Datasets:
metadata
license: cc-by-nc-4.0
OpenDialog
OpenDialog is a 6.8k hours spoken dialogue dataset, introduced in paper ZipVoice-Dialog. 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)