>>24118212It's insane how well-documented everything is. They tag artists, characters, poses, miniscule details (color of clothes or nipples, finger count etc), references, sibling images, certain stylistic choices, languages used if any, color palette, image aspect ratio, publication date, everything.
No wonder furry local diffusion models trained on that data are so capable and accurate out of the box. If only the same was true for t2v or i2v models, which are trained on god knows what.
>Species>canid canine fox mammal >General>5 fingers 5 toes anthro asian clothing big breasts biped blush bodily fluids bottomwear breast squish breasts bulging breasts cleavage cleavage overflow clothed clothing detailed background east asian clothing feet female fingers footwear fundoshi gloves (marking) hair hanging breasts humanoid feet japanese clothing leaning on wall leg markings loincloth looking at viewer markings natural breasts one breast out plantigrade pube peek pubes sandals shoes socks (marking) solo squish standing sweat thick thighs toes underwear >Meta >2017 colored cover cover art cover page digital drawing (artwork) digital media (artwork) full-length portrait hi res portrait Someone filled that in.
