AI Beautiful Woman Portrait: Stable Diffusion Prompt Example
ポーズや表情、背景などを指示して画像を再生成。生成物は右の「生成物」パネルに追加されます。高解像度ダウンロードは¥150(税込)で解放されます。
当記事のステーブルディフュージョンはGoogle colabを利用して作成しています。
【プロンプト】
(best quality:1.2), (masterpiece:1.2), ultra high resolution, (photo-realistic:1.4), 1 beautiful girl, looking directly at the camera, wearing an oversized off-shoulder knit sweater over delicate lingerie, resembling a famous Korean actress, stunningly beautiful and kawaii, cinematic feel, captured with a 35mm lens at f/1.8, accent lighting enhancing her features, ultra-detailed, 8k resolution, luxurious hotel setting, warm ambient lighting, soft and inviting atmosphere.
【ネガティブプロンプト】
(worst quality:2), (low quality:2), (monochrome), (grayscale), visible skin imperfections (acne, blemishes, age spots), (outdoor scene:1.6), (ugly face), nsfw content, poorly defined features (bad hands, bad arms, missing fingers), overly muscular, unnatural proportions, multiple subjects (4-6 people), awkward composition, unwanted elements (watermark, logo, text), sketch-like and rough painting style.
【Model】
Stable Diffusion checkpoint・・・beautifulRealistic_v7.safetensors
Beautiful Realistic Asians – v7 | Stable Diffusion 1.x Checkpoint | Civitai
【その他の設定】
■ Sampling method・・・DPM＋＋2M
■ Schedule type・・・Automatic
■ Sampling steps・・・20
■ Width・・・512
■ Height・・・512
■ CFG Scale・・・7
Prompt Overview
This prompt is a minimal, open-ended image generation request focused on portraying an attractive female subject. Its simplicity makes it an ideal starting point for exploring how different AI image models interpret concepts of beauty, style, and composition by default — before any additional parameters are layered in. Designers, content creators, and hobbyists often use bare prompts like this to benchmark a model's baseline aesthetic tendencies.
To get more targeted results, consider expanding the prompt with details such as lighting style (soft natural light, studio lighting), setting (urban background, neutral studio), mood (confident, serene), age range, or artistic style (photorealistic, oil painting, anime). Models like Midjourney, Stable Diffusion, and DALL-E each interpret open prompts differently, so experimenting across platforms can reveal a wide range of creative outputs from the same simple starting point.
Was this prompt helpful?
Comments
Comments appear after moderation
この記事が役に立ったら投げ銭で応援
Apple Pay / Google Pay / カード (Visa/Mastercard/JCB/Amex) / Link / Alipay / WeChat Pay 対応 · Stripeで安全に決済
開発者が選ぶ最強ツール集
運営者が毎日使っているツール・ガジェット 6選