GPT-based Generation for Classical Chinese Poetry

We present a simple yet effective method for generating high qual- ity classical Chinese poetry with Generative Pre-trained Language Model (GPT)[5]. The method adopts a simple GPT model, without using any human crafted rules or features, or designing any additional neural compo- nents. While the proposed model learns to generate various forms of clas- sical Chinese poems, including Jueju(绝句), Lu ̈shi(律诗), various Cipai(词牌) and Couples(对联), the generated poems are of very high quality. We also propose and implement a method to fine-tune the model to generate acrostic poetry. To the best of our knowledge, this is the first to em- ploy GPT in developing a poetry generation system. We have released an online mini demonstration program on Wechat1 to show the generation capability of the proposed method for classical Chinese poetry.