Built this workflow from scratch using a few different custom nodes for efficiency and a cleaner layout. It has worked well with a variety of models. Separating the positive prompt into two sections has allowed for creating large batches of images of similar styles.
Custom Nodes Used
- https://github.com/SeargeDP/SeargeSDXL
- https://github.com/wallish77/wlsh_nodes
- https://github.com/ltdrdata/ComfyUI-Impact-Pack
Examples:
My Simple Text to Image SDXL Workflow

14