RedAIGC / StoryMaker

StoryMaker: Towards consistent characters in text-to-image generation
410 stars 35 forks source link

How to Perform Pose Control as Demonstrated in Example (E) of Diverse Applications #2

Open endman100 opened 1 week ago

endman100 commented 1 week ago

Hello, thank you for open-sourcing and sharing this research. I would like to ask how to execute pose control, just like the example shown in (E) of the Diverse Applications.

RedAIGC commented 1 week ago

Thanks for your attention, pose control is simple, you can prepare one target pose extracted from another image, and use the StableDiffusionControlNetPipeline and an open-source pose-controlnet model to generate target image.

haoyuying commented 2 days ago

we need to rewrite StableDiffusionXLStoryMakerPipeline that inherits from StableDiffusionXLControlNetPipeline, right? Do you have some advices? thx @RedAIGC