services:
  llamacpp-server:
    build: .
    container_name: llamacpp_${SERVER_PORT}
    ports:
      - "${SERVER_PORT}:${SERVER_PORT}"
    devices:
      - /dev/kfd:/dev/kfd
      - /dev/dri:/dev/dri
    volumes:
      - /opt/models:/models
    command: >
      -m /models/${MODEL_FILENAME}
      --host 0.0.0.0
      --port ${SERVER_PORT}
      -ngl 99
      -c 4096
    restart: unless-stopped