Poetry generation is an interesting research topic in the field of text. As one of the most valuable literary and cultural heritages of, Chinese classical poetry is very familiar and loved by Chinese people generation to generation. It has many particular characteristics in its structure, ranging from form, sound to meaning, thus is regarded as an testing task for text generation. In this paper, we propose a GPT-2 based framework for generating major types of Chinese classical poems. We a unified format for formulating all types of training samples by detailed form information, then present a simple form-stressed method in GPT-2 to strengthen the control to the form of the poems, with special emphasis on those forms with longer body length. experimental results show this enhanced model can generate Chinese poems of major types with high quality in both form and content, the effectiveness of the proposed strategy. The model has been into Jiuge, the most influential Chinese classical poetry system developed by Tsinghua University (Guo et al., 2019).
No takes yet. Share an insight, caveat, or question.
Hu et al. (2020) studied this question.