A small no-nonsense workflow.
Generation times on a 4090 and 64gb ram - Win11 - PyTorch 2.8.0 nightly / Cuda 12.8:
1024x1488: 35s
1024x1024: 24s
The workflow contains a small guide with all the needed models and optional recommendations.
This image should also contain the workflow except if civitai purges metadata:
Have fun!
Discord if you have any questions: