Resized to 49% of original (view original)
Prompt | qswaxycbdgetrfvnhzmujkiö, .lo-päü+#+qwüeprotizua#sädöflgkhjy-x.c, vmbn+püöä#-.ol, mkiujnbfgvhztreqwasdcxy+#äü-.öpol, mkiujnbhztgvcfredxyswqatzghbnvmfjrueidkc, x.dleowpöx-ysä#+üwqay♥ |
---|---|
Negative prompt | unpicturesque, unquality, unnice, unlovely, unpleasant, unpretty, unfresh, unenchanting, undelightful, unclear, unspectacular, undazzling, unbrilliant, unterrific, unsuperb, unmagnific, unsharp, unhot, unanatomy, unsightly, {worst quality, bad quality, very displeasing, displeasing}, url, web address, text, signature, agorgeous, antigorgeous, ungorgeous, disgorgeous, dysgorgeous, agorgeous, antigorgeous, ungorgeous, disgorgeous, dysgorgeous, {{{mismatched pupils, empty eyes, no pupils, @_@}}}, unfinished, jpeg artifacts, jagged, chromatic aberration, simple, blurry, trash, noise#, #{{{mismatched pupils, empty eyes, no pupils, @_@}}}, gigantic, hyper, {{{from behind, from behind, from behind, multiple girls, baby, cropped legs, child, toddler |
Sampler | Euler Ancestral |
Seed | 1277802064 |
Steps | 28 |
Cfg Scale | 8 |
Model Hash | c1e1de52 |
- ‹ prev Search: commentary:untranslated next ›
Aerial Hug
This was mostly a genetic generation, as in there wasn't really any cohesive prompt behind the final results. Instead the prompt was deliberately kept random, then I've went looking for something interesting, and repetively ran it through i2i+VT and moved forward with the images that retained and leaned in on good traits. Also mixed two images manually once.