Dialogue Generative Pre-Trained Transformer
Year: 2,020
Journal: Annual Meeting of the Association for Computational Linguistics
Languages: English
Programming languages: Python
Input data:
sentences
Output data:
text (responses)
Project website: https://github.com/microsoft/DialoGPT
We present a large, tunable neural conversational response generation model, DIALOGPT (dialogue generative pre-trained transformer). Trained on 147M conversation-like exchanges extracted from Reddit comment chains over a period spanning from 2005 through 2017, DialoGPT extends the Hugging Face PyTorch transformer to attain a performance close to human both in terms of automatic and human evaluation in single-turn dialogue settings.