Dialog_Corpus

Why we included this project

Training a chatbot in Chinese or English usually stalls on the same problem: there is no obvious place to find real dialogue data. This repository gathers the scattered public corpora that do exist into one index, linking each entry back to its original source. The notes are what make it useful, because quality varies a lot across the sets. Chinese movie dialogue is noisy and poorly paired, the ChatterBot Chinese data is small but clean, and the English NLP collections may need machine translation before they fit a Chinese bot. The list even flags one corpus, Microsoft Xiaoice, that is rumored to exist but was never released, saving you from chasing it. This is a map of what training data is out there, not a ready-to-use dataset.