Civitai 見下ろし複数少女上半身 / Multiple girls, with from above and upper body.
LORA
Illustrious
LORA
Illustrious
v1.0

Civitai 見下ろし複数少女上半身 / Multiple girls, with from above and upper body.

Creator
⭐ 0.0
⬇ 409 Downloads
👁 2 Views
🖼 1 Images

Trained Words

on back
lying
looking at viewer
from above
outdoors
clenched_hands
grass
on grass
upper body
1girl
solo
2girls
3girls
4girls
5girls
6girls

About this model

草むら背景の見下ろし式複数少女寝そべり構図

所々欠点あり

■構図要素
・垂直 上からの完全見下ろし
・背景草むら
・上半身描写
・複数少女

■推奨適用強度
1.0- 構図維持に高強度必要。下限は0.8辺りか

■対応人数
1-6人 学習内容的にはタグとして1-6人としているが実際には描写成功率がかなり異なる。

1>2>4>>>3>>5=6

といった所か。5人以上は描写領域的に安定せず。
3人以上はそれぞれ1人分ズレて生成される事がある。
実際の利用域としては1-4人対応といった所か

複数人対応していてもキャラの服装等が正確に追従してくれる訳ではなく、
モデル内蔵キャラでも服装が混在してしまう感じが出る。
そこらへん踏まえると安定的な描写は2人描写辺りまでが現実的と言える。

■その他
描写明度というか彩度が高く白っぽくなりがちなのが欠点


Multiple girls lying down in a top-down view with a grassy background

Has some flaws

■Composition Elements

・Vertical, completely top-down view

・Grassy background

・Upper body depiction

・Multiple girls


■Recommended Application Strength

1.0- High strength is required to maintain the composition. The lower limit is around 0.8.


■Supported Number of Characters

1-6 people. While the training content specifies 1-6 people as the tag, the actual rendering success rate varies considerably.

1 > 2 > 4 >>> 3 >> 5 = 6

Something like that. More than 5 people are unstable in terms of rendering area.

More than 3 people may be generated with a shift of one person's width.

The practical range of use is probably 1-4 people.

Even with multiple character support, the characters' clothing doesn't always track accurately,

and even with built-in models, the clothing can appear mixed up.

Considering these points, stable rendering is realistic up to around 2 characters.


■Other drawbacks: The depiction tends to be washed out due to high brightness and saturation.



Related Models

Similar AI models you may like