What This Is

It is a initial merge of MoonMix - Holiday Mad Dream, 【Checkpoint】YesMix, and AniToon V1.0 using ComfyUI so I am not able to prune it. I tested it several times with different samplers to make sure it would actually generate images and let it rip. I even baked the base SD 1.5 VAE into it for you. :)

What This Is Not

It is not a trained merge. I did not put any LoRAs into it beyond what were baked into the original models.

What It Can Do

Anime girls in various styles.

Usage:

Special Trigger Phrase: Anime Edition

CFG: 6.0 - 8.0 in Automatic1111 OR 8.0 - 10.0 in ComfyUI

Steps: 20+

Sampler: Euler A or LCM (with LoRA)

I also recommend using high-res fix on it and avoiding ADetailer except to fix hands.

Styles: (style:weight) where

• Style - Anime, Lineart, Watercolor, Photorealism, Chibi (these are the ones I've discovered)

• Weight - anywhere from 1.2 to 1.5. Any higher tends to give strange results.

• Styles can be combined either as one tag with the same weight (:style1, style2:weight) or as separate tags with different weight (style1:weight1), (style2:weight2)

Level of Detail: any or all of highly detailed, extremely detailed, intricate, extremely intricate, intricate details

Clarity of foreground:

• Persons: detailed skin, detailed pores, perfect skin, beautiful face, perfect face, symmetrical face, detailed eyes (just the ones I've found to work)

• It also works well with many Danbooru tags so feel free to experiment with this.

Clarity of background: any or all of beautiful scenery, epic scenery, detailed scenery, intricate scenery, wonderful scenery, detailed background, simple background, bokeh blur, extremely blurred background (just the ones I've found to have some effect)

Have fun exploring!!

描述:

训练词语: Anime Edition