First off: This repository is based on NMTGMinor. Please see here for its main contributors.
This repo is a sequence-to-sequence toolkit.
It supports machine translation, speech recognition/translation.
- Python version >= 3.7 (most recommended)
- PyTorch >= 1.0.1
- apex when using half- or mixed-precision training
Create virtual environment from environment.yml by:
conda env create -f environment.yml