FROM python:3.11-slim
WORKDIR /code
COPY ./requirements.txt /code/requirements.txt
RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
COPY . .
EXPOSE 7860
# --timeout 120: first request lazily loads BOTH transformer members into RAM,
# which can exceed gunicorn's default 30s worker timeout on free CPU hardware.
CMD ["gunicorn", "-b", "0.0.0.0:7860", "--timeout", "120", "app:app"]