...files for the popular GloVe model for distributed word...
The demo.sh script downloads a small corpus, consisting of the first 100M characters of Wikipedia. It collects unigram counts, constructs and shuffles cooccurrence data, and trains a simple version of the GloVe