docker-compose.yml

# An example docker-compose file for running LlamaVision with CUDA
services:
  &name llavavision:
    container_name: *name
    hostname: *name
    # profiles:
    #   - *name
    build:
      context: https://github.com/lxe/llavavision.git#main
      dockerfile: Dockerfile
    environment:
      NVIDIA_VISIBLE_DEVICES: all
      NVIDIA_DRIVER_CAPABILITIES: all
      TUNNEL: false
      UID: 1001
      CERT_SUBJ: "/C=GB/ST=LON/O=LlamaOrg, Inc./CN=llavavision.local"
      DEBUG: true
      PORT: 5000
      NLG: 35
      TS: "100,0" # For GPU-only, single GPU
    shm_size: "4gb"
    runtime: nvidia
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: all
              capabilities: ["compute", "utility", "graphics"]
    restart: unless-stopped
    security_opt:
      - no-new-privileges:true
    volumes:
      - models:/app/models
    ports:
      - 5000:5000
      # - 8080:8080 # If you want to access llama.cpp from outside the container you can use this port

volumes:
  models: