@Background Pony #4C0D
AI has a bit of a ways to go to entirely emulate artist styles, but there have been some decently good attempts. It’s even seemingly possible to copy an artist’s style with only a single image but it’s not really too accurate, but the potential is there.
@Background Pony #4C0D
I didn’t make the LORA file, but I did make a request of that same creator to do this humanized Queen Chrysalis LORA. That same creator also did the Sonata Dusk one that was released last week that this prompter used in previous images posted here days ago.
From what I know with general LORA creation, they just take a bunch of images of the person regardless of artist or art style, pack it together and run a script to create the LORA model. From there, since it’s made public in CivitAI so that anyone can download and mess around with it in stable diffusion with whatever checkpoint and prompts they desire leading the result that you see here.
Interestingly, depending on how many source images, if you have one artist that contributes large chunk of images for the LORA something along the majority, you’ll sort of get that artist’s style bleeding over to the image generations. It’s an unintended effect though.
AI has a bit of a ways to go to entirely emulate artist styles, but there have been some decently good attempts. It’s even seemingly possible to copy an artist’s style with only a single image but it’s not really too accurate, but the potential is there.
And that is why I made it aware about this being an AI work because of what AI does to an original drawing.
I didn’t make the LORA file, but I did make a request of that same creator to do this humanized Queen Chrysalis LORA. That same creator also did the Sonata Dusk one that was released last week that this prompter used in previous images posted here days ago.