Workflow Name: Style Consistency Multi-View for Characters

【Workflow Introduction】
In this workflow, please upload two images: one as a pose image and the other as your chosen style reference image. In this process, we utilize the Kolors model's IPAdapter for efficient style transfer. At the same time, we employ ControlNet involving skeletons, lines, and depth for control. Finally, facial repair processing is performed to ensure the final result is more refined and maintains style consistency. If you are intrigued by this process, feel free to try it out and explore the potential of these features to give your images a unique artistic style and charm!

【Use Case】
In the creation of multi-view works for game CG and anime, by uploading a beautiful character pose image and a retro-style reference image, and following this workflow, you can achieve satisfying multi-view results.

【Key Nodes】
IPAdapterAdvanced(kolors), ControlNet

【Model Version】
Kolors
Model Name: Kolors diffusion pytorch model.fp16.safetensors

【LoRA Model】
None

【ControlNet Application】
Depth Preprocessor
Start Time: 0
End Time: 1
Strength: 0.6

【ControlNet Application】
Canny Preprocessor
Start Time: 0
End Time: 1
Strength: 0.6

【ControlNet Application】
Openpose Preprocessor
Start Time: 0
End Time: 1
Strength: 1

【K Sampler】
CFG: 3
Sampling Method: dpmpp_sde
Scheduler: sgm uniform
Denoising: 1