docker-compose.yaml

services:
  Qwen3-Embedding-4B:
    container_name: Qwen3-Embedding-4B
    restart: no
    #image: dengcao/vllm-openai:v0.9.2-dev #采用vllm最新的开发版制作的镜像，经测试正常，可放心使用
    #image: dengcao/vllm-openai:v0.9.2rc2
    image: dengcao/vllm-openai:v0.9.2
    ipc: host
    volumes:
      - ./models:/models
    command: ["--model", "/models/Qwen3-Embedding-4B",  "--served-model-name", "Qwen3-Embedding-4B",  "--gpu-memory-utilization", "0.90"]
    ports:
      - 8008:8000
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: all
              capabilities: [gpu]