diff --git a/compose/services/ollama/compose.yaml b/compose/services/ollama/compose.yaml index 5fbc7fe..0e354dc 100644 --- a/compose/services/ollama/compose.yaml +++ b/compose/services/ollama/compose.yaml @@ -20,17 +20,14 @@ services: - homelab # GPU Support (NVIDIA GTX 1070) - # Uncomment the deploy section below to enable GPU acceleration - # Prerequisites: - # 1. Install NVIDIA Container Toolkit on host - # 2. Configure Docker to use nvidia runtime - # deploy: - # resources: - # reservations: - # devices: - # - driver: nvidia - # count: 1 - # capabilities: [gpu] + runtime: nvidia + deploy: + resources: + reservations: + devices: + - driver: nvidia + count: 1 + capabilities: [gpu] labels: # Traefik (API only, no web UI) @@ -44,7 +41,7 @@ services: traefik.http.services.ollama.loadbalancer.server.port: 11434 # SSO Protection for API and restrict to local network - traefik.http.routers.ollama.middlewares: tinyauth,local-only + traefik.http.routers.ollama.middlewares: local-only@docker,authelia@docker # Homarr Discovery homarr.name: Ollama (LLM) diff --git a/compose/services/open-webui/compose.yaml b/compose/services/open-webui/compose.yaml index 697e572..cda5318 100644 --- a/compose/services/open-webui/compose.yaml +++ b/compose/services/open-webui/compose.yaml @@ -19,13 +19,11 @@ services: # Enable RAG (Retrieval-Augmented Generation) - ENABLE_RAG_WEB_SEARCH=true - - RAG_WEB_SEARCH_ENGINE=searxng - - # Web search (optional, requires searxng) - - ENABLE_RAG_WEB_SEARCH=false + - RAG_WEB_SEARCH_ENGINE=duckduckgo + - ENABLE_RAG_WEB_LOADER_SSL_VERIFICATION=false # Default model - - DEFAULT_MODELS=qwen2.5-coder:7b + - DEFAULT_MODELS=qwen2.5:3b networks: - homelab @@ -43,7 +41,6 @@ services: # No SSO - Open WebUI has its own auth system # Uncomment to add SSO protection: - # traefik.http.routers.open-webui.middlewares: tinyauth # Homarr Discovery homarr.name: Open WebUI (AI Chat)