Andrey Sure, this is just me playing around with SD running locally on my GPU (instead of some spy cloud), essentially, i use dreambooth to add a person (f.e my kid) to an existing model, then use it to generate images like
“kidNameSD as a superhero flying over a building with lazer eyes, comicbook, jack kirby” or “kidNameSD riding a dragon, illustration, fantasy theme”
i’m getting mixed results, but occasionally something truly amazing comes out, and results are better the more images i add to the dataset.