This repository is about Chinese Ci(宋词) generation, and the paper has been accepted by AAAI-19.[pdf]
- Python 3.6
- PyTorch 1.5.1（GPU）
We crawl a dataset from a Chinese poetry website: https://sou-yun.com/ 95% of the corpus are contributed by the most popular 314 Cipai’s; the most popular 20 Cipai’s contribute to about 45% of the corpus. The dataset contains 82,724 Ci’s, written for 818 Cipai’s; on average there are 102 Ci’ per Cipai. In total, 3,797 Ci’s in the testing set and 78,927 in the training set.The data can be found in the**/base/data** directory.
Training the model:(in the base directory)
1、modify the parameters in config.py.change the parameter is_train from the default False to True
2、python main.py
Start Ci generation interface:
run /base/generate_ci_interface.py to start the interface
python -m generate_ci_interface -local_port [interface number]
tips:the interface number needs to be between 30011-30020.
If this work is helpful, please cite as:
@article{
Zhang_Liu_Chen_Hu_Xu_Mao_2019,
title={Generating Chinese Ci with Designated Metrical Structure},
volume={33},
url={https://ojs.aaai.org/index.php/AAAI/article/view/4736},
DOI={10.1609/aaai.v33i01.33017459},
number={01},
journal={Proceedings of the AAAI Conference on Artificial Intelligence},
author={Zhang, Richong and Liu, Xinyu and Chen, Xinwei and Hu, Zhiyuan and Xu, Zhaoqing and Mao, Yongyi},
year={2019},
month={Jul.},
pages={7459-7467}
}