Yoon Kim
yhk255@nyu.edu
September 24, 2014

Code for:

Convolutional Neural Networks for Sentence Classification
EMNLP 2014
http://arxiv.org/abs/1408.5882

Runs the model on Pang and Lee's movie review dataset (MR in the paper).
Please cite the original paper when using the data.

Runs on Python 2.7 and Theano 0.6

The code does not use GPUs. It should run much faster if you have a good GPU. See http://deeplearning.net/software/theano/tutorial/using_gpu.html for information on how to make it use GPUs.

(People have have reported up to 20x speed-up by modifying this code to make use of a modern GPU, so this is a must if working with larger corpora.)

Instructions:

1. with all the files in folder, run

python process_data.py -path

where -path points to the word2vec binary file (i.e. GoogleNews-vectors-negative300.bin file). 
Downloadable at https://code.google.com/p/word2vec/
This will create a pickle object called "mr.p" in the same folder, which contains the dataset in the right format.

2. run

python conv_net_sentence.py -nonstatic -rand
python conv_net_sentence.py -static -word2vec
python conv_net_sentence.py -nonstatic -word2vec

This will run the CNN-rand, CNN-static, and CNN-nonstatic models respectively in the paper.

*Note: Step 1 will create the dataset with different fold-assignments than was used in the paper.
You should still be getting a CV score of >81% with CNN-nonstatic model, though.