stable-diffusion container

stable-diffusion shipped as linux container.

Usage

Run the prebuilt container:

docker run --rm --privileged \
	-v `pwd`/models:/stable-diffusion/models \
	-v `pwd`/output:/output \
	mgoltzsche/stable-diffusion \
	--prompt "a photograph of an astronaut riding a horse" --plms --outdir /output

Development

To list the supported targets, run make help.

Prerequisites

• git
• make
• docker 20+
• kubectl

Build the application

To build the application container image using skaffold, run:

make image

Deploy the application

To deploy the application using skaffold, run:

make deploy

To deploy the application in debug mode (debug ports forwarded), stream its logs and redeploy on source code changes automatically, run:

make debug

To undeploy the application, run:

make undeploy

Apply blueprint updates

To apply blueprint updates to the application codebase, update the kpt package:

1. Before updating the package, make sure you don't have uncommitted changes in order to be able to distinguish package update changes from others.
2. Call make blueprint-update or rather kpt pkg update and kpt fn render (applies the configuration within setters.yaml to the manifests and skaffold.yaml).
3. Before committing the changes, review them carefully and make manual changes if necessary.

TL;DR: Variant Constructor Pattern

Release

The release process is driven by Conventional Commits, letting the CI pipeline generate a version and publish a release depending on the commit messages on the main branch.