MattGPT

I am able to run a local GPT platform based on this software. I have also added it to my SSO and I think that's pretty neat. This means if you want access, you can ask me for it.

Here's my Docker Compose.

docker-compose.yaml

services:
  ollama:
    image: ghcr.io/open-webui/open-webui:ollama
    container_name: mattgpt
    volumes:
      - ollama:/root/.ollama
      - open-webui:/app/backend/data
    environment:
      WEBUI_URL: 'https://gpt.domain.com'
      ENABLE_OAUTH_SIGNUP: 'true'
      OAUTH_MERGE_ACCOUNTS_BY_EMAIL: 'true'
      OAUTH_CLIENT_ID: 'gpt-oidc'
      OAUTH_CLIENT_SECRET: 'makeupanicelongstringofletters'
      OPENID_PROVIDER_URL: 'https://auth.domain.com/.well-known/openid-configuration'
      OAUTH_PROVIDER_NAME: 'Authelia'
      OAUTH_SCOPES: 'openid email profile groups'
      ENABLE_OAUTH_ROLE_MANAGEMENT: 'true'
      OAUTH_ALLOWED_ROLES: 'gpt-access,gpt-admin'
      OAUTH_ADMIN_ROLES: 'gpt-admin'
      OAUTH_ROLES_CLAIM: 'groups'
    ports:
       - 8080:8080
    restart: always
    network_mode: bridge
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 2
              capabilities: [gpu]

volumes:
  ollama:
    driver: local # Define the driver and options under the volume name
    driver_opts:
      type: none
      device: ./ollama
      o: bind

  open-webui:
    driver: local # Define the driver and options under the volume name
    driver_opts:
      type: none
      device: ./open-webui
      o: bind

Sample Screenshot.