For best outputs and usage:

Sampler: Euler a.

Steps: around 30, cfg 7 ~ 11, CLIP skip 2.

VAE: none required.

Pos prompt: start it with "score_9, score_8_up, score_7_up, score_6_up", adding "source anime" (or "source_anime") to better mode guiding towards such style. It will not require "source_pony" or "source_furry" but adding them ensures such outputs.

Neg prompt: "score_4, score_5, score_6, low quality, worst quality, monochrome, (censor), 3d, realistic, render, CGI, text, watermark, cartoon, sketch, greyscale, child, kid, loli, underage" plus what ever else you don't want to see.

Curiosity note: this model understands classical emoticons like :3, +_+ or :D.

More info about tags:

The model attends to certain tags in order to create different concepts:

Concept: "nekomata", (cat girl), "kittunemimi" (fox girl), "usausamimi (rabbit girl), "kuronagamimi" (dark elf dragon girl), "slime girl", "fangs out" or "oni".

Characters: apart from that any good pony model is well versed in known characters, this one particularly knows of the next ones by using their "trigger" word:

"olohpnxl", holo, spice and wolf

"nereirfpnxl", frieren

"nurefpnxl", fern (sousou no frieren)

"lebupnxl", ubel (sousou no frieren)

"eihplyspnxl", sylphiette

"siretaryergpnxl", eris greyrat

"yxorpnxl", roxy migurdia

"enialsihgpnxl", ghislaine dedoldia

Checkpoint Merge.