Audio2Face-3D-Dataset-v1.0.0-claire

## Dataset Description:
NVIDIA Audio2Face-3D-dataset-v1.0.0-claire includes audio files, blendshape data, animated geometry caches, geometry files, and transform files.
This dataset is for demonstration purposes and not for production usage.
For source code, documentation, helper scripts, packaged builds, and links to all components in the Audio2Face-3D technology stack, visit the [Audio2Face-3D GitHub repository](https://github.com/NVIDIA/Audio2Face-3D)

## Dataset Owner:
NVIDIA Corporation

## Dataset Creation Date:
July 2025

## License/Terms of Use:
Use of this model is governed by the [NVIDIA AUDIO2FACE SAMPLE DATA LICENSE](https://huggingface.co/datasets/nvidia/Audio2Face-3D-Dataset-v1.0.0-claire/blob/main/LICENSE)

## Intended Usage:
Intended use cases include example training for NVIDIA Audio2Face-3D networks and serving as a reference for custom Audio2Face-3D datasets. Intended users include creators and developers of interactive avatars and game characters.

## Dataset Characterization
Data Collection Method<br>
* Human - 4D facial tracking data and audio <br>

Labeling Method<br>
* Hybrid: Human, Synthetic, Automated <br>

## Dataset Format
audio files (.wav), motion data (.mc, .npz), asset data (.npy, .obj), configs (.json)

## Dataset Quantification
<1 GB total - 54 audio files, 4 blendshape data files, 108 animated geometry caches, 10 geometry files, and 5 transform files.

## Ethical Considerations:
NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. When downloaded or used in accordance with our terms of service, developers should work with their internal model team to ensure this model meets requirements for the relevant industry and use case and addresses unforeseen product misuse.

Please report security vulnerabilities or NVIDIA AI Concerns [here](https://www.nvidia.com/en-us/support/submit-security-vulnerability/).

## 数据集描述：
NVIDIA Audio2Face-3D数据集v1.0.0-claire包含音频文件、融合形变数据（blendshape）、动画几何缓存文件、几何文件以及变换文件。
本数据集仅用于演示用途，不得用于生产环境。
如需获取Audio2Face-3D技术栈所有组件的源代码、文档、辅助脚本、打包构建版本及相关链接，请访问[Audio2Face-3D GitHub仓库](https://github.com/NVIDIA/Audio2Face-3D)

## 数据集所有者：
NVIDIA公司

## 数据集创建日期：
2025年7月

## 许可与使用条款：
本数据集的使用受[NVIDIA AUDIO2FACE示例数据许可协议](https://huggingface.co/datasets/nvidia/Audio2Face-3D-Dataset-v1.0.0-claire/blob/main/LICENSE)约束。

## 预期用途：
本数据集的应用场景包括：作为NVIDIA Audio2Face-3D网络的示例训练数据，以及作为自定义Audio2Face-3D数据集的参考范本。目标用户涵盖交互式虚拟化身与游戏角色的创作者与开发者。

## 数据集特征
### 数据采集方式
- 人工采集：获取4D面部追踪数据与音频
### 标注方式
- 混合标注：结合人工、合成与自动化标注

## 数据集格式
包含音频文件（.wav）、运动数据（.mc、.npz）、资产数据（.npy、.obj）以及配置文件（.json）。

## 数据集规模
总容量不足1GB，共包含54个音频文件、4个融合形变数据文件、108个动画几何缓存文件、10个几何文件以及5个变换文件。

## 伦理考量：
NVIDIA认为，可信人工智能是一项共同责任，我们已建立相关政策与实践规范，以支持各类人工智能应用的开发。开发者在按照服务条款下载或使用本数据集时，应与内部模型团队协作，确保本数据集符合相关行业与应用场景的要求，并防范可能出现的产品误用风险。

请通过[此链接](https://www.nvidia.com/en-us/support/submit-security-vulnerability/)提交安全漏洞或NVIDIA人工智能相关问题的反馈。