File size: 1,196 Bytes
1a08fe3
 
 
 
 
 
 
 
 
8ef7026
af250da
a3f9d09
8ef7026
b69731f
8ef7026
 
 
eeca452
8ef7026
 
 
 
 
 
eeca452
63aeed2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
---
title: README
emoji: 📈
colorFrom: indigo
colorTo: red
sdk: static
pinned: false
---

# LEMAS: A 150K-Hour Large-scale Extensible Multilingual Audio Suite with Generative Speech Models

[**LEMAS**](https://lemas-project.github.io/LEMAS-Project/) is a large-scale extensible multilingual audio suite, providing multilingual speech corpus (**LEMAS-Dataset**) with word-level timestamps, covering over **150,000 hours** across **10 major languages(zh/en/de/fr/es/pt/it/ru/id/vi)**. Built with a rigorous alignment and confidence-based filtering pipeline, LEMAS supports diverse generative paradigms including zero-shot multilingual synthesis (**LEMAS-TTS**)(0.3B) and seamless speech editing (**LEMAS-Edit**)(0.3B).

More technical details can be found in our technical report: [https://arxiv.org/abs/2601.04233](https://arxiv.org/abs/2601.04233)

<details><summary>Citation</summary>
  
```
@article{zhao2026lemas,
  title={LEMAS: A 150K-Hour Large-scale Extensible Multilingual Audio Suite with Generative Speech Models},
  author={Zhao, Zhiyuan and Lin, Lijian and Zhu, Ye and Xie, Kai and Liu, Yunfei and Li, Yu},
  journal={arXiv preprint arXiv:2601.04233},
  year={2026}
}
```
</details>