Generation-based Conversation Dataset

The dataset used in the paper is another dataset containing 1.6 million query-reply pairs for generation-based conversation systems.

BibTex: