A multi-party and multi-round personalized dataset
You can download it on https://drive.google.com/drive/folders/1-8kwWCo6vfmDzGk8eJ1UeMMnuUlyx3mI?usp=sharing
Brief description of dataset format
u_sep：context
p_sep：personas of the speakers in the context
resp_sep：personas of the responder
edge：the relationship between utterances and personas and the relationship between different personas
Average length statistics
setence level text knowledge 7.5
docoment level text knowledge 487.7