Skip to content
chat corpus collection from various open sources
Branch: master
Clone or download
Pull request Compare This branch is even with Marsan-Ma-zz:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.

Chat corpus repository

This is a chat corpus collection from various open sources, all files are composed of question-answer pairs, where odd lines are questions, even lines are answers.

I use them for training chatbot on seq2seq model. theory: implementation:

1. open_subtitles

English movie subtitles parsed from

2. movie_subtitles_en

Cornell Movie-Dialogs Corpus

3. lyrics_zh

lyrics from PTT forum

4. twitter_en

corpus scrap from twitter (700k lines), where odd lines are tweet and even lines are corresponding responded tweets. actually you could scrape your own with my twitter scraper repository

5. twitter_en big

twitter corpus in larger size (5M lines), files splitted to walkaround 100m filesize limit,
just cat them to recover the original gz file. cat twitter_en_big.txt.gz.part* > twitter_en_big.txt.gz

You can’t perform that action at this time.