This script emulates the experiment in Tomas Milokov's Exploiting Similarities among Languages for Machine Translation.
Usage:
Train two models in word2vec using gensim, one in a target langauge, one in a source language. Load both, and run the script to see accuracy @5 and accuracy @1. Refer to paper for full experiment details.
This is a cleaned, comment version of the script in https://github.com/mostafachatillon/word2vec/tree/master/word2vec%20machine%20translation by Mostafa Chatillon.