Z-image-Turbo to SDXL (workflow)
Details
Download Files
Model description
[Ver1.0a]
Z_image_to SDXL+ LLM
Added LLM using LMStudio.
LM Studio - Local AI on your computer
(This function requires LMStudio to be installed and running)
Additional Custom Nodes
ALatentPlace/YANC_LMStudio: A ComfyUI integration to use LMStudio for extending prompts.
[Ver1.0]
Z_image_to SDXL
This is probably an anime workflow.
I've seen refiners like SDXL to Z-image-Turbo, but I tried the opposite workflow.
Z-image-Turbo is used on the front-end Ksampler, and SDXL on the back-end Ksampler. (Here, I'm not particularly recommending WAI-illustrious-SDXL v14.0; it's up to you.)
In this example, the denoise value is set to 0.6 for now, so adjusting it appropriately should produce results.
Positive prompts are common, but negative prompts are only available in SDXL.
As for Optional Positive (for SDXL), use it with some good ideas, but it might not be necessary.
I haven't tried it with NSFW or realistic content yet.
Two custom nodes are used: