Lora Strength: 0.7 - 1.0

Trigger Words

Character A:

• yurieww, blue eyes, brown hair, short hair, low twintails, short twintails, braid, hair behind ear, hair over shoulder, bangs, medium breasts,

Outfit

• turtleneck sweater, ribbed sweater, long sleeves, pleated skirt, kneehighs, loafers,

Character B:

• yuriebww, blue eyes, brown hair, medium hair, twintails, sidelocks, bangs, medium breasts,

• yuriebww, blue eyes, brown hair, long hair, hair down, sidelocks, bangs, medium breasts,

Outfit

• frilled headdress, neck ribbon, crop top, collared shirt, black sleeveless shirt, frilled shirt, center frills, black skirt, arm garter, elbow gloves, white gloves, frilled skirt, miniskirt, brown pantyhose, mary janes,

Since this LoRA was trained on a dataset of monochrome images, it is recommended to add "monochrome, grayscale" to the negative prompt when generating color illustrations.

Additionally, including prompts such as "anime coloring" and specifying hair and outfit colors can increase the likelihood of generating color illustrations.

For example, try prompts like "brown eyes" or "blue skirt" to clearly indicate colors other than black or white.

If the character's trigger word already includes color information, you may not need to worry about this as much.

このLoRAはモノクロ画像のデータセットで学習されているため、カラーイラストを生成する際にはネガティブプロンプトに "monochrome, grayscale" を追加することを推奨します。

さらに、"anime coloring" のようなプロンプトを含めたり、髪や服の色を指定することで、カラーイラストを生成する可能性を高めることができます。

例えば、"brown eyes"や"blue skirt"のように、白や黒以外の色を明確に指定するのが効果的です。

なお、キャラクターのトリガーワードに既にカラー情報が含まれている場合は、特に気にしなくても構いません。