Stack para buildar imagem com CUDA e llama-cpp e subir stack com docker-compose

docker-compose.yml

services:
  cuda-llama-cpp:
    image: imagem-buildada
    ports:
      - 5001:5001
    command: python3 app.py
    deploy:
      resources:
        reservations:
          devices:
          - capabilities: [gpu]

Dockerfile

ARG CUDA_IMAGE="12.5.0-devel-ubuntu22.04"
FROM nvidia/cuda:${CUDA_IMAGE}

WORKDIR /app
ENV ACCEPT_EULA=Y
ENV DEBIAN_FRONTEND=noninteractive
ENV HOST=0.0.0.0
ENV CUDA_DOCKER_ARCH=all
ENV GGML_CUDA=1

RUN apt-get update && apt-get upgrade -y \
    && apt-get install -y git build-essential \
    python3 python3-pip gcc wget \
    libopenblas-dev

COPY ./requirements.txt .

RUN pip install -r requirements.txt
RUN CMAKE_ARGS="-DGGML_CUDA=on" pip install llama-cpp-python

EXPOSE 5001