AI生成街拍美女图像的提示词示例｜Stable Diffusion实用模板

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 res, (photo-realistic:1.4), super fine illustration,vibrant colors, masterpiece, sharp focus, best quality, depth of field, cinematic lighting,ultra detailed,1girl,solo,white shirt,crop top,midriff,denim shorts,large breasts,long hair,silver hair,red eyes,blush,smile,open mouth,street,<lora:more_details:0.5>

【ネガティブプロンプト】

(worst quality:2), (low quality:2), blurry, out of focus, bad anatomy, extra limbs, extra fingers, deformed body, bad hands, bad legs, low resolution, watermark, text, logo, overly muscular, unnatural face, disfigured face, bad proportions, missing limbs, duplicate people, pixelated, low contrast, sitting

【Model】

Stable Diffusion checkpoint・・・beautifulRealistic_v7.safetensors

Beautiful Realistic Asians – v7 | Stable Diffusion 1.x Checkpoint | Civitai

Im currently preparing and collecting dataset for SDXL, Its gonna be huge and a monumental task. I wanna thank everyone for supporting me so…

【その他の設定】

■ Sampling method・・・DPM＋＋2M

■ Schedule type・・・Karras

■ Sampling steps・・・40

■ Width・・・800

■ Height・・・800

■ CFG Scale・・・10