You're looking at a specific version of this model. Jump to the model overview.

zf-kbot /qwen:90c24626

Input schema

The fields you can use to run this model with an API. If you don’t give a value for a field its default value will be used.

Field Type Default value Description
apikey
string
API 密钥
input_image
string
输入图片
prompt
string
修复破损、污渍、划痕
正向提示词
negative_prompt
string
负向提示词
megapixels
number
1
目标百万像素
image_width
integer
1280
图片宽度
image_height
integer
720
图片高度
enable_manual_image_size_switch
boolean
False
启用手动图片尺寸设置
enable_multi_image_edit
boolean
False
启用多图编辑模式
ref_image
string
参考图片(多图编辑模式时使用)
seed
integer
-1
随机种子,-1 为随机
steps
integer
4
生成步数
website_name
string
google.com
网站名称(用于文件名后缀)
image_filename
string
output
输出文件名(不含扩展名)
color_match_switch
boolean
False
启用颜色匹配
output_ext
None
.jpg
输出图片格式
quality
integer
90

Min: 1

Max: 100

输出图片质量 (1-100)
prompt_enhance_switch
boolean
False
启用提示词优化
visual_analysis_prompt
string
你是一个图像分析助手。请分析这张图片,输出以下 JSON: { "category": "portrait|product|landscape|illustration|screenshot|other", "style": "photorealistic|anime|3d_render|flat_design|mixed", "main_subject": "简短描述主体内容", "quality_score": 1-10, "issues": ["blur", "noise", "compression", "low_res", "overexposure", "underexposure"], "preserve": ["color_tone", "art_style", "composition", "lighting_mood"], "detail_areas": "需要重点增强细节的区域描述"
视觉分析提示词
prompt_enhance_prompt
string
你是一个图像增强提示词生成专家。你的任务是根据图像分析结果,生成精准的增强提示词。 ## 输入格式 你会收到一个 JSON,包含图像分析结果: - category: 图像类型 - style: 视觉风格 - main_subject: 主体内容 - quality_score: 质量评分 1-10 - issues: 质量问题列表 - preserve: 需要保留的特征 - detail_areas: 需要重点增强的区域 ## 策略规则 ### 问题修复策略 - blur → "restore sharpness and edge definition" - noise → "reduce grain while preserving natural texture" - compression → "remove compression artifacts, restore smooth gradients" - low_res → "synthesize fine details for higher resolution" - overexposure → "recover highlight details, balance exposure" - underexposure → "lift shadows, reveal hidden details" ### 类型专属策略 - portrait: 强调皮肤质感、毛发细节、眼神光、面部轮廓 - product: 强调材质还原、边缘锐利、光泽准确、背景干净 - landscape: 强调大气透视、远近层次、植被细节、天空渐变 - illustration: 强调线条锐利、色块干净、风格一致性 - screenshot: 强调文字清晰、UI元素锐利、去除摩尔纹 ### 风格保护策略 - photorealistic: 不要艺术化,保持真实感 - anime: 不要写实化,保持动漫风格线条和色彩 - 3d_render: 保持渲染质感,不要添加照片噪点 - flat_design: 保持扁平化,不要添加渐变和阴影 ## 输出要求 生成一段英文增强提示词,结构如下: 1. 开头:明确增强任务和主体 2. 修复指令:根据 issues 生成,最多 3 条核心修复 3. 细节增强:根据 detail_areas 指定重点区域 4. 风格锁定:根据 style 添加风格保护指令 5. 保留指令:根据 preserve 生成保护性约束 ## 约束 - 提示词长度:50-300 词 - 语言:英文 - 不要使用模糊词汇如 "better"、"improve"、"enhance quality" - 使用具体的技术描述 - quality_score < 4 时,修复优先;>= 7 时,细节增强优先 - 只输出提示词本身,不要解释 ## 示例 输入: { "category": "portrait", "style": "photorealistic", "main_subject": "女性半身照,室内自然光", "quality_score": 4, "issues": ["blur", "noise", "low_res"], "preserve": ["color_tone", "lighting_mood"], "detail_areas": "面部五官、头发丝、衣物纹理" } 输出: Upscale and restore this portrait photograph. Sharpen facial features and recover fine skin texture. Reduce luminance noise while maintaining natural skin pores. Synthesize realistic hair strand details and fabric weave patterns. Focus enhancement on eyes, lips, and clothing folds. Maintain the soft natural window lighting and warm color temperature. Do not alter the original composition or add artificial elements.
提示词优化提示词
detect_switch
boolean
False
启用检测
detect_params
string
640;0.7;0.7;0.5;0.7;0.5
检测参数配置

Output schema

The shape of the response you’ll get when you run this model with an API.

Schema
{'items': {'format': 'uri', 'type': 'string'},
 'title': 'Output',
 'type': 'array'}