services:
  vllm:
    image: vllm/vllm-openai:latest
    container_name: vllm
    restart: unless-stopped

    ipc: host
    gpus: all

    ports:
      - "8001:8000"

    volumes:
      - /opt/vLLM/models:/root/.cache/huggingface
      - /opt/vLLM/config:/vllm-workspace/config

    environment:
      HUGGING_FACE_HUB_TOKEN: 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
      NVIDIA_VISIBLE_DEVICES: "0,1"
      NVIDIA_DRIVER_CAPABILITIES: compute,utility

    shm_size: "16g"

    command:
      - "--model"
      - "Qwen/Qwen3-14B-AWQ"
      - "--served-model-name"
      - "qwen3-14b"
      - "--host"
      - "0.0.0.0"
      - "--port"
      - "8000"
      - "--tensor-parallel-size"
      - "2"
      - "--dtype"
      - "half"
      - "--gpu-memory-utilization"
      - "0.95"
      - "--max-model-len"
      - "16384"
      - "--max-num-seqs"
      - "32"
      - "--enable-prefix-caching"
      - "--enable-auto-tool-choice"
      - "--tool-call-parser"
      - "hermes"