Since the image quality of Stable Diffusion 3.5 isn't quite perfect yet, here's my workflow to improve it:

First, an SD 3.5 base image is generated. Then, I enhance the skin using a skin detailer with SD 1.5. The optimized image is sent to the upscaler, which can optionally be bypassed within the group to increase speed. In the final step, the image goes through a small automated post-production process where it is treated with a Kodachrome LUT and given a touch of film grain.

Enjoy the workflow, and feel free to post some of your images here in the gallery if you'd like to support me! ;)