This is a project made with love for my best friend Corey who left this earth on April 15th, 2023 at the age of 36. Motherfucker.

so eventually I'll include everything I'll need to run the backend and the frontend, but at the moment, this repo is only going to feature the front end. For the back end.....

To build this yourself on a new machine... First, let's assume you're me, and have an NVIDIA graphics card, otherwise, idfk. AMD is supported, but you'll have to look around.

First, install wsl ubuntu on your host system (in my case, a windows 11 insider preview build). Then grab the nvidia container toolkit from docker hub and install that inside WSL ubuntu. Check and see if it's working by typing nvidia-smi in your command prompt.

Finally, you need an nvidia supplied base image. I went with Cuda 11.0.3 and ubuntu 20.04. once you find one you like, you'll "pull" it to download it to your system for use in the docker engine

to pull the base image, run this docker pull nvidia/cuda:11.0.3-base-ubuntu20.04 if you pull a more recent one, it'll install a later version of Python that makes the installation fail.

docker run command = docker run --gpus all -d -p 7860:7860 stable-diffusion-crai-v01

Dockerfile

FROM nvidia/cuda:11.0.3-base-ubuntu20.04
ENV DEBIAN_FRONTEND=noninteractive
RUN apt update
RUN apt install -y python3-pip python3 python-is-python3 git-all
WORKDIR /usr/src/app
COPY . .
CMD nvidia-smi \
    && python --version \
    && python launch.py --listen --api
ENV LISTEN 7860
EXPOSE 7860

docker build -t stable-diffusion-crai-v01 .

• deploy in swarm mode to use secrets for the private keys.
• figure out if I can just use DNS from digitalocean