Minori Yamato PonyXL (Deathblight | Crescentia Fortuna) | LoCon

Minori Yamato from the webcomic by Projekt Wolfenstahl "Deathblight". Works best around 0.7-1.0 weight. Training done on Pony Diffusion V6 XL.

Trigger: "mnori, wolf girl, wolf tail, animal ears, multicolored eyes, black hair, large breasts, nail polish, toned" always needed for accuracy. Add "Hair ornament" for her classic hair ribbon that Ferania gave her in the comic. Add "abs, navel" where visible for proper physique. Hairstyles, choose one: "short hair" for her main hairstyle. "long hair" for her long hair from early chapters. "single hair bun" for her hairstyle from earlier chapter.

Suggestions/notes:

Halterneck dress Outfit: "purple halterneck, cleavage, purple dress, short dress, black gloves, fingerless gloves, elbow gloves, bike shorts, black thighhighs, black belt" Somewhat inconsistent due to poor data but can work with some RNG.

Plunging neckline Outfit: "purple dougi, plunging neckline, black choker, black gloves, fingerless gloves, elbow gloves, black thighhighs, belt bow" The bow is a bit random but the data was poor sadly.

Crop top Outfit: "multicolored crop top, japanese clothes, collarbone, sleeveless, midriff, black shorts" Best used with "long hair" style as that was the original in the comic for the outfit. Works well even for the limited data but the top has some RNG.

Very flexible and NSFW able as there was a lot of nude data and the outfits were very normal.

I suggest lower weight to 0.8 when using other outfits or loras. The nails are slightly inconsistent but work better than when I made Ferania. Remove "nail polish" if you dislike them or have issues with the nails.

Minori is best girl so it was a battle to get her working well, I managed to capture face, eyes, hair very well but sadly he outfits lacked good data so do not work as well as I would like. There is some style burn but sadly most of the data is from the comic and there was not much fanart to work with. I hope you all enjoy.

Feedback and reviews are always appreciated. Training data is in the Lora metadata on civitai. Trained on 105 images.