OpenNMT

An open source neural machine translation system.

Features

While OpenNMT initially focused on standard sequence to sequence models applied to machine translation, it has been extended to support many additional models and features. The tables below highlight some of the key features of each implementation.

Tasks

  OpenNMT-py OpenNMT-tf
Image to text ✓ (v1 only)  
Language modeling
Sequence classification  
Sequence tagging  
Sequence to sequence
Speech to text ✓ (v1 only)
Summarization  

Models

  OpenNMT-py OpenNMT-tf
ConvS2S  
DeepSpeech2 ✓ (v1 only)  
GPT-2
Im2Text ✓ (v1 only)  
Listen, Attend and Spell  
RNN with attention
Transformer

Model configuration

  OpenNMT-py OpenNMT-tf
Cascaded and multi-column encoder  
Copy attention  
Coverage attention  
Hybrid models  
Multiple source  
Multiple input features
Relative position representations
Tied embeddings

Training

  OpenNMT-py OpenNMT-tf
Automatic evaluation
Automatic model export  
Contrastive learning  
Data augmentation (e.g. noise)
Early stopping
Gradient accumulation
Supervised alignment
Mixed precision
Moving average
Multi-GPU
Multi-node
On-the-fly tokenization
Pretrained embeddings
Scheduled sampling  
Sentence weighting  
Vocabulary update
Weighted dataset

Decoding

  OpenNMT-py OpenNMT-tf
Beam search
Coverage penalty
CTranslate2 compatibility
Ensemble  
Length penalty
N-best rescoring
N-gram blocking  
Phrase table  
Random noise
Random sampling
Replace unknown

For more details on how to use these features, please refer to the documentation of each project: