Datasets:
license: mit
viewer: false
task_categories:
- translation
language:
- en
tags:
- art
Introduction
The Actor Voice dataset features audio snippets recorded by professional human actors, designed to support the development of large-scale audio generative models. These recordings can be utilized for various research purposes, including training models for audio separation, voice cloning, and speech generation. The dataset is particularly valuable for conditioning generative models to produce speech outputs based on specific voice inputs.
Dataset Statistics
- Total Audio Duration: 55.55 hours
Potential Research Use Cases:
- Audio Separation
- Audio Generation
- Voice Cloning
Acknowledgements
We extend our gratitude to our partner labs and sponsors for their support in making this dataset publicly available.
License
This dataset is released under EU research law for research purposes only. Users must comply with local regulations when using this dataset in their research.
Recombining File
the dataset is provided in splits to make it uploadbale on HF and to recombine them use below command:
cat rname-output.tar.gz.part-* > rname-output.tar.gz