This website requires JavaScript.

TextBox 2.0: A Text Generation Library with Pre-trained Language Models

Tianyi TangJunyi LiZhipeng Chen ...+8 Ji-Rong Wen
Dec 2022
摘要
To facilitate research on text generation, this paper presents acomprehensive and unified library, TextBox 2.0, focusing on the use ofpre-trained language models (PLMs). To be comprehensive, our library covers$13$ common text generation tasks and their corresponding $83$ datasets andfurther incorporates $45$ PLMs covering general, translation, Chinese,dialogue, controllable, distilled, prompting, and lightweight PLMs. We alsoimplement $4$ efficient training strategies and provide $4$ generationobjectives for pre-training new PLMs from scratch. To be unified, we design theinterfaces to support the entire research pipeline (from data loading totraining and evaluation), ensuring that each step can be fulfilled in a unifiedway. Despite the rich functionality, it is easy to use our library, eitherthrough the friendly Python API or command line. To validate the effectivenessof our library, we conduct extensive experiments and exemplify four types ofresearch scenarios. The project is released at the link:https://github.com/RUCAIBox/TextBox.
展开全部
图表提取

暂无人提供速读十问回答

论文十问由沈向洋博士提出,鼓励大家带着这十个问题去阅读论文,用有用的信息构建认知模型。写出自己的十问回答,还有机会在当前页面展示哦。

Q1论文试图解决什么问题?
Q2这是否是一个新的问题?
Q3这篇文章要验证一个什么科学假设?
0
被引用
笔记
问答