AI art "[h]-Dissolve"

[h]-Dissolve

18

例によって調整中ですが新Lora作ってみました⭐ 全然想定とは違う効果なのですが、学習画像から見直さないといけなさそうなので「これはこれ!」で使っちゃいます😋 モヤモヤさせるLoraです💭 もうちょっと霧の中から出てくるみたいにしたかったのですが、学習画像の時点でやり切れていなかったかもしれません ・距離による描き込みの省略 ・遠景のシルエット化 ・曖昧な光源 あたりを覚えさせてあるつもりです

(Masterpiece), (Best quality), <lora:[h]-Dissolve>, yust-hz, hazy, warm-white theme, immersed in moody fog, distance-based fading, ethereal atmosphere, lost edges, vanishing contours, atmospheric perspective applied to character, fading shadow contrast, fog absorbs details, silhouette dissolving into mist, subtle volumetric scattering, shallow depth through atmospheric haze, details disappear with distance, xxx667_illu, 1girl, dutch girl, sitting on windmill roof, feeling the wind, (blonde long hair), leaning back, arched back, hands spreading over head, singing song, small dog, medium dog, large dog, dog style old man, happy, fine day, misty morning, sunrise, side of lighthouse, light and shadow direction, lively location, blurry foreground

Mostra parametri
Immissione richiesta originale
(Masterpiece), (Best quality), <lora:[h]-Dissolve>, yust-hz, hazy, warm-white theme, immersed in moody fog, distance-based fading, ethereal atmosphere, lost edges, vanishing contours, atmospheric perspective applied to character, fading shadow contrast, fog absorbs details, silhouette dissolving into mist, subtle volumetric scattering, shallow depth through atmospheric haze, details disappear with distance, xxx667_illu, fine day, misty morning, sunrise, side of lighthouse, dutch girl is sitting on windmill roof and feeling the wind, (blonde long hair), leaning back, arched back, hands spreading over head, singing song, small dog, medium dog, large dog, dog style old man, happy, light and shadow direction, lively location, blurry foreground, ,
Dimensione
678 x 1024
Passaggi di campionamento
10
Metodo di campionamento
Euler a
Scala CFG
2
Negativo
lowres, worst quality, bad quality, bad anatomy, jpeg artifacts, old, oldest, watermark, anime screenshot, anime coloring, nostrils, nose, paper edge,
Correzione del viso
Helper di Prompt

Modello e LoRA utilizzati