Spaces:
Running
on
Zero
Running
on
Zero
Update README.md
Browse files
README.md
CHANGED
@@ -1,202 +1,56 @@
|
|
1 |
---
|
2 |
-
title:
|
3 |
emoji: 🥸
|
4 |
colorFrom: purple
|
5 |
colorTo: yellow
|
6 |
sdk: gradio
|
7 |
-
sdk_version: 5.
|
8 |
app_file: app.py
|
9 |
pinned: false
|
10 |
license: apache-2.0
|
11 |
---
|
12 |
-
|
13 |
-
|
14 |
-
|
15 |
-
|
16 |
-
|
17 |
-
|
18 |
-
|
19 |
-
|
20 |
-
|
21 |
-
|
22 |
-
|
23 |
-
|
24 |
-
-
|
25 |
-
|
26 |
-
|
27 |
-
|
28 |
-
|
29 |
-
|
30 |
-
|
31 |
-
|
32 |
-
|
33 |
-
|
34 |
-
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
45 |
-
|
46 |
-
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
|
51 |
-
|
52 |
-
|
53 |
-
-
|
54 |
-
-
|
55 |
-
|
56 |
-
|
57 |
-
|
58 |
-
- **Prompt**: Describe the desired scene, style, and atmosphere
|
59 |
-
- **Reference Image**: Upload a clear face photo for feature extraction
|
60 |
-
- **Negative Prompt**: Specify elements to avoid (default: "low quality")
|
61 |
-
- **Seed**: Control randomness (0-2147483647)
|
62 |
-
- **Guidance Scale**: Balance between prompt adherence and creativity (0-10)
|
63 |
-
- **Inference Steps**: Quality vs speed trade-off (10-50 steps)
|
64 |
-
|
65 |
-
### Technical Specifications
|
66 |
-
|
67 |
-
- **Base Model**: Kwai-Kolors/Kolors
|
68 |
-
- **Face Adapter**: Kolors-IP-Adapter-FaceID-Plus
|
69 |
-
- **Face Analysis**: InsightFace AntelopeV2
|
70 |
-
- **Image Encoder**: CLIP-ViT-Large-Patch14-336
|
71 |
-
- **Framework**: Diffusers, PyTorch, Gradio
|
72 |
-
- **Precision**: float16 for optimal performance
|
73 |
-
- **Resolution**: 1024x1024 output images
|
74 |
-
|
75 |
-
### Why Choose Kolors IP-Adapter FaceID Plus?
|
76 |
-
|
77 |
-
1. **Superior Face Preservation**: Maintains facial identity while allowing creative freedom
|
78 |
-
2. **Professional Quality**: Gallery-ready portraits with consistent results
|
79 |
-
3. **Easy to Use**: Simple interface with powerful underlying technology
|
80 |
-
4. **Versatile Styling**: From photorealistic to highly stylized outputs
|
81 |
-
5. **Fast Generation**: Optimized pipeline for quick results
|
82 |
-
6. **Open Technology**: Built on open-source foundations
|
83 |
-
|
84 |
-
### Best Practices for Portrait Generation
|
85 |
-
|
86 |
-
- **Clear Reference Photos**: Use well-lit, front-facing images for best results
|
87 |
-
- **Detailed Prompts**: Describe style, mood, lighting, and environment
|
88 |
-
- **Experiment with Seeds**: Try different seeds for varied interpretations
|
89 |
-
- **Adjust Guidance Scale**: Higher values follow prompts more strictly
|
90 |
-
- **Use Negative Prompts**: Exclude unwanted elements like "blurry, distorted"
|
91 |
-
- **Multiple Attempts**: Generate several versions to find the perfect result
|
92 |
-
|
93 |
-
### Privacy and Ethical Considerations
|
94 |
-
|
95 |
-
- Face data is processed locally during generation
|
96 |
-
- No permanent storage of facial information
|
97 |
-
- Use only with consent for any uploaded images
|
98 |
-
- Respect portrait rights and privacy laws
|
99 |
-
- Avoid creating misleading or harmful content
|
100 |
-
|
101 |
-
### Get Started with AI Portraits
|
102 |
-
|
103 |
-
Transform your photos into stunning AI artwork today. Whether creating personal portraits, character designs, or exploring artistic concepts, Kolors IP-Adapter FaceID Plus provides the tools for professional-quality results.
|
104 |
-
|
105 |
-
**Community**: [Discord - Openfree AI](https://discord.gg/openfreeai) | **More AI Services**: [OpenFree Best AI](https://huggingface.co/spaces/openfree/Best-AI)
|
106 |
-
|
107 |
-
---
|
108 |
-
|
109 |
-
## 🎨 Kolors IP-어댑터 FaceID Plus: 고급 AI 초상화 생성
|
110 |
-
|
111 |
-
### 최첨단 기술로 얼굴을 놀라운 AI 예술작품으로 변환
|
112 |
-
|
113 |
-
**Kolors IP-어댑터 FaceID Plus**에 오신 것을 환영합니다. 얼굴 인식 기술과 고급 확산 모델을 결합한 혁신적인 AI 초상화 생성 시스템입니다. 이 강력한 도구는 참조 사진의 얼굴 특징을 창의적인 텍스트 프롬프트와 원활하게 혼합하여 고품질의 개인화된 이미지를 생성하며, 디지털 아트와 초상화 사진에 새로운 가능성을 열어줍니다.
|
114 |
-
|
115 |
-
### Kolors IP-어댑터 FaceID Plus란?
|
116 |
-
|
117 |
-
Kolors IP-어댑터 FaceID Plus는 Kwai-Kolors의 최첨단 모델과 InsightFace 기술을 활용하는 **고급 AI 이미지 생성 시스템**입니다. 입력 이미지에서 얼굴 특징을 추출하고 텍스트 설명을 기반으로 AI 생성 예술작품에 통합하여 개인화된 초상화를 만드는 데 특화되어 있습니다.
|
118 |
-
|
119 |
-
### 전문 초상화 생성을 위한 주요 기능
|
120 |
-
|
121 |
-
- **🎯 얼굴 인식 통합**: 정확한 얼굴 특징 추출을 위한 고급 InsightFace 기술
|
122 |
-
- **🖼️ 고해상도 출력**: 놀라운 1024x1024 픽셀 초상화 생성
|
123 |
-
- **🎨 Kolors 모델**: Kwai-Kolors의 고급 확산 파이프라인 구동
|
124 |
-
- **👤 얼굴 충실도 제어**: 조정 가능한 얼굴 유사성 강도 (0.8 스케일)
|
125 |
-
- **🔧 CLIP 비전 인코더**: 정교한 얼굴 특징 이해
|
126 |
-
- **⚡ GPU 가속**: CUDA 지원 장치에 최적화
|
127 |
-
- **🎲 시드 제어**: 수동 또는 무작위 시드로 재현 가능한 결과
|
128 |
-
- **📝 네거티브 프롬프트**: 원하지 않는 요소를 제외하여 출력 미세 조정
|
129 |
-
|
130 |
-
### 작동 방식
|
131 |
-
|
132 |
-
#### **지능형 얼굴 처리 파이프라인**
|
133 |
-
1. **얼굴 감지**: AntelopeV2 모델이 얼굴 특징을 식별하고 추출
|
134 |
-
2. **얼굴 임베딩**: 얼굴 특성을 수치 표현으로 변환
|
135 |
-
3. **특징 통합**: 얼굴 임베딩을 텍스트 프롬프트와 혼합
|
136 |
-
4. **이미지 생성**: 확산 기술을 사용하여 개인화된 초상화 생성
|
137 |
-
|
138 |
-
#### **고급 기술 구성 요소**
|
139 |
-
- **ChatGLM 텍스트 인코더**: 정교한 언어 이해
|
140 |
-
- **UNet2D 조건부 모델**: 고품질 이미지 생성
|
141 |
-
- **AutoencoderKL VAE**: 효율적인 잠재 공간 처리
|
142 |
-
- **Euler 이산 스케줄러**: 최적화된 샘플링 전략
|
143 |
-
|
144 |
-
### 완벽한 응용 분야
|
145 |
-
|
146 |
-
- **디지털 초상화**: 사진의 예술적 해석 생성
|
147 |
-
- **캐릭터 디자인**: 이야기나 게임을 위한 일관된 캐릭터 생성
|
148 |
-
- **아바타 생성**: 개인화된 프로필 사진 및 디지털 신원
|
149 |
-
- **패션 시각화**: 가상으로 다양한 스타일과 룩 시도
|
150 |
-
- **역사적 재해석**: 다른 시대나 스타일에 얼굴 배치
|
151 |
-
- **창의적 사진**: 예술적 초상화 개념 탐구
|
152 |
-
|
153 |
-
### 생성 매개변수
|
154 |
-
|
155 |
-
- **프롬프트**: 원하는 장면, 스타일 및 분위기 설명
|
156 |
-
- **참조 이미지**: 특징 추출을 위한 선명한 얼굴 사진 업로드
|
157 |
-
- **네거티브 프롬프트**: 피할 요소 지정 (기본값: "low quality")
|
158 |
-
- **시드**: 무작위성 제어 (0-2147483647)
|
159 |
-
- **가이던스 스케일**: 프롬프트 준수와 창의성 간의 균형 (0-10)
|
160 |
-
- **추론 단계**: 품질 대 속도 균형 (10-50단계)
|
161 |
-
|
162 |
-
### 기술 사양
|
163 |
-
|
164 |
-
- **기본 모델**: Kwai-Kolors/Kolors
|
165 |
-
- **얼굴 어댑터**: Kolors-IP-Adapter-FaceID-Plus
|
166 |
-
- **얼굴 분석**: InsightFace AntelopeV2
|
167 |
-
- **이미지 인코더**: CLIP-ViT-Large-Patch14-336
|
168 |
-
- **프레임워크**: Diffusers, PyTorch, Gradio
|
169 |
-
- **정밀도**: 최적 성능을 위한 float16
|
170 |
-
- **해상도**: 1024x1024 출력 이미지
|
171 |
-
|
172 |
-
### Kolors IP-어댑터 FaceID Plus를 선택해야 하는 이유
|
173 |
-
|
174 |
-
1. **우수한 얼굴 보존**: 창의적 자유를 허용하면서 얼굴 정체성 유지
|
175 |
-
2. **전문가 품질**: 일관된 결과로 갤러리 준비 초상화
|
176 |
-
3. **사용하기 쉬움**: 강력한 기본 기술을 갖춘 간단한 인터페이스
|
177 |
-
4. **다양한 스타일링**: 사실적인 것부터 고도로 양식화된 출력까지
|
178 |
-
5. **빠른 생성**: 빠른 결과를 위한 최적화된 파이프라인
|
179 |
-
6. **개방형 기술**: 오픈소스 기반으로 구축
|
180 |
-
|
181 |
-
### 초상화 생성 모범 사례
|
182 |
-
|
183 |
-
- **선명한 참조 사진**: 최상의 결과를 위해 조명이 좋고 정면을 향한 이미지 사용
|
184 |
-
- **상세한 프롬프트**: 스타일, 분위기, 조명 및 환경 설명
|
185 |
-
- **시드 실험**: 다양한 해석을 위해 다른 시드 시도
|
186 |
-
- **가이던스 스케일 조정**: 높은 값은 프롬프트를 더 엄격하게 따름
|
187 |
-
- **네거티브 프롬프트 사용**: "blurry, distorted"와 같은 원하지 않는 요소 제외
|
188 |
-
- **여러 시도**: 완벽한 결과를 찾기 위해 여러 버전 생성
|
189 |
-
|
190 |
-
### 개인정보 보호 및 윤리적 고려사항
|
191 |
-
|
192 |
-
- 얼굴 데이터는 생성 중에 로컬로 처리됨
|
193 |
-
- 얼굴 정보의 영구 저장 없음
|
194 |
-
- 업로드된 이미지에 대해서만 동의하에 사용
|
195 |
-
- 초상권 및 개인정보 보호법 준수
|
196 |
-
- 오해의 소지가 있거나 해로운 콘텐츠 생성 금지
|
197 |
-
|
198 |
-
### AI 초상화 시작하기
|
199 |
-
|
200 |
-
오늘 사진을 놀라운 AI 예술작품으로 변환하세요. 개인 초상화, 캐릭터 디자인 또는 예술적 개념 탐구 등 무엇을 만들든 Kolors IP-어댑터 FaceID Plus는 전문가 수준의 결과를 위한 도구를 제공합니다.
|
201 |
-
|
202 |
-
**커뮤니티**: [Discord - Openfree AI](https://discord.gg/openfreeai) | **더 많은 AI 서비스**: [OpenFree Best AI](https://huggingface.co/spaces/openfree/Best-AI)
|
|
|
1 |
---
|
2 |
+
title: Open Nano Banana
|
3 |
emoji: 🥸
|
4 |
colorFrom: purple
|
5 |
colorTo: yellow
|
6 |
sdk: gradio
|
7 |
+
sdk_version: 5.43.1
|
8 |
app_file: app.py
|
9 |
pinned: false
|
10 |
license: apache-2.0
|
11 |
---
|
12 |
+
🎨 Open Nano-Banana: Revolution in Ultra-Fast AI Image Editing!
|
13 |
+
|
14 |
+
🚀 Introduction
|
15 |
+
**Open Nano-Banana** is an innovative image editing tool based on the Qwen-Image-Edit model. Experience amazing quality image editing in just 8 steps!
|
16 |
+
|
17 |
+
Heartsync/Nano-Banana
|
18 |
+
|
19 |
+
|
20 |
+
✨ Core Features
|
21 |
+
|
22 |
+
⚡ Lightning-Fast Editing
|
23 |
+
* **8-Step Generation**: Ultra-fast processing with Qwen-Image-Lightning LoRA
|
24 |
+
* **Real-time Editing**: 10x faster than conventional methods
|
25 |
+
* **GPU Optimization**: Maximized memory efficiency with xformers
|
26 |
+
|
27 |
+
🤖 AI Prompt Enhancement
|
28 |
+
* **Automatic Prompt Improvement**: Intelligent rewriting with Cerebras' Qwen3-235B model
|
29 |
+
* **Multilingual Support**: Auto-detection for Korean/Chinese/English
|
30 |
+
* **Context Understanding**: Sophisticated command generation aligned with image context
|
31 |
+
|
32 |
+
🎯 Versatile Editing Functions
|
33 |
+
✅ Add/Delete/Replace objects
|
34 |
+
✅ Text editing and style transformation
|
35 |
+
✅ Person editing (expressions, hairstyles)
|
36 |
+
✅ Vintage restoration and style conversion
|
37 |
+
✅ Background replacement and enhancement
|
38 |
+
|
39 |
+
🛠️ Tech Stack
|
40 |
+
* Base Model: Qwen-Image-Edit
|
41 |
+
* Acceleration: Qwen-Image-Lightning LoRA
|
42 |
+
* Prompt AI: Qwen3-235B (Cerebras)
|
43 |
+
* Framework: Gradio + Diffusers
|
44 |
+
* Optimization: bfloat16 precision
|
45 |
+
|
46 |
+
🌟 Why Open Nano-Banana?
|
47 |
+
* ⚡ Speed: Instant results with 8 steps
|
48 |
+
* 🎨 Quality: Perfect editing with Prompt AI
|
49 |
+
* 🔒 Security: Token-based secure processing
|
50 |
+
* 💜 Design: Beautiful gradient UI
|
51 |
+
|
52 |
+
🏷️ Tags
|
53 |
+
#image-editing #ai-image-generation #qwen-image-edit #image-to-image #diffusers
|
54 |
+
#gradio #huggingface-spaces #lightning-lora #prompt-engineering #cerebras
|
55 |
+
#multilingual #real-time-editing #gpu-optimization #open-source #computer-vision
|
56 |
+
#deep-learning #machine-learning #artificial-intelligence #image-processing #creative-ai
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|