Dockerfile

# Use the official pre-compiled Llama engine
FROM ghcr.io/abetlen/llama-cpp-python:latest

# Change our working directory to /src so we don't conflict with the base image's /app folder
WORKDIR /src

# Copy requirements and install our extra packages
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

# Copy your app.py and other files into /src
COPY . .

# Expose the Hugging Face port
EXPOSE 7860

# Override the default entrypoint to run your custom FastAPI app
ENTRYPOINT []
CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]