services:
  local-llm-gpu:
    profiles:
      - local-gpu # This profile requires both 'local' and 'gpu'
    image: ollama/ollama:latest
    pull_policy: always
    volumes:
      - ${HOME}/.ollama:/root/.ollama
    ports:
      - '11434:11434'
    environment:
      - NVIDIA_DRIVER_CAPABILITIES=all
      - OLLAMA_LOAD_TIMEOUT=-1
      - OLLAMA_KEEP_ALIVE=-1
      - OLLAMA_DEBUG=1
    command: 'serve'
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: all
              capabilities: [gpu]
    healthcheck:
      test: ['CMD', 'curl', '-f', 'http://localhost:11434/']
      interval: 10s
      timeout: 5s
      retries: 5

  local-llm-cpu:
    profiles:
      - local-cpu # This profile requires both 'local' and 'cpu'
    image: ollama/ollama:latest
    pull_policy: always
    volumes:
      - ${HOME}/.ollama:/root/.ollama
    ports:
      - '11434:11434'
    environment:
      - OLLAMA_LOAD_TIMEOUT=-1
      - OLLAMA_KEEP_ALIVE=-1
      - OLLAMA_DEBUG=1
    command: 'serve'
    healthcheck:
      test: ['CMD', 'curl', '-f', 'http://localhost:11434/']
      interval: 10s
      timeout: 5s
      retries: 5