A dataset focused on summarization of dialogs, which represents the rich domain of Twitter customer care conversations.
Refer to the original dataset from TWEETSUM. This pair dataset was re-formatted for fine tuning the text-summarization models into the dialogue summarization task. Dataset was split into 70% of training, 20% testing and 10% evaluation.
Dialogue | Summary |
---|---|
customer: My iPhone is loosing.... | Customer complains........ |