This is advanced modular workflow for Redux and ControlNet.
The idea here is to have more control over conditionings and redux. You can do style transfer, get image variations and mix up to 3 images.
Diffusion and GGUF models.
Image Crop.
ControlNet (depth+canny).
Autoprompting.
Ability to control Prompt conditioning and Redux weights.
Redux models - https://huggingface.co/StableDiffusionVN/Flux/tree/main/Redux
Canny lora - https://huggingface.co/black-forest-labs/FLUX.1-Canny-dev-lora/tree/main
Depth lora - https://huggingface.co/black-forest-labs/FLUX.1-Depth-dev
Thumbnail loras:
https://civarchive.com/models/844226/lea-seydoux-2014-2018-flux1d
https://civarchive.com/models/915918/moriimee-gothic-niji-or-lora-style
Description
FAQ
Comments (5)
very nice workflow. it adds lot of variety. Thanks for sharing
Hello, I would like to ask if this workflow can be connected to load the latest ControlNet Union Pro 2.0 model. Thank you very much.
hi! no this won't work, because mixing controlnet and redux via "conditioning" inputs results in very inconsistent outputs. for this workflow you can't use controlnet that uses "conditioning" as inputs, which is basically any controlnet except the official one from black forest lab, since they use "model" input.
@Nexssar ok thank your answer
Cool workflow, very useful! thx