# Overview

Recent studies show that the end-to-end direct speech translation approaches achieve promising performance and attract a lot of interest. This project will list recent research works on the end-to-end ST and links to available datasets and toolkits. Moreover, we build reproducible and reliable benchmarks for this field and make fair comparisons with advanced approaches, either from their released repositories or from our re-implementations using NeurST (opens new window). We hope our project could facilitate further research works.

Last Updated: 6/6/2021, 1:10:38 AM