Please react to my videos: https://civarchive.com/user/ntc/images?sort=Most+Reactions
#:0.08|~photo of gucci fashion, lips pursed, stern aloof look, profile picture, full torso and head in shot, fashion magazine photoshoot, fashionable hairstyle, cheekbones, hair fluffed and down
https://github.com/ntc-ai/conceptmod
You might belong in Gucci,
Where dwell the brave at heart,
Their daring, nerve, and chivalry,
Set Guccis apart
Description
FAQ
Comments (5)
Forgive my utter ignorance but how does one train without a dataset? I thought you HAD to have some sort of data on which to train. I am quite confused.
Is right there in the description, the github link. It only works on things that SD already has some knowledge of it seems.
The data is within the model, this is a form of self play where the distance between '' and 'some prompt' is estimated by measuring the result of the diffusion process across three points and two models. You end up with a 'concept' vector for 'some prompt' in the latent that can be optimized towards. It's all very experimental.
It's open source if you want to mess with it.
This is brilliant work. I think I'll have a play over the weekend.
from ？ to ？
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.