AI角色设计参考表生成指南：使用inference.sh CLI创建多角度一致角色

character-design-sheet by inferen-sh/skills

6,800 周安装量

182 GitHub Stars

GitHub

安装命令

npx skills add https://github.com/inferen-sh/skills --skill character-design-sheet

AI/机器学习设计

🇨🇳中文介绍

角色设计参考表

通过 inference.sh CLI 在多张 AI 生成的图像中创建一致的角色。

快速开始

需要 inference.sh CLI (infsh)。安装说明

infsh login

# 生成角色概念
infsh app run falai/flux-dev-lora --input '{
  "prompt": "character design reference sheet, front view of a young woman with short red hair, green eyes, wearing a blue jacket and white t-shirt, full body, white background, clean lines, concept art style, character turnaround",
  "width": 1024,
  "height": 1024
}'

一致性问题

即使使用相同的提示词，AI 图像生成每次都会产生看起来不同的角色。这是任何需要在多张图像中使用同一角色的 AI 艺术项目中面临的首要挑战。

解决方案（按效果排序）

技术	一致性	工作量	最适合
（针对角色训练）

广告位招租

在这里展示您的产品或服务

触达数万 AI 开发者，精准高效

联系我们

1. 多角度展示表

从多个角度展示角色：

┌────────┬────────┬────────┬────────┐
│        │        │        │        │
│ FRONT  │  3/4   │  SIDE  │  BACK  │
│  VIEW  │  VIEW  │  VIEW  │  VIEW  │
│        │        │        │        │
└────────┴────────┴────────┴────────┘



# 生成正面视图
infsh app run falai/flux-dev-lora --input '{
  "prompt": "character design, front view, young woman with short asymmetric red hair, bright green eyes, wearing navy blue bomber jacket over white graphic tee, dark jeans, red sneakers, standing in neutral pose, full body, clean white background, concept art, sharp details",
  "width": 768,
  "height": 1024
}' --no-wait

# 生成 3/4 视图（相同描述）
infsh app run falai/flux-dev-lora --input '{
  "prompt": "character design, three-quarter view, young woman with short asymmetric red hair, bright green eyes, wearing navy blue bomber jacket over white graphic tee, dark jeans, red sneakers, standing, full body, clean white background, concept art, sharp details",
  "width": 768,
  "height": 1024
}' --no-wait

# 生成侧面视图
infsh app run falai/flux-dev-lora --input '{
  "prompt": "character design, side profile view, young woman with short asymmetric red hair, bright green eyes, wearing navy blue bomber jacket over white graphic tee, dark jeans, red sneakers, standing, full body, clean white background, concept art, sharp details",
  "width": 768,
  "height": 1024
}' --no-wait

# 生成背面视图
infsh app run falai/flux-dev-lora --input '{
  "prompt": "character design, back view, young woman with short asymmetric red hair, wearing navy blue bomber jacket over white graphic tee, dark jeans, red sneakers, standing, full body, clean white background, concept art, sharp details",
  "width": 768,
  "height": 1024
}' --no-wait

# 拼接成参考表
infsh app run infsh/stitch-images --input '{
  "images": ["front.png", "three-quarter.png", "side.png", "back.png"],
  "direction": "horizontal"
}'

展示角色带有不同情绪的面部：

┌────────┬────────┬────────┐
│NEUTRAL │ HAPPY  │ ANGRY  │
│        │        │        │
├────────┼────────┼────────┤
│  SAD   │SURPRISE│THINKING│
│        │        │        │
└────────┴────────┴────────┘

至少 6 种表情：中性、快乐、愤怒、悲伤、惊讶、思考。

# 中性
infsh app run falai/flux-dev-lora --input '{
  "prompt": "character portrait, close-up face, young woman with short red hair and green eyes, neutral calm expression, clean white background, concept art, consistent character design",
  "width": 512,
  "height": 512
}' --no-wait

# 快乐
infsh app run falai/flux-dev-lora --input '{
  "prompt": "character portrait, close-up face, young woman with short red hair and green eyes, warm genuine smile, happy expression, clean white background, concept art, consistent character design",
  "width": 512,
  "height": 512
}' --no-wait

# 愤怒
infsh app run falai/flux-dev-lora --input '{
  "prompt": "character portrait, close-up face, young woman with short red hair and green eyes, furrowed brows, angry determined expression, clean white background, concept art, consistent character design",
  "width": 512,
  "height": 512
}' --no-wait

# （继续生成悲伤、惊讶、思考等...）

同一角色的多套服装：

服装	描述
休闲装	飞行员夹克、T恤、牛仔裤
工作装	西装外套、衬衫、西裤
运动装	运动内衣、紧身裤、跑鞋
正装	晚礼服、高跟鞋

记录精确颜色以确保一致性：

CHARACTER: Maya Chen

Skin:    ████ #F5D0A9 (warm beige)
Hair:    ████ #C0392B (auburn red)
Eyes:    ████ #27AE60 (emerald green)
Jacket:  ████ #2C3E50 (navy blue)
T-shirt: ████ #ECF0F1 (off-white)
Jeans:   ████ #34495E (dark slate)
Shoes:   ████ #E74C3C (bright red)

最实用的保持一致性技术：编写一个 50+ 字的详细描述，并在每个提示词中完全重复使用它。

[age] [gender] with [hair: color, length, style], [eye color] eyes,
[skin tone], [facial features: any distinctive marks],
wearing [top: specific color and style], [bottom: specific color and style],
[shoes: specific color and style], [accessories: specific items]

young woman in her mid-twenties with short asymmetric auburn red hair
swept to the right side, bright emerald green eyes, light warm skin
with a small beauty mark below her left eye, wearing a fitted navy
blue bomber jacket with silver zipper over a white crew-neck t-shirt,
dark slate slim jeans, and bright red canvas sneakers, small silver
stud earrings

在每个提示词中都使用这个完全相同的文本块，只改变动作/姿势/场景。

风格	头身比	最适合
写实	7.5 : 1	电影、照片级真实感
英雄式	8 : 1	超级英雄、动作
动漫/漫画	5-6 : 1	日本动画风格
风格化	4-5 : 1	西方动画
Q版/超级变形	2-3 : 1	可爱、喜剧、吉祥物

在提示词中包含比例风格："realistic proportions" 对比 "anime style proportions" 对比 "chibi proportions"

使用 LoRA 确保一致性

对于需要同一角色大量图像的项目，训练一个 LoRA：

# 使用带有角色 LoRA 的 FLUX
infsh app run falai/flux-dev-lora --input '{
  "prompt": "maya_chen character, sitting at a cafe reading a book, warm afternoon light, candid photography style",
  "loras": [{"path": "path/to/maya-chen-lora.safetensors", "scale": 0.8}]
}'

LoRA 训练技巧：

需要 10-20 张角色的参考图像（风格一致）
训练特定的触发词（例如，"maya_chen"）
缩放比例 0.7-0.9 可在一致性与提示词灵活性之间取得平衡
缩放比例越低 = 创意自由度越高，越高 = 匹配越严格

常见一致性失败问题

问题	原因	缓解措施
发色漂移	模型每次对"红发"的理解不同	使用特定色调："auburn red #C0392B"
眼睛颜色变化	在生成过程中优先级低	在提示词中尽早提及眼睛颜色
服装不一致	模型创造性地填充细节	明确描述每件衣物
年龄偏移	年龄描述模糊	使用"mid-twenties"而非"young"
面部结构变化	不同生成批次 = 不同面孔	使用 LoRA 或相同种子基础
比例偏移	风格解释不同	指定"7.5 head proportions"

角色设定集模板

对于持续进行的项目，维护一个角色设定集文档：

# Character: Maya Chen

## 视觉描述（在所有提示词中使用）
young woman in her mid-twenties with short asymmetric auburn red hair...
[完整的 50+ 字锚点描述]

## 配色方案
- Skin: #F5D0A9
- Hair: #C0392B
- Eyes: #27AE60
- Primary outfit: Navy #2C3E50
- Accent: Red #E74C3C

## 性格笔记（用于表情/姿势选择）
- Confident but approachable
- Default expression: slight curious smile
- Gestures: talks with hands, leans forward when interested

## 风格关键词
concept art, clean lines, sharp details, [art style reference]

## LoRA（如果已训练）
Path: ./loras/maya-chen-v2.safetensors
Trigger: maya_chen
Recommended scale: 0.8

错误	问题	修复方法
描述模糊	每次都是不同的角色	使用 50+ 字的详细锚点
提示词结构不一致	重点不同 = 结果不同	保持相同结构，只改变动作/场景
只生成一个视图	无法在不同上下文中使用角色	创建完整的多角度参考
没有颜色记录	颜色在不同生成批次间漂移	记录确切的十六进制代码
跳过表情表	角色感觉单一	生成 6+ 种表情
大型项目不使用 LoRA	不一致性会累积	为 10+ 图像项目训练 LoRA

npx skills add inference-sh/skills@ai-image-generation
npx skills add inference-sh/skills@flux-image
npx skills add inference-sh/skills@prompt-engineering

浏览所有应用：infsh app list

🇺🇸English

Character Design Sheet

Create consistent characters across multiple AI-generated images via inference.sh CLI.

Quick Start

Requires inference.sh CLI (infsh). Install instructions

infsh login

# Generate a character concept
infsh app run falai/flux-dev-lora --input '{
  "prompt": "character design reference sheet, front view of a young woman with short red hair, green eyes, wearing a blue jacket and white t-shirt, full body, white background, clean lines, concept art style, character turnaround",
  "width": 1024,
  "height": 1024
}'

The Consistency Problem

AI image generation produces different-looking characters every time, even with the same prompt. This is the #1 challenge in AI art for any project requiring the same character across multiple images.

Solutions (Ranked by Effectiveness)

Technique	Consistency	Effort	Best For
FLUX LoRA (trained on character)	Very high	High (requires training data)	Ongoing projects, many images
Detailed description anchor	Medium-high	Low	Quick projects, few images
Same seed + similar prompt	Medium	Low	Variations of single pose
Image-to-image refinement	Medium	Medium	Refining existing images
Reference image in prompt	Varies	Low	When model supports it

Reference Sheet Types

1. Turnaround Sheet

Shows the character from multiple angles:

┌────────┬────────┬────────┬────────┐
│        │        │        │        │
│ FRONT  │  3/4   │  SIDE  │  BACK  │
│  VIEW  │  VIEW  │  VIEW  │  VIEW  │
│        │        │        │        │
└────────┴────────┴────────┴────────┘



# Generate front view
infsh app run falai/flux-dev-lora --input '{
  "prompt": "character design, front view, young woman with short asymmetric red hair, bright green eyes, wearing navy blue bomber jacket over white graphic tee, dark jeans, red sneakers, standing in neutral pose, full body, clean white background, concept art, sharp details",
  "width": 768,
  "height": 1024
}' --no-wait

# Generate 3/4 view (same description)
infsh app run falai/flux-dev-lora --input '{
  "prompt": "character design, three-quarter view, young woman with short asymmetric red hair, bright green eyes, wearing navy blue bomber jacket over white graphic tee, dark jeans, red sneakers, standing, full body, clean white background, concept art, sharp details",
  "width": 768,
  "height": 1024
}' --no-wait

# Generate side view
infsh app run falai/flux-dev-lora --input '{
  "prompt": "character design, side profile view, young woman with short asymmetric red hair, bright green eyes, wearing navy blue bomber jacket over white graphic tee, dark jeans, red sneakers, standing, full body, clean white background, concept art, sharp details",
  "width": 768,
  "height": 1024
}' --no-wait

# Generate back view
infsh app run falai/flux-dev-lora --input '{
  "prompt": "character design, back view, young woman with short asymmetric red hair, wearing navy blue bomber jacket over white graphic tee, dark jeans, red sneakers, standing, full body, clean white background, concept art, sharp details",
  "width": 768,
  "height": 1024
}' --no-wait

# Stitch into reference sheet
infsh app run infsh/stitch-images --input '{
  "images": ["front.png", "three-quarter.png", "side.png", "back.png"],
  "direction": "horizontal"
}'

2. Expression Sheet

Shows the character's face with different emotions:

┌────────┬────────┬────────┐
│NEUTRAL │ HAPPY  │ ANGRY  │
│        │        │        │
├────────┼────────┼────────┤
│  SAD   │SURPRISE│THINKING│
│        │        │        │
└────────┴────────┴────────┘

Minimum 6 expressions: neutral, happy, angry, sad, surprised, thinking.

# Neutral
infsh app run falai/flux-dev-lora --input '{
  "prompt": "character portrait, close-up face, young woman with short red hair and green eyes, neutral calm expression, clean white background, concept art, consistent character design",
  "width": 512,
  "height": 512
}' --no-wait

# Happy
infsh app run falai/flux-dev-lora --input '{
  "prompt": "character portrait, close-up face, young woman with short red hair and green eyes, warm genuine smile, happy expression, clean white background, concept art, consistent character design",
  "width": 512,
  "height": 512
}' --no-wait

# Angry
infsh app run falai/flux-dev-lora --input '{
  "prompt": "character portrait, close-up face, young woman with short red hair and green eyes, furrowed brows, angry determined expression, clean white background, concept art, consistent character design",
  "width": 512,
  "height": 512
}' --no-wait

# (Continue for sad, surprised, thinking...)

3. Outfit/Costume Sheet

Multiple outfits for the same character:

Outfit	Description
Casual	Bomber jacket, t-shirt, jeans
Work	Blazer, button-down, slacks
Athletic	Sports bra, leggings, running shoes
Formal	Evening dress, heels

4. Color Palette Sheet

Document exact colors for consistency:

CHARACTER: Maya Chen

Skin:    ████ #F5D0A9 (warm beige)
Hair:    ████ #C0392B (auburn red)
Eyes:    ████ #27AE60 (emerald green)
Jacket:  ████ #2C3E50 (navy blue)
T-shirt: ████ #ECF0F1 (off-white)
Jeans:   ████ #34495E (dark slate)
Shoes:   ████ #E74C3C (bright red)

The Description Anchor Technique

The most practical consistency technique: write a 50+ word detailed description and reuse it exactly in every prompt.

Template

[age] [gender] with [hair: color, length, style], [eye color] eyes,
[skin tone], [facial features: any distinctive marks],
wearing [top: specific color and style], [bottom: specific color and style],
[shoes: specific color and style], [accessories: specific items]

Example

young woman in her mid-twenties with short asymmetric auburn red hair
swept to the right side, bright emerald green eyes, light warm skin
with a small beauty mark below her left eye, wearing a fitted navy
blue bomber jacket with silver zipper over a white crew-neck t-shirt,
dark slate slim jeans, and bright red canvas sneakers, small silver
stud earrings

Use this exact block in EVERY prompt for this character, only changing the action/pose/scene.

Proportion Guide

Style	Head-to-Body Ratio	Best For
Realistic	7.5 : 1	Film, photorealistic
Heroic	8 : 1	Superheroes, action
Anime/Manga	5-6 : 1	Japanese animation style
Stylized	4-5 : 1	Western animation
Chibi/Super-deformed	2-3 : 1	Cute, comedic, mascots

Include proportion style in your prompts: "realistic proportions" vs "anime style proportions" vs "chibi proportions"

Using LoRA for Consistency

For projects requiring many images of the same character, train a LoRA:

# Use FLUX with a character LoRA
infsh app run falai/flux-dev-lora --input '{
  "prompt": "maya_chen character, sitting at a cafe reading a book, warm afternoon light, candid photography style",
  "loras": [{"path": "path/to/maya-chen-lora.safetensors", "scale": 0.8}]
}'

LoRA Training Tips:

Need 10-20 reference images of the character (consistent style)
Train on specific trigger word (e.g., "maya_chen")
Scale 0.7-0.9 balances consistency with prompt flexibility
Lower scale = more creative freedom, higher = more strict matching

Common Consistency Failures

Issue	Why It Happens	Mitigation
Hair color drift	Model interprets "red hair" differently each time	Use specific shade: "auburn red #C0392B"
Eye color change	Low priority in generation	Mention eye color early in prompt
Outfit inconsistency	Model fills in details creatively	Describe every clothing item explicitly
Age shift	Vague age description	Use "mid-twenties" not "young"
Face structure change	Different generations = different faces	Use LoRA or same seed base
Proportion shift	Style interpretation varies	Specify "7.5 head proportions"

Character Bible Template

For ongoing projects, maintain a character bible document:

# Character: Maya Chen

## Visual Description (use in all prompts)
young woman in her mid-twenties with short asymmetric auburn red hair...
[full 50+ word anchor description]

## Color Palette
- Skin: #F5D0A9
- Hair: #C0392B
- Eyes: #27AE60
- Primary outfit: Navy #2C3E50
- Accent: Red #E74C3C

## Personality Notes (for expression/pose choices)
- Confident but approachable
- Default expression: slight curious smile
- Gestures: talks with hands, leans forward when interested

## Style Keywords
concept art, clean lines, sharp details, [art style reference]

## LoRA (if trained)
Path: ./loras/maya-chen-v2.safetensors
Trigger: maya_chen
Recommended scale: 0.8

Common Mistakes

Mistake	Problem	Fix
Vague descriptions	Different character every time	50+ word detailed anchor
Inconsistent prompt structure	Varying emphasis = varying results	Same structure, only change action/scene
Generating one view only	Can't use character in different contexts	Create full turnaround reference
No color documentation	Colors drift across generations	Record exact hex codes
Skipping expression sheet	Character feels one-dimensional	Generate 6+ expressions
Not using LoRA for big projects	Inconsistency compounds	Train LoRA for 10+ image projects

Related Skills

npx skills add inference-sh/skills@ai-image-generation
npx skills add inference-sh/skills@flux-image
npx skills add inference-sh/skills@prompt-engineering

Browse all apps: infsh app list

Weekly Installs

6.8K

Repository

inferen-sh/skills

GitHub Stars

182

First Seen

12 days ago

Security Audits

Gen Agent Trust HubPass SocketWarn SnykPass

Installed on

claude-code5.5K

gemini-cli4.9K

codex4.9K

amp4.9K

opencode4.9K

github-copilot4.9K