/vt/ - Virtual Youtubers

>>69192727
>Wouldn't I need at least 20 good images for that and then tag them each manually?
There are auto-taggers and stuff. if you just make a dataset with just the pictures, archive and box it, I can take care of the rest for you for animagine or 1.5 models. It would be preferable to have the outfit pics in different styles/sd models to not get the style trained into the lora.Can't help with pony.
>Wow! That looks awesome, is it a complex process to make that Fauna picrel?
Not really, It was pretty random. With comfy I have it automated, with A1111 it would be a bit more involved as in sending it to img2img and running SD Ultimate Upscale on it after using Adetailer - person.
Here's the schizo prompt if you think it would help you:( \n means new line)
>masterpiece, best quality, \n1girl, ceres fauna, hololive, antlers, green hair, hair flower, hair ornament, mole under eye, yellow eyes,\n\n\nin absolute black darkness, ais-darkpartz,Dreamyvibes Artstyle,official art,unity 8k wallpaper,ultra detailed,beautiful and aesthetic, super closeup portrait, (detailed, masterpiece, 8k, 8k wallpaper),\nUltra detailed illustration of the light silhouette of a woman,phantasmagorical figure,oil on canvas, small brushstroke, (supreme masterpiece by Lucien Pissarro and Master W. B. and Allegretto Nuzi:1.2), mysterious glow gleaming light, cosmic andinfernal and striking atmosphere, (vigorous diarylide yellow and bubblegumpink colors:0.1), superb details, sharp focus,\n\n 8k, absurdres,