A small word2vec project to illustrate its power when trained even on the smallest datasets.
Project is Python 2.7 and uses gensim to build word2vec model on a small dataset of Rock music entries in Wikipedia.
The dataset can be downloaded from here.
See the blog post with more information here.
http://localhost:5000/api/v1/rock/similar?pos=<comma separate list>&neg=<comma separated list>
For example: