Skip to content

GPU support for OpenFaaS Edge

At the time of speaking, OpenFaaS Pro on Kubernetes supports GPUs for functions via Profiles.

OpenFaaS Edge supports Nvidia GPUs for core services only.

Example:

services:  
  ollama:
    image: docker.io/ollama/ollama:latest
    command:
      - "ollama"
      - "serve"
    volumes:
      - type: bind
        source: ./ollama
        target: /root/.ollama
    ports:
      - "127.0.0.1:11434:11434"
    gpus: all
    deploy:
      restart: always

Learn more: How to Protect Your Data with Self-Hosted LLMs and OpenFaaS Edge