Hi friends! TLDR: This is a Wan2.1 test and you can use that model in the cloud here (it’s not cheap though). All unedited videos are here and the prompts are here. Also, Latina RC keyframes are almost completely done! I mentioned earlier that I wanted to run head-to-head tests on these new NSFW models using the same keyframes, but I’m relatively sure that Wan2.1 is the best. I still think Hunyuan text-to-image is great but none of the current image-to-video implementations (official, Skyreels, Dashtoon) seem to work as well as Wan2.1. My current plan is to try the documentary TG script with this model after I finish the Latina RC.
Speaking of the Latina RC, I’m very close to finishing the hires keyframes! All that remains is harmonizing them – ensuring the backgrounds match perfectly, the clothing colors are as consistent as possible, ect. It will probably take longer than I expect, but it’s possible I’ll be able to start generating the animation this weekend! I will post a rough, no-sound version for tippers as soon as it’s done. Realistically, the full release probably won’t be until early April. I really, really underestimated the time it would take me to generate multiple stages of a transformation from multiple angles, so I apologize for the delays. This will be the most detailed transformation I’ve ever tried, and going forward I probably will try to use a 3D render -> AI workflow to hopefully make things easier/faster.
As always, thank you for the support and patience. I’m very excited for you to see what I’ve been working on!
Blankage
2025-03-22 13:13:41 +0000 UTCIstmael
2025-03-22 11:07:35 +0000 UTC