/ollama-webui-stack

Primary LanguageJust

Ollama OpenWebUI Docker Compose Configuration

This docker-compose.yml file defines the services and volumes required to run OpenWebUI, an AI chatbot interface with GPU acceleration.

Services:

  • open-webui:
    • Exposes port 5555 to the host's 8080
    • Uses the ghcr.io/open-webui/open-webui image with GPU support
    • Depends on the ollama service
    • Stores data in the open-webui volume
    • Starts automatically on boot
  • ollama:
    • Uses the ollama/ollama image with GPU support
    • Stores data in the ollama volume
    • Exposes port 11434 to the host's 11434

Volumes:

  • ollama:
    • External volume shared across all services
    • Persistent storage for Ollama data
  • open-webui:
    • External volume specific to OpenWebUI data
    • Persistent storage for OpenWebUI data

Usage:

  1. Save this file as docker-compose.yml in the root directory of your project.
  2. Run the following command to start the services:
docker-compose up -d
  1. Access OpenWebUI at localhost:5555.

Notes:

  • This configuration requires a GPU-enabled host.
  • The ghcr.io/open-webui/open-webui image is built from the latest OpenWebUI code.
  • The ollama/ollama image provides the language model capabilities for OpenWebUI.
  • The resources.reservations.devices section ensures that all available GPUs are allocated to the services.

Additional Information: