volkancirik / eoffset

eOffset - Syntactic/Semantic Similarity Tasks for Word Vectors

Basic benchmarks for Mikolov's work.

Here is a better and general version : http://wordvectors.org/suite.php

First unzip the sample word vector file :

cd embeddings
zcat rnn80.gz > rnn80.embeddings

Compile the binaries

cd run
make bin

Then run the bencmark

make rnn80.result

Take a look at the results. It should be exactly the same as in this paper.