XNMT: The eXtensible Neural Machine Translation Toolkit

TL;DR

XNMT is an open-source, modular toolkit for neural machine translation that facilitates rapid research iteration and reliable, reproducible results across multiple language and speech tasks.

Contribution

It introduces a highly modular design and experiment configuration system, enhancing flexibility and reproducibility in NMT research.

Findings

Demonstrated utility on machine translation, speech recognition, and multi-task learning

Facilitated rapid experimentation and iteration

Produced reliable and reproducible results

Abstract

This paper describes XNMT, the eXtensible Neural Machine Translation toolkit. XNMT distin- guishes itself from other open-source NMT toolkits by its focus on modular code design, with the purpose of enabling fast iteration in research and replicable, reliable results. In this paper we describe the design of XNMT and its experiment configuration system, and demonstrate its utility on the tasks of machine translation, speech recognition, and multi-tasked machine translation/parsing. XNMT is available open-source at https://github.com/neulab/xnmt