services:
  vllm-gptoss20b:
    image: vllm/vllm-openai:latest
    ports: ["5777:8000"]
    volumes: ["./models/gpt-oss-20b:/m:ro"]
    environment: { HF_HUB_OFFLINE: "1", TRANSFORMERS_OFFLINE: "1" }
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              capabilities: [gpu]
              device_ids: ["0"]
    command: >-
      --host 0.0.0.0 --port 8000
      --model /m
      --served-model-name gpt-oss-20b
      --tensor-parallel-size 1
      --gpu-memory-utilization 0.9
      --max-model-len 131072
      --max-num-batched-tokens 4096
      --max-num-seqs 8
      --enable-prefix-caching
      --enable-chunked-prefill
      --disable-log-requests