Workflow Name: Multi-Head View Role Table
【Workflow Introduction】
When you open this workflow, you first need to upload 3 images. The first image on the left is a portrait used for face swapping, the second image on the left serves as a reference image for prompts, and the third image can be a skeleton image or a multi-view image, which needs to be processed using DWopenpose. Next, the face swapping operation is performed using PuLID, the skeleton function of controlnet is utilized for control, and the Flux model is used for image generation. Currently, the built-in parameters of the workflow have been set, so you can experience it right away!

【Usage Scenario】
According to this workflow, when you upload a Harry Potter image as a reference for your own style prompts, and then upload a portrait of yourself, you can obtain a similar multi-view styling effect. This will undoubtedly bring you great surprises.

【Key Node】
PuLID

【Model Version】
FLUX
Model Name: fux1 dev Q8 0.gguf

【LoRA Model】
None

【ControlNet Application】
FLUX.1 dev ControINet Union Pro Shakker Labs.safetensors
DW Pose Preprocessor
Start Time: 0
End Time: 0.7
Intensity: 0.65

【Custom Sampler (Advanced)】
Sampling Method: deis
Scheduler: beta
Denoising: 1