Embedded Files
Create a stunning portrait sketch of a person with delicate features, rendered in graphite pencil medium, set against a soft, creamy white background that subtly gradates to a light gray towards the edges, evoking a sense of intimacy and subtlety, with bold, expressive lines that capture the subject's essence, and gentle shading that adds depth and volume to the overall piece, with the sketch itself occupying most of the composition, leaving a slim border around the edges to create a sense of breathability and space.
A hyper-realistic miniature scene featuring tiny construction workers dressed in white shirts, brown pants, and white helmets working on a stack of elegant beige macarons with a rich chocolate filling. Some workers are standing on ladders, carefully placing small chocolate chunks on top, while others carry ingredients and tools. The macarons are set on a white plate dusted with powdered sugar, surrounded by small caramelized crumbs. The background is softly blurred with a cool, moody blue tone, creating a stylish and artistic atmosphere."
訓練模型的使用
1.訓練自己的模型Training&Dataset。到Element>新增，選資料集MIXKuplus(28張相片)>Start Training, 三類選Character, 加入敘述，開始訓練。
2.完成後的模型，Type為Element (之前為Finetuned Model)。該如何使用，以Eements的方式加入自己的訓練模型。
3.搭配喜歡的圖片的提示詞(注意：使用英文詞更佳)，做必要修正，可生成接近自己面部的圖片。
4.範例 Model/Preset: Cinematic Kino; Style:Fashion; PhotoReal:ON Negative Prompt: ON
4.相較於提示詞搭配，Image Guidance 的 Character Reference；以及Legacy Mode(傳統介面)的Image Guidance 等上傳參考照片的方式 更為接近提供的相片集。 *單一上傳參考照，leonardo的生成結果不如預期。 **也就是訓練模型(使用ELEMENT)跟參考的差異。
5.Element的功能，帶入比例更多。Elemens強度可調，以1為基礎。Elements即LoRA。
6.可使用清楚好看的單張人物照(例如演藝人員與明星)，或多張照片訓練，測試結果。
7.Tensor.Art的訓練結果?
8.定義
LoRA:
FineTune:
Leonardo.ai 使用Element範例
[Not Leagcy Model]
Preset:Cinematic Kino, Element: ON, Preset: Fashion, Generation Mode: Quality
Advanced Settings= Model: Leonardo Kino XL, PhotoReal: ON
Drawing on whitCool badass futuristic man in armor, blend of medieval and cyberpunk styles, USA colors and themes , realistic, posing heroically, USA flag cape, snow, realistic, cinematic, depth of field, intricate details, holding sworde background, mountain , with a sunrise, realistic, ultra realistic, ultra HD
修改提示詞
[Not Leagcy Model]
Preset:Cinematic Kino, Element: ON, Preset: Fashion, Generation Mode: Quality
Advanced Settings= Model: Leonardo Kino XL, PhotoReal: ON
Based on whitCool's badass futuristic suit man, on a college campus, teaching data analysis courses, incorporating fashion style, rich colors and themes, bright scenes, bright lights, realistic photos, heroic poses, bright, high brightness, light colors, atmospheric scenes, depth of field, clean details, realistic, surreal, ultra-high definition.
Leonardo AI提示詞
Leonardo AI提示詞
Data visualization, art style, modern style, EXCEL, POWER BI, TABLEAU, infographics, ASIA COLLEGE YOUNG beautiful female, pictures used in teaching and commercial websites.
Data visualization, art style, modern style, EXCEL, POWER BI, TABLEAU, infographics, business people, pictures used in teaching and commercial websites.
Fashion design, textile technology, modern style, English lettering title "TEXTILES & CLOTHING", knitwear design, fiber science, infographics, young beautiful women in Asian college, pictures for educational and business websites.
Fashion Design, Textile Technology, modern style, 英文字母"Textiles and Clothing", Knitting Design, Fiber Science, infographics, ASIA COLLEGE YOUNG beautiful female, pictures used in teaching and commercial websites.
Model/Preset: Flux Schnell / Flux Dev
Style: Acrylic Prompt Enhance: Auto
Content Reference
Data visualization, art style, modern style, EXCEL, POWER BI, TABLEAU, infographics, ASIA COLLEGE YOUNG beautiful female, pictures used in teaching and commercial websites.
Page updated
Report abuse