I made these for myself but I figured I'd share them for other sickos, in the hopes those sickos also share stuff they make so I can see it.

All metadata should be intact, so you can recreate my methods if you want to create your own.

arranged roughly in order of creation. In the future I'll have better file name sorting methods.
stuff generally gets more guro farther towards the end, as I found new loras.

includes: bdsm, bestiality(horse, werewolf), monster rape, goblins, fellatio, impalement, needle torture, amputee(bloody, and non bloody), stocks, public humiliation, branding with hot irons, occasional flesh wounds.

characters include
Tatsumaki,
Saruei (vtuber),
dva (overwatch)
yoruichi shihouin
soifon
Raven (DC)
Etna (disgaea)
ohisashiburi's lady bat (Mermaid Melody Pichi Pichi Pitch)
Shuten_Douji_(Fate)
anubis
micchan (ohisashiburi)
ifrit_(arknights)
rebecca (cyberpunk edgerunners)
cat girl, cat
nagatoro hayase
au ra, horns
y'shtola rhul
casca (berserk)
futaba sakura (persona 5)
lilith aensland (stalkers)
matoi ryuuko
witch
high elf archer (goblin slayer)
elf (dragon's crown)
amazon (dragon's crown)
jinx (league of legends)
Tanya Degurechaff (Youjo Senki)
jane doe (zenless zone zero)
iori_(blue_archive)
ramlethal valentine (guilty gear)
Kasumi Yoshizawa (persona 5)
mitsuru (persona 3)
nadia la arwall
bambietta basterbine (bleach)
lolly (bleach)
tier_harribel
mirko (my hero academia)
cithis (dungoen meshi)
platinum (arknights)
soukaku (zzz)
kneesocks (psg)
perrine (strike witches)
riley fairfeather (sort of)
abigail jones
demon girl
dragon girl,
dryad elf
vampire girl
soldier girl
and more

more，plz

Base +4

感觉较为一般，p站一些大佬烧烤母畜的模型已经炼得很成熟了

Base +6, wantedman007 +6

直接搞死的反而不过过瘾啊

Base +6

thanks share the keyword

Base +6

这类模型有分享的地方吗？

Base +4

not sure what you mean by keyword.
the loras and models mostly either came from civitai, with a few from gurochan ai board.
these were also generated in text2ikg then upscaled using tiled diffusion by 1.5 in img2img
the entire prompts, model, seed, etc is attached to each image as a chunk of metadata and can be seen if you put an image into sd webui's png info, or use tweakpng or other tool.

not sure what you mean by keyword.
the loras and models mostly either came from civitai, with a few from gurochan ai board.
these were also generated in text2ikg then upscaled using tiled diffusion by 1.5 in img2img
the entire prompts, model, seed, etc is attached to each image as a chunk of metadata and can be seen if you put an image into sd webui's png info, or use tweakpng or other tool.

Base +6