Uploaded by Feather Skylark
3744x4480 JPG 618 kB Description
No description provided.
Tags
+-SH safe15570 +-SH prompter:feather_skylark7 +-SH fluttershy4024 +-SH pegasus8240 +-SH pony21832 +-SH g438530 +-SH crepuscular rays93 +-SH cute4872 +-SH female42902 +-SH floppy ears1123 +-SH flower1188 +-SH folded wings880 +-SH full body608 +-SH grass963 +-SH looking up443 +-SH lying down4633 +-SH lying on the ground18 +-SH mare13997 +-SH prone907 +-SH shyabetes432 +-SH solo34161 +-SH subsurface scattering26 +-SH wings12367
Source
not provided yet
Loading...
Loading...
ah, that might be part of my struggles. i use a self-managed local ComfyUI server that pipes into a Krita plugin for in- and output. it’s “free” and private that way, (and also scuffed), but it also does nothing to help me get nicer results, and alternate styles are one of the things i haven’t figured out so far, so i just get the model’s default look. and if i add a Lora for a certain pose or whatnot there’s a good chance it just messes up the generation with duplicated anatomy and/or subjects/objects clipping into each other
The app I’m using to generate these images provides a variety of art styles to choose from. Some styles are surely from certain artists, and I guess that’s why it feels more natural than some models like Pony Diffusion
it’s a lovely result, it doesnt even look machine generated. there probably are details that suggest its real origin, especially to an artist, but to me it looks very natural. i wish i knew how to get output like this from the models i play with
It is purely AI work, actually
Edited