nmt-list

KO | EN

Minimalist Open Source Repository Search Engine

gitlite — search

#typescript #ai-agents #cli #open-source #ai #claude-code #codex #deepseek-harness #dsh-plugin #react #windows #developer-tools

★ 364 Open GitHub ↗

A list of Neural MT implementations

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Neural Machine Translation Implementations

## Contributions

Feel free add to this via a [pull request](https://help.github.com/articles/creating-a-pull-request), with each section alphabetically ordered.

## Project Stats
* **23** project names include *nmt*
* **6** project names include *seq2seq*
* **2** project names include *dl4mt*
* **3** projects are derived from Kyunghyun Cho's tutorial at the 2015 [DL4MT winter school](http://dl4mt.computing.dcu.ie)

## The Lists

### [Caffe2](https://github.com/caffe2/caffe2)-based NMT
| Software | Who | Comments |
| -------- | --- | -------- |
| [seq2seq](https://github.com/caffe2/caffe2/tree/master/caffe2/python/models/seq2seq) | [Deepak](https://github.com/deepakg92), [Yury](https://github.com/urikz) |  |

### [Chainer](https://github.com/pfnet/chainer)-based NMT
| Software | Who | Comments |
| -------- | --- | -------- |
| [attention_is_all_you_need](https://github.com/soskek/attention_is_all_you_need) | [Sosuke](https://github.com/soskek) | impements the [Transformer](https://arxiv.org/abs/1706.03762) |
| [chainn](https://github.com/philip30/chainn) | [Philip](http://isw3.naist.jp/~philip-a/index.html) |  |
| [KyotoNMT](https://github.com/fabiencro/knmt) | [Fabien](https://github.com/fabiencro) |  |

### Custom C/C++ NMT
| Software | Who | Comments |
| -------- | --- | -------- |
| [EUREKA-MangoNMT](https://github.com/jiajunzhangnlp/EUREKA-MangoNMT) | [Jiajun](https://github.com/jiajunzhangnlp) |  |
| [Mantidae](https://github.com/duyvuleo/Mantidae) | [Vu](https://github.com/duyvuleo) | A successor of [Mantis](https://github.com/trevorcohn/mantis) |
| [Marian](https://github.com/marian-nmt/marian) | [Marcin](https://github.com/emjotde), [Tomasz](https://github.com/tomekd) | Compatible with Nematus models |
| [N3LP](https://github.com/hassyGo/N3LP) | [Kazuma](https://github.com/hassyGo), [Akiko](https://github.com/tempra28) |  |
| [Zoph_RNN](https://github.com/isi-nlp/Zoph_RNN) | [Barret](https://github.com/barretzoph) | Supports multi-GPU parallelism |

### [DyNet](https://github.com/clab/dynet)-based NMT
| Software | Who | Comments |
| -------- | --- | -------- |
| [dynmt-py](https://github.com/roeeaharoni/dynmt-py) | [Roee](https://github.com/roeeaharoni) | Python |
| [lamtram](https://github.com/neubig/lamtram) | [Graham](https://github.com/neubig) | C++ |
| [mantis](https://github.com/trevorcohn/mantis) | [Trevor](https://github.com/trevorcohn) | C++ |
| [NMTKit](https://github.com/odashi/nmtkit) | [Yusuke](https://github.com/odashi) | C++ |
| [Transformer-DyNet](https://github.com/duyvuleo/Transformer-DyNet) | [Vu](https://github.com/duyvuleo) | C++; implements the [Transformer](https://arxiv.org/abs/1706.03762) |
| [xnmt](https://github.com/neulab/xnmt) | [NeuLab](https://github.com/neulab) | Python |

### [Keras](https://github.com/fchollet/keras)-based NMT
| Software | Who | Comments |
| -------- | --- | -------- |
| Keras [seq2seq](https://github.com/farizrahman4u/seq2seq) | [Fariz](https://github.com/farizrahman4u) |  |
| [NMT-Keras](https://github.com/lvapeab/nmt-keras) | [Álvaro](https://github.com/lvapeab) |  |

### [MXNet](http://mxnet.io)-based NMT
| Software | Who | Comments |
| -------- | --- | -------- |
| [MXNMT](https://github.com/magic282/MXNMT) | [magic282](https://github.com/magic282) |  |
| [sockeye](https://github.com/awslabs/sockeye) | [AWS Labs](https://github.com/awslabs) | |

### [PyTorch](https://github.com/pytorch/pytorch)-based NMT
| Software | Who | Comments |
| -------- | --- | -------- |
| [attention-is-all-you-need-pytorch](https://github.com/jadore801120/attention-is-all-you-need-pytorch) | [Yu-Hsiang](https://github.com/jadore801120) |  |
| [fairseq-py](https://github.com/facebookresearch/fairseq-py) | [Myle](https://github.com/myleott) & [Facebook Research](https://github.com/facebookresearch) |  |
| [Joey NMT](https://github.com/joeynmt/joeynmt) | [Heidelberg University](https://www.cl.uni-heidelberg.de/statnlpgroup/) & University of Amsterdam | minimalist, beginner-friendly |
| [NJUNMT-pytorch](https://github.com/whr94621/NJUNMT-pytorch) | [Haoran Wei](https://github.com/whr94621) & [Zaixiang Zheng](https://github.com/zhengzx-nlp) |  |
| [nmtpytorch](https://github.com/lium-lst/nmtpytorch) | [Ozan](https://github.com/ozancaglayan) |  |
| [nonauto-nmt](https://github.com/salesforce/nonauto-nmt) | [Salesforce Research](https://github.com/salesforce) |  |
| [OpenNMT-py](https://github.com/OpenNMT/OpenNMT-py) | [Harvard](http://nlp.seas.harvard.edu/), [Systran](http://www.systrangroup.com) |  |
| [pytorch_nmt](https://github.com/pcyin/pytorch_nmt) | [Pengcheng](https://github.com/pcyin) |  |
| [seq2seq.pytorch](https://github.com/eladhoffer/seq2seq.pytorch) | [Elad](https://github.com/eladhoffer) |  |
| [UNdreaMT](https://github.com/artetxem/undreamt) | [Mikel](https://github.com/artetxem) | Unsupervised |

### [TensorFlow](https://github.com/tensorflow/tensorflow)-based NMT
| Software | Who | Comments |
| -------- | --- | -------- |
| [byteNet-tensorflow](https://github.com/paarthneekhara/byteNet-tensorflow) | [Paarth](https://github.com/paarthneekhara) | |
| [bytenet_translation](https://github.com/Kyubyong/bytenet_translation) | [Kyubyong](https://github.com/Kyubyong) |  |
| [Neural Monkey](https://github.com/ufal/neuralmonkey) | [ÚFAL](https://github.com/ufal) |  |
| [NJUNMT-tf](https://github.com/zhaocq-nlp/NJUNMT-tf) | [ZhaoChengqi](https://github.com/zhaocq-nlp) |  |
| [NMT](https://github.com/tensorflow/nmt) | [Thang](https://github.com/lmthang), [Eugene](https://github.com/ebrevdo), [Rui](https://github.com/ruizhaogit) |  |
| [OpenNMT-tf](https://github.com/OpenNMT/OpenNMT-tf) | [Guillaume](https://github.com/guillaumekln) |  |
| [OpenSeq2Seq](https://github.com/NVIDIA/OpenSeq2Seq) | [Kuchaiev](https://github.com/okuchaiev) |  |
| [RNNsearch](https://github.com/XMUNLP/RNNsearch) | [Playinf](https://github.com/Playinf), [Xiamen University](https://github.com/XMUNLP) |  |
| [seq2seq](https://github.com/eske/seq2seq) | [Alexandre](https://github.com/eske) |  |
| [Tensor2Tensor](https://github.com/tensorflow/tensor2tensor) | [Google](https://github.com/tensorflow) | Implements many models |
| [tf-seq2seq](https://github.com/google/seq2seq) | [Denny](https://github.com/dennybritz) | [See also tutorial](https://www.tensorflow.org/versions/master/tutorials/seq2seq/index.html) |

### [Theano](https://github.com/Theano/Theano)-based NMT
Note: Theano will be [discontinued](https://groups.google.com/forum/#!topic/theano-users/7Poq8BZutbY)

| Software | Who | Comments |
| -------- | --- | -------- |
| [DCNMT](https://github.com/swordyork/dcnmt) | [SwordYork](https://github.com/SwordYork) |  |
| [dl4mt-tutorial](https://github.com/nyu-dl/dl4mt-tutorial) | [Cho](https://github.com/kyunghyuncho), [Orhan](https://github.com/orhanf) |  |
| [dl4mt-c2c](https://github.com/nyu-dl/dl4mt-c2c) | [Jason](https://github.com/jasonleeinf), [Cho](https://github.com/kyunghyuncho) |  |
| [HNMT](https://github.com/robertostling/hnmt) | [Robert](https://github.com/robertostling) |  |
| [Nematus](https://github.com/EdinburghNLP/nematus) | [Rico](https://github.com/rsennrich), [Cho](https://github.com/kyunghyuncho), [Orhan](https://github.com/orhanf) |  |
| [neuralmt](https://github.com/zomux/neuralmt) | [Raphael](https://github.com/zomux) |  |
| [NMT](https://github.com/tuzhaopeng/NMT) | [Zhaopeng](https://github.com/tuzhaopeng) | Adds coverage and context gate |
| [nmtpy](https://github.com/lium-lst/nmtpy) | [LIUM](https://github.com/lium-lst) | refactored dl4mt-tutorial |
| [SGNMT](https://github.com/ucam-smt/sgnmt) | [Cambridge](https://github.com/ucam-smt) | Syntactically-guided NMT |
| [THUMT](https://github.com/thumt/THUMT) | [Tsinghua University](https://github.com/thumt) | |

### [Torch](https://github.com/torch/distro)-based NMT
| Software | Who | Comments |
| -------- | --- | -------- |
| [fairseq](https://github.com/facebookresearch/fairseq) | [Facebook](https://github.com/facebookresearch) | Novel CNN+attention |
| [nmt-android](https://github.com/harvardnlp/nmt-android) | [Yoon](https://github.com/yoonkim), [Sasha](https://github.com/srush) |  |
| [NPMT](https://github.com/posenhuang/NPMT) | [Po-Sen](https://github.com/posenhuang), [Fairseq team](https://github.com/facebookresearch) | Phrase-based NMT |
| [OpenNMT](https://github.com/OpenNMT/OpenNMT) | [Harvard](http://nlp.seas.harvard.edu/), [Systran](http://www.systrangroup.com) |  |
| [seq2seq-attn](https://github.com/harvardnlp/seq2seq-attn) | [Yoon](https://github.com/yoonkim) | Superseded by OpenNMT, except for character-level stuff |

### Other
| Software | Who | Backend | Comments |
| -------- | --- | ------- | -------- |
| [nmt.hybrid](https://github.com/lmthang/nmt.hybrid) | [Thang](https://github.com/lmthang) | Matlab | also [word-based](https://github.com/lmthang/nmt.matlab) |
| [textsum](https://github.com/tensorflow/models/tree/master/research/textsum) | Google | TensorFlow | text summarization demo using seq2seq+attn |

nmt-list

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README