3.11-slim

WORKDIR /app

# Install system dependencies for llama-cpp-python
RUN apt-get update && apt-get install -y \
    curl wget cmake build-essential git \
    && rm -rf /var/lib/apt/lists/*

# Install Python dependencies
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

# Copy app files
COPY download_model.py .
COPY server.py .

# Download the GGUF model at build time
RUN python download_model.py

EXPOSE 7860

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