Search code examples
theanoword2vecdeep-learning

How to compute a language model with word2vec tool?

I'm trying to build a neural network language model and it seems that word2vec tool by Mikolov et al is a good tool for this purpose. I tried that but it just produces word representations. Does anybody know how i can produce a language model by that tool or any other reasonable deep learning framework?

Solution

  • Doc2Vec implemented in Gensim does the job. The trick is that they use the document ID as a context word, which is present in all window sizes of all the words in the document.

    Code is here in Python/Gensim