Spaces:
Running
on
Zero
Running
on
Zero
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
---
|
| 2 |
-
title: FramePack
|
| 3 |
emoji: 🎬
|
| 4 |
colorFrom: indigo
|
| 5 |
colorTo: purple
|
|
@@ -9,65 +9,3 @@ app_file: app.py
|
|
| 9 |
pinned: false
|
| 10 |
license: mit
|
| 11 |
---
|
| 12 |
-
|
| 13 |
-
# FramePack - Image to Video Generation
|
| 14 |
-
|
| 15 |
-
This is a modified version of the FramePack model with a 5-second maximum video length limit.
|
| 16 |
-
|
| 17 |
-
## Features
|
| 18 |
-
|
| 19 |
-
- Generate realistic videos from still images
|
| 20 |
-
- Simple and intuitive interface
|
| 21 |
-
- Bilingual support (English/Chinese)
|
| 22 |
-
- Maximum video length of 5 seconds to ensure quick generation times
|
| 23 |
-
|
| 24 |
-
## Usage
|
| 25 |
-
|
| 26 |
-
1. Upload an image
|
| 27 |
-
2. Enter a prompt describing the desired motion
|
| 28 |
-
3. Adjust parameters if needed (seed, video length, etc.)
|
| 29 |
-
4. Click "Generate" and wait for the result
|
| 30 |
-
|
| 31 |
-
## Technical Details
|
| 32 |
-
|
| 33 |
-
This application uses the HunyuanVideo transformer model for image-to-video generation. The model has been optimized to work efficiently with videos up to 5 seconds in length.
|
| 34 |
-
|
| 35 |
-
## Credits
|
| 36 |
-
|
| 37 |
-
Based on the original FramePack model by lllyasviel.
|
| 38 |
-
|
| 39 |
-
## 特点
|
| 40 |
-
|
| 41 |
-
- 使用单张图片生成流畅的动作视频
|
| 42 |
-
- 基于HunyuanVideo和FramePack架构
|
| 43 |
-
- 支持低显存GPU(最低6GB)运行
|
| 44 |
-
- 可以生成最长5秒的视频
|
| 45 |
-
- 使用TeaCache技术加速生成过程
|
| 46 |
-
|
| 47 |
-
## 使用方法
|
| 48 |
-
|
| 49 |
-
1. 上传一张人物图像
|
| 50 |
-
2. 输入描述所需动作的提示词
|
| 51 |
-
3. 设置所需视频长度(最大5秒)
|
| 52 |
-
4. 点击"开始生成"按钮
|
| 53 |
-
5. 等待视频生成(生成过程是渐进式的,会不断扩展视频长度)
|
| 54 |
-
|
| 55 |
-
## 示例提示词
|
| 56 |
-
|
| 57 |
-
- "The girl dances gracefully, with clear movements, full of charm."
|
| 58 |
-
- "A character doing some simple body movements."
|
| 59 |
-
- "The man dances energetically, leaping mid-air with fluid arm swings and quick footwork."
|
| 60 |
-
|
| 61 |
-
## 注意事项
|
| 62 |
-
|
| 63 |
-
- 视频生成是倒序进行的,结束动作将先于开始动作生成
|
| 64 |
-
- 如果需要高质量结果,建议关闭TeaCache选项
|
| 65 |
-
- 如果遇到内存不足错误,可以增加"GPU推理保留内存"的值
|
| 66 |
-
|
| 67 |
-
## 技术细节
|
| 68 |
-
|
| 69 |
-
此应用基于[FramePack](https://github.com/lllyasviel/FramePack)项目,使用了Hunyuan Video模型和FramePack技术进行视频生成。该技术可以将输入上下文压缩为固定长度,使生成工作量与视频长度无关,从而在笔记本电脑GPU上也能处理大量帧。
|
| 70 |
-
|
| 71 |
-
---
|
| 72 |
-
|
| 73 |
-
原项目链接:[FramePack GitHub](https://github.com/lllyasviel/FramePack)
|
|
|
|
| 1 |
---
|
| 2 |
+
title: FramePack image to video
|
| 3 |
emoji: 🎬
|
| 4 |
colorFrom: indigo
|
| 5 |
colorTo: purple
|
|
|
|
| 9 |
pinned: false
|
| 10 |
license: mit
|
| 11 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|