Update README.md

README.md

@@ -1,100 +1,99 @@
----
-frameworks:
-- Pytorch
-license: Apache License 2.0
-tasks:
-- text-to-video-synthesis
-
-
-# FantasyPortrait：基于表情增强扩散变换器的多角色肖像动画生成
+# FantasyPortrait: Enhancing Multi-Character Portrait Animation with Expression-Augmented Diffusion Transformers
 
 [![Home Page](https://img.shields.io/badge/Project-FantasyPortrait-blue.svg)](https://fantasy-amap.github.io/fantasy-portrait/)
 [![arXiv](https://img.shields.io/badge/Arxiv-2507.12956-b31b1b.svg?logo=arXiv)](https://arxiv.org/abs/2507.12956)
-[![hf_dataset](https://img.shields.io/badge/🤗%20Dataset-FantasyPortrait-yellow.svg)](https://huggingface.co/datasets/acvlab/FantasyPortrait)
+[![hf_dataset](https://img.shields.io/badge/🤗%20Dataset-FantasyPortrait-yellow.svg)](https://huggingface.co/datasets/acvlab/FantasyPortrait-Multi-Expr)
+[![hf_model](https://img.shields.io/badge/🤗%20Model-FantasyPortrait-green.svg)](https://huggingface.co/acvlab/FantasyPortrait)
 [![hf_paper](https://img.shields.io/badge/🤗-FantasyPortrait-red.svg)](https://huggingface.co/papers/2507.12956)
+[![ms_model](https://img.shields.io/badge/ModelScope-Model-9cf.svg)](https://modelscope.cn/models/amap_cvlab/FantasyPortrait)
+[![ms_dataset](https://img.shields.io/badge/ModelScope-Dataset-ff69b4.svg)](https://www.modelscope.cn/datasets/amap_cvlab/FantasyPortrait-Multi-Expr)
 
-## 🔥 最新动态！！
-* 2025年8月10日：我们已发布推理代码、模型权重和数据集。
+## 🔥 Latest News!!
+* August 12, 2025: We released the inference code, model weights and datasets.
 
-## 演示
-更多有趣的结果，请访问我们的[网站](https://fantasy-amap.github.io/fantasy-portrait/)。
+## Demo
+For more interesting results, please visit our [website](https://fantasy-amap.github.io/fantasy-portrait/).
 
 | ![单人示例](./danren_1.gif) | ![对比](./duibi.gif) |
 | :---: | :---: |
 | ![动物](./dongwu.gif) | ![双人1](./shuangren_1.gif) |
 | ![双人2](./shuangren_2.gif) | ![三人](./sanren.gif) |
 
-## 快速开始
-### 🛠️ 安装
 
-克隆仓库：
+## Quickstart
+### 🛠️Installation
+
+Clone the repo:
 
 ```
 git clone https://github.com/Fantasy-AMAP/fantasy-portrait.git
 cd fantasy-portrait
 ```
 
-安装依赖：
+Install dependencies:
 ```
 apt-get install ffmpeg
-# 确保 torch >= 2.0.0
+# Ensure torch >= 2.0.0
 pip install -r requirements.txt
-# 注意：必须安装 flash attention
+# Note: flash attention must be installed
 pip install flash_attn
 ```
 
-### 📦 Multi-Expr 数据集
-我们公开了首个多人肖像面部表情视频数据集 **Multi-Expr Dataset**，请通过以下链接下载：
+### 📦Multi-Expr Dataset
+We make public the first multi-portrait facial expression video dataset **Multi-Expr Dataset**, Please download it via the [ModelScope](https://www.modelscope.cn/datasets/amap_cvlab/FantasyPortrait-Multi-Expr) or [Huggingface](https://huggingface.co/datasets/acvlab/FantasyPortrait-Multi-Expr).
 
-
-### 🧱 模型下载
-| 模型        |                       下载链接                                           |    说明                      |
+### 🧱Model Download
+| Models        |                       Download Link                                           |    Notes                      |
 | --------------|-------------------------------------------------------------------------------|-------------------------------|
-| Wan2.1-I2V-14B-720P  |      🤗 [Huggingface](https://huggingface.co/Wan-AI/Wan2.1-I2V-14B-720P)    🤖 [ModelScope](https://www.modelscope.cn/models/Wan-AI/Wan2.1-I2V-14B-720P)     | 基础模型
-| FantasyPortrait      |      🤗 [Huggingface](https://huggingface.co/acvlab/FantasyPortrait/)     🤖 [ModelScope](https://www.modelscope.cn/models/amap_cvlab/FantasyPortrait/)         | 我们的表情条件权重
+| Wan2.1-I2V-14B-720P  |      🤗 [Huggingface](https://huggingface.co/Wan-AI/Wan2.1-I2V-14B-720P)    🤖 [ModelScope](https://www.modelscope.cn/models/Wan-AI/Wan2.1-I2V-14B-720P)     | Base model
+| FantasyPortrait      |      🤗 [Huggingface](https://huggingface.co/acvlab/FantasyPortrait/)     🤖 [ModelScope](https://www.modelscope.cn/models/amap_cvlab/FantasyPortrait/)         | Our emo condition weights
 
-使用 huggingface-cli 下载模型：
+Download models using huggingface-cli:
 ``` sh
 pip install "huggingface_hub[cli]"
 huggingface-cli download Wan-AI/Wan2.1-I2V-14B-720P --local-dir ./models/Wan2.1-I2V-14B-720P
 huggingface-cli download acvlab/FantasyPortrait --local-dir ./models
 ```
 
-使用 modelscope-cli 下载模型：
+Download models using modelscope-cli:
 ``` sh
 pip install modelscope
 modelscope download Wan-AI/Wan2.1-I2V-14B-720P --local_dir ./models/Wan2.1-I2V-14B-720P
 modelscope download amap_cvlab/FantasyPortrait  --local_dir ./models
 ```
 
-### 🔑 单人肖像推理
+### 🔑 Single-Portrait Inference
 ``` sh
 bash infer_single.sh
 ```
 
-### 🔑 多人肖像推理
+### 🔑 Multi-Portrait Inference
+If you use input image and drive videos with multiple people, you can run as follows:
 ``` sh
 bash infer_multi.sh
 ```
+If you use input image with multiple people and different multiple single-human driven videos, you can run as follows:
+```sh
+bash infer_multi_diff.sh
+```
 
+### 📦Speed and VRAM Usage
+We present a detailed table here. The model is tested on a single A100.
 
-### 📦 速度与显存占用
-我们在此提供详细表格。模型在单张A100上进行测试。
-
-|`torch_dtype`|`num_persistent_param_in_dit`|速度|所需显存|
+|`torch_dtype`|`num_persistent_param_in_dit`|Speed|Required VRAM|
 |-|-|-|-|
-|torch.bfloat16|None (无限制)|15.5秒/迭代|40G|
-|torch.bfloat16|7*10**9 (7B)|32.8秒/迭代|20G|
-|torch.bfloat16|0|42.6秒/迭代|5G|
+|torch.bfloat16|None (unlimited)|15.5s/it|40G|
+|torch.bfloat16|7*10**9 (7B)|32.8s/it|20G|
+|torch.bfloat16|0|42.6s/it|5G|
 
 
 
-## 🧩 社区贡献
-我们 ❤️ 来自开源社区的贡献！如果您的工作改进了 FantasyPortrait，请告知我们。
-您也可以直接发送邮件至 [frank.jf@alibaba-inc.com](mailto://frank.jf@alibaba-inc.com)。我们很乐意引用您的项目，方便大家使用。
+## 🧩 Community Works
+We ❤️ contributions from the open-source community! If your work has improved FantasyPortrait, please inform us.
+Or you can directly e-mail [frank.jf@alibaba-inc.com](mailto://frank.jf@alibaba-inc.com). We are happy to reference your project for everyone's convenience.
 
-## 🔗 引用
-如果本仓库对您有帮助，请考虑给我们一个 star ⭐ 并引用以下论文：
+## 🔗Citation
+If you find this repository useful, please consider giving a star ⭐ and citation
 ```
 @article{wang2025fantasyportrait,
   title={FantasyPortrait: Enhancing Multi-Character Portrait Animation with Expression-Augmented Diffusion Transformers},
@@ -104,5 +103,5 @@ bash infer_multi.sh
 }
 ```
 
-## 致谢
-感谢 [Wan2.1](https://github.com/Wan-Video/Wan2.1)、[PD-FGC](https://github.com/Dorniwang/PD-FGC-inference) 和 [DiffSynth-Studio](https://github.com/modelscope/DiffSynth-Studio) 开源他们的模型和代码，为本项目提供了宝贵的参考和支持。我们非常感谢他们对开源社区的贡献。
+## Acknowledgments
+Thanks to [Wan2.1](https://github.com/Wan-Video/Wan2.1), [PD-FGC](https://github.com/Dorniwang/PD-FGC-inference) and [DiffSynth-Studio](https://github.com/modelscope/DiffSynth-Studio) for open-sourcing their models and code, which provided valuable references and support for this project. Their contributions to the open-source community are truly appreciated.