docker-compose-intel.yml

version: "3.8"

services:
  ollama:
    image: intelanalytics/ipex-llm-inference-cpp-xpu:latest
    command: bash -c "mkdir /llm/ollama && cd /llm/ollama && /usr/local/bin/init-ollama && ./ollama serve"
    ports:
      - "11434:11434"
    volumes:
      - ./ollama:/root/.ollama
    devices:
      - /dev/dri/card0:/dev/dri/card0
      - /dev/dri/renderD128:/dev/dri/renderD128
    environment:
      - OLLAMA_HOST=0.0.0.0
      - USE_XETLA=OFF
      - SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=1
      - SYCL_CACHE_PERSISTENT=1
    networks:
      - llm-network
    healthcheck:
      test: ["CMD-SHELL", "curl -f http://localhost:11434"]
      interval: 5s
      timeout: 2s
      retries: 5

  openwebui:
    image: ghcr.io/open-webui/open-webui:main
    ports:
      - "8080:8080"
    depends_on:
      ollama:
        condition: service_healthy
    environment:
      - WEBUI_URL=http://localhost:8080
      - WEBUI_AUTH=False
      - OLLAMA_BASE_URL=http://ollama:11434
    volumes:
      - ./open-webui:/app/backend/data
    networks:
      - llm-network

networks:
  llm-network: