Datasets:

Modalities:
Audio
Text
Formats:
webdataset
ArXiv:
Libraries:
Datasets
WebDataset
License:
File size: 407 Bytes
df494c7
 
 
 
396cf74
 
 
 
 
 
9dc9ce2
1
2
3
4
5
6
7
8
9
10
11
---
license: cc-by-nc-4.0
---

# OpenDialog

OpenDialog is a 6.8k hours spoken dialogue dataset, introduced in paper [ZipVoice-Dialog](https://arxiv.org/abs/2507.09318). Check the paper for details

It consists of 1759 hours of Chinese data and 5074 hours of English data. 

This dataset is also available at https://www.modelscope.cn/datasets/k2-fsa/OpenDialog (more friendly for users from China mainland)