イラストリアス（IL）版の統合LoRA

・タオパイパイの乗ってたピンク柱
・筋斗雲2人乗り（1人乗り可）
・上記同時描写（学習元は１人柱+１人筋斗雲）

時間無くてAnima自体まだ使えてないなぁ
タグ記載か自然言語記載かでずっと悩みそう

ドラゴンボール　タオパイパイの柱乗り＆筋斗雲。同時描写可能版。
同時描写側の柱乗りは縦乗り（普通の柱に乗ってる）に化ける事がある
on pillarをタグに使ってるとまぁそうもなるか
横倒ししてると説明書いても微妙だったので本来的には横倒し柱のタグでも入れとくか、そもそもon pillar使わない方が良かったかもしれない

■推奨適用強度
1.0 下限的には0.8辺り
ピンク柱＆筋斗雲は並走する形で人物描写サイズを同等に出来ないと
描写がどうしても潰れがちになりやすいかもしれない

■その他
記述が甘いせいかサンプル画像、自然言語で筋斗雲二人の前後関係を制御出来てない

Integrated LoRA for Illustrious (IL) version

- The pink pillar Tao Pai Pai rode

- Two-person riding on the Flying Nimbus (single-person riding possible)

- Simultaneous depiction of the above (learned from one person on the pillar + one person on the Flying Nimbus)

I haven't had time to use Anima yet.

I'll probably keep agonizing over whether to use tags or natural language descriptions.

Dragon Ball - Tao Pai Pai's pillar riding & Flying Nimbus. Simultaneous depiction possible version.

The pillar riding in the simultaneous depiction can sometimes turn into vertical riding (riding on a normal pillar).

Well, that happens when you use the tag "on pillar".

Even with instructions, it's a bit awkward when it's lying on its side, so maybe I should have included a tag for the horizontal pillar, or maybe it would have been better not to use "on pillar" at all.

■ Recommended application strength

1.0, lower limit around 0.8

If the pink pillar and Flying Nimbus aren't drawn side-by-side with the character sizes equal,

the depiction tends to get distorted.

■Other:
Perhaps due to insufficient description, the sample image and natural language fail to control the context between the two characters on the Flying Nimbus.