Tensorflow demo using the Large Movie Review Dataset from http://ai.stanford.edu/~amaas/data/sentiment/

All of these models predict sentiment from movie reviews.

python3, tensorflow-0.8, nltk

To set up:

1) clone this directory
2) cd USF/
3) wget http://ai.stanford.edu/~amaas/data/sentiment/aclImdb_v1.tar.gz
4) tar -zxvf aclImdb_v1.tar.gz
5) pip install -R requirements.txt
6) open the python3 terminal and import nltk
7) nltk.download()
8) Enter: d
9) Enter: punkt
10) Enter: q
11) exit python
12) python3 preprocess.py
13) cp default_config.py config.py

To run a model: python runner.py [command] [-d]

Where command is either: tokrnn to run the token (word embedding) RNN charrnn to run the character level rnn tokconv to run a token conv net charconv to run a character conv net chartokrnn to run the hierarchical char -> token RNN model

Options -d runs the model in debug mode, which will make all of the size parameters small so the model compiles fast. This is especially important when setting up a large RNN because compiling the graph takes a long time. Run it with -d to make sure that nothing breaks and then stop and run the real model.

This will run on a GPU or CPU.

@InProceedings{maas-EtAl:2011:ACL-HLT2011, author = {Maas, Andrew L. and Daly, Raymond E. and Pham, Peter T. and Huang, Dan and Ng, Andrew Y. and Potts, Christopher}, title = {Learning Word Vectors for Sentiment Analysis}, booktitle = {Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies}, month = {June}, year = {2011}, address = {Portland, Oregon, USA}, publisher = {Association for Computational Linguistics}, pages = {142--150}, url = {http://www.aclweb.org/anthology/P11-1015} }

References

Potts, Christopher. 2011. On the negativity of negation. In Nan Li and David Lutz, eds., Proceedings of Semantics and Linguistic Theory 20, 636-659.

Contact

For questions/comments/corrections please contact Andrew Maas amaas@cs.stanford.edu