<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/1fe83225-1662-43fe-8aa2-d47495881b0c/width=525/1fe83225-1662-43fe-8aa2-d47495881b0c.jpeg" />Here is my Wan2.2 Workflow. I'm still improving it but I think its a nice starting point if you don't have a working workflow for Wan2.2. I got this workflow from someone else but I added some nice customization that I think make it worth sharing.As a rule I always use <a target="_blank" rel="ugc" href="https://civarchive.com/models/1307155">(wan 2.2 experimental) WAN General NSFW</a> Loras even though I'm making SFW, because it helps with prompt adherence. Or at least it helps with Danbooru tag prompting.Also I tend to always use this <a target="_blank" rel="ugc" href="https://civarchive.com/models/1626704?modelVersionId=1971181">Boob Physics</a> Lora because it seems to add a bit of bounce to everything.If you are just getting started I recommend lowering the height and width to 400 each so you can experiment. 1600 pixels (400*400=1600) is lowres but 2400 pixels (600*400) produces nice results. Push the resolution higher if you have time and VRAM. 3000+ is a pretty good balance between speed and results.PS if you use Windows you will probably need to reverse all the "/" I used in the nodes that make the file paths. (Also you should consider dropping that spyware before Windows 11 starts literally recording everything you are doing ¯\_(ツ)_/¯ )PPS Credit to Olivio Sarikas and this <a target="_blank" rel="ugc" href="https://youtu.be/geSIepK8ekQ?si=OBtby8tDig7NoGdf">Youtube</a> as the source workflow that I built on.

Files with SHA256: d38d41768c0f1f6eb514fd5af3b69bf1c702a539e7488ee51bb6b27fd9260dfc