FROM nvidia/cuda:12.4.1-cudnn-devel-ubuntu22.04 RUN apt-get update && apt-get install -y \ python3.10 python3.10-venv python3-pip git \ && rm -rf /var/lib/apt/lists/* RUN useradd -m -u 1000 user && \ mkdir -p /app /tmp/.triton && \ chown 1000:1000 /app /tmp/.triton WORKDIR /app COPY requirements.txt . RUN --mount=type=cache,target=/root/.cache/pip \ python3.10 -m pip install --upgrade pip setuptools wheel && \ python3.10 -m pip install --no-cache-dir -r requirements.txt \ gradio[oauth,mcp]==6.1.0 \ "uvicorn>=0.14.0" \ "websockets>=10.4" \ spaces COPY --chown=1000:1000 . . ENV PYTHONUNBUFFERED=1 \ TRITON_CACHE_DIR=/tmp/.triton \ HOME=/home/user \ MAX_CONCURRENT_JOBS=4 USER 1000 CMD ["python3.10", "-m", "uvicorn", "server:app", "--host", "0.0.0.0", "--port", "7860"]