Stable Diffusion - 幻想 (Fantasy) 风格与糖果世界 (Candy Land) 的人物图像提示词配置

欢迎关注我的CSDN:https://spike.blog.csdn.net/

本文地址:https://spike.blog.csdn.net/article/details/132212193

图像由 DreamShaper8 模型生成,融合糖果世界。

幻想 (Fantasy) 风格图像是一种以想象力为主导的艺术形式,创造了一些超越现实的场景、角色、物品和故事,通常具有神秘、奇幻、魔幻或科幻的色彩。图像特点,如下:

  • 主题多样:涵盖各种主题,如古代神话、中世纪传奇、东方神话、未来科技、异世界探险等,只要有想象力,就可以创造出无限的可能性。
  • 色彩丰富:使用各种色彩来营造不同的氛围和情感,如明亮的色彩可以表现欢乐和活力,暗淡的色彩可以表现悲伤和沉重,鲜艳的色彩可以表现奇异和惊艳等。
  • 形式多变:采用不同的形式来展现想象力,如写实的形式可以表现细节和真实感,抽象的形式可以表现概念和意境,卡通的形式可以表现夸张和幽默等。
  • 风格独特:有不同的风格来区别于其他类型的艺术,如霓虹风可以表现科幻和未来感,哥特风可以表现黑暗和恐怖感,水彩风可以表现柔和和浪漫感等。

参考:Candy Land (糖果世界) LoRA 提示词配置与效果展示

1. 提示词

幻想风格 与 糖果世界 (Candy Land) 融合,正向提示词:

bash 复制代码
fantasy,dynamic,shiny,details,perfect,beautiful,dreamy,heart,colorful,vivid,
candyland,candy print,candy clothing,rainbow,syrup,cookie,good figure proportions,chocolate bar,glitter,various colors,gradients,dripping,glaze,beads,
(hyperrealistic:0.2),(bloom:0.2),
1girl,solo,full body,pose,beautiful woman,looking at viewer,depth of field,sharp focus,
(thick wrap hip colorful pantyhose),detailed pantyhose,detailed shoes,
updo,curls,long hair,ponytail,smile,naughty face,purple eyes,
frilled skirt,center frills,heels,long sleeves,(colorful turtleneck),
head tilt,arms up,(relax sitting),knees separation,
good anatomy,(good figure proportions),good pose,(perfect legs),(pair shoes:1.3),
(nice hand),nice figure,
sunshine bright light,outdoors,full background,candy's fairy tale castle,
[(colorful explosion psychedelic paint colors:1.2)::0.05],
(masterpiece, best quality:1.2),highly detailed,extremely detailed,8k,
<lora:more_details:0.4>,<lora:candyland-10:0.85>,

具体分析:

  • fantasy,dynamic,shiny,details,perfect,beautiful,dreamy,heart,colorful,vivid,,幻想风格
  • candyland,candy print,candy clothing,rainbow,syrup,cookie,macaron,chocolate bar,glitter,various colors,gradients,dripping,glaze,beads,,糖果世界
  • (hyperrealistic:0.2),(bloom:0.2),,糖果的辅助,权重较低
  • 1girl,solo,full body,pose,beautiful woman,looking at viewer,depth of field,sharp focus,,构图
  • thick wrap hip colorful pantyhose,detailed pantyhose,detailed shoes,,袜子
  • updo,curls,long hair,ponytail,smile,naughty face,purple eyes,,脸部
  • frilled skirt,center frills,heels,long sleeves,turtleneck,,服装,long sleeves,turtleneck,避免过于暴漏
  • head tilt,arms up,(relax sitting),knees separation,,坐姿
  • good anatomy,good proportions,good pose,perfect legs,(pair shoes,pair legs:1.3),,坐姿质量,(pair shoes,pair legs:1.3), 有效解决腿部的异常问题,尤其是 RevAnimated 模型
  • nice hand,nice figure,,与负向结合,手部质量
  • sunshine bright light,outdoors,full background,candy's fairy tale castle,,室外风格,注意,背景提示词排序在后,避免影响主题,室内室外均可选择
  • [(colorful explosion psychedelic paint colors:1.2)::0.05],,颜色扩散
  • (masterpiece, best quality:1.2),highly detailed,extremely detailed,8k,,标准质量
  • <lora:more_details:0.4>,<lora:candyland-10:0.85>,,糖果世界与细节丰富的LoRA,注意 more_details 权重较低,避免坐姿肢体异常

负向提示词:

bash 复制代码
(negative_hand,ng_deepnegative_v1_75t:1.3),(negative_hand),(badhandv4),
bad anatomy,disfigure,(deformed limbs and arms and legs),(strange hands and fingers),toes,
cropped,border,(out of frame:1.1),bad composition,blurry,debris,messy,logo,text,
spikes,sepia,clutter,monochrome,greyscale,oversaturated,
taut,ugly,skin tight,dirt face,
photographic,realistic,realism,35mm film,dslr,cropped,frame,text,deformed,glitch,noise,noisy,off-center,deformed,cross-eyed,closed eyes,bad anatomy,ugly,disfigured,sloppy,duplicate,mutated,black and white,
(nsfw),breast,cleavage,
(worst quality, bad quality:1.1),jpeg artifacts,lowres,normal quality,low quality,
EasyNegative,

具体分析:

  • (negative_hand,ng_deepnegative_v1_75t:1.3),(negative_hand),(badhandv4),,负向质量 Embeddings,避免 RevAnimated 的手部较差
  • bad anatomy,disfigure,(deformed limbs and arms and legs),(strange hands and fingers),toes,,肢体质量
  • cropped,border,(out of frame:1.1),bad composition,blurry,debris,messy,logo,text,,构图质量
  • spikes,sepia,clutter,monochrome,greyscale,oversaturated,,色彩质量
  • taut,ugly,skin tight,dirt face,,皮肤质量
  • photographic,realistic,realism,35mm film,dslr,cropped,frame,text,deformed,glitch,noise,noisy,off-center,deformed,cross-eyed,closed eyes,bad anatomy,ugly,disfigured,sloppy,duplicate,mutated,black and white,,幻想风格
  • (nsfw),breast,cleavage,,安全,避免 RevAnimated 衣着较少
  • (worst quality, bad quality:1.1),jpeg artifacts,lowres,normal quality,low quality,,标准质量
  • EasyNegative,,提示词补助 Embeddings

使用 ADetailer 脸部修复,关键词:

bash 复制代码
asian girl, make up, beautiful face,

其他配置如下:

2. 图像效果

测试 DreamShaper 8 模型,支持风格较多,室外 (Outdoors) 背景,图像如下:

室内 (Indoors) 背景,只修改提示词 outdoorsindoors,图像如下:

图像示例的提示词与配置,如下:

bash 复制代码
fantasy,dynamic,shiny,details,perfect,beautiful,dreamy,heart,colorful,vivid,
candyland,candy print,candy clothing,rainbow,syrup,cookie,macaron,chocolate bar,glitter,various colors,gradients,dripping,glaze,beads,
(hyperrealistic:0.2),(bloom:0.2),
sunshine bright light,outdoors,
full background,candy's fairy tale castle,
1girl,solo,full body,pose,beautiful woman,looking at viewer,depth of field,sharp focus,
thick wrap hip colorful pantyhose,detailed pantyhose,detailed shoes,
updo,curls,long hair,ponytail,smile,naughty face,purple eyes,
frilled skirt,center frills,heels,long sleeves,turtleneck,
head tilt,arms up,(relax sitting),knees separation,
good anatomy,good proportions,good pose,perfect legs,(pair shoes,pair legs:1.2),
nice hand,nice figure,
[(colorful explosion psychedelic paint colors:1.2)::0.05],
(masterpiece, best quality:1.2),highly detailed,extremely detailed,8k,
<lora:more_details:0.4>,<lora:candyland-10:0.85>,
Negative prompt: (negative_hand,ng_deepnegative_v1_75t:1.3),badhandv4,
bad anatomy,disfigure,(deformed limbs and arms and legs),(strange hands and fingers),toes,
cropped,border,(out of frame:1.1),bad composition,blurry,debris,messy,logo,text,
spikes,sepia,clutter,monochrome,greyscale,oversaturated,
taut,ugly,skin tight,dirt face,
photographic,realistic,realism,35mm film,dslr,cropped,frame,text,deformed,glitch,noise,noisy,off-center,deformed,cross-eyed,closed eyes,bad anatomy,ugly,disfigured,sloppy,duplicate,mutated,black and white,
(worst quality, bad quality:1.1),jpeg artifacts,lowres,normal quality,low quality,
EasyNegative,
Steps: 30, Sampler: DPM++ 2M SDE Karras, CFG scale: 5, Seed: 1881339644, Size: 512x768, Model hash: 879db523c3, Model: Dreamshaper_8, Denoising strength: 0.35, Clip skip: 2, ADetailer model: face_yolov8n.pt, ADetailer prompt: "asian girl, make up, beautiful face,", ADetailer confidence: 0.3, ADetailer dilate/erode: 4, ADetailer mask blur: 4, ADetailer denoising strength: 0.4, ADetailer inpaint only masked: True, ADetailer inpaint padding: 0, ADetailer ControlNet model: control_v11p_sd15_inpaint [ebff9138], ADetailer ControlNet module: inpaint_global_harmonious, ADetailer version: 23.7.6, Hires upscale: 2, Hires steps: 5, Hires upscaler: 4x-UltraSharp, Lora hashes: "more_details: 3b8aa1d351ef, candyland-10: 80249b96f2ea", TI hashes: "negative_hand: 73b524a2da12, ng_deepnegative_v1_75t: 54e7e4826d53, badhandv4: 5e40d722fc3d, EasyNegative: 66a7279a88dd", ControlNet 0: "preprocessor: inpaint_global_harmonious, model: control_v11p_sd15_inpaint [ebff9138], weight: 1.0, starting/ending: (0.0, 1.0), resize mode: ResizeMode.INNER_FIT, pixel perfect: True, control mode: ControlMode.BALANCED, preprocessor params: (-1, -1, -1)", Version: v1.5.1

测试 RevAnimated_v122 模型,更偏动画,整体丰富度优于 DreamShaper8,提示词依赖较强,效果不稳定,手部和腿部异常较多,无法做到全部正确。

相关推荐
空白诗3 天前
CANN ops-nn 算子解读:Stable Diffusion 图像生成中的 Conv2D 卷积实现
深度学习·计算机视觉·stable diffusion
学易3 天前
第十五节.别人的工作流,如何使用和调试(上)?(2类必现报错/缺失节点/缺失模型/思路/实操/通用调试步骤)
人工智能·ai作画·stable diffusion·报错·comfyui·缺失节点
心疼你的一切3 天前
基于CANN仓库算力手把手实现Stable Diffusion图像生成(附完整代码+流程图)
数据仓库·深度学习·stable diffusion·aigc·流程图·cann
Niuguangshuo4 天前
DALL-E 3:如何通过重构“文本描述“革新图像生成
人工智能·深度学习·计算机视觉·stable diffusion·重构·transformer
Niuguangshuo4 天前
深入解析 Stable Diffusion XL(SDXL):改进潜在扩散模型,高分辨率合成突破
stable diffusion
Niuguangshuo4 天前
深入解析Stable Diffusion基石——潜在扩散模型(LDMs)
人工智能·计算机视觉·stable diffusion
迈火4 天前
SD - Latent - Interposer:解锁Stable Diffusion潜在空间的创意工具
人工智能·gpt·计算机视觉·stable diffusion·aigc·语音识别·midjourney
迈火12 天前
Facerestore CF (Code Former):ComfyUI人脸修复的卓越解决方案
人工智能·gpt·计算机视觉·stable diffusion·aigc·语音识别·midjourney
重启编程之路13 天前
Stable Diffusion 参数记录
stable diffusion
孤狼warrior16 天前
图像生成 Stable Diffusion模型架构介绍及使用代码 附数据集批量获取
人工智能·python·深度学习·stable diffusion·cnn·transformer·stablediffusion