This repository contains the code for the Partial Shuffle method, and a modified version of the DOC language model that utilizes this method.
If you'd like to run the DOC + Partial Shuffle models, use the same commands as in the original DOC model, presented here.
The code for the Partial Shuffle method itself is in partial_shuffle.py
. If you'd like to use this method in your own language model, simply import partial_shuffle.py
, and call it before each epoch, as in line 196 in main.py
. No other modifications are required.
If you found this code useful, please cite the following paper:
@article{press2019partially,
title={Partially Shuffling the Training Data to Improve Language Models},
author={Press, Ofir},
journal={arXiv preprint arXiv:1903.04167},
year={2019}
}