Based on workflow from civitai,I created this text to image workflow with face detailer(Flux/illustrious),and you can compare the result before/after the face refine.
I made 3 versions for Flux dev,gguf and Illustrious checkpoints.
I tested with several checkpoints,this workflow works well with these checkpoints:
For dev version:
Edge of Reality 真实边缘 - 3.2 Beyond Reality | Flux Checkpoint | Civitai
FLUX - Dev | Flux Checkpoint | Civitai
Real Dream - Flux.1 - V1 | Flux Checkpoint | Civitai
majicFlus 麦橘超然 - v1.0 | Flux Checkpoint | Civitai
For gguf version:
Real Dream - Flux 1 V1 GGUF Q4_K_M | Flux Checkpoint | Civitai
Real Dream - Flux 1 V1 GGUF Q5_K_M | Flux Checkpoint | Civitai
For illustrious version:
Prefect illustrious XL - v2.0p | Illustrious Checkpoint | Civitai
Vauxz Merge 2.5D - 3D - ILLU v.Sarah | Illustrious Checkpoint | Civitai
Description
This version is for gguf checkpoints.
FAQ
Comments (2)
Where do I go to adjust the resolution of the image?
If under img2img mode then you may need to develop the current workflow since according to this workflow img2img will keep the width and height of the original pic.But under the txt2img mode you may adjust the width and height just by typing.