nmhkahn / word2vec-tensorflow

Implement word2vec in tensorflow

word2vec implementation in tensorflow

This code first crawl Korean Law data in here, and do word2vec and visualization.
Basically each laws is separated by symbol <END>. If your text has no <END> symbol (just a single document) it will be no problem too.

Basic Usage

1. To crawl law data, use command python word2vec.py crawl.
2. python word2vec.py word2vec will excute preprocess of data and word2vec.
   You can also use python script.py that I made for test.
3. Visualization using t-SNE can perform with python word2vec.py vis.

For more information about arguemts, press python word2vec [crawl|word2vec|vis] --help or see parse_argument function in util.py.