[Workflow Introduction]
This is a workflow that can generate a Tang Dynasty Princess style blind box based on the uploaded portrait. By inputting a photo, first randomly extract a posture picture, then fix the body parts through the controlnet function, and use the ip-adapter to make the face of the blind box more like the uploaded picture, and finally generate a higher quality final image through the picture-to-picture. We have set the built-in parameters. Now you only need to upload a picture to customize your blind box. Come and try it!

[Usage scenario]
The advantage of this workflow is its simplicity and ease of use and artistic processing capabilities. You can customize the blind box by uploading a picture, eliminating complicated steps. You can use this tool to make blind box peripherals, increase entertainment and creativity, and improve efficiency.

[Key nodes]
Controlnet, InstantID

[Model version]
SDXL
Model name: WildCardX-XL-ANIMATON.safetensors

[LoRA model]
None

[Controlnet application]
controlnet-depth-sdxl-1.0-xinsir.safetensors
Start time: 0.000
End time: 0.617
Intensity: 0.55

controinet-canny-sdx1-1.0-v2-xinsir.safetensors

Start time: 0.000
End time: 0.815
Intensity: 0.85

[Sampler]
Sampling method: dpmpp_2m
Scheduler: simple
Noise reduction: 0.52