I wanted to replace the old Midjourney v1 art I had (which was the whole reason I got into GenAI) with some Flux generations, but I thought to myself, why not use Wan2.1 to also make the cards animated. And so I did!
My workflow is to first generate a flux image, then I switch to a wan2.1 workflow and import that image as the source, and use a Live Wallpaper Style LoRa to add a simple idle animation.
I then import it into Hypnagonia and if an animation exists, then the card art will loop that animation. I will also allow a switch to disable it, or prevent looping as well.
The best part about making a surrealist/dreamlike game is that the GenAI imperfections in art and animation are totally in-theme for the world.
PS: Anyone have a good method to create perfect loops in Wan2.1?
How do you run these video models? I’ve been using Krita AI Diffusion for still images, but I’m interested in getting into video
If you have at least 24g vram I can send you my comfyui workflows
Please do. I don’t have that much VRAM, but I do have slow RAM I can use.
This is the workflow and plugins I used. Try to set this up. You will need to download some custom modules.
I must have missed, which app are you using to run the models?
ComfyUI
Note that it took me 5 minutes per vid on a 4090. On ram expect hours per vid
