A machine transliterator for Hindi-English transliteration using CRF's and RNN's.
This project was developed as a final project for the CS880: Machine Learning for Sequences and Text course at the Univeristy of New Hampshire
Instructions on running this project.
The project is divided into parts: 1) CRF 2) Neural Networks There are two jupyter notebooks corresponding to each part with self-explanatory names.
There are a few custom utility python files which are used for data preprocessing. These files should not be changed.
HOW TO RUN:
To run the CRF model, open the transliteratorCRF.ipynb notebook using jupyter or colab. To run the NN models, open the transliteratorRnnSeq2Seq.ipynb notebook using jupyter or colab.
On the first run, uncomment the installation commands to install the required libraries and dependencies.
After installing, run each code cell one by one to get to the final evaluation cell.