docker-compose.yml

networks:
  ai-network:
    driver: bridge
services:
  tabbyapi:
    image: local/tabbyapi:exl3-fixed
    container_name: tabby-tavern-tabbyapi-1
    restart: unless-stopped
    ports:
    - 5000:5000
    volumes:
    - ./tabby_models:/app/models
    - ./tabby_config/config.yml:/app/config.yml
    - ./tabby_config/api_tokens.yml:/app/api_tokens.yml
    environment:
    - CUDA_VISIBLE_DEVICES=0
    - CUDA_DEVICE_ORDER=PCI_BUS_ID
    - NVIDIA_VISIBLE_DEVICES=all
    - PYTORCH_CUDA_ALLOC_CONF=max_split_size_mb:128
    - OMP_NUM_THREADS=8
    - EXLLAMA_GPU_LAYERS=999
    - EXLLAMA_KV_CACHE=q8_0
    - EXLLAMA_FLASH_ATTENTION=1
    deploy:
      resources:
        reservations:
          devices:
          - driver: nvidia
            count: all
            capabilities:
            - gpu
            - compute
            - utility
    shm_size: 16g
    networks:
    - ai-network
  sillytavern:
    image: ghcr.io/sillytavern/sillytavern:latest
    container_name: tabby-tavern-sillytavern-1
    restart: unless-stopped
    ports:
    - 8000:8000
    volumes:
    - ./sillytavern_config:/home/node/app/config
    - ./sillytavern_data:/home/node/app/data
    environment:
    - WHITELIST_MODE=false
    networks:
    - ai-network
  ollama:
    image: ollama/ollama:latest
    container_name: tabby-tavern-ollama-1
    restart: unless-stopped
    ports:
    - 11434:11434
    volumes:
    - ./ollama_data:/root/.ollama
    environment:
    - CUDA_VISIBLE_DEVICES=0
    - CUDA_DEVICE_ORDER=PCI_BUS_ID
    - NVIDIA_VISIBLE_DEVICES=all
    - OMP_NUM_THREADS=8
    - OLLAMA_FLASH_ATTENTION=1
    - OLLAMA_KV_CACHE_TYPE=q8_0
    - OLLAMA_NUM_PARALLEL=1
    deploy:
      resources:
        reservations:
          devices:
          - driver: nvidia
            count: all
            capabilities:
            - gpu
            - compute
            - utility
    shm_size: 16g
    networks:
    - ai-network
  open-webui:
    image: ghcr.io/open-webui/open-webui:main
    container_name: tabby-tavern-open-webui-1
    restart: unless-stopped
    ports:
    - 3000:8080
    volumes:
    - ./openwebui_data:/app/backend/data
    environment:
    - OLLAMA_BASE_URL=http://ollama:11434
    networks:
    - ai-network
  searxng:
    container_name: tabby-tavern-searxng-1
    image: docker.io/searxng/searxng:latest
    restart: unless-stopped
    ports:
    - 8080:8080
    volumes:
    - ./searxng_config:/etc/searxng
    environment:
    - SEARXNG_BASE_URL=http://localhost:8080/
    - INSTANCE_NAME=Tabby-Tavern-Search
    - UWSGI_WORKERS=4
    - UWSGI_THREADS=4
    cap_drop:
    - ALL
    cap_add:
    - CHOWN
    - SETGID
    - SETUID
    networks:
    - ai-network
    depends_on:
    - searxng-redis
  searxng-redis:
    container_name: tabby-tavern-searxng-redis-1
    image: docker.io/library/redis:alpine
    command: redis-server --save "" --appendonly "no"
    restart: unless-stopped
    networks:
    - ai-network