This Lora is intended to create key frames of a growth sequence, NOT smooth animation. From these, you can either just take frame clips and create a "comic" or sequence of still images, or you can use these key frames in a First Frame, Last Frame workflow<a rel="ugc" href="https://civarchive.com/models/1951971/body-growthexpansion"> with my other Lora</a> (or without) to create a smoother growth animation there.V 0.9Workflow Example: <a target="_blank" rel="ugc" href="https://pastebin.com/GZZsiDHT">https://pastebin.com/GZZsiDHT</a>I feel like I can eventually do a better job, but sharing initial attempt in case people wanted to try it. <ul><li>Due to being trained on NSFW sequences, but WAN 2.2 not being natively great for NSFW, there tends to be a lot of weird bulge or appendage coming from the crotch. This can be somewhat remedied by increasing the CFG of the High lora and including clothing in positive prompt or "naked" in the negative prompt. Used image also can affect this. I'll probably need to exclude most of the NSFW stuff in the future to prevent this.</li><li>If you are using a more photorealistic image, then you need to increase the High lora strength</li><li>I attempted to split generation between <code>single-panel comic</code> and <code>multi-panel comic</code> with those key words in the prompt, but the resulting video seems to mostly just depend on whether you passed in a regular image/single panel or a multi-panel image to start with and it kind of ignores those two terms. Still, try them out for different effects.</li></ul>

Files with SHA256: ed40cc78dd93e873ff38fc81397c9ba48d90e143f2b4602c21572f6232e764e9