This came about as an experiment. I recently made a lora of actress Poppy Drayton, and generated several images of her as Amberle from the Shanarra Chronicles. One of the images had a dress that I absolutely fell in love with. I tried reproducing it in other generations, but it was never quite right, and Flux doesn't really have a good Reference Control Net to use. So I needed a way to replicate it.
This is the result of taking that photo, and making a ten second video of the character walking around on Kling. I capped 16 frames from that video and used those to make this lora. Since Kling plays havoc with faces, the face in this Lora is inconsistent and doesn't really resemble Poppy. Run together with my PoppyD lora however, both at 0.8 has produced some great results.
So I'm posting it to see if anyone has any use for it, and to just talk about the process.
Description
FAQ
Comments (1)
Actually, I like the fact that you talked about the process you used. I have a feeling that when I start doing loras I will be taking videos and pulling stills to maximize the effects I want. It looks like one of the best ways to maximize the effects/colors/ poses/ details that are wanted.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.







