Python interface to Google word2vec.
Training is done using the original C code plus some patches, the other funcionality is pure python + numpy.
pip install word2vec
I highly recommend the use the Anaconda python distribution
The easiest way is to look at this examples: word2vec and word clusters
The default functionality from word2vec is also available from the command line:
- word2vec
- word2phrase
- w2v-distance
- w2v-word-analogy
- w2v-compute-accuracy
Some people reported that they needed to do this if running OS X: http://stackoverflow.com/questions/15590169/ld-library-not-found-for-lgfortran-mac-symlink-issue
- create a
bin
directory - run the make file inside
word2vec-c
- export the new
bin
directory to PATH
Tests require py.test