Open Source Neural Machine Translation in PyTorch
Switch branches/tags
Nothing to show
Clone or download
Pull request Compare This branch is 96 commits ahead of hpzhao:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.

OpenNMT-py: Open-Source Neural Machine Translation

This is a Pytorch port of OpenNMT, an open-source (MIT) neural machine translation system. It is designed to be research friendly to try out new ideas in translation, summary, image-to-text, morphology, and many other domains.

OpenNMT-py is run as a collaborative open-source project. It is currently maintained by Sasha Rush (Cambridge, MA), Ben Peters (Saarbrücken), and Jianyu Zhan (Shenzhen). The original code was written by Adam Lerer (NYC). Codebase is nearing a stable 0.1 version. We currently recommend forking if you want stable code.

We love contributions. Please consult the Issues page for any Contributions Welcome tagged post.

Table of Contents


pip install -r requirements.txt


The following OpenNMT features are implemented:

  • multi-layer bidirectional RNNs with attention and dropout
  • data preprocessing
  • saving and loading from checkpoints
  • Inference (translation) with batching and beam search
  • Context gate
  • Multiple source and target RNN (lstm/gru) types and attention (dotprod/mlp) types
  • TensorBoard/Crayon logging
  • Source word features

Beta Features (committed):

  • multi-GPU
  • Image-to-text processing
  • "Attention is all you need"
  • Copy, coverage
  • Structured attention
  • Conv2Conv convolution model
  • SRU "RNNs faster than CNN" paper
  • Inference time loss functions.


Step 1: Preprocess the data

python -train_src data/src-train.txt -train_tgt data/tgt-train.txt -valid_src data/src-val.txt -valid_tgt data/tgt-val.txt -save_data data/demo

We will be working with some example data in data/ folder.

The data consists of parallel source (src) and target (tgt) data containing one sentence per line with tokens separated by a space:

  • src-train.txt
  • tgt-train.txt
  • src-val.txt
  • tgt-val.txt

Validation files are required and used to evaluate the convergence of the training. It usually contains no more than 5000 sentences.

After running the preprocessing, the following files are generated:

  • serialized PyTorch file containing training data
  • serialized PyTorch file containing validation data
  • serialized PyTorch file containing vocabulary data

Internally the system never touches the words themselves, but uses these indices.

Step 2: Train the model

python -data data/demo -save_model demo-model

The main train command is quite simple. Minimally it takes a data file and a save file. This will run the default model, which consists of a 2-layer LSTM with 500 hidden units on both the encoder/decoder. You can also add -gpuid 1 to use (say) GPU 1.

Step 3: Translate

python -model -src data/src-test.txt -output pred.txt -replace_unk -verbose

Now you have a model which you can use to predict on new data. We do this by running beam search. This will output predictions into pred.txt.

!!! note "Note" The predictions are going to be quite terrible, as the demo dataset is small. Try running on some larger datasets! For example you can download millions of parallel sentences for translation or summarization.

Some useful tools:

Full Translation Example

The example below uses the Moses tokenizer ( to prepare the data and the moses BLEU script for evaluation.

sed -i "s/$RealBin\/..\/share\/nonbreaking_prefixes//" tokenizer.perl

WMT'16 Multimodal Translation: Multi30k (de-en)

An example of training for the WMT'16 Multimodal Translation task (

0) Download the data.

mkdir -p data/multi30k
wget &&  tar -xf training.tar.gz -C data/multi30k && rm training.tar.gz
wget && tar -xf validation.tar.gz -C data/multi30k && rm validation.tar.gz
wget && tar -xf mmt16_task1_test.tgz -C data/multi30k && rm mmt16_task1_test.tgz

1) Preprocess the data.

# Delete the last line of val and training files.
for l in en de; do for f in data/multi30k/*.$l; do if [[ "$f" != *"test"* ]]; then sed -i "$ d" $f; fi;  done; done
for l in en de; do for f in data/multi30k/*.$l; do perl tokenizer.perl -a -no-escape -l $l -q  < $f > $f.atok; done; done
python -train_src data/multi30k/train.en.atok -train_tgt data/multi30k/ -valid_src data/multi30k/val.en.atok -valid_tgt data/multi30k/ -save_data data/multi30k.atok.low -lower

2) Train the model.

python -data data/multi30k.atok.low -save_model multi30k_model -gpuid 0

3) Translate sentences.

python -gpu 0 -model multi30k_model_* -src data/multi30k/test.en.atok -tgt data/multi30k/ -replace_unk -verbose -output multi30k.test.pred.atok

4) Evaluate.

perl tools/multi-bleu.perl data/multi30k/ < multi30k.test.pred.atok

Pretrained embeddings (e.g. GloVe)

Go to tutorial: How to use GloVe pre-trained embeddings in OpenNMT-py

Pretrained Models

The following pretrained models can be downloaded and used with (These were trained with an older version of the code; they will be updated soon).

  • onmt_model_en_de_200k: An English-German translation model based on the 200k sentence dataset at OpenNMT/IntegrationTesting. Perplexity: 20.
  • onmt_model_en_fr_b1M (coming soon): An English-French model trained on benchmark-1M. Perplexity: 4.85.


OpenNMT technical report

  author    = {Guillaume Klein and
               Yoon Kim and
               Yuntian Deng and
               Jean Senellart and
               Alexander M. Rush},
  title     = {OpenNMT: Open-Source Toolkit for Neural Machine Translation},
  booktitle = {Proc. ACL},
  year      = {2017},
  url       = {},
  doi       = {10.18653/v1/P17-4012}