Ro-Style

詳細

ファイルをダウンロード (1)

このバージョンについて

NOTES:

Because this lora was trained on art with panels (doujinshi), use negative prompts like: comic panel, manga, monochrome, multiple girls

The reason for the negative prompt of "multiple girls" is because the AI often confuses panels containing the same character as multiple characters.

Some basic possitive prompts you should add are: masterpiece, best quality, recent, newest, absurdres, highres, 1girl

Obviously if you want multiple girls remove the positive/negative prompts for those.

This lora was optimized for illustrious, and works well with Nova Furry XL (the illustrious version)

モデル説明

"Ro" is Doujinshi artist known for their furry/shortstack art style. This Lora was trained on his works. I made this because I wanted to make Ankha generations, but couldn't get the right body proportions. Ro's art is basically exactly the style that I wanted, sooo here we are.

GENERATION TIPS: important!!! you must include the trigger prompt: Ro-Style

This Lora was trained on doujinshi panels, so unless you want your generations to contain panels, you need negative prompts like: comic panel, manga, monochrome, multiple girls

Basic positive prompts are: masterpiece, best quality, recent, newest, absurdres, highres, 1girl

The reason behind specifying negative prompt "multiple girls" or the positive prompt "1girl" is that the lora cannot understand that panels containing the same character aren't multiple girls (if that makes sense). Obviously remove this if you actually want multiple girls.

This Lora was optimized for Illustrious and works well with Nova Furry XL V.5 (the illustrious version)

このモデルで生成された画像

並び替え

モデルタイプ	LORA
ベースモデル	Illustrious
公開日	2025-03-30
トレーニングワード	Ro-Style