/MODE-LSTM

The repository for EMNLP 2020 accepted paper "MODE-LSTM: A Parameter-efficient Recurrent Network with Multi-Scale for Sentence Classification"

Primary LanguagePython

MODE-LSTM

The repository for EMNLP 2020 accepted paper "MODE-LSTM: A Parameter-efficient Recurrent Network with Multi-Scale for Sentence Classification".

Dependencies

  • python 2.7
  • tensorflow 1.14.0
  • keras 2.2.4

Data Preprocessing

The folder data contains the dataset SST5 for testing. As for other datasets, the IE can be downloaded according to paper "Dynamic Compositional Neural Networks over Tree Structure", and the remaining datasets can refer to repository TextCNN.

Here, we present a case how to process SST5. First, you should download the pretrain word embedding glove.840B.300d.txt from glove, and place it under folder data.

To process the raw data SST5, run the command

python text_process.py

Running model

You can run the command

python main.py

When you run this command, please sure you have run the data preprocessing file text_process.py.

Reference

@inproceedings{ma-etal-2020-mode,
    title = "{MODE}-{LSTM}: A Parameter-efficient Recurrent Network with Multi-Scale for Sentence Classification",
    author = "Ma, Qianli  and
      Lin, Zhenxi  and
      Yan, Jiangyue  and
      Chen, Zipeng  and
      Yu, Liuhong",
    booktitle = "Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)",
    month = nov,
    year = "2020",
    publisher = "Association for Computational Linguistics",
    pages = "6705--6715"
}