The workflow doesn't give good results for me.

[Yagsuwu]

As you can see in the video, the character morphs into someone else with a green background and the end image is not even considered. Your given workflow doesn't use Wan In2p fun models. How to get better quality like your result? I am also sharing your modified wf from which i got the result.

WanSEI_00003_cgckx_1744475011.mp4

Blank Workflow.json