Authors
Yaoming Zhu, Sidi Lu, Lei Zheng, Jiaxian Guo, Weinan Zhang, Jun Wang, and Yong Yu
Overview
Texygen is an open-source benchmarking platform for open-domain text generation. It brings representative generation models and evaluation metrics into one reproducible environment, making it easier to compare methods under consistent settings.
The benchmark covers complementary measurements of generation quality, diversity, and consistency rather than reducing model comparison to a single score.
Results
The platform provides shared implementations and standardized evaluation procedures for neural text generation research. Its purpose is to make empirical comparisons more reproducible and to reduce duplicated implementation work across projects.
Paper
Status
SIGIR 2018 Short Paper.
Citation
Yaoming Zhu, Sidi Lu, Lei Zheng, Jiaxian Guo, Weinan Zhang, Jun Wang, and Yong Yu. “Texygen: A Benchmarking Platform for Text Generation Models.” Proceedings of the 41st International ACM SIGIR Conference on Research & Development in Information Retrieval, pages 1097–1100, 2018.