I've made so many different style loras of artists who does TSF and it got me thinking, how about I just make a concept so that you can make sequences out of any styles?
Well, this is it. Trained with a few hundred sequences, this "Should" work as intended with various different style loras and whatnot, though you might wish to use it at a strength around 0.3 to 0.5.
Tested with some of the style loras' that I've made and holy shit, it works.
Might need repeated attempts but this is quite simple to use :
"Your style lora, <lora:Linear_TSF_Sequence:0.4>Linear_TSF_Sequence, linear transformation, transformation , sequence,
1boy, outfit and all that,
1girl, outfit and all that,"
Might wanna put "dissolving clothes" for extra styles, also try bigger resolutions for more stages, while smaller ones have better fingers and faces.
Description
FAQ
Comments (3)
Any plans on creating Anima ver for TSF sequence? You would be a lifesaver!
Oh ho ho, yes buuuuut I will need to re-caption everything to be suitable for natural language so anima model will do its magic , the illustrious one that’s just the trigger and “1boy , clothing and all that, dissolving clothes and all that” followed by “1girl, clothing and all that” just won’t do for anima , I’ll try to sort of tag it in such a way that every frame or stage is clearly labelled . Might take quite a while though , since I’ve used about 200 images for this one , and 200 x 4 is about 800 images to tag , at the very least
welp, here you go, compiled, captioned and trained, https://civitai.red/models/2820631/animalineartsfsequence ,this one however have an instruction sheet in the model description , it might be a bit of a pickle to work with
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.