Software of the Bergamot Project

• Our recipe for training efficient translation systems
• A machine translation server for Marian with a REST API and more features than the default server.
• Ptakopět: a system for quality estimation and outbound translation
• intgemm: a library for 8-bit and 16-bit matrix multiplication on x86_64 CPUs

Partner Software Relevant for the Bergamot Project

University of Edinburgh

• Marian: Fast Neural Machine Translation in C++, Marcin Junczys-Dowmunt, Roman Grundkiewicz, Tomasz Dwojak, Hieu Hoang, Kenneth Heafield, Tom Neckermann, Frank Seide, Ulrich Germann, Alham Fikri Aji, Nikolay Bogoychev, Andre ́ F. T. Martins, and Alexandra Birch.

Charles University

• Neural Monkey: An Open-source Tool for Sequence Learning, Jindřich Helcl, Jindřich Libovický. In The Prague Bulletin of Mathematical Linguistics, 2017.

University of Sheffield

• Multi-level Translation Quality Prediction with QuEst++, Lucia Specia, Gustavo Henrique Paetzold and Carolina Scarton. In the Proceedings of ACL-IJCNLP 2015 System Demonstrations, Beijing, China, pp. 110-120. 2015.

Mozilla

• Firefox, popular web browser.
• Deep Speech: Deep Leaning STT Engine, localizable speech-to-text engine that uses deep learning techniques to transcribe speech to text. It has more than 9,400 stars, and over 1,600 forks on GitHub repository.
• TTS: Deep Leaning TTS Engine, Localizable text-to-speech engine that uses deep learning techniques to speak written text.