Datasets:
| 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) |