Gutenberg Dialog Dataset

This is a high-quality dataset consisting of 14.8M utterances in English, extracted from processed dialogues from publicly available online books.

Source: https://github.com/ricsinaruto/gutenberg-dialog