Skip to content

Latest commit

 

History

History
17 lines (10 loc) · 708 Bytes

File metadata and controls

17 lines (10 loc) · 708 Bytes

CNN_Sentence_Classification

This is a decent implemenation of the paper "Convolutional Neural Networks for Sentence Classification"(EMNLP2014)

Original implementation is here: https://github.com/yoonkim/CNN_sentence.git

To make the code concise and easy understand. I use Lasagne, which is a deep learning framework based on Theano.

---Usage---

  1. Follow the command in https://github.com/stanfordnlp/treelstm.git to get Stanford Parser, Stanford POS Tagger.
  2. Before run preprocess.py, you should set your own glove path and stanford parser lib path in this script.
  3. python preprocess.py
  4. ./run.sh

Btw, it is very easy to run the code on GPU, just follow the instructions in Theano website.