Hypernetwork of adult Rosemary Winters from RE:Village. Used edgeofRealism to train and generate.
Description
FAQ
Comments (3)
So these smaller models, are they a way to sort of limit your prompt specifically to only what is provided in the model. So rather than an extension its a limitation? Thats great if so, but couldnt you just use any other model and just prompt for this 'rosemary'? Is that because it might not be able to do it as well? Basicohlly im still trying to get a hang what this is all about.
@dadadies
LORA / dreambooth = teach a model something new, something that it does NOT know until you teach it. This can be an object, person, very specific face, pose, or a style.
Hypernetwork = train the model to use things it already knows in a specific way, usually best to set up the image at first. Fairly similar to TI / Embeddings, but takes much longer to train.
Textual inversion / Embeddings = train the model to use things it already knows to make a specific thing in an image - like training a face. If the model already knows faces, it's just a shortcut to prompt with one word for a very specific nose / chin / mouth / eyes combo that you could get with a long complicated prompt, since the model already knows all of the pieces. It takes longer to train than a LORA, and if the model doesn't already have enough of the features of what you are training it can only put out things that are kind of close to the training subject.
(Summary found on Reddit: https://www.reddit.com/r/StableDiffusion/comments/12gfu41/comment/jfl2pu9/?utm_source=share&utm_medium=web2x&context=3)
Is it possible to make "The Baby" hallucination version of her?