marian/README.md

81 lines
4.1 KiB
Markdown
Raw Normal View History

2017-06-04 16:07:24 +03:00
Marian
======
2018-09-16 13:02:46 +03:00
2021-11-21 20:06:01 +03:00
[![Build Status CUDA 10](https://img.shields.io/jenkins/s/http/vali.inf.ed.ac.uk/jenkins/view/marian/job/marian-dev-cuda-10.2.svg?label=CUDA%2010.2)](http://vali.inf.ed.ac.uk/jenkins/job/marian-dev-cuda-10.2/)
[![Build Status CUDA 11](https://img.shields.io/jenkins/s/http/vali.inf.ed.ac.uk/jenkins/view/marian/job/marian-dev-cuda-11.4.svg?label=CUDA%2011.4)](http://vali.inf.ed.ac.uk/jenkins/job/marian-dev-cuda-11.4/)
[![Build Status CPU](https://img.shields.io/jenkins/s/http/vali.inf.ed.ac.uk/jenkins/view/marian/job/marian-dev-cpu.svg?label=CPU)](http://vali.inf.ed.ac.uk/jenkins/job/marian-dev-cpu/)
2018-10-13 11:25:52 +03:00
[![Tests Status](https://img.shields.io/jenkins/s/http/vali.inf.ed.ac.uk/jenkins/view/marian/job/marian-regression-tests.svg?label=tests)](http://vali.inf.ed.ac.uk/jenkins/job/marian-regression-tests/)
2018-09-20 18:40:25 +03:00
[![Latest release](https://img.shields.io/github/release/marian-nmt/marian.svg?label=release)](https://github.com/marian-nmt/marian/releases)
2018-09-18 16:54:53 +03:00
[![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](./LICENSE.md)
2018-09-18 16:53:23 +03:00
[![Twitter](https://img.shields.io/twitter/follow/marian_nmt.svg?style=social)](https://twitter.com/intent/follow?screen_name=marian_nmt)
2016-09-19 16:02:30 +03:00
2019-01-25 13:09:56 +03:00
*Marian* is an efficient Neural Machine Translation framework written in pure
C++ with minimal dependencies.
2018-09-18 16:53:23 +03:00
2019-01-25 13:09:56 +03:00
Named in honour of Marian Rejewski, a Polish mathematician and cryptologist.
2018-09-16 12:56:46 +03:00
2019-01-25 13:09:56 +03:00
Main features:
2018-09-16 12:56:46 +03:00
2019-01-25 13:09:56 +03:00
- Efficient pure C++ implementation
- Fast multi-GPU training and GPU/CPU translation
- State-of-the-art NMT architectures: deep RNN and transformer
- Permissive open source license (MIT)
- [more detail...](https://marian-nmt.github.io/features)
2018-09-16 12:56:46 +03:00
If you use this, please cite:
2018-09-18 16:53:23 +03:00
Marcin Junczys-Dowmunt, Roman Grundkiewicz, Tomasz Dwojak, Hieu Hoang, Kenneth
Heafield, Tom Neckermann, Frank Seide, Ulrich Germann, Alham Fikri Aji, Nikolay
Bogoychev, André F. T. Martins, Alexandra Birch (2018). Marian: Fast Neural
Machine Translation in C++ (http://www.aclweb.org/anthology/P18-4020)
2018-09-16 12:56:46 +03:00
@InProceedings{mariannmt,
title = {Marian: Fast Neural Machine Translation in {C++}},
author = {Junczys-Dowmunt, Marcin and Grundkiewicz, Roman and
Dwojak, Tomasz and Hoang, Hieu and Heafield, Kenneth and
Neckermann, Tom and Seide, Frank and Germann, Ulrich and
Fikri Aji, Alham and Bogoychev, Nikolay and
Martins, Andr\'{e} F. T. and Birch, Alexandra},
booktitle = {Proceedings of ACL 2018, System Demonstrations},
pages = {116--121},
publisher = {Association for Computational Linguistics},
year = {2018},
month = {July},
address = {Melbourne, Australia},
url = {http://www.aclweb.org/anthology/P18-4020}
}
2018-09-16 13:05:32 +03:00
## Amun
2019-01-25 13:09:56 +03:00
The handwritten decoder for RNN models compatible with Marian and Nematus has
been superseded by the Marian decoder. The code is available in a separate
repository: https://github.com/marian-nmt/amun
2018-09-16 13:05:32 +03:00
2017-06-04 16:07:24 +03:00
## Website
2016-09-13 15:04:23 +03:00
2017-07-18 11:47:32 +03:00
More information on https://marian-nmt.github.io
2017-06-04 16:07:24 +03:00
2018-09-16 12:56:46 +03:00
- [Quick start](https://marian-nmt.github.io/quickstart)
- [Installation and usage documentation](https://marian-nmt.github.io/docs)
- [Usage examples](https://marian-nmt.github.io/examples)
2017-10-15 17:34:52 +03:00
## Acknowledgements
2017-06-04 16:07:24 +03:00
2017-06-14 15:32:31 +03:00
The development of Marian received funding from the European Union's
_Horizon 2020 Research and Innovation Programme_ under grant agreements
688139 ([SUMMA](http://www.summa-project.eu); 2016-2019),
2018-02-02 12:46:00 +03:00
645487 ([Modern MT](http://www.modernmt.eu); 2015-2017),
2017-06-14 15:32:31 +03:00
644333 ([TraMOOC](http://tramooc.eu/); 2015-2017),
2018-02-02 12:46:00 +03:00
644402 ([HiML](http://www.himl.eu/); 2015-2017),
2019-04-25 12:23:18 +03:00
825303 ([Bergamot](https://browser.mt/); 2019-2021),
2022-02-09 20:05:48 +03:00
the European Union's Connecting Europe Facility project
2019-EU-IA-0045 ([User-focused Marian](https://marian-project.eu); 2020-2022),
2018-03-30 12:56:04 +03:00
the Amazon Academic Research Awards program,
the World Intellectual Property Organization,
and is based upon work supported in part by the Office of the Director of
National Intelligence (ODNI), Intelligence Advanced Research Projects Activity
(IARPA), via contract #FA8650-17-C-9117.
This software contains source code provided by NVIDIA Corporation.