llm_vector_server/docker-compose.yaml

services:
  app:
    build:
      context: .
      dockerfile: Dockerfile
    container_name: vector-server-app
    ports:
      - "8000:8000"
    volumes:
      - .:/app
      - ./models/Qwen3-Embedding-4B:/app/model/Qwen3-Embedding-4B  # 挂载模型
    environment:
      - PYTHONUNBUFFERED=1
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: all
              capabilities: [gpu]