Ordered Memory
This repository contains the code used for Ordered Memory.
The code comes with instructions for experiments:
If you use this code or our results in your research, please cite as appropriate:
@incollection{NIPS2019_8748,
title = {Ordered Memory},
author = {Shen, Yikang and Tan, Shawn and Hosseini, Arian and Lin, Zhouhan and Sordoni, Alessandro and Courville, Aaron C},
booktitle = {Advances in Neural Information Processing Systems 32},
editor = {H. Wallach and H. Larochelle and A. Beygelzimer and F. d\textquotesingle Alch\'{e}-Buc and E. Fox and R. Garnett},
pages = {5038--5049},
year = {2019},
publisher = {Curran Associates, Inc.},
url = {http://papers.nips.cc/paper/8748-ordered-memory.pdf}
}
Software Requirements
Python 3, PyTorch 1.2, and torchtext are required for the current codebase.
Experiments
Propositional Logic
python -u proplog.py --cuda --save logic.pt
ListOps
python -u listops.py --cuda --name listops.pt
SST
python -u main.py --subtrees --cuda --name sentiment.pt --glove/--elmo (--fine-grained)