Shakker.aiによる画像生成の記録簿

最近はShakker.aiで画像生成するようになりモデルとLoraの相性を検証しつつプロンプトの書き方も変えたのでプロンプトの検証もしながら画像生成をしている
その記録簿として、ここに残しておく事にする

モデル、Lora

モデル：Anime_illstration-XL-pundit

Anime_illstration-XL-pundit-Checkpoint-AIGC_JIAMO-Shakker

Lora：Han's fashion wardrobe
　　　Photo to Oil Painting Style Portrait -lora-XL-pundit V1.0
　　　Two Dimensional Color Wind -lora-XL-pundit

Han's fashion wardrobe-LoRA-cowcat-Shakker

Photo transfer Oil painting Wind Portrait -lora-XL-pundit-LoRA-AIGC_JIAMO-Shakker

Two Dimensional Color Wind -lora-XL-pundit-LoRA-AIGC_JIAMO-Shakker

パラメーター等の設定

チェックポイント（モデル）・Clip skip

Lora・Sampler・Step・Hiresfix・解像度（Width、Height）

Hiresfixの設定値・CFG・Seed・Extra設定値

プロンプト

プロンプト

prompt:,
- quality: "(masterpiece, 4K, 8k, high resolution, ultra-detailed:1.2),(realistic, photorealistic, photo-realistic:1.37)",
- description: "A young and cute girl playing the violin",
- style: "Photorealistic,ultra-detailed,Leica M11-style focus and bokeh,daytime scene",
- details:,
- "Low-angle perspective with Tokyo Tower against a clear blue sky,captured with Leica M11-style focus and bokeh,unblurred and detailed",
- "Accurate details of hands and fingers",
- "Natural pose holding the violin",
- "Joyful and innocent expression,eyes sparkling,big smile with raised corners",
- "Ponytail hairstyle,hair moving slightly in the breeze,each strand detailed",
- "Eyes are bright and sparkling,with detailed depiction of the irises",
- "High-resolution details capturing skin,hair,eyes,and her joyful smile",
- "Sharp,clear contrast of Tokyo Tower and the blue sky",
- attire: "Skirt,black skirt,black shoes,high heels,socks,long-sleeve white shirt,black ascot tie",
- facial_expression: "Fully enjoying her performance,with eyes sparkling and a big,innocent smile",
- lighting: "Natural daylight softly illuminating her face and hair,creating a bright and warm atmosphere",
- color_tone: "Vibrant contrast of the blue sky with Tokyo Tower",
- model_pose: "Energetic,joyful pose while happily playing the violin",
colorful,YHFG,Oil painting style

太字はLoraのトリガーワード

ネガティブプロント

- negative_prompt:,
- "Eyes completely closed",
- "Unnatural smile",
- "Blurry hair",
- "Lack of hair detail",
- "Blurry eyes",
- "Tokyo Tower is blurry",
- "Tokyo Tower with lighting",
- "Awkward or distorted hands and fingers",
- "Incorrect finger positioning",
- "Blurry details on hands and fingers",
- "Extra limbs",
- "Unnatural posture",
- "Artificial lines or shadows",
- "Disproportionate features",
- "Oversaturated colors",
sketches,(worst quality:2),(low quality:2),(normal quality:2),lowers,normal quality,((monochrome)),((grayscale)),skin spots,acnes,skin blemishes,bad anatomy,DeepNegative,(fat:1.2),facing away,looking away,tilted head,bad anatomy,bad hands,text,error,missing fingers,extra digit,fewer digits,cropped,worst quality,low quality,normal quality,jpeg artifacts,signature,watermark,username,blurry,bad feet,cropped,poorly drawn hands,poorly drawn face,mutation,deformed,worst quality,low quality,normal quality,jpeg artifacts,signature,watermark,extra fingers,fewer digits,extra limbs,extra arms,extra legs,malformed limbs,fused fingers,too many fingers,long neck,cross-eyed,mutated hands,bad body,bad proportions,gross proportions,text,error,missing fingers,missing arms,missing legs,extra digit,extra arms,extra leg,extra foot,

太字はSD系で使われるネガティブワード

その他

• 画像生成に使っているパッケージ：Shakker.ai内で使用できる「A1111　Web UI」

• プロンプト作成：Chat GPT-4o

• HiresfixのDenoisingは上げすぎると生成された元画像から大きく変わるので「０.３」くらいがベスト

終わり

ここまで読んでくれた方、ありがとうございました。