References & Citations
Computer Science > Computation and Language
Title: DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation
(Submitted on 1 Nov 2019 (this version), latest version 2 May 2020 (v3))
Abstract: 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. We show that conversational systems that leverage DialoGPT generate more relevant, contentful and context-consistent responses than strong baseline systems. The pre-trained model and training pipeline are publicly released to facilitate research into neural response generation and the development of more intelligent open-domain dialogue systems.
Submission history
From: Yizhe Zhang [view email][v1] Fri, 1 Nov 2019 18:16:54 GMT (323kb,D)
[v2] Tue, 28 Apr 2020 05:45:19 GMT (422kb,D)
[v3] Sat, 2 May 2020 07:09:50 GMT (325kb,D)
Link back to: arXiv, form interface, contact.