ManishRoy9/LowRes-Assamese-English-NMT

OpenNMTpy-based repository for low-resource Assamese-English NMT. Implements iterative back-translation for accurate translation. Boosts accessibility of Assamese language. # LowRes-Assamese-English-NMT OpenNMTpy-based repository for low-resource Assamese-English NMT. Implements iterative back-translation for accurate translation. Boosts accessibility of Assamese language. This repository provides an implementation of a low-resource Assamese to English Neural Machine Translation (NMT) system, leveraging the OpenNMT-py framework and employing an iterative back-translation approach. The project focuses on addressing the challenges of limited training data and resources, aiming to achieve accurate translations in a low-resource setting. Key Features: ● Utilizes OpenNMT-py, a powerful and flexible NMT toolkit, for efficient model training and translation. ● Implements an iterative back-translation technique, leveraging parallel data augmentation to enhance translation performance. ● Incorporates pre-processing steps such as tokenization, subword segmentation, and data cleaning to optimize the input data for training. ● Offers comprehensive documentation, including detailed instructions on setting up the environment, training the NMT model, and performing translations. Provides customizable configuration files and scripts for easy experimentation and adaptation to specific project requirements. ● Encourages community collaboration, enabling researchers and developers to contribute enhancements, share insights, and further improve the Assamese to English translation capability in low-resource scenarios. ● By utilizing this repository, researchers and practitioners can leverage the power of OpenNMT-py and the effectiveness of iterative back-translation to develop robust Assamese to English translation models, even with limited resources, advancing the accessibility and understanding of the Assamese language.