追加
いいね 34
eru
277 Followers·1647 Posts
言語モデルを活用しマルチモーダルに意味空間を拡張した多様体に基づいた画像生成に興味があります(キリッ) また、画像を用いることにより、知恵を利用したより面白い知識を選択したより意外性のある面白い小説を生成できるかどうか、にも興味があります
typescript形式はdalle3以外でも有効だろう
related url
https://www.aipictors.com/works/381162/masterpiece, best quality, extremely detailed, anime, girl,interface ImageDescription background: structure: string; details: string; lighting: string; ; character: pose: string; expression: string; hair: color: string; style: string; length: string; ; outfit: top: style: string; color: string; accessories: string; ; bottom: style: string; color: string; ; ; accessories: string[]; emotion: string; ; props: string[];const imageDescription: ImageDescription = background: structure: "Gothic architecture with large cross and stained glass window", details: "Detailed shadows and highlights emphasizing depth", lighting: "Soft, ethereal light filtering through the window" , character: pose: "Sitting on a stone platform, one leg bent, the other relaxed", expression: "Sad, reflective with slightly downcast eyes", hair: color: "Light blonde", style: "Straight, with some strands falling over the face", length: "Shoulder-length" , outfit: top: style: "Black halter top with cross emblem", color: "Black", accessories: "Black choker with cross" , bottom: style: "Short black skirt", color: "Black" , accessories: ["Black choker with cross", "Bandage on arm"], emotion: "Melancholy and introspective" , props: ["Stone platform", "Large cross in background", "Stained glass window"];,

