This model is a merged model of the checkpoint groups that I used separately when I was creating my illustrations.

Without LoRA and with Textual Inversion, the line is relatively thin and adjusted around glamorous women.

I try to combine character LoRA and painting style LoRA to create an anime style, try to reproduce textures similar to the original, and create NSFW illustrations.

〇Merged Models

gy动漫010 v1

seizaMix v2

FiaMix++ Dynamic v4.0

【Checkpoint】YesMix v1.6(Original)

〇Notice

The model itself is available for commercial use, but basically check the model from which you are merging.

There are currently no models that restrict commercial use.

このモデルは私がイラストを作るときに、別々に使っていたCheckpointをマージしたモデルです。

Textual Inversionあり、LoRAなしのモデルでは線細めで、グラマラスな美女が生成されやすくなっています。

画風LoRAや固定のTextual Inversionを駆使して、アニメ調だったり、原作再現だったり、エッチな画像を生成しています。

このモデル自体は商用利用可能ですが、基本的にはマージ元の方針に従います。

※多分、商用利用不可のモデルは使っていないはず・・・

〇Recommended Settings

Sampling Method：　DPM++ 2M Karras

Sampling Steps：　arround 20

CFG Scale： 7 - 8.5

Hires Fix：Basicaly it is not nesessary. please repeat inpainting.

　　　　　If you use, please refer below

　Upscaler：R-ESRGAN 4x+ Anime6B

　Hires Steps：20

　Denoising Strength：0.4

　Upscale by 1.5

〇Recommended Textual Inversion

easynegative

bad-artist

badhandsv4

bad-picture75v

deep negative 1.75t