OpenNMT project is composed of 3 main repositories:
- OpenNMT-Lua (a.k.a. OpenNMT): the main project developed with LuaTorch.
Optimized and stable code for production and large scale experiments.
- OpenNMT-py: light version of OpenNMT using PyTorch.
Initially created by the Facebook AI research team as a sample project for PyTorch, this version is easier to extend and is suited for research purpose but does not include all features.
- OpenNMT-C (a.k.a. CTranslate): C++ inference engine for OpenNMT models.
OpenNMT is a generic deep learning framework mainly specialized in sequence-to-sequence models covering a variety of tasks such as machine translation, summarization, image to text, and speech recognition. The framework has also been extended for other non sequence-to-sequence tasks like language modelling and sequence tagging.
The framework is implemented to be as generic as possible and can be used either via command line applications, client-server, or libraries.
The project is self-contained and ready to use for both research and production.
OpenNMT project is an open-source initiative derivated from seq2seq-attn, initially created by Kim Yoon at HarvardNLP group.
You can find additional help or tutorials in the following resources:
If you find an error in this documentation, please consider opening an issue or directly submitting a modification by clicking on the edit button at the top of a page.