Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,72 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
base_model:
|
| 6 |
+
- Qwen/Qwen-Image-Edit-2509
|
| 7 |
+
tags:
|
| 8 |
+
- image-generation
|
| 9 |
+
- lora
|
| 10 |
+
- Qwen-Image
|
| 11 |
+
pipeline_tag: image-to-image
|
| 12 |
+
library_name: diffusers
|
| 13 |
+
widget:
|
| 14 |
+
- text: >-
|
| 15 |
+
保持整个照片的布局和内容不变，把这个照片做旧，每个格子中的照片色调变成颗粒感滤镜，整个画面具有老式照片的朦胧感
|
| 16 |
+
output:
|
| 17 |
+
url: images/InstantCameraTriptych (1).webp
|
| 18 |
+
- text: >-
|
| 19 |
+
保持整个照片的布局和内容不变，把这个照片做旧，每个格子中的照片色调变成颗粒感滤镜，整个画面具有老式照片的朦胧感
|
| 20 |
+
output:
|
| 21 |
+
url: images/InstantCameraTriptych (2).webp
|
| 22 |
+
- text: >-
|
| 23 |
+
保持整个照片的布局和内容不变，把这个照片做旧，每个格子中的照片色调变成颗粒感滤镜，整个画面具有老式照片的朦胧感
|
| 24 |
+
output:
|
| 25 |
+
url: images/InstantCameraTriptych (3).webp
|
| 26 |
+
---
|
| 27 |
+
# valiantcat Qwen-Image-Edit-2509 LoRA
|
| 28 |
+
|
| 29 |
+
<Gallery />
|
| 30 |
+
|
| 31 |
+
|
| 32 |
+
## 🎨 效果对比展示
|
| 33 |
+
|
| 34 |
+
|
| 35 |
+
| 控制图1 | 控制图2 | 生成结果 |
|
| 36 |
+
|:---:|:---:|:---:|
|
| 37 |
+
|  |  |  |
|
| 38 |
+
|  |  |  |
|
| 39 |
+
|  |  |  |
|
| 40 |
+
|
| 41 |
+
<Gallery />
|
| 42 |
+
|
| 43 |
+
## Model Card for Model ID
|
| 44 |
+
|
| 45 |
+
因为原生的Qwen-Edit-2509的多图输入的结果中人物一致性并不是很好，特别是在一些指定的场景和指定的分辨率下。本模型重新针对多图输入的合照场景对Qwen-Edit-2509进行重新训练，保持高度人物一致性，输入两张图的图像为他们生成合照。
|
| 46 |
+
The native Qwen-Edit-2509 model often struggles with consistency in multi-image inputs, particularly at specific resolutions and scenes. This model is a fine-tuned version, specifically optimized for generating group photos from multiple portraits. It maintains high character consistency, allowing you to input individual photos and produce a coherent composite image.
|
| 47 |
+
|
| 48 |
+
|
| 49 |
+
## Model Card for Model ID
|
| 50 |
+
|
| 51 |
+
将图像输入的多图生成合照。。
|
| 52 |
+
|
| 53 |
+
## Prompts phrase
|
| 54 |
+
|
| 55 |
+
```两人相互对视并牵手的照片，保持两人人脸、服饰、装扮高度一致，背景在温暖的客厅室内，95年代复古风格。整个画面具有老式照片的不清晰感，颗粒感，严重的晃动感，复古的滤镜。```
|
| 56 |
+
```两人相互接吻的照片，保持两人人脸、服饰、装扮高度一致，背景是平坦的高原山脉，80年代复古风格。整个画面具有老式照片的不清晰感，颗粒感，严重的晃动感，复古的滤镜。```
|
| 57 |
+
```两人相互对视并牵手的照片，保持两人人脸、服饰、装扮高度一致，背景在温暖的客厅室内，95年代复古风格。整个画面具有老式照片的不清晰感，颗粒感，严重的晃动感，复古的滤镜。```
|
| 58 |
+
```两人站在街头合照并挽着手的照片，保持两人人脸高度一致，男士穿着夹克外套，露出南瓜头的T恤，女士穿着时尚的偏橘黄色的装扮，背景是万圣节日的欧洲小镇的马路中央，万圣节节日风格```
|
| 59 |
+
```两人站在一个石头拱门墙前的酷酷合拍，保持两人人脸高度一致，两人穿着黑色的时尚帅气装扮，拱门后的天气阴暗，还有稀疏的枯木。```
|
| 60 |
+
```两人在室内的合照，保持两人人脸高度一致，两人穿着有万圣节节日气息的小饰品，背后有是奶油风的室内场景，有一些万圣节背景元素。```
|
| 61 |
+
|
| 62 |
+
There is no fixed trigger word. The specific photous prompt needs to be tested more.
|
| 63 |
+
|
| 64 |
+
## Download model
|
| 65 |
+
|
| 66 |
+
Weights for this model are available in Safetensors format.
|
| 67 |
+
|
| 68 |
+
[Download](https://huggingface.co/valiantcat/Qwen-Image-Edit-2509-photous)
|
| 69 |
+
|
| 70 |
+
## Training at Chongqing Valiant Cat
|
| 71 |
+
|
| 72 |
+
This model was trained by the AI Laboratory of Chongqing Valiant Cat Technology Co., LTD(```https://vvicat.com/```).Business cooperation is welcome
|